aboutsummaryrefslogtreecommitdiffstats
path: root/arch/powerpc/boot
Commit message (Collapse)AuthorAge
* powerpc/p1022ds: fix audio-related properties in the device treeTimur Tabi2011-06-22
| | | | | | | | | | | | | | | | | On the Freescale P1022DS reference board, the SSI audio controller is connected in "asynchronous" mode to the codec's clocks, so the device tree needs an "fsl,ssi-asynchronous" property. Also remove the clock-frequency property from the wm8776 node, because the clock is enabled only if U-Boot enables it, and U-Boot will set the property if the clock is enabled. A future version of the P1022DS audio driver will configure the clock itself, but for now, the driver should not be told that the clock is running when it isn't. Also fix the FIFO depth to 15, instead of 16. Signed-off-by: Timur Tabi <timur@freescale.com> Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
* dtc/powerpc: remove obsolete .gitignore entriesWolfram Sang2011-06-08
| | | | | | | | | dtc was moved and .gitignores have been added to the new location. So, we can delete the old, forgotten ones. Signed-off-by: Wolfram Sang <w.sang@pengutronix.de> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
* powerpc/4xx: Adding PCIe MSI supportRupjyoti Sarmah2011-05-26
| | | | | | | | | This patch adds MSI support for 440SPe, 460Ex, 460Sx and 405Ex. Signed-off-by: Rupjyoti Sarmah <rsarmah@apm.com> Signed-off-by: Tirumala R Marri <tmarri@apm.com> Acked-by: Josh Boyer <jwboyer@linux.vnet.ibm.com> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
* ptp: Added a clock that uses the eTSEC found on the MPC85xx.Richard Cochran2011-05-23
| | | | | | | | | | | The eTSEC includes a PTP clock with quite a few features. This patch adds support for the basic clock adjustment functions, plus two external time stamps, one alarm, and the PPS callback. Signed-off-by: Richard Cochran <richard.cochran@omicron.at> Acked-by: David S. Miller <davem@davemloft.net> Acked-by: John Stultz <john.stultz@linaro.org> Signed-off-by: John Stultz <john.stultz@linaro.org>
* Merge git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6Linus Torvalds2011-05-20
|\ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | * git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6: (45 commits) crypto: caam - add support for sha512 variants of existing AEAD algorithms crypto: caam - remove unused authkeylen from caam_ctx crypto: caam - fix decryption shared vs. non-shared key setting crypto: caam - platform_bus_type migration crypto: aesni-intel - fix aesni build on i386 crypto: aesni-intel - Merge with fpu.ko crypto: mv_cesa - make count_sgs() null-pointer proof crypto: mv_cesa - copy remaining bytes to SRAM only when needed crypto: mv_cesa - move digest state initialisation to a better place crypto: mv_cesa - fill inner/outer IV fields only in HMAC case crypto: mv_cesa - refactor copy_src_to_buf() crypto: mv_cesa - no need to save digest state after the last chunk crypto: mv_cesa - print a warning when registration of AES algos fail crypto: mv_cesa - drop this call to mv_hash_final from mv_hash_finup crypto: mv_cesa - the descriptor pointer register needs to be set just once crypto: mv_cesa - use ablkcipher_request_cast instead of the manual container_of crypto: caam - fix printk recursion for long error texts crypto: caam - remove unused keylen from session context hwrng: amd - enable AMD hw rnd driver for Maple PPC boards hwrng: amd - manage resource allocation ...
| * crypto: caam - de-CHIP-ify device tree compatiblesKim Phillips2011-03-26
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | - all the integration parameters have been captured by the binding. - the block name really uniquely identifies this hardware. Some advocate putting SoC names everywhere in case software needs to work around some chip-specific bug, but more precise SoC information already exists in SVR, and board information already exists in the top-level device tree node. Note that sometimes the SoC name is a worse identifier than the block version, as the block version can change between revisions of the same SoC. As a matter of historical reference, neither SEC versions 2.x nor 3.x (driven by talitos) ever needed CHIP references. Signed-off-by: Kim Phillips <kim.phillips@freescale.com> Cc: Kumar Gala <kumar.gala@freescale.com> Cc: Scott Wood <scottwood@freescale.com> Acked-off-by: Grant Likely <grant.likely@secretlab.ca> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
| * crypto: caam - standardize device tree naming convention to utilize '-vX.Y'Kim Phillips2011-03-26
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Help clarify that the number trailing in compatible nomenclature is the version number of the device, i.e., change: "fsl,p4080-sec4.0", "fsl,sec4.0"; to: "fsl,p4080-sec-v4.0", "fsl,sec-v4.0"; Signed-off-by: Kim Phillips <kim.phillips@freescale.com> Cc: Kumar Gala <kumar.gala@freescale.com> Cc: Steve Cornelius <sec@pobox.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
| * crypto: caam - Add support for the Freescale SEC4/CAAMKim Phillips2011-03-26
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The SEC4 supercedes the SEC2.x/3.x as Freescale's Integrated Security Engine. Its programming model is incompatible with all prior versions of the SEC (talitos). The SEC4 is also known as the Cryptographic Accelerator and Assurance Module (CAAM); this driver is named caam. This initial submission does not include support for Data Path mode operation - AEAD descriptors are submitted via the job ring interface, while the Queue Interface (QI) is enabled for use by others. Only AEAD algorithms are implemented at this time, for use with IPsec. Many thanks to the Freescale STC team for their contributions to this driver. Signed-off-by: Steve Cornelius <sec@pobox.com> Signed-off-by: Kim Phillips <kim.phillips@freescale.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
* | Merge remote branch 'kumar/next' into nextBenjamin Herrenschmidt2011-05-19
|\ \
| * | powerpc/85xx: Create dts of each core in CAMP mode for P1020RDBPrabhakar Kushwaha2011-05-19
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Create the dts files for each core and splits the devices between the two cores for P1020RDB. Core0 has core0 to have memory, l2, i2c, spi, gpio, tdm, dma, usb, eth1, eth2, sdhc, crypto, global-util, message, pci0, pci1, msi. Core1 has l2, eth0, crypto. MPIC is shared between two cores but each core will protect its interrupts from other core by using "protected-sources" of mpic. Fix compatible property for global-util node of P1020si.dtsi. Signed-off-by: Prabhakar Kushwaha <prabhakar@freescale.com> Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
| * | powerpc/85xx: Fix PCIe IDSEL for Px020RDBPrabhakar Kushwaha2011-05-19
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | PCIe device in legacy mode can trigger interrupts using the wires #INTA, #INTB ,#INTC and #INTD. PCI devices are obligated to use #INTx for interrupts under legacy mode. Each PCI slot or device is typically wired to different inputs on the interrupt controller. So, Define interrupt-map and interrupt-map-mask properties for device tree to of map each PCI interrupt signal to the inputs of the interrupt controller. Signed-off-by: Prabhakar Kushwaha <prabhakar@freescale.com> Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
| * | powerpc/85xx: P2020 DTS: re-organize dts filesPrabhakar Kushwaha2011-05-19
| | | | | | | | | | | | | | | | | | | | | | | | Creates P2020si.dtsi, containing information for P2020 SoC. Modifies dts files for P2020 based systems to use dtsi file. Signed-off-by: Prabhakar Kushwaha <prabhakar@freescale.com> Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
| * | powerpc/85xx: P1020 DTS : re-organize dts filesPrabhakar Kushwaha2011-05-19
| | | | | | | | | | | | | | | | | | | | | | | | | | | Creates P1020si.dtsi, containing information for the P1020 SoC. Modifies dts files for P1020 based systems to use dtsi file Signed-off-by: Prabhakar Kushwaha <prabhakar@freescale.com> Acked-by: Grant Likely <grant.likelY@secretlab.ca> Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
| * | powerpc/p1022ds: fix broken mpic timer nodeScott Wood2011-05-19
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | There is no hardware interrupt 0xf7. But now we can express the timer interrupt using 4-cell interrupts. This requires converting all of the other interrupt specifiers in the tree as well. Also add the second timer group, and fix the reg property to only describe the timer registers. Signed-off-by: Scott Wood <scottwood@freescale.com> Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
* | | powerpc: Use the deterministic mode of arMichal Marek2011-05-19
| | | | | | | | | | | | | | | Signed-off-by: Michal Marek <mmarek@suse.cz> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
* | | powerpc: Call gzip with -nMichal Marek2011-05-19
|/ / | | | | | | | | | | | | | | | | | | The timestamps recorded in the .gz files add no value. Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Paul Mackerras <paulus@samba.org> Cc: linuxppc-dev@lists.ozlabs.org Signed-off-by: Michal Marek <mmarek@suse.cz> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
* | powerpc/boot: Add an ePAPR compliant boot wrapperDavid Gibson2011-04-20
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This is a first cut at making bootwrapper code which will produce a zImage compliant with the requirements set down by ePAPR. This is a very simple bootwrapper, taking the device tree blob supplied by the ePAPR boot program and passing it on to the kernel. It builds on the earlier patch to build a relocatable ET_DYN zImage to meet the other ePAPR image requirements. For good measure we have some paranoid checks which will generate warnings if some of the ePAPR entry condition guarantees are not met. Signed-off-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
* | powerpc/boot: Allow building the zImage wrapper as a relocatable ET_DYNMichael Ellerman2011-04-20
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This patch adds code, linker script and makefile support to allow building the zImage wrapper around the kernel as a position independent executable. This results in an ET_DYN instead of an ET_EXEC ELF output file, which can be loaded at any location by the firmware and will process its own relocations to work correctly at the loaded address. This is of interest particularly since the standard ePAPR image format must be an ET_DYN (although this patch alone is not sufficient to produce a fully ePAPR compliant boot image). Note for now we don't enable building with -pie for anything. Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
* | powerpc/85xx: Update dts for PCIe memory maps to match u-boot of Px020RDBPrabhakar Kushwaha2011-04-04
|/ | | | | | | | | | PCIe memory address space is 1:1 mapped with u-boot. Update dts of Px020RDB i.e. P1020RDB and P2020RDB to match the address map changes in u-boot. Signed-off-by: Prabhakar Kushwaha <prabhakar@freescale.com> Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
* Merge remote branch 'jwb/next' into nextBenjamin Herrenschmidt2011-03-17
|\
| * powerpc/44x: PHY fixup for USB on canyonlands boardRupjyoti Sarmah2011-02-02
| | | | | | | | | | | | | | | | This fix is a reset for USB PHY that requires some amount of time for power to be stable on Canyonlands. Signed-off-by: Rupjyoti Sarmah <rsarmah@apm.com> Signed-off-by: Josh Boyer <jwboyer@linux.vnet.ibm.com>
| * powerpc/44x: Add USB DWC DTS entry to Canyonlands boardTirumala Marri2011-02-02
| | | | | | | | | | | | | | Add Synopsys Designware DTS entry for 460EX based Canyonlands board. Signed-off-by: Tirumala R Marri<tmarri@apm.com> Signed-off-by: Josh Boyer <jwboyer@linux.vnet.ibm.com>
* | powerpc/85xx: Update sata controller compatible for p1022ds boardXulei2011-03-15
| | | | | | | | | | | | | | | | | | | | Update p1022 sata compatible to "fsl,p1022-sata", "fsl,pq-sata-v2". p1022ds sata controller is v2 version comparing previous FSL sata controller, for example, mpc8536. Signed-off-by: Lei Xu <B33228@freescale.com> Signed-off-by: Roy Zang <tie-fei.zang@freescale.com> Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
* | powerpc/8xx: remove obsolete mgsuvd boardHolger Brunck2011-03-15
| | | | | | | | | | | | | | | | | | | | | | | | The MPC852 based mgsuvd board from Keymile was initially ported, but later on not developed further. This patch removes the respective files to decrease merging conflicts and unneeded maintenance. Signed-off-by: Holger Brunck <holger.brunck@keymile.com> Acked-by: Heiko Schocher<hs@denx.de> Cc: Vitaly Bordug <vitb@kernel.crashing.org> Cc: Marcelo Tosatti <marcelo@kvack.org> Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
* | powerpc/82xx: rename and update mgcoge board supportHolger Brunck2011-03-15
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The mgcoge board from keymile is now base for some other similar boards. Therefore the board specific name mgcoge was renamed to a generic name km82xx. Additionally some enhancements were made: - rework partition table in dts file - add cpm2_pio_c gpio controller in dts file - update defconfig - add pin description for SCC1 - add pin description and configuration for USB Signed-off-by: Holger Brunck <holger.brunck@keymile.com> Acked-by: Heiko Schocher <hs@denx.de> CC: Benjamin Herrenschmidt <benh@kernel.crashing.org> CC: Heiko Schocher <hs@denx.de> Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
* | powerpc/83xx: rename and update kmeter1Holger Brunck2011-03-15
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Beside the MPC 8360 based board kmeter1 other km83xx boards from keymile will follow. Therefore the board specific naming kmeter1 for functions and files were replaced with km83xx. Additionally some updates were made: - update defconfig for 2.6.38 - rework flash partitioning in dts file - add gpio controller for qe_pio_c in dts Signed-off-by: Holger Brunck <holger.brunck@keymile.com> Acked-by: Heiko Schocher <hs@denx.de> CC: Benjamin Herrenschmidt <benh@kernel.crashing.org> CC: Heiko Schocher <hs@denx.de> Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
* | powerpc/boot/dts: Install dts from the right directoryBen Hutchings2011-01-20
| | | | | | | | | | | | | | | | | | | | | | | | The dts-installed variable is initialised using a wildcard path that will be expanded relative to the build directory. Use the existing variable dtstree to generate an absolute wildcard path that will work when building in a separate directory. Reported-by: Gerhard Pircher <gerhard_pircher@gmx.net> Signed-off-by: Ben Hutchings <ben@decadent.org.uk> Tested-by: Gerhard Pircher <gerhard_pircher@gmx.net> [against 2.6.32] Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
* | Merge remote branch 'kumar/next' into mergeBenjamin Herrenschmidt2011-01-20
|\ \
| * | powerpc/85xx: fix compatible properties of the P1022DS DMA nodes used for audioTimur Tabi2011-01-12
| | | | | | | | | | | | | | | | | | | | | | | | | | | In order to prevent the fsl_dma driver from claiming the DMA channels that the P1022DS audio driver needs, the compatible properties for those nodes must say "fsl,ssi-dma-channel" instead of "fsl,eloplus-dma-channel". Signed-off-by: Timur Tabi <timur@freescale.com> Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
| * | powerpc/mpc8308: fix USB DR controller initializationIlya Yanok2011-01-12
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | MPC8308 has ULPI pin muxing settings in SICRH register, bits 17-18 which is different from both MPC8313 and MPC8315. Also MPC8308 doesn't have REFSEL, UTMI_PHY_EN and OTG_PORT fields in the USB DR controller CONTROL register. Signed-off-by: Ilya Yanok <yanok@emcraft.com> Tested-by: Wolfgang Denk <wd@denx.de> Acked-by: Wolfgang Denk <wd@denx.de> Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
* | | Merge branch 'next' of ↵Linus Torvalds2011-01-11
|\ \ \ | | |/ | |/| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc * 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc: (72 commits) powerpc/pseries: Fix build of topology stuff without CONFIG_NUMA powerpc/pseries: Fix VPHN build errors on non-SMP systems powerpc/83xx: add mpc8308_p1m DMA controller device-tree node powerpc/83xx: add DMA controller to mpc8308 device-tree node powerpc/512x: try to free dma descriptors in case of allocation failure powerpc/512x: add MPC8308 dma support powerpc/512x: fix the hanged dma transfer issue powerpc/512x: scatter/gather dma fix powerpc/powermac: Make auto-loading of therm_pm72 possible of/address: Use propper endianess in get_flags powerpc/pci: Use printf extension %pR for struct resource powerpc: Remove unnecessary casts of void ptr powerpc: Disable VPHN polling during a suspend operation powerpc/pseries: Poll VPA for topology changes and update NUMA maps powerpc: iommu: Add device name to iommu error printks powerpc: Record vma->phys_addr in ioremap() powerpc: Update compat_arch_ptrace powerpc: Fix PPC_PTRACE_SETHWDEBUG on PPC_BOOK3S powerpc/time: printk time stamp init not correct powerpc: Minor cleanups for machdep.h ...
| * | Merge remote branch 'gcl/next' into nextBenjamin Herrenschmidt2011-01-10
| |\ \
| | * | powerpc/83xx: add mpc8308_p1m DMA controller device-tree nodeIlya Yanok2010-12-30
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | MPC8308 has DMA controller compatible with mpc512x_dma driver. This patch adds device-tree node to support DMA controller on MPC8308 P1M board. Signed-off-by: Ilya Yanok <yanok@emcraft.com> Acked-by: Wolfgang Denk <wd@denx.de> Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
| | * | powerpc/83xx: add DMA controller to mpc8308 device-tree nodeIlya Yanok2010-12-30
| | |/ | | | | | | | | | | | | | | | | | | | | | | | | | | | MPC8308 has DMA controller compatible with mpc512x_dma driver. This patch adds device-tree node to support DMA controller on MPC8308RDB board. Signed-off-by: Ilya Yanok <yanok@emcraft.com> Acked-by: Wolfgang Denk <wd@denx.de> Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
| * | powerpc/4xx: Add Canyonlands suspend and idle supportVictor Gallardo2010-11-29
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | - Add Clock Power Management (CPM) node to dts tree - Add idle-doze entry in CPM node - Add standby entry in CPM node - Add PM and SUSPEND support by default in defconfig - Remove UART2 and UART3 as they are unused, this will allow CPM to put unused-units (UART2 and UART3) to sleep. Signed-off-by: Victor Gallardo <vgallardo@apm.com> Signed-off-by: Josh Boyer <jwboyer@linux.vnet.ibm.com>
| * | powerpc/4xx: Add Kilauea suspend and idle supportVictor Gallardo2010-11-29
| |/ | | | | | | | | | | | | | | | | | | | | | | - Add Clock Power Management (CPM) node to dts tree - Add idle-doze entry in CPM node - Add standby entry in CPM node - Add PM and SUSPEND support by default in defconfig - Add NO_HZ and CONFIG_HIGH_RES_TIMERS support by default in defconfig Signed-off-by: Victor Gallardo <vgallardo@apm.com> Signed-off-by: Josh Boyer <jwboyer@linux.vnet.ibm.com>
* | powerpc/405: Fix missing #{address,size}-cells in i2c nodeGrant Likely2011-01-03
| | | | | | | | Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
* | powerpc/5200: dts: refactor dts filesJohn Bonesio2011-01-03
| | | | | | | | | | | | | | | | | | This patch creates mpc5200b.dtsi containing the information for the MPC5200b SoC then modifies all of the dts files for MPC5200b based systems to use mpc5200b.dtsi. Signed-off-by: John Bonesio <bones@secretlab.ca> Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
* | powerpc/5200: dts: Change combatible strings on localbusJohn Bonesio2011-01-03
| | | | | | | | | | | | | | | | This patch changes some incorrect compatible strings on the local plus bus node in dts files for MPC5200b based systems. Signed-off-by: John Bonesio <bones@secretlab.ca> Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
* | powerpc/5200: dts: remove unused propertiesJohn Bonesio2011-01-03
| | | | | | | | | | | | | | | | This patch remove unused properties in dts files in preparation of refactoring the dts files for MPC5200b based boards. Signed-off-by: John Bonesio <bones@secretlab.ca> Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
* | powerpc/5200: dts: rename nodes to prepare for refactoring dts filesJohn Bonesio2011-01-03
| | | | | | | | | | | | | | | | | | This patch renames nodes in dts fils for MPC5200b files to prepare for refactoring of these files later. When refactoring it will be easier to verify the results if the node names aren't changing at the same time. Signed-off-by: John Bonesio <bones@secretlab.ca> Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
* | powerpc/dts: fix syntax bugs in bluestone.dtsGrant Likely2011-01-03
| | | | | | | | Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
* | powerpc/pcm{030,032}: add pagesize to dtsWolfram Sang2010-12-24
| | | | | | | | | | Signed-off-by: Wolfram Sang <w.sang@pengutronix.de> Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
* | of/powerpc: Use generic rule to build dtb'sDirk Brandewie2010-12-23
|/ | | | | | | | Modify arch/powerpc/boot/Makefile to use dtc command in scripts/Makefile.lib Signed-off-by: Dirk Brandewie <dirk.brandewie@gmail.com> Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
* powerpc: Fix div64 in bootloaderBenjamin Herrenschmidt2010-11-17
| | | | | | | | The code is missing a fix that went into the main kernel variant (we should try to share that code again at some stage) Reported-by: Albert Cahalan <acahalan@gmail.com> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
* Merge branch 'for-linus' of ↵Linus Torvalds2010-10-25
|\ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6 * 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6: (365 commits) ALSA: hda - Disable sticky PCM stream assignment for AD codecs ALSA: usb - Creative USB X-Fi volume knob support ALSA: ca0106: Use card specific dac id for mute controls. ALSA: ca0106: Allow different sound cards to use different SPI channel mappings. ALSA: ca0106: Create a nice spot for mapping channels to dacs. ALSA: ca0106: Move enabling of front dac out of hardcoded setup sequence. ALSA: ca0106: Pull out dac powering routine into separate function. ALSA: ca0106 - add Sound Blaster 5.1vx info. ASoC: tlv320dac33: Use usleep_range for delays ALSA: usb-audio: add Novation Launchpad support ALSA: hda - Add workarounds for CT-IBG controllers ALSA: hda - Fix wrong TLV mute bit for STAC/IDT codecs ASoC: tpa6130a2: Error handling for broken chip ASoC: max98088: Staticise m98088_eq_band ASoC: soc-core: Fix codec->name memory leak ALSA: hda - Apply ideapad quirk to Acer laptops with Cxt5066 ALSA: hda - Add some workarounds for Creative IBG ALSA: hda - Fix wrong SPDIF NID assignment for CA0110 ALSA: hda - Fix codec rename rules for ALC662-compatible codecs ALSA: hda - Add alc_init_jacks() call to other codecs ...
| * Merge branch 'topic/asoc' into for-linusTakashi Iwai2010-10-25
| |\ | | | | | | | | | | | | Conflicts: arch/powerpc/platforms/85xx/p1022_ds.c
| | * Merge branch 'for-2.6.36' into for-2.6.37Mark Brown2010-08-16
| | |\ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Fairly simple conflicts, the most serious ones are the i.MX ones which I suspect now need another rename. Conflicts: arch/arm/mach-mx2/clock_imx27.c arch/arm/mach-mx2/devices.c arch/arm/mach-omap2/board-rx51-peripherals.c arch/arm/mach-omap2/board-zoom2.c sound/soc/fsl/mpc5200_dma.c sound/soc/fsl/mpc5200_dma.h sound/soc/fsl/mpc8610_hpcd.c sound/soc/pxa/spitz.c
| | * | asoc/multi-component: fsl: add support for disabled SSI nodesTimur Tabi2010-08-12
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Add support for adding "status = disabled" to an SSI node to incidate that it is not wired on the board. This replaces the not-so-intuitive previous method of omitting a codec-handle property. Signed-off-by: Timur Tabi <timur@freescale.com> Acked-by: Kumar Gala <galak@kernel.crashing.org> Acked-by: Mark Brown <broonie@opensource.wolfsonmicro.com> Signed-off-by: Liam Girdwood <lrg@slimlogic.co.uk>
* | | | Merge branch 'next' of ↵Linus Torvalds2010-10-22
|\ \ \ \ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc * 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc: (71 commits) powerpc/44x: Update ppc44x_defconfig powerpc/watchdog: Make default timeout for Book-E watchdog a Kconfig option fsl_rio: Add comments for sRIO registers. powerpc/fsl-booke: Add e55xx (64-bit) smp defconfig powerpc/fsl-booke: Add p5020 DS board support powerpc/fsl-booke64: Use TLB CAMs to cover linear mapping on FSL 64-bit chips powerpc/fsl-booke: Add support for FSL Arch v1.0 MMU in setup_page_sizes powerpc/fsl-booke: Add support for FSL 64-bit e5500 core powerpc/85xx: add cache-sram support powerpc/85xx: add ngPIXIS FPGA device tree node to the P1022DS board powerpc: Fix compile error with paca code on ppc64e powerpc/fsl-booke: Add p3041 DS board support oprofile/fsl emb: Don't set MSR[PMM] until after clearing the interrupt. powerpc/fsl-booke: Add PCI device ids for P2040/P3041/P5010/P5020 QoirQ chips powerpc/mpc8xxx_gpio: Add support for 'qoriq-gpio' controllers powerpc/fsl_booke: Add support to boot from core other than 0 powerpc/p1022: Add probing for individual DMA channels powerpc/fsl_soc: Search all global-utilities nodes for rstccr powerpc: Fix invalid page flags in create TLB CAM path for PTE_64BIT powerpc/mpc83xx: Support for MPC8308 P1M board ... Fix up conflict with the generic irq_work changes in arch/powerpc/kernel/time.c
2684 2685 2686 2687 2688 2689 2690 2691 2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702 2703 2704 2705 2706 2707 2708 2709 2710 2711 2712 2713 2714 2715 2716 2717 2718 2719 2720 2721 2722 2723 2724 2725 2726 2727 2728 2729 2730 2731 2732 2733 2734 2735 2736 2737 2738 2739 2740 2741 2742 2743 2744 2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 2766 2767 2768 2769 2770 2771 2772 2773 2774 2775 2776 2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791 2792 2793 2794 2795 2796 2797 2798 2799 2800 2801 2802 2803 2804 2805 2806 2807 2808 2809 2810 2811 2812 2813 2814 2815 2816 2817 2818 2819 2820 2821 2822 2823 2824 2825 2826 2827 2828 2829 2830 2831 2832 2833 2834 2835 2836 2837 2838 2839 2840 2841 2842 2843 2844 2845 2846 2847 2848 2849 2850 2851 2852 2853 2854 2855 2856 2857 2858 2859 2860 2861 2862 2863 2864 2865 2866 2867 2868 2869 2870 2871 2872 2873 2874 2875 2876 2877 2878 2879 2880 2881 2882 2883 2884 2885 2886 2887 2888 2889 2890 2891 2892 2893 2894 2895 2896 2897 2898 2899 2900 2901 2902 2903 2904 2905 2906 2907 2908 2909 2910 2911 2912 2913 2914 2915 2916 2917 2918 2919 2920 2921 2922 2923 2924 2925 2926 2927 2928 2929 2930 2931 2932 2933 2934 2935 2936 2937 2938 2939 2940 2941 2942 2943 2944 2945 2946 2947 2948 2949 2950 2951 2952 2953 2954 2955 2956 2957 2958 2959 2960 2961 2962 2963 2964 2965 2966 2967 2968 2969 2970 2971 2972 2973 2974 2975 2976 2977 2978 2979 2980 2981 2982 2983 2984 2985 2986 2987 2988 2989 2990 2991 2992 2993 2994 2995 2996 2997 2998 2999 3000 3001 3002 3003 3004 3005 3006 3007 3008 3009 3010 3011 3012 3013 3014 3015 3016 3017 3018 3019 3020 3021 3022 3023 3024 3025 3026 3027 3028 3029 3030 3031 3032 3033 3034 3035 3036 3037 3038 3039 3040 3041 3042 3043 3044 3045 3046 3047 3048 3049 3050 3051 3052 3053 3054 3055 3056 3057 3058 3059 3060 3061 3062 3063 3064 3065 3066 3067 3068 3069 3070 3071 3072 3073 3074 3075 3076 3077 3078 3079 3080 3081 3082 3083 3084 3085 3086 3087 3088 3089 3090 3091 3092 3093 3094 3095 3096 3097 3098 3099 3100 3101 3102 3103 3104 3105 3106 3107 3108 3109 3110 3111 3112 3113 3114 3115 3116 3117 3118 3119 3120 3121 3122 3123 3124 3125 3126 3127 3128 3129 3130 3131 3132 3133 3134 3135 3136 3137 3138 3139 3140 3141 3142 3143 3144 3145 3146 3147 3148 3149 3150 3151 3152 3153 3154 3155 3156 3157 3158 3159 3160 3161 3162 3163 3164 3165 3166 3167 3168 3169 3170 3171 3172 3173 3174 3175 3176 3177 3178 3179 3180 3181 3182 3183 3184 3185 3186 3187 3188 3189 3190 3191 3192 3193 3194 3195 3196 3197 3198 3199 3200 3201 3202 3203 3204 3205 3206 3207 3208 3209 3210 3211 3212 3213 3214 3215 3216 3217 3218 3219 3220 3221 3222 3223 3224 3225 3226 3227 3228 3229 3230 3231 3232 3233 3234 3235 3236 3237 3238 3239 3240 3241 3242 3243 3244 3245 3246 3247 3248 3249 3250 3251 3252 3253 3254 3255 3256 3257 3258 3259 3260 3261 3262 3263 3264 3265 3266 3267 3268 3269 3270 3271 3272 3273 3274 3275 3276 3277 3278 3279 3280 3281 3282 3283 3284 3285 3286 3287 3288 3289 3290 3291 3292 3293 3294 3295 3296 3297 3298 3299 3300 3301 3302 3303 3304 3305 3306 3307 3308 3309 3310 3311 3312 3313 3314 3315 3316 3317 3318 3319 3320 3321 3322 3323 3324 3325 3326 3327 3328 3329 3330 3331 3332 3333 3334 3335 3336 3337 3338 3339 3340 3341 3342 3343 3344 3345 3346 3347 3348 3349 3350 3351 3352 3353 3354 3355 3356 3357 3358 3359 3360 3361 3362 3363 3364 3365 3366 3367 3368 3369 3370 3371 3372 3373 3374 3375 3376 3377 3378 3379 3380 3381 3382 3383 3384 3385 3386 3387 3388 3389 3390 3391 3392 3393 3394 3395 3396 3397 3398 3399 3400 3401 3402 3403 3404 3405 3406 3407 3408 3409 3410 3411 3412 3413 3414 3415 3416 3417 3418 3419 3420 3421 3422 3423 3424 3425 3426 3427 3428 3429 3430 3431 3432 3433 3434 3435 3436 3437 3438 3439 3440 3441 3442 3443 3444 3445 3446 3447 3448 3449 3450 3451 3452 3453 3454 3455 3456 3457 3458 3459 3460 3461 3462 3463 3464 3465 3466 3467 3468 3469 3470 3471 3472 3473 3474 3475 3476 3477 3478 3479 3480 3481 3482 3483 3484 3485 3486 3487 3488 3489 3490 3491 3492 3493 3494 3495 3496 3497 3498 3499 3500 3501 3502 3503 3504 3505 3506 3507 3508 3509 3510 3511 3512 3513 3514 3515 3516 3517 3518 3519 3520 3521 3522 3523 3524 3525 3526 3527 3528 3529 3530 3531 3532 3533 3534 3535 3536 3537 3538 3539 3540 3541 3542 3543 3544 3545 3546 3547 3548 3549 3550 3551 3552 3553 3554 3555 3556 3557 3558 3559 3560 3561 3562 3563 3564 3565 3566 3567 3568 3569 3570 3571 3572 3573 3574 3575 3576 3577 3578 3579 3580 3581 3582 3583 3584 3585 3586 3587 3588 3589 3590 3591 3592 3593 3594 3595 3596 3597 3598 3599 3600 3601 3602 3603 3604 3605 3606 3607 3608 3609 3610 3611 3612 3613 3614 3615 3616 3617 3618 3619 3620 3621 3622 3623 3624 3625 3626 3627 3628 3629 3630 3631 3632 3633 3634 3635 3636 3637 3638 3639 3640 3641 3642 3643 3644 3645 3646 3647 3648 3649 3650 3651 3652 3653 3654 3655 3656 3657 3658 3659 3660 3661 3662 3663 3664 3665 3666 3667 3668 3669 3670 3671 3672 3673 3674 3675 3676 3677 3678 3679 3680 3681 3682 3683 3684 3685 3686 3687 3688 3689 3690 3691 3692 3693 3694 3695 3696 3697 3698 3699 3700 3701 3702 3703 3704 3705 3706 3707 3708 3709 3710 3711 3712 3713 3714 3715 3716 3717 3718 3719 3720 3721 3722 3723 3724 3725 3726 3727 3728 3729 3730 3731 3732 3733 3734 3735 3736 3737 3738 3739 3740 3741 3742 3743 3744 3745 3746 3747 3748 3749 3750 3751 3752 3753 3754 3755 3756 3757 3758 3759 3760 3761 3762 3763 3764 3765 3766 3767 3768 3769 3770 3771 3772 3773 3774 3775 3776 3777 3778 3779 3780 3781 3782 3783 3784 3785 3786 3787 3788 3789 3790 3791 3792 3793 3794 3795 3796 3797 3798 3799 3800 3801 3802 3803 3804 3805 3806 3807 3808 3809 3810 3811 3812 3813 3814 3815 3816 3817 3818 3819 3820 3821 3822 3823 3824 3825 3826 3827 3828 3829 3830 3831 3832 3833 3834 3835 3836 3837 3838 3839 3840 3841 3842 3843 3844 3845 3846 3847 3848 3849 3850 3851 3852 3853 3854 3855 3856 3857 3858 3859 3860 3861 3862 3863 3864 3865 3866 3867 3868 3869 3870 3871 3872 3873 3874 3875 3876 3877 3878 3879 3880 3881 3882 3883 3884 3885 3886 3887 3888 3889 3890 3891 3892 3893 3894 3895 3896 3897 3898 3899 3900 3901 3902 3903 3904 3905 3906 3907 3908 3909 3910 3911 3912 3913 3914 3915 3916 3917 3918 3919 3920 3921 3922 3923 3924 3925 3926 3927 3928 3929 3930 3931 3932 3933 3934 3935 3936 3937 3938 3939 3940 3941 3942 3943 3944 3945 3946 3947 3948 3949 3950 3951 3952 3953 3954 3955 3956 3957 3958 3959 3960 3961 3962 3963 3964 3965 3966 3967 3968 3969 3970 3971 3972 3973 3974 3975 3976 3977 3978 3979 3980 3981 3982 3983 3984 3985 3986 3987 3988 3989 3990 3991 3992 3993 3994 3995 3996 3997 3998 3999 4000 4001 4002 4003 4004 4005 4006 4007 4008 4009 4010 4011 4012 4013 4014 4015 4016 4017 4018 4019 4020 4021 4022 4023 4024 4025 4026 4027 4028 4029 4030 4031 4032 4033 4034 4035 4036 4037 4038 4039 4040 4041 4042 4043 4044 4045 4046 4047 4048 4049 4050 4051 4052 4053 4054 4055 4056 4057 4058 4059 4060 4061 4062 4063 4064 4065 4066 4067 4068 4069 4070 4071 4072 4073 4074 4075 4076 4077 4078 4079 4080 4081 4082 4083 4084 4085 4086 4087 4088 4089 4090 4091 4092 4093 4094 4095 4096 4097 4098 4099 4100 4101 4102 4103 4104 4105 4106 4107 4108 4109 4110 4111 4112 4113 4114 4115 4116 4117 4118 4119 4120 4121 4122 4123 4124 4125 4126 4127 4128 4129 4130 4131 4132 4133 4134 4135 4136 4137 4138 4139 4140 4141 4142 4143 4144 4145 4146 4147 4148 4149 4150 4151 4152 4153 4154 4155 4156 4157 4158 4159 4160 4161 4162 4163 4164 4165 4166 4167 4168 4169 4170 4171 4172 4173 4174 4175 4176 4177 4178 4179 4180 4181 4182 4183 4184 4185 4186 4187 4188 4189 4190 4191 4192 4193 4194 4195 4196 4197 4198 4199 4200 4201 4202 4203 4204 4205 4206 4207 4208 4209 4210 4211 4212 4213 4214 4215 4216 4217 4218 4219 4220 4221 4222 4223 4224 4225 4226 4227 4228 4229 4230 4231 4232 4233 4234 4235 4236 4237 4238 4239 4240 4241 4242 4243 4244 4245 4246 4247 4248 4249 4250 4251 4252 4253 4254 4255 4256 4257 4258 4259 4260 4261 4262 4263 4264 4265 4266 4267 4268 4269 4270 4271 4272 4273 4274 4275 4276 4277 4278 4279 4280 4281 4282 4283 4284 4285 4286 4287 4288 4289 4290 4291 4292 4293 4294 4295 4296 4297 4298 4299 4300 4301 4302 4303 4304 4305 4306 4307 4308 4309 4310 4311 4312 4313 4314 4315 4316 4317 4318 4319 4320 4321 4322 4323 4324 4325 4326 4327 4328 4329 4330 4331 4332 4333 4334 4335 4336 4337 4338 4339 4340 4341 4342 4343 4344 4345 4346 4347 4348 4349 4350 4351 4352 4353 4354 4355 4356 4357 4358 4359 4360 4361 4362 4363 4364 4365 4366 4367 4368 4369 4370 4371 4372 4373 4374 4375 4376 4377 4378 4379 4380 4381 4382 4383 4384 4385 4386 4387 4388 4389 4390 4391 4392 4393 4394 4395 4396 4397 4398 4399 4400 4401 4402 4403 4404 4405 4406 4407 4408 4409 4410 4411 4412 4413 4414 4415 4416 4417 4418 4419 4420 4421 4422 4423 4424 4425 4426 4427 4428 4429 4430 4431 4432 4433 4434 4435 4436 4437 4438 4439 4440 4441 4442 4443 4444 4445 4446 4447 4448 4449 4450 4451 4452 4453 4454 4455 4456 4457 4458 4459 4460 4461 4462 4463 4464 4465 4466 4467 4468 4469 4470 4471 4472 4473 4474 4475 4476 4477 4478 4479 4480 4481 4482 4483 4484 4485 4486 4487 4488 4489 4490 4491 4492 4493 4494 4495 4496 4497 4498 4499 4500 4501 4502 4503 4504 4505 4506 4507 4508 4509 4510 4511 4512 4513 4514 4515 4516 4517 4518 4519 4520 4521 4522 4523 4524 4525 4526 4527 4528 4529 4530 4531 4532 4533 4534 4535 4536 4537 4538 4539 4540 4541 4542 4543 4544 4545 4546 4547 4548 4549 4550 4551 4552 4553 4554 4555 4556 4557 4558 4559 4560 4561 4562 4563 4564 4565 4566 4567 4568 4569 4570 4571 4572 4573 4574 4575 4576 4577 4578 4579 4580 4581 4582 4583 4584 4585 4586 4587 4588 4589 4590 4591 4592 4593 4594 4595 4596 4597 4598 4599 4600 4601 4602 4603 4604 4605 4606 4607 4608 4609 4610 4611 4612 4613 4614 4615 4616 4617 4618 4619 4620 4621 4622 4623 4624 4625 4626 4627 4628 4629 4630 4631 4632 4633 4634 4635 4636 4637 4638 4639 4640 4641 4642 4643 4644 4645 4646 4647 4648 4649 4650 4651 4652 4653 4654 4655 4656 4657 4658 4659 4660 4661 4662 4663 4664 4665 4666 4667 4668 4669 4670 4671 4672 4673 4674 4675 4676 4677 4678 4679 4680 4681 4682 4683 4684 4685 4686 4687 4688 4689 4690 4691 4692 4693 4694 4695 4696 4697 4698 4699 4700 4701 4702 4703 4704 4705 4706 4707 4708 4709 4710 4711 4712 4713 4714 4715 4716 4717 4718 4719 4720 4721 4722 4723 4724 4725 4726 4727 4728 4729 4730 4731 4732 4733 4734 4735 4736 4737 4738 4739 4740 4741 4742 4743 4744 4745 4746 4747 4748 4749 4750 4751 4752 4753 4754 4755 4756 4757 4758 4759 4760 4761 4762 4763 4764 4765 4766 4767 4768 4769 4770 4771 4772 4773 4774 4775 4776 4777 4778 4779 4780 4781 4782 4783 4784 4785 4786 4787 4788 4789 4790 4791 4792 4793 4794 4795 4796 4797 4798 4799 4800 4801 4802 4803 4804 4805 4806 4807 4808 4809 4810 4811 4812 4813 4814 4815 4816 4817 4818 4819 4820 4821 4822 4823 4824 4825 4826 4827 4828 4829 4830 4831 4832 4833 4834 4835 4836 4837 4838 4839 4840 4841 4842 4843 4844 4845 4846 4847 4848 4849 4850 4851 4852 4853 4854 4855 4856 4857 4858 4859 4860 4861 4862 4863 4864 4865 4866 4867 4868 4869 4870 4871 4872 4873 4874 4875 4876 4877 4878 4879 4880 4881 4882 4883 4884 4885 4886 4887 4888 4889 4890 4891 4892 4893 4894 4895 4896 4897 4898 4899 4900 4901 4902 4903 4904 4905 4906 4907 4908 4909 4910 4911 4912 4913 4914 4915 4916 4917 4918 4919 4920 4921 4922 4923 4924 4925 4926 4927 4928 4929 4930 4931 4932 4933 4934 4935 4936 4937 4938 4939 4940 4941 4942 4943 4944 4945 4946 4947 4948 4949 4950 4951 4952 4953 4954 4955 4956 4957 4958 4959 4960 4961 4962 4963 4964 4965 4966 4967 4968 4969 4970 4971 4972 4973 4974 4975 4976 4977 4978 4979 4980 4981 4982 4983 4984 4985 4986 4987 4988 4989 4990 4991 4992 4993 4994 4995 4996 4997 4998 4999 5000 5001 5002 5003 5004 5005 5006 5007 5008 5009 5010 5011 5012 5013 5014 5015 5016 5017 5018 5019 5020 5021 5022 5023 5024 5025 5026 5027 5028 5029 5030 5031 5032 5033 5034 5035 5036 5037 5038 5039 5040 5041 5042 5043 5044 5045 5046 5047 5048 5049 5050 5051 5052 5053 5054 5055 5056 5057 5058 5059 5060 5061 5062 5063 5064 5065 5066 5067 5068 5069 5070 5071 5072 5073 5074 5075 5076 5077 5078 5079 5080 5081 5082 5083 5084 5085 5086 5087 5088 5089 5090 5091 5092 5093 5094 5095 5096 5097 5098 5099 5100 5101 5102 5103 5104 5105 5106 5107 5108 5109 5110 5111 5112 5113 5114 5115 5116 5117 5118 5119 5120 5121 5122 5123 5124 5125 5126 5127 5128 5129 5130 5131 5132 5133 5134 5135 5136 5137 5138 5139 5140 5141 5142 5143 5144 5145 5146 5147 5148 5149 5150 5151 5152 5153 5154 5155 5156 5157 5158 5159 5160 5161 5162 5163 5164 5165 5166 5167 5168 5169 5170 5171 5172 5173 5174 5175 5176 5177 5178 5179 5180 5181 5182 5183 5184 5185 5186 5187 5188 5189 5190 5191 5192 5193 5194 5195 5196 5197 5198 5199 5200 5201 5202 5203 5204 5205 5206 5207 5208 5209 5210 5211 5212 5213 5214 5215 5216 5217 5218 5219 5220 5221 5222 5223 5224 5225 5226 5227 5228 5229 5230 5231 5232 5233 5234 5235 5236 5237 5238 5239 5240 5241 5242 5243 5244 5245 5246 5247 5248 5249 5250 5251 5252 5253 5254 5255 5256 5257 5258 5259 5260 5261 5262 5263 5264 5265 5266 5267 5268 5269 5270 5271 5272 5273 5274 5275 5276 5277 5278 5279 5280 5281 5282 5283 5284 5285 5286 5287 5288 5289 5290 5291 5292 5293 5294 5295 5296 5297 5298 5299 5300 5301 5302 5303 5304 5305 5306 5307 5308 5309 5310 5311 5312 5313 5314 5315 5316 5317 5318 5319 5320 5321 5322 5323 5324 5325 5326 5327 5328 5329 5330 5331 5332 5333 5334 5335 5336 5337 5338 5339 5340 5341 5342 5343 5344 5345 5346 5347 5348 5349 5350 5351 5352 5353 5354 5355 5356 5357 5358 5359 5360 5361 5362 5363 5364 5365 5366 5367 5368 5369 5370 5371 5372 5373 5374 5375 5376 5377 5378 5379 5380 5381 5382 5383 5384 5385 5386 5387 5388 5389 5390 5391 5392 5393 5394 5395 5396 5397 5398 5399 5400 5401 5402 5403 5404 5405 5406 5407 5408 5409 5410 5411 5412 5413 5414 5415 5416 5417 5418 5419 5420 5421 5422 5423 5424 5425 5426 5427 5428 5429 5430 5431 5432 5433 5434 5435 5436 5437 5438 5439 5440 5441 5442 5443 5444 5445 5446 5447 5448 5449 5450 5451 5452 5453 5454 5455 5456 5457 5458 5459 5460 5461 5462 5463 5464 5465 5466 5467 5468 5469 5470 5471 5472 5473 5474 5475 5476 5477 5478 5479 5480 5481 5482 5483 5484 5485 5486 5487 5488 5489 5490 5491 5492 5493 5494 5495 5496 5497 5498 5499 5500 5501 5502 5503 5504 5505 5506 5507 5508 5509 5510 5511 5512 5513 5514 5515 5516 5517 5518 5519 5520 5521 5522 5523 5524 5525 5526 5527 5528 5529 5530 5531 5532 5533 5534 5535 5536 5537 5538 5539 5540 5541 5542 5543 5544 5545 5546 5547 5548 5549 5550 5551 5552 5553 5554 5555 5556 5557 5558 5559 5560 5561 5562 5563 5564 5565 5566 5567 5568 5569 5570 5571 5572 5573 5574 5575 5576 5577 5578 5579 5580 5581 5582 5583 5584 5585 5586 5587 5588 5589 5590 5591 5592 5593 5594 5595 5596 5597 5598 5599 5600 5601 5602 5603 5604 5605 5606 5607 5608 5609 5610 5611 5612 5613 5614 5615 5616 5617 5618 5619 5620 5621 5622 5623 5624 5625 5626 5627 5628 5629 5630 5631 5632 5633 5634 5635 5636 5637 5638 5639 5640 5641 5642 5643 5644 5645 5646 5647 5648 5649 5650 5651 5652 5653 5654 5655 5656 5657 5658 5659 5660 5661 5662 5663 5664 5665 5666 5667 5668 5669 5670 5671 5672 5673 5674 5675 5676 5677 5678 5679 5680 5681 5682 5683 5684 5685 5686 5687 5688 5689 5690 5691 5692 5693 5694 5695 5696 5697 5698 5699 5700 5701 5702 5703 5704 5705 5706 5707 5708 5709 5710 5711 5712 5713 5714 5715 5716 5717 5718 5719 5720 5721 5722 5723 5724 5725 5726 5727 5728 5729 5730 5731 5732 5733 5734 5735 5736 5737 5738 5739 5740 5741 5742 5743 5744 5745 5746 5747 5748 5749 5750 5751 5752 5753 5754 5755 5756 5757 5758 5759 5760 5761 5762 5763 5764 5765 5766 5767 5768 5769 5770 5771 5772 5773 5774 5775 5776 5777 5778 5779 5780 5781 5782 5783 5784 5785 5786 5787 5788 5789 5790 5791 5792 5793 5794 5795 5796 5797 5798 5799 5800 5801 5802 5803 5804 5805 5806 5807 5808 5809 5810 5811 5812 5813 5814 5815 5816 5817 5818 5819 5820 5821 5822 5823 5824 5825 5826 5827 5828 5829 5830 5831 5832 5833 5834 5835 5836 5837 5838 5839 5840 5841 5842 5843 5844 5845 5846 5847 5848 5849 5850 5851 5852 5853 5854 5855 5856 5857 5858 5859 5860 5861 5862 5863 5864 5865 5866 5867 5868 5869 5870 5871 5872 5873 5874 5875 5876 5877 5878 5879 5880 5881 5882 5883 5884 5885 5886 5887 5888 5889 5890 5891 5892 5893 5894 5895 5896 5897 5898 5899 5900 5901 5902 5903 5904 5905 5906 5907 5908 5909 5910 5911 5912 5913 5914 5915 5916 5917 5918 5919 5920 5921 5922 5923 5924 5925 5926 5927 5928 5929 5930 5931 5932 5933 5934 5935 5936 5937 5938 5939 5940 5941 5942 5943 5944 5945 5946 5947 5948 5949 5950 5951 5952 5953 5954 5955 5956 5957 5958 5959 5960 5961 5962 5963 5964 5965 5966 5967 5968 5969 5970 5971 5972 5973 5974 5975 5976 5977 5978 5979 5980 5981 5982 5983 5984 5985 5986 5987 5988 5989 5990 5991 5992 5993 5994 5995 5996 5997 5998 5999 6000 6001 6002 6003 6004 6005 6006 6007 6008 6009 6010 6011 6012 6013 6014 6015 6016 6017 6018 6019 6020 6021 6022 6023 6024 6025 6026 6027 6028 6029 6030 6031 6032 6033 6034 6035 6036 6037 6038 6039 6040 6041 6042 6043 6044 6045 6046 6047 6048 6049 6050 6051 6052 6053 6054 6055 6056 6057 6058 6059 6060 6061 6062 6063 6064 6065 6066 6067 6068 6069 6070 6071 6072 6073 6074 6075 6076 6077 6078 6079 6080 6081 6082 6083 6084 6085 6086 6087 6088 6089 6090 6091 6092 6093 6094 6095 6096 6097 6098 6099 6100 6101 6102 6103 6104 6105 6106 6107 6108 6109 6110 6111 6112 6113 6114 6115 6116 6117 6118 6119 6120 6121 6122 6123 6124 6125 6126 6127 6128 6129 6130 6131 6132 6133 6134 6135 6136 6137 6138 6139 6140 6141 6142 6143 6144 6145 6146 6147 6148 6149 6150 6151 6152 6153 6154 6155 6156 6157 6158 6159 6160 6161 6162 6163 6164 6165 6166 6167 6168 6169 6170 6171 6172 6173 6174 6175 6176 6177 6178 6179 6180 6181 6182 6183 6184 6185 6186 6187 6188 6189 6190 6191 6192 6193 6194 6195 6196 6197 6198 6199 6200 6201 6202 6203 6204 6205 6206 6207 6208 6209 6210 6211 6212 6213 6214 6215 6216 6217 6218 6219 6220 6221 6222 6223 6224 6225 6226 6227 6228 6229 6230 6231 6232 6233 6234 6235 6236 6237 6238 6239 6240 6241 6242 6243 6244 6245 6246 6247 6248 6249 6250 6251 6252 6253 6254 6255 6256 6257 6258 6259 6260 6261 6262 6263 6264 6265 6266 6267 6268 6269 6270 6271 6272 6273 6274 6275 6276 6277 6278 6279 6280 6281 6282 6283 6284 6285 6286 6287 6288 6289 6290 6291 6292 6293 6294 6295 6296 6297 6298 6299 6300 6301 6302 6303 6304 6305 6306 6307 6308 6309 6310 6311 6312 6313 6314 6315 6316 6317 6318 6319 6320 6321 6322 6323 6324 6325 6326 6327 6328 6329 6330 6331 6332 6333 6334 6335 6336 6337 6338 6339 6340 6341 6342 6343 6344 6345 6346 6347 6348 6349 6350 6351 6352 6353 6354 6355 6356 6357 6358 6359 6360 6361 6362 6363 6364 6365 6366 6367 6368 6369 6370 6371 6372 6373 6374 6375 6376 6377 6378 6379 6380 6381 6382 6383 6384 6385 6386 6387 6388 6389 6390 6391 6392 6393 6394 6395 6396 6397 6398 6399 6400 6401 6402 6403 6404 6405 6406 6407 6408 6409 6410 6411 6412 6413 6414 6415 6416 6417 6418 6419 6420 6421 6422 6423 6424 6425 6426 6427 6428 6429 6430 6431 6432 6433 6434 6435 6436 6437 6438 6439 6440 6441 6442 6443 6444 6445 6446 6447 6448 6449 6450 6451 6452 6453 6454 6455 6456 6457 6458 6459 6460 6461 6462 6463 6464 6465 6466 6467 6468 6469 6470 6471 6472 6473 6474 6475 6476 6477 6478 6479 6480 6481 6482 6483 6484 6485 6486 6487 6488 6489 6490 6491 6492 6493 6494 6495 6496 6497 6498 6499 6500 6501 6502 6503 6504 6505 6506 6507 6508 6509 6510 6511 6512 6513 6514 6515 6516 6517 6518 6519 6520 6521 6522 6523 6524 6525 6526 6527 6528 6529 6530 6531 6532 6533 6534 6535 6536 6537 6538 6539 6540 6541 6542 6543 6544 6545 6546 6547 6548 6549 6550 6551 6552 6553 6554 6555 6556 6557 6558 6559 6560 6561 6562 6563 6564 6565 6566 6567 6568 6569 6570 6571 6572 6573 6574 6575 6576 6577 6578 6579 6580 6581 6582 6583 6584 6585 6586 6587 6588 6589 6590 6591 6592 6593 6594 6595 6596 6597 6598 6599 6600 6601 6602 6603 6604 6605 6606 6607 6608 6609 6610 6611 6612 6613 6614 6615 6616 6617 6618 6619 6620 6621 6622 6623 6624 6625 6626 6627 6628 6629 6630 6631 6632 6633 6634 6635 6636 6637 6638 6639 6640 6641 6642 6643 6644 6645 6646 6647 6648 6649 6650 6651 6652 6653 6654 6655 6656 6657 6658 6659 6660 6661 6662 6663 6664 6665 6666 6667 6668 6669 6670 6671 6672 6673 6674 6675 6676 6677 6678 6679 6680 6681 6682 6683 6684 6685 6686 6687 6688 6689 6690 6691 6692 6693 6694 6695 6696 6697 6698 6699 6700 6701 6702 6703 6704 6705 6706 6707 6708 6709 6710 6711 6712 6713 6714 6715 6716 6717 6718 6719 6720 6721 6722 6723 6724 6725 6726 6727 6728 6729 6730 6731 6732 6733 6734 6735 6736 6737 6738 6739 6740 6741 6742 6743 6744 6745 6746 6747 6748 6749 6750 6751 6752 6753 6754 6755 6756 6757 6758 6759 6760 6761 6762 6763 6764 6765 6766 6767 6768 6769 6770 6771 6772 6773 6774 6775 6776 6777 6778 6779 6780 6781 6782 6783 6784 6785 6786 6787 6788 6789 6790 6791 6792 6793 6794 6795 6796 6797 6798 6799 6800 6801 6802 6803 6804 6805 6806 6807 6808 6809 6810 6811 6812 6813 6814 6815 6816 6817 6818 6819 6820 6821 6822 6823 6824 6825 6826 6827 6828 6829 6830 6831 6832 6833 6834 6835 6836 6837 6838 6839 6840 6841 6842 6843 6844 6845 6846 6847 6848 6849 6850 6851 6852 6853 6854 6855 6856 6857 6858 6859 6860 6861 6862 6863 6864 6865 6866 6867 6868 6869 6870 6871 6872 6873 6874 6875 6876 6877 6878 6879 6880 6881 6882 6883 6884 6885 6886 6887 6888 6889 6890 6891 6892 6893 6894 6895 6896 6897 6898 6899 6900 6901 6902 6903 6904 6905 6906 6907 6908 6909 6910 6911 6912 6913 6914 6915 6916 6917 6918 6919 6920 6921 6922 6923 6924 6925 6926 6927 6928 6929 6930 6931 6932 6933 6934 6935 6936 6937 6938 6939 6940 6941 6942 6943 6944 6945 6946 6947 6948 6949 6950 6951 6952 6953 6954 6955 6956 6957 6958 6959 6960 6961 6962 6963 6964 6965 6966 6967 6968 6969 6970 6971 6972 6973 6974 6975 6976 6977 6978 6979 6980 6981 6982 6983 6984 6985 6986 6987 6988 6989 6990 6991 6992 6993 6994 6995 6996 6997 6998 6999 7000 7001 7002 7003 7004 7005 7006 7007 7008 7009 7010 7011 7012 7013 7014 7015 7016 7017 7018 7019 7020 7021 7022 7023 7024 7025 7026 7027 7028 7029 7030 7031 7032 7033 7034 7035 7036 7037 7038 7039 7040 7041 7042 7043 7044 7045 7046 7047 7048 7049 7050 7051 7052 7053 7054 7055 7056 7057 7058 7059 7060 7061 7062 7063 7064 7065 7066 7067 7068 7069 7070 7071 7072 7073 7074 7075 7076 7077 7078 7079 7080 7081 7082 7083 7084 7085 7086 7087 7088 7089 7090 7091 7092 7093 7094 7095 7096 7097 7098 7099 7100 7101 7102 7103 7104 7105 7106 7107 7108 7109 7110 7111 7112 7113 7114 7115 7116 7117 7118 7119 7120 7121 7122 7123 7124 7125 7126 7127 7128 7129 7130 7131 7132 7133 7134 7135 7136 7137 7138 7139 7140 7141 7142 7143 7144 7145 7146 7147 7148 7149 7150 7151 7152 7153 7154 7155 7156 7157 7158 7159 7160 7161 7162 7163 7164 7165 7166 7167 7168 7169 7170 7171 7172 7173 7174 7175 7176 7177 7178 7179 7180
/*
 * Copyright (C) 2007 Oracle.  All rights reserved.
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public
 * License v2 as published by the Free Software Foundation.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * General Public License for more details.
 *
 * You should have received a copy of the GNU General Public
 * License along with this program; if not, write to the
 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
 * Boston, MA 021110-1307, USA.
 */
#include <linux/sched.h>
#include <linux/pagemap.h>
#include <linux/writeback.h>
#include <linux/blkdev.h>
#include <linux/sort.h>
#include <linux/rcupdate.h>
#include "compat.h"
#include "hash.h"
#include "ctree.h"
#include "disk-io.h"
#include "print-tree.h"
#include "transaction.h"
#include "volumes.h"
#include "locking.h"
#include "free-space-cache.h"

static int update_reserved_extents(struct btrfs_root *root,
				   u64 bytenr, u64 num, int reserve);
static int update_block_group(struct btrfs_trans_handle *trans,
			      struct btrfs_root *root,
			      u64 bytenr, u64 num_bytes, int alloc,
			      int mark_free);
static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
				struct btrfs_root *root,
				u64 bytenr, u64 num_bytes, u64 parent,
				u64 root_objectid, u64 owner_objectid,
				u64 owner_offset, int refs_to_drop,
				struct btrfs_delayed_extent_op *extra_op);
static void __run_delayed_extent_op(struct btrfs_delayed_extent_op *extent_op,
				    struct extent_buffer *leaf,
				    struct btrfs_extent_item *ei);
static int alloc_reserved_file_extent(struct btrfs_trans_handle *trans,
				      struct btrfs_root *root,
				      u64 parent, u64 root_objectid,
				      u64 flags, u64 owner, u64 offset,
				      struct btrfs_key *ins, int ref_mod);
static int alloc_reserved_tree_block(struct btrfs_trans_handle *trans,
				     struct btrfs_root *root,
				     u64 parent, u64 root_objectid,
				     u64 flags, struct btrfs_disk_key *key,
				     int level, struct btrfs_key *ins);

static int do_chunk_alloc(struct btrfs_trans_handle *trans,
			  struct btrfs_root *extent_root, u64 alloc_bytes,
			  u64 flags, int force);

static int block_group_bits(struct btrfs_block_group_cache *cache, u64 bits)
{
	return (cache->flags & bits) == bits;
}

/*
 * this adds the block group to the fs_info rb tree for the block group
 * cache
 */
static int btrfs_add_block_group_cache(struct btrfs_fs_info *info,
				struct btrfs_block_group_cache *block_group)
{
	struct rb_node **p;
	struct rb_node *parent = NULL;
	struct btrfs_block_group_cache *cache;

	spin_lock(&info->block_group_cache_lock);
	p = &info->block_group_cache_tree.rb_node;

	while (*p) {
		parent = *p;
		cache = rb_entry(parent, struct btrfs_block_group_cache,
				 cache_node);
		if (block_group->key.objectid < cache->key.objectid) {
			p = &(*p)->rb_left;
		} else if (block_group->key.objectid > cache->key.objectid) {
			p = &(*p)->rb_right;
		} else {
			spin_unlock(&info->block_group_cache_lock);
			return -EEXIST;
		}
	}

	rb_link_node(&block_group->cache_node, parent, p);
	rb_insert_color(&block_group->cache_node,
			&info->block_group_cache_tree);
	spin_unlock(&info->block_group_cache_lock);

	return 0;
}

/*
 * This will return the block group at or after bytenr if contains is 0, else
 * it will return the block group that contains the bytenr
 */
static struct btrfs_block_group_cache *
block_group_cache_tree_search(struct btrfs_fs_info *info, u64 bytenr,
			      int contains)
{
	struct btrfs_block_group_cache *cache, *ret = NULL;
	struct rb_node *n;
	u64 end, start;

	spin_lock(&info->block_group_cache_lock);
	n = info->block_group_cache_tree.rb_node;

	while (n) {
		cache = rb_entry(n, struct btrfs_block_group_cache,
				 cache_node);
		end = cache->key.objectid + cache->key.offset - 1;
		start = cache->key.objectid;

		if (bytenr < start) {
			if (!contains && (!ret || start < ret->key.objectid))
				ret = cache;
			n = n->rb_left;
		} else if (bytenr > start) {
			if (contains && bytenr <= end) {
				ret = cache;
				break;
			}
			n = n->rb_right;
		} else {
			ret = cache;
			break;
		}
	}
	if (ret)
		atomic_inc(&ret->count);
	spin_unlock(&info->block_group_cache_lock);

	return ret;
}

/*
 * this is only called by cache_block_group, since we could have freed extents
 * we need to check the pinned_extents for any extents that can't be used yet
 * since their free space will be released as soon as the transaction commits.
 */
static int add_new_free_space(struct btrfs_block_group_cache *block_group,
			      struct btrfs_fs_info *info, u64 start, u64 end)
{
	u64 extent_start, extent_end, size;
	int ret;

	while (start < end) {
		ret = find_first_extent_bit(&info->pinned_extents, start,
					    &extent_start, &extent_end,
					    EXTENT_DIRTY);
		if (ret)
			break;

		if (extent_start == start) {
			start = extent_end + 1;
		} else if (extent_start > start && extent_start < end) {
			size = extent_start - start;
			ret = btrfs_add_free_space(block_group, start,
						   size);
			BUG_ON(ret);
			start = extent_end + 1;
		} else {
			break;
		}
	}

	if (start < end) {
		size = end - start;
		ret = btrfs_add_free_space(block_group, start, size);
		BUG_ON(ret);
	}

	return 0;
}

static int remove_sb_from_cache(struct btrfs_root *root,
				struct btrfs_block_group_cache *cache)
{
	u64 bytenr;
	u64 *logical;
	int stripe_len;
	int i, nr, ret;

	for (i = 0; i < BTRFS_SUPER_MIRROR_MAX; i++) {
		bytenr = btrfs_sb_offset(i);
		ret = btrfs_rmap_block(&root->fs_info->mapping_tree,
				       cache->key.objectid, bytenr, 0,
				       &logical, &nr, &stripe_len);
		BUG_ON(ret);
		while (nr--) {
			btrfs_remove_free_space(cache, logical[nr],
						stripe_len);
		}
		kfree(logical);
	}
	return 0;
}

static int cache_block_group(struct btrfs_root *root,
			     struct btrfs_block_group_cache *block_group)
{
	struct btrfs_path *path;
	int ret = 0;
	struct btrfs_key key;
	struct extent_buffer *leaf;
	int slot;
	u64 last;

	if (!block_group)
		return 0;

	root = root->fs_info->extent_root;

	if (block_group->cached)
		return 0;

	path = btrfs_alloc_path();
	if (!path)
		return -ENOMEM;

	path->reada = 2;
	/*
	 * we get into deadlocks with paths held by callers of this function.
	 * since the alloc_mutex is protecting things right now, just
	 * skip the locking here
	 */
	path->skip_locking = 1;
	last = max_t(u64, block_group->key.objectid, BTRFS_SUPER_INFO_OFFSET);
	key.objectid = last;
	key.offset = 0;
	btrfs_set_key_type(&key, BTRFS_EXTENT_ITEM_KEY);
	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
	if (ret < 0)
		goto err;

	while (1) {
		leaf = path->nodes[0];
		slot = path->slots[0];
		if (slot >= btrfs_header_nritems(leaf)) {
			ret = btrfs_next_leaf(root, path);
			if (ret < 0)
				goto err;
			if (ret == 0)
				continue;
			else
				break;
		}
		btrfs_item_key_to_cpu(leaf, &key, slot);
		if (key.objectid < block_group->key.objectid)
			goto next;

		if (key.objectid >= block_group->key.objectid +
		    block_group->key.offset)
			break;

		if (btrfs_key_type(&key) == BTRFS_EXTENT_ITEM_KEY) {
			add_new_free_space(block_group, root->fs_info, last,
					   key.objectid);

			last = key.objectid + key.offset;
		}
next:
		path->slots[0]++;
	}

	add_new_free_space(block_group, root->fs_info, last,
			   block_group->key.objectid +
			   block_group->key.offset);

	block_group->cached = 1;
	remove_sb_from_cache(root, block_group);
	ret = 0;
err:
	btrfs_free_path(path);
	return ret;
}

/*
 * return the block group that starts at or after bytenr
 */
static struct btrfs_block_group_cache *
btrfs_lookup_first_block_group(struct btrfs_fs_info *info, u64 bytenr)
{
	struct btrfs_block_group_cache *cache;

	cache = block_group_cache_tree_search(info, bytenr, 0);

	return cache;
}

/*
 * return the block group that contains the given bytenr
 */
struct btrfs_block_group_cache *btrfs_lookup_block_group(
						 struct btrfs_fs_info *info,
						 u64 bytenr)
{
	struct btrfs_block_group_cache *cache;

	cache = block_group_cache_tree_search(info, bytenr, 1);

	return cache;
}

void btrfs_put_block_group(struct btrfs_block_group_cache *cache)
{
	if (atomic_dec_and_test(&cache->count))
		kfree(cache);
}

static struct btrfs_space_info *__find_space_info(struct btrfs_fs_info *info,
						  u64 flags)
{
	struct list_head *head = &info->space_info;
	struct btrfs_space_info *found;

	rcu_read_lock();
	list_for_each_entry_rcu(found, head, list) {
		if (found->flags == flags) {
			rcu_read_unlock();
			return found;
		}
	}
	rcu_read_unlock();
	return NULL;
}

/*
 * after adding space to the filesystem, we need to clear the full flags
 * on all the space infos.
 */
void btrfs_clear_space_info_full(struct btrfs_fs_info *info)
{
	struct list_head *head = &info->space_info;
	struct btrfs_space_info *found;

	rcu_read_lock();
	list_for_each_entry_rcu(found, head, list)
		found->full = 0;
	rcu_read_unlock();
}

static u64 div_factor(u64 num, int factor)
{
	if (factor == 10)
		return num;
	num *= factor;
	do_div(num, 10);
	return num;
}

u64 btrfs_find_block_group(struct btrfs_root *root,
			   u64 search_start, u64 search_hint, int owner)
{
	struct btrfs_block_group_cache *cache;
	u64 used;
	u64 last = max(search_hint, search_start);
	u64 group_start = 0;
	int full_search = 0;
	int factor = 9;
	int wrapped = 0;
again:
	while (1) {
		cache = btrfs_lookup_first_block_group(root->fs_info, last);
		if (!cache)
			break;

		spin_lock(&cache->lock);
		last = cache->key.objectid + cache->key.offset;
		used = btrfs_block_group_used(&cache->item);

		if ((full_search || !cache->ro) &&
		    block_group_bits(cache, BTRFS_BLOCK_GROUP_METADATA)) {
			if (used + cache->pinned + cache->reserved <
			    div_factor(cache->key.offset, factor)) {
				group_start = cache->key.objectid;
				spin_unlock(&cache->lock);
				btrfs_put_block_group(cache);
				goto found;
			}
		}
		spin_unlock(&cache->lock);
		btrfs_put_block_group(cache);
		cond_resched();
	}
	if (!wrapped) {
		last = search_start;
		wrapped = 1;
		goto again;
	}
	if (!full_search && factor < 10) {
		last = search_start;
		full_search = 1;
		factor = 10;
		goto again;
	}
found:
	return group_start;
}

/* simple helper to search for an existing extent at a given offset */
int btrfs_lookup_extent(struct btrfs_root *root, u64 start, u64 len)
{
	int ret;
	struct btrfs_key key;
	struct btrfs_path *path;

	path = btrfs_alloc_path();
	BUG_ON(!path);
	key.objectid = start;
	key.offset = len;
	btrfs_set_key_type(&key, BTRFS_EXTENT_ITEM_KEY);
	ret = btrfs_search_slot(NULL, root->fs_info->extent_root, &key, path,
				0, 0);
	btrfs_free_path(path);
	return ret;
}

/*
 * Back reference rules.  Back refs have three main goals:
 *
 * 1) differentiate between all holders of references to an extent so that
 *    when a reference is dropped we can make sure it was a valid reference
 *    before freeing the extent.
 *
 * 2) Provide enough information to quickly find the holders of an extent
 *    if we notice a given block is corrupted or bad.
 *
 * 3) Make it easy to migrate blocks for FS shrinking or storage pool
 *    maintenance.  This is actually the same as #2, but with a slightly
 *    different use case.
 *
 * There are two kinds of back refs. The implicit back refs is optimized
 * for pointers in non-shared tree blocks. For a given pointer in a block,
 * back refs of this kind provide information about the block's owner tree
 * and the pointer's key. These information allow us to find the block by
 * b-tree searching. The full back refs is for pointers in tree blocks not
 * referenced by their owner trees. The location of tree block is recorded
 * in the back refs. Actually the full back refs is generic, and can be
 * used in all cases the implicit back refs is used. The major shortcoming
 * of the full back refs is its overhead. Every time a tree block gets
 * COWed, we have to update back refs entry for all pointers in it.
 *
 * For a newly allocated tree block, we use implicit back refs for
 * pointers in it. This means most tree related operations only involve
 * implicit back refs. For a tree block created in old transaction, the
 * only way to drop a reference to it is COW it. So we can detect the
 * event that tree block loses its owner tree's reference and do the
 * back refs conversion.
 *
 * When a tree block is COW'd through a tree, there are four cases:
 *
 * The reference count of the block is one and the tree is the block's
 * owner tree. Nothing to do in this case.
 *
 * The reference count of the block is one and the tree is not the
 * block's owner tree. In this case, full back refs is used for pointers
 * in the block. Remove these full back refs, add implicit back refs for
 * every pointers in the new block.
 *
 * The reference count of the block is greater than one and the tree is
 * the block's owner tree. In this case, implicit back refs is used for
 * pointers in the block. Add full back refs for every pointers in the
 * block, increase lower level extents' reference counts. The original
 * implicit back refs are entailed to the new block.
 *
 * The reference count of the block is greater than one and the tree is
 * not the block's owner tree. Add implicit back refs for every pointer in
 * the new block, increase lower level extents' reference count.
 *
 * Back Reference Key composing:
 *
 * The key objectid corresponds to the first byte in the extent,
 * The key type is used to differentiate between types of back refs.
 * There are different meanings of the key offset for different types
 * of back refs.
 *
 * File extents can be referenced by:
 *
 * - multiple snapshots, subvolumes, or different generations in one subvol
 * - different files inside a single subvolume
 * - different offsets inside a file (bookend extents in file.c)
 *
 * The extent ref structure for the implicit back refs has fields for:
 *
 * - Objectid of the subvolume root
 * - objectid of the file holding the reference
 * - original offset in the file
 * - how many bookend extents
 *
 * The key offset for the implicit back refs is hash of the first
 * three fields.
 *
 * The extent ref structure for the full back refs has field for:
 *
 * - number of pointers in the tree leaf
 *
 * The key offset for the implicit back refs is the first byte of
 * the tree leaf
 *
 * When a file extent is allocated, The implicit back refs is used.
 * the fields are filled in:
 *
 *     (root_key.objectid, inode objectid, offset in file, 1)
 *
 * When a file extent is removed file truncation, we find the
 * corresponding implicit back refs and check the following fields:
 *
 *     (btrfs_header_owner(leaf), inode objectid, offset in file)
 *
 * Btree extents can be referenced by:
 *
 * - Different subvolumes
 *
 * Both the implicit back refs and the full back refs for tree blocks
 * only consist of key. The key offset for the implicit back refs is
 * objectid of block's owner tree. The key offset for the full back refs
 * is the first byte of parent block.
 *
 * When implicit back refs is used, information about the lowest key and
 * level of the tree block are required. These information are stored in
 * tree block info structure.
 */

#ifdef BTRFS_COMPAT_EXTENT_TREE_V0
static int convert_extent_item_v0(struct btrfs_trans_handle *trans,
				  struct btrfs_root *root,
				  struct btrfs_path *path,
				  u64 owner, u32 extra_size)
{
	struct btrfs_extent_item *item;
	struct btrfs_extent_item_v0 *ei0;
	struct btrfs_extent_ref_v0 *ref0;
	struct btrfs_tree_block_info *bi;
	struct extent_buffer *leaf;
	struct btrfs_key key;
	struct btrfs_key found_key;
	u32 new_size = sizeof(*item);
	u64 refs;
	int ret;

	leaf = path->nodes[0];
	BUG_ON(btrfs_item_size_nr(leaf, path->slots[0]) != sizeof(*ei0));

	btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
	ei0 = btrfs_item_ptr(leaf, path->slots[0],
			     struct btrfs_extent_item_v0);
	refs = btrfs_extent_refs_v0(leaf, ei0);

	if (owner == (u64)-1) {
		while (1) {
			if (path->slots[0] >= btrfs_header_nritems(leaf)) {
				ret = btrfs_next_leaf(root, path);
				if (ret < 0)
					return ret;
				BUG_ON(ret > 0);
				leaf = path->nodes[0];
			}
			btrfs_item_key_to_cpu(leaf, &found_key,
					      path->slots[0]);
			BUG_ON(key.objectid != found_key.objectid);
			if (found_key.type != BTRFS_EXTENT_REF_V0_KEY) {
				path->slots[0]++;
				continue;
			}
			ref0 = btrfs_item_ptr(leaf, path->slots[0],
					      struct btrfs_extent_ref_v0);
			owner = btrfs_ref_objectid_v0(leaf, ref0);
			break;
		}
	}
	btrfs_release_path(root, path);

	if (owner < BTRFS_FIRST_FREE_OBJECTID)
		new_size += sizeof(*bi);

	new_size -= sizeof(*ei0);
	ret = btrfs_search_slot(trans, root, &key, path,
				new_size + extra_size, 1);
	if (ret < 0)
		return ret;
	BUG_ON(ret);

	ret = btrfs_extend_item(trans, root, path, new_size);
	BUG_ON(ret);

	leaf = path->nodes[0];
	item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
	btrfs_set_extent_refs(leaf, item, refs);
	/* FIXME: get real generation */
	btrfs_set_extent_generation(leaf, item, 0);
	if (owner < BTRFS_FIRST_FREE_OBJECTID) {
		btrfs_set_extent_flags(leaf, item,
				       BTRFS_EXTENT_FLAG_TREE_BLOCK |
				       BTRFS_BLOCK_FLAG_FULL_BACKREF);
		bi = (struct btrfs_tree_block_info *)(item + 1);
		/* FIXME: get first key of the block */
		memset_extent_buffer(leaf, 0, (unsigned long)bi, sizeof(*bi));
		btrfs_set_tree_block_level(leaf, bi, (int)owner);
	} else {
		btrfs_set_extent_flags(leaf, item, BTRFS_EXTENT_FLAG_DATA);
	}
	btrfs_mark_buffer_dirty(leaf);
	return 0;
}
#endif

static u64 hash_extent_data_ref(u64 root_objectid, u64 owner, u64 offset)
{
	u32 high_crc = ~(u32)0;
	u32 low_crc = ~(u32)0;
	__le64 lenum;

	lenum = cpu_to_le64(root_objectid);
	high_crc = crc32c(high_crc, &lenum, sizeof(lenum));
	lenum = cpu_to_le64(owner);
	low_crc = crc32c(low_crc, &lenum, sizeof(lenum));
	lenum = cpu_to_le64(offset);
	low_crc = crc32c(low_crc, &lenum, sizeof(lenum));

	return ((u64)high_crc << 31) ^ (u64)low_crc;
}

static u64 hash_extent_data_ref_item(struct extent_buffer *leaf,
				     struct btrfs_extent_data_ref *ref)
{
	return hash_extent_data_ref(btrfs_extent_data_ref_root(leaf, ref),
				    btrfs_extent_data_ref_objectid(leaf, ref),
				    btrfs_extent_data_ref_offset(leaf, ref));
}

static int match_extent_data_ref(struct extent_buffer *leaf,
				 struct btrfs_extent_data_ref *ref,
				 u64 root_objectid, u64 owner, u64 offset)
{
	if (btrfs_extent_data_ref_root(leaf, ref) != root_objectid ||
	    btrfs_extent_data_ref_objectid(leaf, ref) != owner ||
	    btrfs_extent_data_ref_offset(leaf, ref) != offset)
		return 0;
	return 1;
}

static noinline int lookup_extent_data_ref(struct btrfs_trans_handle *trans,
					   struct btrfs_root *root,
					   struct btrfs_path *path,
					   u64 bytenr, u64 parent,
					   u64 root_objectid,
					   u64 owner, u64 offset)
{
	struct btrfs_key key;
	struct btrfs_extent_data_ref *ref;
	struct extent_buffer *leaf;
	u32 nritems;
	int ret;
	int recow;
	int err = -ENOENT;

	key.objectid = bytenr;
	if (parent) {
		key.type = BTRFS_SHARED_DATA_REF_KEY;
		key.offset = parent;
	} else {
		key.type = BTRFS_EXTENT_DATA_REF_KEY;
		key.offset = hash_extent_data_ref(root_objectid,
						  owner, offset);
	}
again:
	recow = 0;
	ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
	if (ret < 0) {
		err = ret;
		goto fail;
	}

	if (parent) {
		if (!ret)
			return 0;
#ifdef BTRFS_COMPAT_EXTENT_TREE_V0
		key.type = BTRFS_EXTENT_REF_V0_KEY;
		btrfs_release_path(root, path);
		ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
		if (ret < 0) {
			err = ret;
			goto fail;
		}
		if (!ret)
			return 0;
#endif
		goto fail;
	}

	leaf = path->nodes[0];
	nritems = btrfs_header_nritems(leaf);
	while (1) {
		if (path->slots[0] >= nritems) {
			ret = btrfs_next_leaf(root, path);
			if (ret < 0)
				err = ret;
			if (ret)
				goto fail;

			leaf = path->nodes[0];
			nritems = btrfs_header_nritems(leaf);
			recow = 1;
		}

		btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
		if (key.objectid != bytenr ||
		    key.type != BTRFS_EXTENT_DATA_REF_KEY)
			goto fail;

		ref = btrfs_item_ptr(leaf, path->slots[0],
				     struct btrfs_extent_data_ref);

		if (match_extent_data_ref(leaf, ref, root_objectid,
					  owner, offset)) {
			if (recow) {
				btrfs_release_path(root, path);
				goto again;
			}
			err = 0;
			break;
		}
		path->slots[0]++;
	}
fail:
	return err;
}

static noinline int insert_extent_data_ref(struct btrfs_trans_handle *trans,
					   struct btrfs_root *root,
					   struct btrfs_path *path,
					   u64 bytenr, u64 parent,
					   u64 root_objectid, u64 owner,
					   u64 offset, int refs_to_add)
{
	struct btrfs_key key;
	struct extent_buffer *leaf;
	u32 size;
	u32 num_refs;
	int ret;

	key.objectid = bytenr;
	if (parent) {
		key.type = BTRFS_SHARED_DATA_REF_KEY;
		key.offset = parent;
		size = sizeof(struct btrfs_shared_data_ref);
	} else {
		key.type = BTRFS_EXTENT_DATA_REF_KEY;
		key.offset = hash_extent_data_ref(root_objectid,
						  owner, offset);
		size = sizeof(struct btrfs_extent_data_ref);
	}

	ret = btrfs_insert_empty_item(trans, root, path, &key, size);
	if (ret && ret != -EEXIST)
		goto fail;

	leaf = path->nodes[0];
	if (parent) {
		struct btrfs_shared_data_ref *ref;
		ref = btrfs_item_ptr(leaf, path->slots[0],
				     struct btrfs_shared_data_ref);
		if (ret == 0) {
			btrfs_set_shared_data_ref_count(leaf, ref, refs_to_add);
		} else {
			num_refs = btrfs_shared_data_ref_count(leaf, ref);
			num_refs += refs_to_add;
			btrfs_set_shared_data_ref_count(leaf, ref, num_refs);
		}
	} else {
		struct btrfs_extent_data_ref *ref;
		while (ret == -EEXIST) {
			ref = btrfs_item_ptr(leaf, path->slots[0],
					     struct btrfs_extent_data_ref);
			if (match_extent_data_ref(leaf, ref, root_objectid,
						  owner, offset))
				break;
			btrfs_release_path(root, path);
			key.offset++;
			ret = btrfs_insert_empty_item(trans, root, path, &key,
						      size);
			if (ret && ret != -EEXIST)
				goto fail;

			leaf = path->nodes[0];
		}
		ref = btrfs_item_ptr(leaf, path->slots[0],
				     struct btrfs_extent_data_ref);
		if (ret == 0) {
			btrfs_set_extent_data_ref_root(leaf, ref,
						       root_objectid);
			btrfs_set_extent_data_ref_objectid(leaf, ref, owner);
			btrfs_set_extent_data_ref_offset(leaf, ref, offset);
			btrfs_set_extent_data_ref_count(leaf, ref, refs_to_add);
		} else {
			num_refs = btrfs_extent_data_ref_count(leaf, ref);
			num_refs += refs_to_add;
			btrfs_set_extent_data_ref_count(leaf, ref, num_refs);
		}
	}
	btrfs_mark_buffer_dirty(leaf);
	ret = 0;
fail:
	btrfs_release_path(root, path);
	return ret;
}

static noinline int remove_extent_data_ref(struct btrfs_trans_handle *trans,
					   struct btrfs_root *root,
					   struct btrfs_path *path,
					   int refs_to_drop)
{
	struct btrfs_key key;
	struct btrfs_extent_data_ref *ref1 = NULL;
	struct btrfs_shared_data_ref *ref2 = NULL;
	struct extent_buffer *leaf;
	u32 num_refs = 0;
	int ret = 0;

	leaf = path->nodes[0];
	btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);

	if (key.type == BTRFS_EXTENT_DATA_REF_KEY) {
		ref1 = btrfs_item_ptr(leaf, path->slots[0],
				      struct btrfs_extent_data_ref);
		num_refs = btrfs_extent_data_ref_count(leaf, ref1);
	} else if (key.type == BTRFS_SHARED_DATA_REF_KEY) {
		ref2 = btrfs_item_ptr(leaf, path->slots[0],
				      struct btrfs_shared_data_ref);
		num_refs = btrfs_shared_data_ref_count(leaf, ref2);
#ifdef BTRFS_COMPAT_EXTENT_TREE_V0
	} else if (key.type == BTRFS_EXTENT_REF_V0_KEY) {
		struct btrfs_extent_ref_v0 *ref0;
		ref0 = btrfs_item_ptr(leaf, path->slots[0],
				      struct btrfs_extent_ref_v0);
		num_refs = btrfs_ref_count_v0(leaf, ref0);
#endif
	} else {
		BUG();
	}

	BUG_ON(num_refs < refs_to_drop);
	num_refs -= refs_to_drop;

	if (num_refs == 0) {
		ret = btrfs_del_item(trans, root, path);
	} else {
		if (key.type == BTRFS_EXTENT_DATA_REF_KEY)
			btrfs_set_extent_data_ref_count(leaf, ref1, num_refs);
		else if (key.type == BTRFS_SHARED_DATA_REF_KEY)
			btrfs_set_shared_data_ref_count(leaf, ref2, num_refs);
#ifdef BTRFS_COMPAT_EXTENT_TREE_V0
		else {
			struct btrfs_extent_ref_v0 *ref0;
			ref0 = btrfs_item_ptr(leaf, path->slots[0],
					struct btrfs_extent_ref_v0);
			btrfs_set_ref_count_v0(leaf, ref0, num_refs);
		}
#endif
		btrfs_mark_buffer_dirty(leaf);
	}
	return ret;
}

static noinline u32 extent_data_ref_count(struct btrfs_root *root,
					  struct btrfs_path *path,
					  struct btrfs_extent_inline_ref *iref)
{
	struct btrfs_key key;
	struct extent_buffer *leaf;
	struct btrfs_extent_data_ref *ref1;
	struct btrfs_shared_data_ref *ref2;
	u32 num_refs = 0;

	leaf = path->nodes[0];
	btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
	if (iref) {
		if (btrfs_extent_inline_ref_type(leaf, iref) ==
		    BTRFS_EXTENT_DATA_REF_KEY) {
			ref1 = (struct btrfs_extent_data_ref *)(&iref->offset);
			num_refs = btrfs_extent_data_ref_count(leaf, ref1);
		} else {
			ref2 = (struct btrfs_shared_data_ref *)(iref + 1);
			num_refs = btrfs_shared_data_ref_count(leaf, ref2);
		}
	} else if (key.type == BTRFS_EXTENT_DATA_REF_KEY) {
		ref1 = btrfs_item_ptr(leaf, path->slots[0],
				      struct btrfs_extent_data_ref);
		num_refs = btrfs_extent_data_ref_count(leaf, ref1);
	} else if (key.type == BTRFS_SHARED_DATA_REF_KEY) {
		ref2 = btrfs_item_ptr(leaf, path->slots[0],
				      struct btrfs_shared_data_ref);
		num_refs = btrfs_shared_data_ref_count(leaf, ref2);
#ifdef BTRFS_COMPAT_EXTENT_TREE_V0
	} else if (key.type == BTRFS_EXTENT_REF_V0_KEY) {
		struct btrfs_extent_ref_v0 *ref0;
		ref0 = btrfs_item_ptr(leaf, path->slots[0],
				      struct btrfs_extent_ref_v0);
		num_refs = btrfs_ref_count_v0(leaf, ref0);
#endif
	} else {
		WARN_ON(1);
	}
	return num_refs;
}

static noinline int lookup_tree_block_ref(struct btrfs_trans_handle *trans,
					  struct btrfs_root *root,
					  struct btrfs_path *path,
					  u64 bytenr, u64 parent,
					  u64 root_objectid)
{
	struct btrfs_key key;
	int ret;

	key.objectid = bytenr;
	if (parent) {
		key.type = BTRFS_SHARED_BLOCK_REF_KEY;
		key.offset = parent;
	} else {
		key.type = BTRFS_TREE_BLOCK_REF_KEY;
		key.offset = root_objectid;
	}

	ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
	if (ret > 0)
		ret = -ENOENT;
#ifdef BTRFS_COMPAT_EXTENT_TREE_V0
	if (ret == -ENOENT && parent) {
		btrfs_release_path(root, path);
		key.type = BTRFS_EXTENT_REF_V0_KEY;
		ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
		if (ret > 0)
			ret = -ENOENT;
	}
#endif
	return ret;
}

static noinline int insert_tree_block_ref(struct btrfs_trans_handle *trans,
					  struct btrfs_root *root,
					  struct btrfs_path *path,
					  u64 bytenr, u64 parent,
					  u64 root_objectid)
{
	struct btrfs_key key;
	int ret;

	key.objectid = bytenr;
	if (parent) {
		key.type = BTRFS_SHARED_BLOCK_REF_KEY;
		key.offset = parent;
	} else {
		key.type = BTRFS_TREE_BLOCK_REF_KEY;
		key.offset = root_objectid;
	}

	ret = btrfs_insert_empty_item(trans, root, path, &key, 0);
	btrfs_release_path(root, path);
	return ret;
}

static inline int extent_ref_type(u64 parent, u64 owner)
{
	int type;
	if (owner < BTRFS_FIRST_FREE_OBJECTID) {
		if (parent > 0)
			type = BTRFS_SHARED_BLOCK_REF_KEY;
		else
			type = BTRFS_TREE_BLOCK_REF_KEY;
	} else {
		if (parent > 0)
			type = BTRFS_SHARED_DATA_REF_KEY;
		else
			type = BTRFS_EXTENT_DATA_REF_KEY;
	}
	return type;
}

static int find_next_key(struct btrfs_path *path, int level,
			 struct btrfs_key *key)

{
	for (; level < BTRFS_MAX_LEVEL; level++) {
		if (!path->nodes[level])
			break;
		if (path->slots[level] + 1 >=
		    btrfs_header_nritems(path->nodes[level]))
			continue;
		if (level == 0)
			btrfs_item_key_to_cpu(path->nodes[level], key,
					      path->slots[level] + 1);
		else
			btrfs_node_key_to_cpu(path->nodes[level], key,
					      path->slots[level] + 1);
		return 0;
	}
	return 1;
}

/*
 * look for inline back ref. if back ref is found, *ref_ret is set
 * to the address of inline back ref, and 0 is returned.
 *
 * if back ref isn't found, *ref_ret is set to the address where it
 * should be inserted, and -ENOENT is returned.
 *
 * if insert is true and there are too many inline back refs, the path
 * points to the extent item, and -EAGAIN is returned.
 *
 * NOTE: inline back refs are ordered in the same way that back ref
 *	 items in the tree are ordered.
 */
static noinline_for_stack
int lookup_inline_extent_backref(struct btrfs_trans_handle *trans,
				 struct btrfs_root *root,
				 struct btrfs_path *path,
				 struct btrfs_extent_inline_ref **ref_ret,
				 u64 bytenr, u64 num_bytes,
				 u64 parent, u64 root_objectid,
				 u64 owner, u64 offset, int insert)
{
	struct btrfs_key key;
	struct extent_buffer *leaf;
	struct btrfs_extent_item *ei;
	struct btrfs_extent_inline_ref *iref;
	u64 flags;
	u64 item_size;
	unsigned long ptr;
	unsigned long end;
	int extra_size;
	int type;
	int want;
	int ret;
	int err = 0;

	key.objectid = bytenr;
	key.type = BTRFS_EXTENT_ITEM_KEY;
	key.offset = num_bytes;

	want = extent_ref_type(parent, owner);
	if (insert) {
		extra_size = btrfs_extent_inline_ref_size(want);
		path->keep_locks = 1;
	} else
		extra_size = -1;
	ret = btrfs_search_slot(trans, root, &key, path, extra_size, 1);
	if (ret < 0) {
		err = ret;
		goto out;
	}
	BUG_ON(ret);

	leaf = path->nodes[0];
	item_size = btrfs_item_size_nr(leaf, path->slots[0]);
#ifdef BTRFS_COMPAT_EXTENT_TREE_V0
	if (item_size < sizeof(*ei)) {
		if (!insert) {
			err = -ENOENT;
			goto out;
		}
		ret = convert_extent_item_v0(trans, root, path, owner,
					     extra_size);
		if (ret < 0) {
			err = ret;
			goto out;
		}
		leaf = path->nodes[0];
		item_size = btrfs_item_size_nr(leaf, path->slots[0]);
	}
#endif
	BUG_ON(item_size < sizeof(*ei));

	ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
	flags = btrfs_extent_flags(leaf, ei);

	ptr = (unsigned long)(ei + 1);
	end = (unsigned long)ei + item_size;

	if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) {
		ptr += sizeof(struct btrfs_tree_block_info);
		BUG_ON(ptr > end);
	} else {
		BUG_ON(!(flags & BTRFS_EXTENT_FLAG_DATA));
	}

	err = -ENOENT;
	while (1) {
		if (ptr >= end) {
			WARN_ON(ptr > end);
			break;
		}
		iref = (struct btrfs_extent_inline_ref *)ptr;
		type = btrfs_extent_inline_ref_type(leaf, iref);
		if (want < type)
			break;
		if (want > type) {
			ptr += btrfs_extent_inline_ref_size(type);
			continue;
		}

		if (type == BTRFS_EXTENT_DATA_REF_KEY) {
			struct btrfs_extent_data_ref *dref;
			dref = (struct btrfs_extent_data_ref *)(&iref->offset);
			if (match_extent_data_ref(leaf, dref, root_objectid,
						  owner, offset)) {
				err = 0;
				break;
			}
			if (hash_extent_data_ref_item(leaf, dref) <
			    hash_extent_data_ref(root_objectid, owner, offset))
				break;
		} else {
			u64 ref_offset;
			ref_offset = btrfs_extent_inline_ref_offset(leaf, iref);
			if (parent > 0) {
				if (parent == ref_offset) {
					err = 0;
					break;
				}
				if (ref_offset < parent)
					break;
			} else {
				if (root_objectid == ref_offset) {
					err = 0;
					break;
				}
				if (ref_offset < root_objectid)
					break;
			}
		}
		ptr += btrfs_extent_inline_ref_size(type);
	}
	if (err == -ENOENT && insert) {
		if (item_size + extra_size >=
		    BTRFS_MAX_EXTENT_ITEM_SIZE(root)) {
			err = -EAGAIN;
			goto out;
		}
		/*
		 * To add new inline back ref, we have to make sure
		 * there is no corresponding back ref item.
		 * For simplicity, we just do not add new inline back
		 * ref if there is any kind of item for this block
		 */
		if (find_next_key(path, 0, &key) == 0 &&
		    key.objectid == bytenr &&
		    key.type < BTRFS_BLOCK_GROUP_ITEM_KEY) {
			err = -EAGAIN;
			goto out;
		}
	}
	*ref_ret = (struct btrfs_extent_inline_ref *)ptr;
out:
	if (insert) {
		path->keep_locks = 0;
		btrfs_unlock_up_safe(path, 1);
	}
	return err;
}

/*
 * helper to add new inline back ref
 */
static noinline_for_stack
int setup_inline_extent_backref(struct btrfs_trans_handle *trans,
				struct btrfs_root *root,
				struct btrfs_path *path,
				struct btrfs_extent_inline_ref *iref,
				u64 parent, u64 root_objectid,
				u64 owner, u64 offset, int refs_to_add,
				struct btrfs_delayed_extent_op *extent_op)
{
	struct extent_buffer *leaf;
	struct btrfs_extent_item *ei;
	unsigned long ptr;
	unsigned long end;
	unsigned long item_offset;
	u64 refs;
	int size;
	int type;
	int ret;

	leaf = path->nodes[0];
	ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
	item_offset = (unsigned long)iref - (unsigned long)ei;

	type = extent_ref_type(parent, owner);
	size = btrfs_extent_inline_ref_size(type);

	ret = btrfs_extend_item(trans, root, path, size);
	BUG_ON(ret);

	ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
	refs = btrfs_extent_refs(leaf, ei);
	refs += refs_to_add;
	btrfs_set_extent_refs(leaf, ei, refs);
	if (extent_op)
		__run_delayed_extent_op(extent_op, leaf, ei);

	ptr = (unsigned long)ei + item_offset;
	end = (unsigned long)ei + btrfs_item_size_nr(leaf, path->slots[0]);
	if (ptr < end - size)
		memmove_extent_buffer(leaf, ptr + size, ptr,
				      end - size - ptr);

	iref = (struct btrfs_extent_inline_ref *)ptr;
	btrfs_set_extent_inline_ref_type(leaf, iref, type);
	if (type == BTRFS_EXTENT_DATA_REF_KEY) {
		struct btrfs_extent_data_ref *dref;
		dref = (struct btrfs_extent_data_ref *)(&iref->offset);
		btrfs_set_extent_data_ref_root(leaf, dref, root_objectid);
		btrfs_set_extent_data_ref_objectid(leaf, dref, owner);
		btrfs_set_extent_data_ref_offset(leaf, dref, offset);
		btrfs_set_extent_data_ref_count(leaf, dref, refs_to_add);
	} else if (type == BTRFS_SHARED_DATA_REF_KEY) {
		struct btrfs_shared_data_ref *sref;
		sref = (struct btrfs_shared_data_ref *)(iref + 1);
		btrfs_set_shared_data_ref_count(leaf, sref, refs_to_add);
		btrfs_set_extent_inline_ref_offset(leaf, iref, parent);
	} else if (type == BTRFS_SHARED_BLOCK_REF_KEY) {
		btrfs_set_extent_inline_ref_offset(leaf, iref, parent);
	} else {
		btrfs_set_extent_inline_ref_offset(leaf, iref, root_objectid);
	}
	btrfs_mark_buffer_dirty(leaf);
	return 0;
}

static int lookup_extent_backref(struct btrfs_trans_handle *trans,
				 struct btrfs_root *root,
				 struct btrfs_path *path,
				 struct btrfs_extent_inline_ref **ref_ret,
				 u64 bytenr, u64 num_bytes, u64 parent,
				 u64 root_objectid, u64 owner, u64 offset)
{
	int ret;

	ret = lookup_inline_extent_backref(trans, root, path, ref_ret,
					   bytenr, num_bytes, parent,
					   root_objectid, owner, offset, 0);
	if (ret != -ENOENT)
		return ret;

	btrfs_release_path(root, path);
	*ref_ret = NULL;

	if (owner < BTRFS_FIRST_FREE_OBJECTID) {
		ret = lookup_tree_block_ref(trans, root, path, bytenr, parent,
					    root_objectid);
	} else {
		ret = lookup_extent_data_ref(trans, root, path, bytenr, parent,
					     root_objectid, owner, offset);
	}
	return ret;
}

/*
 * helper to update/remove inline back ref
 */
static noinline_for_stack
int update_inline_extent_backref(struct btrfs_trans_handle *trans,
				 struct btrfs_root *root,
				 struct btrfs_path *path,
				 struct btrfs_extent_inline_ref *iref,
				 int refs_to_mod,
				 struct btrfs_delayed_extent_op *extent_op)
{
	struct extent_buffer *leaf;
	struct btrfs_extent_item *ei;
	struct btrfs_extent_data_ref *dref = NULL;
	struct btrfs_shared_data_ref *sref = NULL;
	unsigned long ptr;
	unsigned long end;
	u32 item_size;
	int size;
	int type;
	int ret;
	u64 refs;

	leaf = path->nodes[0];
	ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
	refs = btrfs_extent_refs(leaf, ei);
	WARN_ON(refs_to_mod < 0 && refs + refs_to_mod <= 0);
	refs += refs_to_mod;
	btrfs_set_extent_refs(leaf, ei, refs);
	if (extent_op)
		__run_delayed_extent_op(extent_op, leaf, ei);

	type = btrfs_extent_inline_ref_type(leaf, iref);

	if (type == BTRFS_EXTENT_DATA_REF_KEY) {
		dref = (struct btrfs_extent_data_ref *)(&iref->offset);
		refs = btrfs_extent_data_ref_count(leaf, dref);
	} else if (type == BTRFS_SHARED_DATA_REF_KEY) {
		sref = (struct btrfs_shared_data_ref *)(iref + 1);
		refs = btrfs_shared_data_ref_count(leaf, sref);
	} else {
		refs = 1;
		BUG_ON(refs_to_mod != -1);
	}

	BUG_ON(refs_to_mod < 0 && refs < -refs_to_mod);
	refs += refs_to_mod;

	if (refs > 0) {
		if (type == BTRFS_EXTENT_DATA_REF_KEY)
			btrfs_set_extent_data_ref_count(leaf, dref, refs);
		else
			btrfs_set_shared_data_ref_count(leaf, sref, refs);
	} else {
		size =  btrfs_extent_inline_ref_size(type);
		item_size = btrfs_item_size_nr(leaf, path->slots[0]);
		ptr = (unsigned long)iref;
		end = (unsigned long)ei + item_size;
		if (ptr + size < end)
			memmove_extent_buffer(leaf, ptr, ptr + size,
					      end - ptr - size);
		item_size -= size;
		ret = btrfs_truncate_item(trans, root, path, item_size, 1);
		BUG_ON(ret);
	}
	btrfs_mark_buffer_dirty(leaf);
	return 0;
}

static noinline_for_stack
int insert_inline_extent_backref(struct btrfs_trans_handle *trans,
				 struct btrfs_root *root,
				 struct btrfs_path *path,
				 u64 bytenr, u64 num_bytes, u64 parent,
				 u64 root_objectid, u64 owner,
				 u64 offset, int refs_to_add,
				 struct btrfs_delayed_extent_op *extent_op)
{
	struct btrfs_extent_inline_ref *iref;
	int ret;

	ret = lookup_inline_extent_backref(trans, root, path, &iref,
					   bytenr, num_bytes, parent,
					   root_objectid, owner, offset, 1);
	if (ret == 0) {
		BUG_ON(owner < BTRFS_FIRST_FREE_OBJECTID);
		ret = update_inline_extent_backref(trans, root, path, iref,
						   refs_to_add, extent_op);
	} else if (ret == -ENOENT) {
		ret = setup_inline_extent_backref(trans, root, path, iref,
						  parent, root_objectid,
						  owner, offset, refs_to_add,
						  extent_op);
	}
	return ret;
}

static int insert_extent_backref(struct btrfs_trans_handle *trans,
				 struct btrfs_root *root,
				 struct btrfs_path *path,
				 u64 bytenr, u64 parent, u64 root_objectid,
				 u64 owner, u64 offset, int refs_to_add)
{
	int ret;
	if (owner < BTRFS_FIRST_FREE_OBJECTID) {
		BUG_ON(refs_to_add != 1);
		ret = insert_tree_block_ref(trans, root, path, bytenr,
					    parent, root_objectid);
	} else {
		ret = insert_extent_data_ref(trans, root, path, bytenr,
					     parent, root_objectid,
					     owner, offset, refs_to_add);
	}
	return ret;
}

static int remove_extent_backref(struct btrfs_trans_handle *trans,
				 struct btrfs_root *root,
				 struct btrfs_path *path,
				 struct btrfs_extent_inline_ref *iref,
				 int refs_to_drop, int is_data)
{
	int ret;

	BUG_ON(!is_data && refs_to_drop != 1);
	if (iref) {
		ret = update_inline_extent_backref(trans, root, path, iref,
						   -refs_to_drop, NULL);
	} else if (is_data) {
		ret = remove_extent_data_ref(trans, root, path, refs_to_drop);
	} else {
		ret = btrfs_del_item(trans, root, path);
	}
	return ret;
}

#ifdef BIO_RW_DISCARD
static void btrfs_issue_discard(struct block_device *bdev,
				u64 start, u64 len)
{
	blkdev_issue_discard(bdev, start >> 9, len >> 9, GFP_KERNEL);
}
#endif

static int btrfs_discard_extent(struct btrfs_root *root, u64 bytenr,
				u64 num_bytes)
{
#ifdef BIO_RW_DISCARD
	int ret;
	u64 map_length = num_bytes;
	struct btrfs_multi_bio *multi = NULL;

	/* Tell the block device(s) that the sectors can be discarded */
	ret = btrfs_map_block(&root->fs_info->mapping_tree, READ,
			      bytenr, &map_length, &multi, 0);
	if (!ret) {
		struct btrfs_bio_stripe *stripe = multi->stripes;
		int i;

		if (map_length > num_bytes)
			map_length = num_bytes;

		for (i = 0; i < multi->num_stripes; i++, stripe++) {
			btrfs_issue_discard(stripe->dev->bdev,
					    stripe->physical,
					    map_length);
		}
		kfree(multi);
	}

	return ret;
#else
	return 0;
#endif
}

int btrfs_inc_extent_ref(struct btrfs_trans_handle *trans,
			 struct btrfs_root *root,
			 u64 bytenr, u64 num_bytes, u64 parent,
			 u64 root_objectid, u64 owner, u64 offset)
{
	int ret;
	BUG_ON(owner < BTRFS_FIRST_FREE_OBJECTID &&
	       root_objectid == BTRFS_TREE_LOG_OBJECTID);

	if (owner < BTRFS_FIRST_FREE_OBJECTID) {
		ret = btrfs_add_delayed_tree_ref(trans, bytenr, num_bytes,
					parent, root_objectid, (int)owner,
					BTRFS_ADD_DELAYED_REF, NULL);
	} else {
		ret = btrfs_add_delayed_data_ref(trans, bytenr, num_bytes,
					parent, root_objectid, owner, offset,
					BTRFS_ADD_DELAYED_REF, NULL);
	}
	return ret;
}

static int __btrfs_inc_extent_ref(struct btrfs_trans_handle *trans,
				  struct btrfs_root *root,
				  u64 bytenr, u64 num_bytes,
				  u64 parent, u64 root_objectid,
				  u64 owner, u64 offset, int refs_to_add,
				  struct btrfs_delayed_extent_op *extent_op)
{
	struct btrfs_path *path;
	struct extent_buffer *leaf;
	struct btrfs_extent_item *item;
	u64 refs;
	int ret;
	int err = 0;

	path = btrfs_alloc_path();
	if (!path)
		return -ENOMEM;

	path->reada = 1;
	path->leave_spinning = 1;
	/* this will setup the path even if it fails to insert the back ref */
	ret = insert_inline_extent_backref(trans, root->fs_info->extent_root,
					   path, bytenr, num_bytes, parent,
					   root_objectid, owner, offset,
					   refs_to_add, extent_op);
	if (ret == 0)
		goto out;

	if (ret != -EAGAIN) {
		err = ret;
		goto out;
	}

	leaf = path->nodes[0];
	item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
	refs = btrfs_extent_refs(leaf, item);
	btrfs_set_extent_refs(leaf, item, refs + refs_to_add);
	if (extent_op)
		__run_delayed_extent_op(extent_op, leaf, item);

	btrfs_mark_buffer_dirty(leaf);
	btrfs_release_path(root->fs_info->extent_root, path);

	path->reada = 1;
	path->leave_spinning = 1;

	/* now insert the actual backref */
	ret = insert_extent_backref(trans, root->fs_info->extent_root,
				    path, bytenr, parent, root_objectid,
				    owner, offset, refs_to_add);
	BUG_ON(ret);
out:
	btrfs_free_path(path);
	return err;
}

static int run_delayed_data_ref(struct btrfs_trans_handle *trans,
				struct btrfs_root *root,
				struct btrfs_delayed_ref_node *node,
				struct btrfs_delayed_extent_op *extent_op,
				int insert_reserved)
{
	int ret = 0;
	struct btrfs_delayed_data_ref *ref;
	struct btrfs_key ins;
	u64 parent = 0;
	u64 ref_root = 0;
	u64 flags = 0;

	ins.objectid = node->bytenr;
	ins.offset = node->num_bytes;
	ins.type = BTRFS_EXTENT_ITEM_KEY;

	ref = btrfs_delayed_node_to_data_ref(node);
	if (node->type == BTRFS_SHARED_DATA_REF_KEY)
		parent = ref->parent;
	else
		ref_root = ref->root;

	if (node->action == BTRFS_ADD_DELAYED_REF && insert_reserved) {
		if (extent_op) {
			BUG_ON(extent_op->update_key);
			flags |= extent_op->flags_to_set;
		}
		ret = alloc_reserved_file_extent(trans, root,
						 parent, ref_root, flags,
						 ref->objectid, ref->offset,
						 &ins, node->ref_mod);
		update_reserved_extents(root, ins.objectid, ins.offset, 0);
	} else if (node->action == BTRFS_ADD_DELAYED_REF) {
		ret = __btrfs_inc_extent_ref(trans, root, node->bytenr,
					     node->num_bytes, parent,
					     ref_root, ref->objectid,
					     ref->offset, node->ref_mod,
					     extent_op);
	} else if (node->action == BTRFS_DROP_DELAYED_REF) {
		ret = __btrfs_free_extent(trans, root, node->bytenr,
					  node->num_bytes, parent,
					  ref_root, ref->objectid,
					  ref->offset, node->ref_mod,
					  extent_op);
	} else {
		BUG();
	}
	return ret;
}

static void __run_delayed_extent_op(struct btrfs_delayed_extent_op *extent_op,
				    struct extent_buffer *leaf,
				    struct btrfs_extent_item *ei)
{
	u64 flags = btrfs_extent_flags(leaf, ei);
	if (extent_op->update_flags) {
		flags |= extent_op->flags_to_set;
		btrfs_set_extent_flags(leaf, ei, flags);
	}

	if (extent_op->update_key) {
		struct btrfs_tree_block_info *bi;
		BUG_ON(!(flags & BTRFS_EXTENT_FLAG_TREE_BLOCK));
		bi = (struct btrfs_tree_block_info *)(ei + 1);
		btrfs_set_tree_block_key(leaf, bi, &extent_op->key);
	}
}

static int run_delayed_extent_op(struct btrfs_trans_handle *trans,
				 struct btrfs_root *root,
				 struct btrfs_delayed_ref_node *node,
				 struct btrfs_delayed_extent_op *extent_op)
{
	struct btrfs_key key;
	struct btrfs_path *path;
	struct btrfs_extent_item *ei;
	struct extent_buffer *leaf;
	u32 item_size;
	int ret;
	int err = 0;

	path = btrfs_alloc_path();
	if (!path)
		return -ENOMEM;

	key.objectid = node->bytenr;
	key.type = BTRFS_EXTENT_ITEM_KEY;
	key.offset = node->num_bytes;

	path->reada = 1;
	path->leave_spinning = 1;
	ret = btrfs_search_slot(trans, root->fs_info->extent_root, &key,
				path, 0, 1);
	if (ret < 0) {
		err = ret;
		goto out;
	}
	if (ret > 0) {
		err = -EIO;
		goto out;
	}

	leaf = path->nodes[0];
	item_size = btrfs_item_size_nr(leaf, path->slots[0]);
#ifdef BTRFS_COMPAT_EXTENT_TREE_V0
	if (item_size < sizeof(*ei)) {
		ret = convert_extent_item_v0(trans, root->fs_info->extent_root,
					     path, (u64)-1, 0);
		if (ret < 0) {
			err = ret;
			goto out;
		}
		leaf = path->nodes[0];
		item_size = btrfs_item_size_nr(leaf, path->slots[0]);
	}
#endif
	BUG_ON(item_size < sizeof(*ei));
	ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
	__run_delayed_extent_op(extent_op, leaf, ei);

	btrfs_mark_buffer_dirty(leaf);
out:
	btrfs_free_path(path);
	return err;
}

static int run_delayed_tree_ref(struct btrfs_trans_handle *trans,
				struct btrfs_root *root,
				struct btrfs_delayed_ref_node *node,
				struct btrfs_delayed_extent_op *extent_op,
				int insert_reserved)
{
	int ret = 0;
	struct btrfs_delayed_tree_ref *ref;
	struct btrfs_key ins;
	u64 parent = 0;
	u64 ref_root = 0;

	ins.objectid = node->bytenr;
	ins.offset = node->num_bytes;
	ins.type = BTRFS_EXTENT_ITEM_KEY;

	ref = btrfs_delayed_node_to_tree_ref(node);
	if (node->type == BTRFS_SHARED_BLOCK_REF_KEY)
		parent = ref->parent;
	else
		ref_root = ref->root;

	BUG_ON(node->ref_mod != 1);
	if (node->action == BTRFS_ADD_DELAYED_REF && insert_reserved) {
		BUG_ON(!extent_op || !extent_op->update_flags ||
		       !extent_op->update_key);
		ret = alloc_reserved_tree_block(trans, root,
						parent, ref_root,
						extent_op->flags_to_set,
						&extent_op->key,
						ref->level, &ins);
		update_reserved_extents(root, ins.objectid, ins.offset, 0);
	} else if (node->action == BTRFS_ADD_DELAYED_REF) {
		ret = __btrfs_inc_extent_ref(trans, root, node->bytenr,
					     node->num_bytes, parent, ref_root,
					     ref->level, 0, 1, extent_op);
	} else if (node->action == BTRFS_DROP_DELAYED_REF) {
		ret = __btrfs_free_extent(trans, root, node->bytenr,
					  node->num_bytes, parent, ref_root,
					  ref->level, 0, 1, extent_op);
	} else {
		BUG();
	}
	return ret;
}


/* helper function to actually process a single delayed ref entry */
static int run_one_delayed_ref(struct btrfs_trans_handle *trans,
			       struct btrfs_root *root,
			       struct btrfs_delayed_ref_node *node,
			       struct btrfs_delayed_extent_op *extent_op,
			       int insert_reserved)
{
	int ret;
	if (btrfs_delayed_ref_is_head(node)) {
		struct btrfs_delayed_ref_head *head;
		/*
		 * we've hit the end of the chain and we were supposed
		 * to insert this extent into the tree.  But, it got
		 * deleted before we ever needed to insert it, so all
		 * we have to do is clean up the accounting
		 */
		BUG_ON(extent_op);
		head = btrfs_delayed_node_to_head(node);
		if (insert_reserved) {
			if (head->is_data) {
				ret = btrfs_del_csums(trans, root,
						      node->bytenr,
						      node->num_bytes);
				BUG_ON(ret);
			}
			btrfs_update_pinned_extents(root, node->bytenr,
						    node->num_bytes, 1);
			update_reserved_extents(root, node->bytenr,
						node->num_bytes, 0);
		}
		mutex_unlock(&head->mutex);
		return 0;
	}

	if (node->type == BTRFS_TREE_BLOCK_REF_KEY ||
	    node->type == BTRFS_SHARED_BLOCK_REF_KEY)
		ret = run_delayed_tree_ref(trans, root, node, extent_op,
					   insert_reserved);
	else if (node->type == BTRFS_EXTENT_DATA_REF_KEY ||
		 node->type == BTRFS_SHARED_DATA_REF_KEY)
		ret = run_delayed_data_ref(trans, root, node, extent_op,
					   insert_reserved);
	else
		BUG();
	return ret;
}

static noinline struct btrfs_delayed_ref_node *
select_delayed_ref(struct btrfs_delayed_ref_head *head)
{
	struct rb_node *node;
	struct btrfs_delayed_ref_node *ref;
	int action = BTRFS_ADD_DELAYED_REF;
again:
	/*
	 * select delayed ref of type BTRFS_ADD_DELAYED_REF first.
	 * this prevents ref count from going down to zero when
	 * there still are pending delayed ref.
	 */
	node = rb_prev(&head->node.rb_node);
	while (1) {
		if (!node)
			break;
		ref = rb_entry(node, struct btrfs_delayed_ref_node,
				rb_node);
		if (ref->bytenr != head->node.bytenr)
			break;
		if (ref->action == action)
			return ref;
		node = rb_prev(node);
	}
	if (action == BTRFS_ADD_DELAYED_REF) {
		action = BTRFS_DROP_DELAYED_REF;
		goto again;
	}
	return NULL;
}

static noinline int run_clustered_refs(struct btrfs_trans_handle *trans,
				       struct btrfs_root *root,
				       struct list_head *cluster)
{
	struct btrfs_delayed_ref_root *delayed_refs;
	struct btrfs_delayed_ref_node *ref;
	struct btrfs_delayed_ref_head *locked_ref = NULL;
	struct btrfs_delayed_extent_op *extent_op;
	int ret;
	int count = 0;
	int must_insert_reserved = 0;

	delayed_refs = &trans->transaction->delayed_refs;
	while (1) {
		if (!locked_ref) {
			/* pick a new head ref from the cluster list */
			if (list_empty(cluster))
				break;

			locked_ref = list_entry(cluster->next,
				     struct btrfs_delayed_ref_head, cluster);

			/* grab the lock that says we are going to process
			 * all the refs for this head */
			ret = btrfs_delayed_ref_lock(trans, locked_ref);

			/*
			 * we may have dropped the spin lock to get the head
			 * mutex lock, and that might have given someone else
			 * time to free the head.  If that's true, it has been
			 * removed from our list and we can move on.
			 */
			if (ret == -EAGAIN) {
				locked_ref = NULL;
				count++;
				continue;
			}
		}

		/*
		 * record the must insert reserved flag before we
		 * drop the spin lock.
		 */
		must_insert_reserved = locked_ref->must_insert_reserved;
		locked_ref->must_insert_reserved = 0;

		extent_op = locked_ref->extent_op;
		locked_ref->extent_op = NULL;

		/*
		 * locked_ref is the head node, so we have to go one
		 * node back for any delayed ref updates
		 */
		ref = select_delayed_ref(locked_ref);
		if (!ref) {
			/* All delayed refs have been processed, Go ahead
			 * and send the head node to run_one_delayed_ref,
			 * so that any accounting fixes can happen
			 */
			ref = &locked_ref->node;

			if (extent_op && must_insert_reserved) {
				kfree(extent_op);
				extent_op = NULL;
			}

			if (extent_op) {
				spin_unlock(&delayed_refs->lock);

				ret = run_delayed_extent_op(trans, root,
							    ref, extent_op);
				BUG_ON(ret);
				kfree(extent_op);

				cond_resched();
				spin_lock(&delayed_refs->lock);
				continue;
			}

			list_del_init(&locked_ref->cluster);
			locked_ref = NULL;
		}

		ref->in_tree = 0;
		rb_erase(&ref->rb_node, &delayed_refs->root);
		delayed_refs->num_entries--;

		spin_unlock(&delayed_refs->lock);

		ret = run_one_delayed_ref(trans, root, ref, extent_op,
					  must_insert_reserved);
		BUG_ON(ret);

		btrfs_put_delayed_ref(ref);
		kfree(extent_op);
		count++;

		cond_resched();
		spin_lock(&delayed_refs->lock);
	}
	return count;
}

/*
 * this starts processing the delayed reference count updates and
 * extent insertions we have queued up so far.  count can be
 * 0, which means to process everything in the tree at the start
 * of the run (but not newly added entries), or it can be some target
 * number you'd like to process.
 */
int btrfs_run_delayed_refs(struct btrfs_trans_handle *trans,
			   struct btrfs_root *root, unsigned long count)
{
	struct rb_node *node;
	struct btrfs_delayed_ref_root *delayed_refs;
	struct btrfs_delayed_ref_node *ref;
	struct list_head cluster;
	int ret;
	int run_all = count == (unsigned long)-1;
	int run_most = 0;

	if (root == root->fs_info->extent_root)
		root = root->fs_info->tree_root;

	delayed_refs = &trans->transaction->delayed_refs;
	INIT_LIST_HEAD(&cluster);
again:
	spin_lock(&delayed_refs->lock);
	if (count == 0) {
		count = delayed_refs->num_entries * 2;
		run_most = 1;
	}
	while (1) {
		if (!(run_all || run_most) &&
		    delayed_refs->num_heads_ready < 64)
			break;

		/*
		 * go find something we can process in the rbtree.  We start at
		 * the beginning of the tree, and then build a cluster
		 * of refs to process starting at the first one we are able to
		 * lock
		 */
		ret = btrfs_find_ref_cluster(trans, &cluster,
					     delayed_refs->run_delayed_start);
		if (ret)
			break;

		ret = run_clustered_refs(trans, root, &cluster);
		BUG_ON(ret < 0);

		count -= min_t(unsigned long, ret, count);

		if (count == 0)
			break;
	}

	if (run_all) {
		node = rb_first(&delayed_refs->root);
		if (!node)
			goto out;
		count = (unsigned long)-1;

		while (node) {
			ref = rb_entry(node, struct btrfs_delayed_ref_node,
				       rb_node);
			if (btrfs_delayed_ref_is_head(ref)) {
				struct btrfs_delayed_ref_head *head;

				head = btrfs_delayed_node_to_head(ref);
				atomic_inc(&ref->refs);

				spin_unlock(&delayed_refs->lock);
				mutex_lock(&head->mutex);
				mutex_unlock(&head->mutex);

				btrfs_put_delayed_ref(ref);
				cond_resched();
				goto again;
			}
			node = rb_next(node);
		}
		spin_unlock(&delayed_refs->lock);
		schedule_timeout(1);
		goto again;
	}
out:
	spin_unlock(&delayed_refs->lock);
	return 0;
}

int btrfs_set_disk_extent_flags(struct btrfs_trans_handle *trans,
				struct btrfs_root *root,
				u64 bytenr, u64 num_bytes, u64 flags,
				int is_data)
{
	struct btrfs_delayed_extent_op *extent_op;
	int ret;

	extent_op = kmalloc(sizeof(*extent_op), GFP_NOFS);
	if (!extent_op)
		return -ENOMEM;

	extent_op->flags_to_set = flags;
	extent_op->update_flags = 1;
	extent_op->update_key = 0;
	extent_op->is_data = is_data ? 1 : 0;

	ret = btrfs_add_delayed_extent_op(trans, bytenr, num_bytes, extent_op);
	if (ret)
		kfree(extent_op);
	return ret;
}

static noinline int check_delayed_ref(struct btrfs_trans_handle *trans,
				      struct btrfs_root *root,
				      struct btrfs_path *path,
				      u64 objectid, u64 offset, u64 bytenr)
{
	struct btrfs_delayed_ref_head *head;
	struct btrfs_delayed_ref_node *ref;
	struct btrfs_delayed_data_ref *data_ref;
	struct btrfs_delayed_ref_root *delayed_refs;
	struct rb_node *node;
	int ret = 0;

	ret = -ENOENT;
	delayed_refs = &trans->transaction->delayed_refs;
	spin_lock(&delayed_refs->lock);
	head = btrfs_find_delayed_ref_head(trans, bytenr);
	if (!head)
		goto out;

	if (!mutex_trylock(&head->mutex)) {
		atomic_inc(&head->node.refs);
		spin_unlock(&delayed_refs->lock);

		btrfs_release_path(root->fs_info->extent_root, path);

		mutex_lock(&head->mutex);
		mutex_unlock(&head->mutex);
		btrfs_put_delayed_ref(&head->node);
		return -EAGAIN;
	}

	node = rb_prev(&head->node.rb_node);
	if (!node)
		goto out_unlock;

	ref = rb_entry(node, struct btrfs_delayed_ref_node, rb_node);

	if (ref->bytenr != bytenr)
		goto out_unlock;

	ret = 1;
	if (ref->type != BTRFS_EXTENT_DATA_REF_KEY)
		goto out_unlock;

	data_ref = btrfs_delayed_node_to_data_ref(ref);

	node = rb_prev(node);
	if (node) {
		ref = rb_entry(node, struct btrfs_delayed_ref_node, rb_node);
		if (ref->bytenr == bytenr)
			goto out_unlock;
	}

	if (data_ref->root != root->root_key.objectid ||
	    data_ref->objectid != objectid || data_ref->offset != offset)
		goto out_unlock;

	ret = 0;
out_unlock:
	mutex_unlock(&head->mutex);
out:
	spin_unlock(&delayed_refs->lock);
	return ret;
}

static noinline int check_committed_ref(struct btrfs_trans_handle *trans,
					struct btrfs_root *root,
					struct btrfs_path *path,
					u64 objectid, u64 offset, u64 bytenr)
{
	struct btrfs_root *extent_root = root->fs_info->extent_root;
	struct extent_buffer *leaf;
	struct btrfs_extent_data_ref *ref;
	struct btrfs_extent_inline_ref *iref;
	struct btrfs_extent_item *ei;
	struct btrfs_key key;
	u32 item_size;
	int ret;

	key.objectid = bytenr;
	key.offset = (u64)-1;
	key.type = BTRFS_EXTENT_ITEM_KEY;

	ret = btrfs_search_slot(NULL, extent_root, &key, path, 0, 0);
	if (ret < 0)
		goto out;
	BUG_ON(ret == 0);

	ret = -ENOENT;
	if (path->slots[0] == 0)
		goto out;

	path->slots[0]--;
	leaf = path->nodes[0];
	btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);

	if (key.objectid != bytenr || key.type != BTRFS_EXTENT_ITEM_KEY)
		goto out;

	ret = 1;
	item_size = btrfs_item_size_nr(leaf, path->slots[0]);
#ifdef BTRFS_COMPAT_EXTENT_TREE_V0
	if (item_size < sizeof(*ei)) {
		WARN_ON(item_size != sizeof(struct btrfs_extent_item_v0));
		goto out;
	}
#endif
	ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);

	if (item_size != sizeof(*ei) +
	    btrfs_extent_inline_ref_size(BTRFS_EXTENT_DATA_REF_KEY))
		goto out;

	if (btrfs_extent_generation(leaf, ei) <=
	    btrfs_root_last_snapshot(&root->root_item))
		goto out;

	iref = (struct btrfs_extent_inline_ref *)(ei + 1);
	if (btrfs_extent_inline_ref_type(leaf, iref) !=
	    BTRFS_EXTENT_DATA_REF_KEY)
		goto out;

	ref = (struct btrfs_extent_data_ref *)(&iref->offset);
	if (btrfs_extent_refs(leaf, ei) !=
	    btrfs_extent_data_ref_count(leaf, ref) ||
	    btrfs_extent_data_ref_root(leaf, ref) !=
	    root->root_key.objectid ||
	    btrfs_extent_data_ref_objectid(leaf, ref) != objectid ||
	    btrfs_extent_data_ref_offset(leaf, ref) != offset)
		goto out;

	ret = 0;
out:
	return ret;
}

int btrfs_cross_ref_exist(struct btrfs_trans_handle *trans,
			  struct btrfs_root *root,
			  u64 objectid, u64 offset, u64 bytenr)
{
	struct btrfs_path *path;
	int ret;
	int ret2;

	path = btrfs_alloc_path();
	if (!path)
		return -ENOENT;

	do {
		ret = check_committed_ref(trans, root, path, objectid,
					  offset, bytenr);
		if (ret && ret != -ENOENT)
			goto out;

		ret2 = check_delayed_ref(trans, root, path, objectid,
					 offset, bytenr);
	} while (ret2 == -EAGAIN);

	if (ret2 && ret2 != -ENOENT) {
		ret = ret2;
		goto out;
	}

	if (ret != -ENOENT || ret2 != -ENOENT)
		ret = 0;
out:
	btrfs_free_path(path);
	return ret;
}

#if 0
int btrfs_cache_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root,
		    struct extent_buffer *buf, u32 nr_extents)
{
	struct btrfs_key key;
	struct btrfs_file_extent_item *fi;
	u64 root_gen;
	u32 nritems;
	int i;
	int level;
	int ret = 0;
	int shared = 0;

	if (!root->ref_cows)
		return 0;

	if (root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID) {
		shared = 0;
		root_gen = root->root_key.offset;
	} else {
		shared = 1;
		root_gen = trans->transid - 1;
	}

	level = btrfs_header_level(buf);
	nritems = btrfs_header_nritems(buf);

	if (level == 0) {
		struct btrfs_leaf_ref *ref;
		struct btrfs_extent_info *info;

		ref = btrfs_alloc_leaf_ref(root, nr_extents);
		if (!ref) {
			ret = -ENOMEM;
			goto out;
		}

		ref->root_gen = root_gen;
		ref->bytenr = buf->start;
		ref->owner = btrfs_header_owner(buf);
		ref->generation = btrfs_header_generation(buf);
		ref->nritems = nr_extents;
		info = ref->extents;

		for (i = 0; nr_extents > 0 && i < nritems; i++) {
			u64 disk_bytenr;
			btrfs_item_key_to_cpu(buf, &key, i);
			if (btrfs_key_type(&key) != BTRFS_EXTENT_DATA_KEY)
				continue;
			fi = btrfs_item_ptr(buf, i,
					    struct btrfs_file_extent_item);
			if (btrfs_file_extent_type(buf, fi) ==
			    BTRFS_FILE_EXTENT_INLINE)
				continue;
			disk_bytenr = btrfs_file_extent_disk_bytenr(buf, fi);
			if (disk_bytenr == 0)
				continue;

			info->bytenr = disk_bytenr;
			info->num_bytes =
				btrfs_file_extent_disk_num_bytes(buf, fi);
			info->objectid = key.objectid;
			info->offset = key.offset;
			info++;
		}

		ret = btrfs_add_leaf_ref(root, ref, shared);
		if (ret == -EEXIST && shared) {
			struct btrfs_leaf_ref *old;
			old = btrfs_lookup_leaf_ref(root, ref->bytenr);
			BUG_ON(!old);
			btrfs_remove_leaf_ref(root, old);
			btrfs_free_leaf_ref(root, old);
			ret = btrfs_add_leaf_ref(root, ref, shared);
		}
		WARN_ON(ret);
		btrfs_free_leaf_ref(root, ref);
	}
out:
	return ret;
}

/* when a block goes through cow, we update the reference counts of
 * everything that block points to.  The internal pointers of the block
 * can be in just about any order, and it is likely to have clusters of
 * things that are close together and clusters of things that are not.
 *
 * To help reduce the seeks that come with updating all of these reference
 * counts, sort them by byte number before actual updates are done.
 *
 * struct refsort is used to match byte number to slot in the btree block.
 * we sort based on the byte number and then use the slot to actually
 * find the item.
 *
 * struct refsort is smaller than strcut btrfs_item and smaller than
 * struct btrfs_key_ptr.  Since we're currently limited to the page size
 * for a btree block, there's no way for a kmalloc of refsorts for a
 * single node to be bigger than a page.
 */
struct refsort {
	u64 bytenr;
	u32 slot;
};

/*
 * for passing into sort()
 */
static int refsort_cmp(const void *a_void, const void *b_void)
{
	const struct refsort *a = a_void;
	const struct refsort *b = b_void;

	if (a->bytenr < b->bytenr)
		return -1;
	if (a->bytenr > b->bytenr)
		return 1;
	return 0;
}
#endif

static int __btrfs_mod_ref(struct btrfs_trans_handle *trans,
			   struct btrfs_root *root,
			   struct extent_buffer *buf,
			   int full_backref, int inc)
{
	u64 bytenr;
	u64 num_bytes;
	u64 parent;
	u64 ref_root;
	u32 nritems;
	struct btrfs_key key;
	struct btrfs_file_extent_item *fi;
	int i;
	int level;
	int ret = 0;
	int (*process_func)(struct btrfs_trans_handle *, struct btrfs_root *,
			    u64, u64, u64, u64, u64, u64);

	ref_root = btrfs_header_owner(buf);
	nritems = btrfs_header_nritems(buf);
	level = btrfs_header_level(buf);

	if (!root->ref_cows && level == 0)
		return 0;

	if (inc)
		process_func = btrfs_inc_extent_ref;
	else
		process_func = btrfs_free_extent;

	if (full_backref)
		parent = buf->start;
	else
		parent = 0;

	for (i = 0; i < nritems; i++) {
		if (level == 0) {
			btrfs_item_key_to_cpu(buf, &key, i);
			if (btrfs_key_type(&key) != BTRFS_EXTENT_DATA_KEY)
				continue;
			fi = btrfs_item_ptr(buf, i,
					    struct btrfs_file_extent_item);
			if (btrfs_file_extent_type(buf, fi) ==
			    BTRFS_FILE_EXTENT_INLINE)
				continue;
			bytenr = btrfs_file_extent_disk_bytenr(buf, fi);
			if (bytenr == 0)
				continue;

			num_bytes = btrfs_file_extent_disk_num_bytes(buf, fi);
			key.offset -= btrfs_file_extent_offset(buf, fi);
			ret = process_func(trans, root, bytenr, num_bytes,
					   parent, ref_root, key.objectid,
					   key.offset);
			if (ret)
				goto fail;
		} else {
			bytenr = btrfs_node_blockptr(buf, i);
			num_bytes = btrfs_level_size(root, level - 1);
			ret = process_func(trans, root, bytenr, num_bytes,
					   parent, ref_root, level - 1, 0);
			if (ret)
				goto fail;
		}
	}
	return 0;
fail:
	BUG();
	return ret;
}

int btrfs_inc_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root,
		  struct extent_buffer *buf, int full_backref)
{
	return __btrfs_mod_ref(trans, root, buf, full_backref, 1);
}

int btrfs_dec_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root,
		  struct extent_buffer *buf, int full_backref)
{
	return __btrfs_mod_ref(trans, root, buf, full_backref, 0);
}

static int write_one_cache_group(struct btrfs_trans_handle *trans,
				 struct btrfs_root *root,
				 struct btrfs_path *path,
				 struct btrfs_block_group_cache *cache)
{
	int ret;
	struct btrfs_root *extent_root = root->fs_info->extent_root;
	unsigned long bi;
	struct extent_buffer *leaf;

	ret = btrfs_search_slot(trans, extent_root, &cache->key, path, 0, 1);
	if (ret < 0)
		goto fail;
	BUG_ON(ret);

	leaf = path->nodes[0];
	bi = btrfs_item_ptr_offset(leaf, path->slots[0]);
	write_extent_buffer(leaf, &cache->item, bi, sizeof(cache->item));
	btrfs_mark_buffer_dirty(leaf);
	btrfs_release_path(extent_root, path);
fail:
	if (ret)
		return ret;
	return 0;

}

int btrfs_write_dirty_block_groups(struct btrfs_trans_handle *trans,
				   struct btrfs_root *root)
{
	struct btrfs_block_group_cache *cache, *entry;
	struct rb_node *n;
	int err = 0;
	int werr = 0;
	struct btrfs_path *path;
	u64 last = 0;

	path = btrfs_alloc_path();
	if (!path)
		return -ENOMEM;

	while (1) {
		cache = NULL;
		spin_lock(&root->fs_info->block_group_cache_lock);
		for (n = rb_first(&root->fs_info->block_group_cache_tree);
		     n; n = rb_next(n)) {
			entry = rb_entry(n, struct btrfs_block_group_cache,
					 cache_node);
			if (entry->dirty) {
				cache = entry;
				break;
			}
		}
		spin_unlock(&root->fs_info->block_group_cache_lock);

		if (!cache)
			break;

		cache->dirty = 0;
		last += cache->key.offset;

		err = write_one_cache_group(trans, root,
					    path, cache);
		/*
		 * if we fail to write the cache group, we want
		 * to keep it marked dirty in hopes that a later
		 * write will work
		 */
		if (err) {
			werr = err;
			continue;
		}
	}
	btrfs_free_path(path);
	return werr;
}

int btrfs_extent_readonly(struct btrfs_root *root, u64 bytenr)
{
	struct btrfs_block_group_cache *block_group;
	int readonly = 0;

	block_group = btrfs_lookup_block_group(root->fs_info, bytenr);
	if (!block_group || block_group->ro)
		readonly = 1;
	if (block_group)
		btrfs_put_block_group(block_group);
	return readonly;
}

static int update_space_info(struct btrfs_fs_info *info, u64 flags,
			     u64 total_bytes, u64 bytes_used,
			     struct btrfs_space_info **space_info)
{
	struct btrfs_space_info *found;

	found = __find_space_info(info, flags);
	if (found) {
		spin_lock(&found->lock);
		found->total_bytes += total_bytes;
		found->bytes_used += bytes_used;
		found->full = 0;
		spin_unlock(&found->lock);
		*space_info = found;
		return 0;
	}
	found = kzalloc(sizeof(*found), GFP_NOFS);
	if (!found)
		return -ENOMEM;

	INIT_LIST_HEAD(&found->block_groups);
	init_rwsem(&found->groups_sem);
	spin_lock_init(&found->lock);
	found->flags = flags;
	found->total_bytes = total_bytes;
	found->bytes_used = bytes_used;
	found->bytes_pinned = 0;
	found->bytes_reserved = 0;
	found->bytes_readonly = 0;
	found->bytes_delalloc = 0;
	found->full = 0;
	found->force_alloc = 0;
	*space_info = found;
	list_add_rcu(&found->list, &info->space_info);
	return 0;
}

static void set_avail_alloc_bits(struct btrfs_fs_info *fs_info, u64 flags)
{
	u64 extra_flags = flags & (BTRFS_BLOCK_GROUP_RAID0 |
				   BTRFS_BLOCK_GROUP_RAID1 |
				   BTRFS_BLOCK_GROUP_RAID10 |
				   BTRFS_BLOCK_GROUP_DUP);
	if (extra_flags) {
		if (flags & BTRFS_BLOCK_GROUP_DATA)
			fs_info->avail_data_alloc_bits |= extra_flags;
		if (flags & BTRFS_BLOCK_GROUP_METADATA)
			fs_info->avail_metadata_alloc_bits |= extra_flags;
		if (flags & BTRFS_BLOCK_GROUP_SYSTEM)
			fs_info->avail_system_alloc_bits |= extra_flags;
	}
}

static void set_block_group_readonly(struct btrfs_block_group_cache *cache)
{
	spin_lock(&cache->space_info->lock);
	spin_lock(&cache->lock);
	if (!cache->ro) {
		cache->space_info->bytes_readonly += cache->key.offset -
					btrfs_block_group_used(&cache->item);
		cache->ro = 1;
	}
	spin_unlock(&cache->lock);
	spin_unlock(&cache->space_info->lock);
}

u64 btrfs_reduce_alloc_profile(struct btrfs_root *root, u64 flags)
{
	u64 num_devices = root->fs_info->fs_devices->rw_devices;

	if (num_devices == 1)
		flags &= ~(BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_RAID0);
	if (num_devices < 4)
		flags &= ~BTRFS_BLOCK_GROUP_RAID10;

	if ((flags & BTRFS_BLOCK_GROUP_DUP) &&
	    (flags & (BTRFS_BLOCK_GROUP_RAID1 |
		      BTRFS_BLOCK_GROUP_RAID10))) {
		flags &= ~BTRFS_BLOCK_GROUP_DUP;
	}

	if ((flags & BTRFS_BLOCK_GROUP_RAID1) &&
	    (flags & BTRFS_BLOCK_GROUP_RAID10)) {
		flags &= ~BTRFS_BLOCK_GROUP_RAID1;
	}

	if ((flags & BTRFS_BLOCK_GROUP_RAID0) &&
	    ((flags & BTRFS_BLOCK_GROUP_RAID1) |
	     (flags & BTRFS_BLOCK_GROUP_RAID10) |
	     (flags & BTRFS_BLOCK_GROUP_DUP)))
		flags &= ~BTRFS_BLOCK_GROUP_RAID0;
	return flags;
}

static u64 btrfs_get_alloc_profile(struct btrfs_root *root, u64 data)
{
	struct btrfs_fs_info *info = root->fs_info;
	u64 alloc_profile;

	if (data) {
		alloc_profile = info->avail_data_alloc_bits &
			info->data_alloc_profile;
		data = BTRFS_BLOCK_GROUP_DATA | alloc_profile;
	} else if (root == root->fs_info->chunk_root) {
		alloc_profile = info->avail_system_alloc_bits &
			info->system_alloc_profile;
		data = BTRFS_BLOCK_GROUP_SYSTEM | alloc_profile;
	} else {
		alloc_profile = info->avail_metadata_alloc_bits &
			info->metadata_alloc_profile;
		data = BTRFS_BLOCK_GROUP_METADATA | alloc_profile;
	}

	return btrfs_reduce_alloc_profile(root, data);
}

void btrfs_set_inode_space_info(struct btrfs_root *root, struct inode *inode)
{
	u64 alloc_target;

	alloc_target = btrfs_get_alloc_profile(root, 1);
	BTRFS_I(inode)->space_info = __find_space_info(root->fs_info,
						       alloc_target);
}

/*
 * for now this just makes sure we have at least 5% of our metadata space free
 * for use.
 */
int btrfs_check_metadata_free_space(struct btrfs_root *root)
{
	struct btrfs_fs_info *info = root->fs_info;
	struct btrfs_space_info *meta_sinfo;
	u64 alloc_target, thresh;
	int committed = 0, ret;

	/* get the space info for where the metadata will live */
	alloc_target = btrfs_get_alloc_profile(root, 0);
	meta_sinfo = __find_space_info(info, alloc_target);

again:
	spin_lock(&meta_sinfo->lock);
	if (!meta_sinfo->full)
		thresh = meta_sinfo->total_bytes * 80;
	else
		thresh = meta_sinfo->total_bytes * 95;

	do_div(thresh, 100);

	if (meta_sinfo->bytes_used + meta_sinfo->bytes_reserved +
	    meta_sinfo->bytes_pinned + meta_sinfo->bytes_readonly > thresh) {
		struct btrfs_trans_handle *trans;
		if (!meta_sinfo->full) {
			meta_sinfo->force_alloc = 1;
			spin_unlock(&meta_sinfo->lock);

			trans = btrfs_start_transaction(root, 1);
			if (!trans)
				return -ENOMEM;

			ret = do_chunk_alloc(trans, root->fs_info->extent_root,
					     2 * 1024 * 1024, alloc_target, 0);
			btrfs_end_transaction(trans, root);
			goto again;
		}
		spin_unlock(&meta_sinfo->lock);

		if (!committed) {
			committed = 1;
			trans = btrfs_join_transaction(root, 1);
			if (!trans)
				return -ENOMEM;
			ret = btrfs_commit_transaction(trans, root);
			if (ret)
				return ret;
			goto again;
		}
		return -ENOSPC;
	}
	spin_unlock(&meta_sinfo->lock);

	return 0;
}

/*
 * This will check the space that the inode allocates from to make sure we have
 * enough space for bytes.
 */
int btrfs_check_data_free_space(struct btrfs_root *root, struct inode *inode,
				u64 bytes)
{
	struct btrfs_space_info *data_sinfo;
	int ret = 0, committed = 0;

	/* make sure bytes are sectorsize aligned */
	bytes = (bytes + root->sectorsize - 1) & ~((u64)root->sectorsize - 1);

	data_sinfo = BTRFS_I(inode)->space_info;
again:
	/* make sure we have enough space to handle the data first */
	spin_lock(&data_sinfo->lock);
	if (data_sinfo->total_bytes - data_sinfo->bytes_used -
	    data_sinfo->bytes_delalloc - data_sinfo->bytes_reserved -
	    data_sinfo->bytes_pinned - data_sinfo->bytes_readonly -
	    data_sinfo->bytes_may_use < bytes) {
		struct btrfs_trans_handle *trans;

		/*
		 * if we don't have enough free bytes in this space then we need
		 * to alloc a new chunk.
		 */
		if (!data_sinfo->full) {
			u64 alloc_target;

			data_sinfo->force_alloc = 1;
			spin_unlock(&data_sinfo->lock);

			alloc_target = btrfs_get_alloc_profile(root, 1);
			trans = btrfs_start_transaction(root, 1);
			if (!trans)
				return -ENOMEM;

			ret = do_chunk_alloc(trans, root->fs_info->extent_root,
					     bytes + 2 * 1024 * 1024,
					     alloc_target, 0);
			btrfs_end_transaction(trans, root);
			if (ret)
				return ret;
			goto again;
		}
		spin_unlock(&data_sinfo->lock);

		/* commit the current transaction and try again */
		if (!committed) {
			committed = 1;
			trans = btrfs_join_transaction(root, 1);
			if (!trans)
				return -ENOMEM;
			ret = btrfs_commit_transaction(trans, root);
			if (ret)
				return ret;
			goto again;
		}

		printk(KERN_ERR "no space left, need %llu, %llu delalloc bytes"
		       ", %llu bytes_used, %llu bytes_reserved, "
		       "%llu bytes_pinned, %llu bytes_readonly, %llu may use "
		       "%llu total\n", (unsigned long long)bytes,
		       (unsigned long long)data_sinfo->bytes_delalloc,
		       (unsigned long long)data_sinfo->bytes_used,
		       (unsigned long long)data_sinfo->bytes_reserved,
		       (unsigned long long)data_sinfo->bytes_pinned,
		       (unsigned long long)data_sinfo->bytes_readonly,
		       (unsigned long long)data_sinfo->bytes_may_use,
		       (unsigned long long)data_sinfo->total_bytes);
		return -ENOSPC;
	}
	data_sinfo->bytes_may_use += bytes;
	BTRFS_I(inode)->reserved_bytes += bytes;
	spin_unlock(&data_sinfo->lock);

	return btrfs_check_metadata_free_space(root);
}

/*
 * if there was an error for whatever reason after calling
 * btrfs_check_data_free_space, call this so we can cleanup the counters.
 */
void btrfs_free_reserved_data_space(struct btrfs_root *root,
				    struct inode *inode, u64 bytes)
{
	struct btrfs_space_info *data_sinfo;

	/* make sure bytes are sectorsize aligned */
	bytes = (bytes + root->sectorsize - 1) & ~((u64)root->sectorsize - 1);

	data_sinfo = BTRFS_I(inode)->space_info;
	spin_lock(&data_sinfo->lock);
	data_sinfo->bytes_may_use -= bytes;
	BTRFS_I(inode)->reserved_bytes -= bytes;
	spin_unlock(&data_sinfo->lock);
}

/* called when we are adding a delalloc extent to the inode's io_tree */
void btrfs_delalloc_reserve_space(struct btrfs_root *root, struct inode *inode,
				  u64 bytes)
{
	struct btrfs_space_info *data_sinfo;

	/* get the space info for where this inode will be storing its data */
	data_sinfo = BTRFS_I(inode)->space_info;

	/* make sure we have enough space to handle the data first */
	spin_lock(&data_sinfo->lock);
	data_sinfo->bytes_delalloc += bytes;

	/*
	 * we are adding a delalloc extent without calling
	 * btrfs_check_data_free_space first.  This happens on a weird
	 * writepage condition, but shouldn't hurt our accounting
	 */
	if (unlikely(bytes > BTRFS_I(inode)->reserved_bytes)) {
		data_sinfo->bytes_may_use -= BTRFS_I(inode)->reserved_bytes;
		BTRFS_I(inode)->reserved_bytes = 0;
	} else {
		data_sinfo->bytes_may_use -= bytes;
		BTRFS_I(inode)->reserved_bytes -= bytes;
	}

	spin_unlock(&data_sinfo->lock);
}

/* called when we are clearing an delalloc extent from the inode's io_tree */
void btrfs_delalloc_free_space(struct btrfs_root *root, struct inode *inode,
			      u64 bytes)
{
	struct btrfs_space_info *info;

	info = BTRFS_I(inode)->space_info;

	spin_lock(&info->lock);
	info->bytes_delalloc -= bytes;
	spin_unlock(&info->lock);
}

static void force_metadata_allocation(struct btrfs_fs_info *info)
{
	struct list_head *head = &info->space_info;
	struct btrfs_space_info *found;

	rcu_read_lock();
	list_for_each_entry_rcu(found, head, list) {
		if (found->flags & BTRFS_BLOCK_GROUP_METADATA)
			found->force_alloc = 1;
	}
	rcu_read_unlock();
}

static int do_chunk_alloc(struct btrfs_trans_handle *trans,
			  struct btrfs_root *extent_root, u64 alloc_bytes,
			  u64 flags, int force)
{
	struct btrfs_space_info *space_info;
	struct btrfs_fs_info *fs_info = extent_root->fs_info;
	u64 thresh;
	int ret = 0;

	mutex_lock(&fs_info->chunk_mutex);

	flags = btrfs_reduce_alloc_profile(extent_root, flags);

	space_info = __find_space_info(extent_root->fs_info, flags);
	if (!space_info) {
		ret = update_space_info(extent_root->fs_info, flags,
					0, 0, &space_info);
		BUG_ON(ret);
	}
	BUG_ON(!space_info);

	spin_lock(&space_info->lock);
	if (space_info->force_alloc) {
		force = 1;
		space_info->force_alloc = 0;
	}
	if (space_info->full) {
		spin_unlock(&space_info->lock);
		goto out;
	}

	thresh = space_info->total_bytes - space_info->bytes_readonly;
	thresh = div_factor(thresh, 6);
	if (!force &&
	   (space_info->bytes_used + space_info->bytes_pinned +
	    space_info->bytes_reserved + alloc_bytes) < thresh) {
		spin_unlock(&space_info->lock);
		goto out;
	}
	spin_unlock(&space_info->lock);

	/*
	 * if we're doing a data chunk, go ahead and make sure that
	 * we keep a reasonable number of metadata chunks allocated in the
	 * FS as well.
	 */
	if (flags & BTRFS_BLOCK_GROUP_DATA) {
		fs_info->data_chunk_allocations++;
		if (!(fs_info->data_chunk_allocations %
		      fs_info->metadata_ratio))
			force_metadata_allocation(fs_info);
	}

	ret = btrfs_alloc_chunk(trans, extent_root, flags);
	if (ret)
		space_info->full = 1;
out:
	mutex_unlock(&extent_root->fs_info->chunk_mutex);
	return ret;
}

static int update_block_group(struct btrfs_trans_handle *trans,
			      struct btrfs_root *root,
			      u64 bytenr, u64 num_bytes, int alloc,
			      int mark_free)
{
	struct btrfs_block_group_cache *cache;
	struct btrfs_fs_info *info = root->fs_info;
	u64 total = num_bytes;
	u64 old_val;
	u64 byte_in_group;

	/* block accounting for super block */
	spin_lock(&info->delalloc_lock);
	old_val = btrfs_super_bytes_used(&info->super_copy);
	if (alloc)
		old_val += num_bytes;
	else
		old_val -= num_bytes;
	btrfs_set_super_bytes_used(&info->super_copy, old_val);

	/* block accounting for root item */
	old_val = btrfs_root_used(&root->root_item);
	if (alloc)
		old_val += num_bytes;
	else
		old_val -= num_bytes;
	btrfs_set_root_used(&root->root_item, old_val);
	spin_unlock(&info->delalloc_lock);

	while (total) {
		cache = btrfs_lookup_block_group(info, bytenr);
		if (!cache)
			return -1;
		byte_in_group = bytenr - cache->key.objectid;
		WARN_ON(byte_in_group > cache->key.offset);

		spin_lock(&cache->space_info->lock);
		spin_lock(&cache->lock);
		cache->dirty = 1;
		old_val = btrfs_block_group_used(&cache->item);
		num_bytes = min(total, cache->key.offset - byte_in_group);
		if (alloc) {
			old_val += num_bytes;
			cache->space_info->bytes_used += num_bytes;
			if (cache->ro)
				cache->space_info->bytes_readonly -= num_bytes;
			btrfs_set_block_group_used(&cache->item, old_val);
			spin_unlock(&cache->lock);
			spin_unlock(&cache->space_info->lock);
		} else {
			old_val -= num_bytes;
			cache->space_info->bytes_used -= num_bytes;
			if (cache->ro)
				cache->space_info->bytes_readonly += num_bytes;
			btrfs_set_block_group_used(&cache->item, old_val);
			spin_unlock(&cache->lock);
			spin_unlock(&cache->space_info->lock);
			if (mark_free) {
				int ret;

				ret = btrfs_discard_extent(root, bytenr,
							   num_bytes);
				WARN_ON(ret);

				ret = btrfs_add_free_space(cache, bytenr,
							   num_bytes);
				WARN_ON(ret);
			}
		}
		btrfs_put_block_group(cache);
		total -= num_bytes;
		bytenr += num_bytes;
	}
	return 0;
}

static u64 first_logical_byte(struct btrfs_root *root, u64 search_start)
{
	struct btrfs_block_group_cache *cache;
	u64 bytenr;

	cache = btrfs_lookup_first_block_group(root->fs_info, search_start);
	if (!cache)
		return 0;

	bytenr = cache->key.objectid;
	btrfs_put_block_group(cache);

	return bytenr;
}

int btrfs_update_pinned_extents(struct btrfs_root *root,
				u64 bytenr, u64 num, int pin)
{
	u64 len;
	struct btrfs_block_group_cache *cache;
	struct btrfs_fs_info *fs_info = root->fs_info;

	if (pin) {
		set_extent_dirty(&fs_info->pinned_extents,
				bytenr, bytenr + num - 1, GFP_NOFS);
	} else {
		clear_extent_dirty(&fs_info->pinned_extents,
				bytenr, bytenr + num - 1, GFP_NOFS);
	}

	while (num > 0) {
		cache = btrfs_lookup_block_group(fs_info, bytenr);
		BUG_ON(!cache);
		len = min(num, cache->key.offset -
			  (bytenr - cache->key.objectid));
		if (pin) {
			spin_lock(&cache->space_info->lock);
			spin_lock(&cache->lock);
			cache->pinned += len;
			cache->space_info->bytes_pinned += len;
			spin_unlock(&cache->lock);
			spin_unlock(&cache->space_info->lock);
			fs_info->total_pinned += len;
		} else {
			spin_lock(&cache->space_info->lock);
			spin_lock(&cache->lock);
			cache->pinned -= len;
			cache->space_info->bytes_pinned -= len;
			spin_unlock(&cache->lock);
			spin_unlock(&cache->space_info->lock);
			fs_info->total_pinned -= len;
			if (cache->cached)
				btrfs_add_free_space(cache, bytenr, len);
		}
		btrfs_put_block_group(cache);
		bytenr += len;
		num -= len;
	}
	return 0;
}

static int update_reserved_extents(struct btrfs_root *root,
				   u64 bytenr, u64 num, int reserve)
{
	u64 len;
	struct btrfs_block_group_cache *cache;
	struct btrfs_fs_info *fs_info = root->fs_info;

	while (num > 0) {
		cache = btrfs_lookup_block_group(fs_info, bytenr);
		BUG_ON(!cache);
		len = min(num, cache->key.offset -
			  (bytenr - cache->key.objectid));

		spin_lock(&cache->space_info->lock);
		spin_lock(&cache->lock);
		if (reserve) {
			cache->reserved += len;
			cache->space_info->bytes_reserved += len;
		} else {
			cache->reserved -= len;
			cache->space_info->bytes_reserved -= len;
		}
		spin_unlock(&cache->lock);
		spin_unlock(&cache->space_info->lock);
		btrfs_put_block_group(cache);
		bytenr += len;
		num -= len;
	}
	return 0;
}

int btrfs_copy_pinned(struct btrfs_root *root, struct extent_io_tree *copy)
{
	u64 last = 0;
	u64 start;
	u64 end;
	struct extent_io_tree *pinned_extents = &root->fs_info->pinned_extents;
	int ret;

	while (1) {
		ret = find_first_extent_bit(pinned_extents, last,
					    &start, &end, EXTENT_DIRTY);
		if (ret)
			break;
		set_extent_dirty(copy, start, end, GFP_NOFS);
		last = end + 1;
	}
	return 0;
}

int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans,
			       struct btrfs_root *root,
			       struct extent_io_tree *unpin)
{
	u64 start;
	u64 end;
	int ret;

	while (1) {
		ret = find_first_extent_bit(unpin, 0, &start, &end,
					    EXTENT_DIRTY);
		if (ret)
			break;

		ret = btrfs_discard_extent(root, start, end + 1 - start);

		/* unlocks the pinned mutex */
		btrfs_update_pinned_extents(root, start, end + 1 - start, 0);
		clear_extent_dirty(unpin, start, end, GFP_NOFS);

		cond_resched();
	}
	return ret;
}

static int pin_down_bytes(struct btrfs_trans_handle *trans,
			  struct btrfs_root *root,
			  struct btrfs_path *path,
			  u64 bytenr, u64 num_bytes, int is_data,
			  struct extent_buffer **must_clean)
{
	int err = 0;
	struct extent_buffer *buf;

	if (is_data)
		goto pinit;

	buf = btrfs_find_tree_block(root, bytenr, num_bytes);
	if (!buf)
		goto pinit;

	/* we can reuse a block if it hasn't been written
	 * and it is from this transaction.  We can't
	 * reuse anything from the tree log root because
	 * it has tiny sub-transactions.
	 */
	if (btrfs_buffer_uptodate(buf, 0) &&
	    btrfs_try_tree_lock(buf)) {
		u64 header_owner = btrfs_header_owner(buf);
		u64 header_transid = btrfs_header_generation(buf);
		if (header_owner != BTRFS_TREE_LOG_OBJECTID &&
		    header_transid == trans->transid &&
		    !btrfs_header_flag(buf, BTRFS_HEADER_FLAG_WRITTEN)) {
			*must_clean = buf;
			return 1;
		}
		btrfs_tree_unlock(buf);
	}
	free_extent_buffer(buf);
pinit:
	btrfs_set_path_blocking(path);
	/* unlocks the pinned mutex */
	btrfs_update_pinned_extents(root, bytenr, num_bytes, 1);

	BUG_ON(err < 0);
	return 0;
}


static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
				struct btrfs_root *root,
				u64 bytenr, u64 num_bytes, u64 parent,
				u64 root_objectid, u64 owner_objectid,
				u64 owner_offset, int refs_to_drop,
				struct btrfs_delayed_extent_op *extent_op)
{
	struct btrfs_key key;
	struct btrfs_path *path;
	struct btrfs_fs_info *info = root->fs_info;
	struct btrfs_root *extent_root = info->extent_root;
	struct extent_buffer *leaf;
	struct btrfs_extent_item *ei;
	struct btrfs_extent_inline_ref *iref;
	int ret;
	int is_data;
	int extent_slot = 0;
	int found_extent = 0;
	int num_to_del = 1;
	u32 item_size;
	u64 refs;

	path = btrfs_alloc_path();
	if (!path)
		return -ENOMEM;

	path->reada = 1;
	path->leave_spinning = 1;

	is_data = owner_objectid >= BTRFS_FIRST_FREE_OBJECTID;
	BUG_ON(!is_data && refs_to_drop != 1);

	ret = lookup_extent_backref(trans, extent_root, path, &iref,
				    bytenr, num_bytes, parent,
				    root_objectid, owner_objectid,
				    owner_offset);
	if (ret == 0) {
		extent_slot = path->slots[0];
		while (extent_slot >= 0) {
			btrfs_item_key_to_cpu(path->nodes[0], &key,
					      extent_slot);
			if (key.objectid != bytenr)
				break;
			if (key.type == BTRFS_EXTENT_ITEM_KEY &&
			    key.offset == num_bytes) {
				found_extent = 1;
				break;
			}
			if (path->slots[0] - extent_slot > 5)
				break;
			extent_slot--;
		}
#ifdef BTRFS_COMPAT_EXTENT_TREE_V0
		item_size = btrfs_item_size_nr(path->nodes[0], extent_slot);
		if (found_extent && item_size < sizeof(*ei))
			found_extent = 0;
#endif
		if (!found_extent) {
			BUG_ON(iref);
			ret = remove_extent_backref(trans, extent_root, path,
						    NULL, refs_to_drop,
						    is_data);
			BUG_ON(ret);
			btrfs_release_path(extent_root, path);
			path->leave_spinning = 1;

			key.objectid = bytenr;
			key.type = BTRFS_EXTENT_ITEM_KEY;
			key.offset = num_bytes;

			ret = btrfs_search_slot(trans, extent_root,
						&key, path, -1, 1);
			if (ret) {
				printk(KERN_ERR "umm, got %d back from search"
				       ", was looking for %llu\n", ret,
				       (unsigned long long)bytenr);
				btrfs_print_leaf(extent_root, path->nodes[0]);
			}
			BUG_ON(ret);
			extent_slot = path->slots[0];
		}
	} else {
		btrfs_print_leaf(extent_root, path->nodes[0]);
		WARN_ON(1);
		printk(KERN_ERR "btrfs unable to find ref byte nr %llu "
		       "parent %llu root %llu  owner %llu offset %llu\n",
		       (unsigned long long)bytenr,
		       (unsigned long long)parent,
		       (unsigned long long)root_objectid,
		       (unsigned long long)owner_objectid,
		       (unsigned long long)owner_offset);
	}

	leaf = path->nodes[0];
	item_size = btrfs_item_size_nr(leaf, extent_slot);
#ifdef BTRFS_COMPAT_EXTENT_TREE_V0
	if (item_size < sizeof(*ei)) {
		BUG_ON(found_extent || extent_slot != path->slots[0]);
		ret = convert_extent_item_v0(trans, extent_root, path,
					     owner_objectid, 0);
		BUG_ON(ret < 0);

		btrfs_release_path(extent_root, path);
		path->leave_spinning = 1;

		key.objectid = bytenr;
		key.type = BTRFS_EXTENT_ITEM_KEY;
		key.offset = num_bytes;

		ret = btrfs_search_slot(trans, extent_root, &key, path,
					-1, 1);
		if (ret) {
			printk(KERN_ERR "umm, got %d back from search"
			       ", was looking for %llu\n", ret,
			       (unsigned long long)bytenr);
			btrfs_print_leaf(extent_root, path->nodes[0]);
		}
		BUG_ON(ret);
		extent_slot = path->slots[0];
		leaf = path->nodes[0];
		item_size = btrfs_item_size_nr(leaf, extent_slot);
	}
#endif
	BUG_ON(item_size < sizeof(*ei));
	ei = btrfs_item_ptr(leaf, extent_slot,
			    struct btrfs_extent_item);
	if (owner_objectid < BTRFS_FIRST_FREE_OBJECTID) {
		struct btrfs_tree_block_info *bi;
		BUG_ON(item_size < sizeof(*ei) + sizeof(*bi));
		bi = (struct btrfs_tree_block_info *)(ei + 1);
		WARN_ON(owner_objectid != btrfs_tree_block_level(leaf, bi));
	}

	refs = btrfs_extent_refs(leaf, ei);
	BUG_ON(refs < refs_to_drop);
	refs -= refs_to_drop;

	if (refs > 0) {
		if (extent_op)
			__run_delayed_extent_op(extent_op, leaf, ei);
		/*
		 * In the case of inline back ref, reference count will
		 * be updated by remove_extent_backref
		 */
		if (iref) {
			BUG_ON(!found_extent);
		} else {
			btrfs_set_extent_refs(leaf, ei, refs);
			btrfs_mark_buffer_dirty(leaf);
		}
		if (found_extent) {
			ret = remove_extent_backref(trans, extent_root, path,
						    iref, refs_to_drop,
						    is_data);
			BUG_ON(ret);
		}
	} else {
		int mark_free = 0;
		struct extent_buffer *must_clean = NULL;

		if (found_extent) {
			BUG_ON(is_data && refs_to_drop !=
			       extent_data_ref_count(root, path, iref));
			if (iref) {
				BUG_ON(path->slots[0] != extent_slot);
			} else {
				BUG_ON(path->slots[0] != extent_slot + 1);
				path->slots[0] = extent_slot;
				num_to_del = 2;
			}
		}

		ret = pin_down_bytes(trans, root, path, bytenr,
				     num_bytes, is_data, &must_clean);
		if (ret > 0)
			mark_free = 1;
		BUG_ON(ret < 0);
		/*
		 * it is going to be very rare for someone to be waiting
		 * on the block we're freeing.  del_items might need to
		 * schedule, so rather than get fancy, just force it
		 * to blocking here
		 */
		if (must_clean)
			btrfs_set_lock_blocking(must_clean);

		ret = btrfs_del_items(trans, extent_root, path, path->slots[0],
				      num_to_del);
		BUG_ON(ret);
		btrfs_release_path(extent_root, path);

		if (must_clean) {
			clean_tree_block(NULL, root, must_clean);
			btrfs_tree_unlock(must_clean);
			free_extent_buffer(must_clean);
		}

		if (is_data) {
			ret = btrfs_del_csums(trans, root, bytenr, num_bytes);
			BUG_ON(ret);
		} else {
			invalidate_mapping_pages(info->btree_inode->i_mapping,
			     bytenr >> PAGE_CACHE_SHIFT,
			     (bytenr + num_bytes - 1) >> PAGE_CACHE_SHIFT);
		}

		ret = update_block_group(trans, root, bytenr, num_bytes, 0,
					 mark_free);
		BUG_ON(ret);
	}
	btrfs_free_path(path);
	return ret;
}

/*
 * when we free an extent, it is possible (and likely) that we free the last
 * delayed ref for that extent as well.  This searches the delayed ref tree for
 * a given extent, and if there are no other delayed refs to be processed, it
 * removes it from the tree.
 */
static noinline int check_ref_cleanup(struct btrfs_trans_handle *trans,
				      struct btrfs_root *root, u64 bytenr)
{
	struct btrfs_delayed_ref_head *head;
	struct btrfs_delayed_ref_root *delayed_refs;
	struct btrfs_delayed_ref_node *ref;
	struct rb_node *node;
	int ret;

	delayed_refs = &trans->transaction->delayed_refs;
	spin_lock(&delayed_refs->lock);
	head = btrfs_find_delayed_ref_head(trans, bytenr);
	if (!head)
		goto out;

	node = rb_prev(&head->node.rb_node);
	if (!node)
		goto out;

	ref = rb_entry(node, struct btrfs_delayed_ref_node, rb_node);

	/* there are still entries for this ref, we can't drop it */
	if (ref->bytenr == bytenr)
		goto out;

	if (head->extent_op) {
		if (!head->must_insert_reserved)
			goto out;
		kfree(head->extent_op);
		head->extent_op = NULL;
	}

	/*
	 * waiting for the lock here would deadlock.  If someone else has it
	 * locked they are already in the process of dropping it anyway
	 */
	if (!mutex_trylock(&head->mutex))
		goto out;

	/*
	 * at this point we have a head with no other entries.  Go
	 * ahead and process it.
	 */
	head->node.in_tree = 0;
	rb_erase(&head->node.rb_node, &delayed_refs->root);

	delayed_refs->num_entries--;

	/*
	 * we don't take a ref on the node because we're removing it from the
	 * tree, so we just steal the ref the tree was holding.
	 */
	delayed_refs->num_heads--;
	if (list_empty(&head->cluster))
		delayed_refs->num_heads_ready--;

	list_del_init(&head->cluster);
	spin_unlock(&delayed_refs->lock);

	ret = run_one_delayed_ref(trans, root->fs_info->tree_root,
				  &head->node, head->extent_op,
				  head->must_insert_reserved);
	BUG_ON(ret);
	btrfs_put_delayed_ref(&head->node);
	return 0;
out:
	spin_unlock(&delayed_refs->lock);
	return 0;
}

int btrfs_free_extent(struct btrfs_trans_handle *trans,
		      struct btrfs_root *root,
		      u64 bytenr, u64 num_bytes, u64 parent,
		      u64 root_objectid, u64 owner, u64 offset)
{
	int ret;

	/*
	 * tree log blocks never actually go into the extent allocation
	 * tree, just update pinning info and exit early.
	 */
	if (root_objectid == BTRFS_TREE_LOG_OBJECTID) {
		WARN_ON(owner >= BTRFS_FIRST_FREE_OBJECTID);
		/* unlocks the pinned mutex */
		btrfs_update_pinned_extents(root, bytenr, num_bytes, 1);
		update_reserved_extents(root, bytenr, num_bytes, 0);
		ret = 0;
	} else if (owner < BTRFS_FIRST_FREE_OBJECTID) {
		ret = btrfs_add_delayed_tree_ref(trans, bytenr, num_bytes,
					parent, root_objectid, (int)owner,
					BTRFS_DROP_DELAYED_REF, NULL);
		BUG_ON(ret);
		ret = check_ref_cleanup(trans, root, bytenr);
		BUG_ON(ret);
	} else {
		ret = btrfs_add_delayed_data_ref(trans, bytenr, num_bytes,
					parent, root_objectid, owner,
					offset, BTRFS_DROP_DELAYED_REF, NULL);
		BUG_ON(ret);
	}
	return ret;
}

static u64 stripe_align(struct btrfs_root *root, u64 val)
{
	u64 mask = ((u64)root->stripesize - 1);
	u64 ret = (val + mask) & ~mask;
	return ret;
}

/*
 * walks the btree of allocated extents and find a hole of a given size.
 * The key ins is changed to record the hole:
 * ins->objectid == block start
 * ins->flags = BTRFS_EXTENT_ITEM_KEY
 * ins->offset == number of blocks
 * Any available blocks before search_start are skipped.
 */
static noinline int find_free_extent(struct btrfs_trans_handle *trans,
				     struct btrfs_root *orig_root,
				     u64 num_bytes, u64 empty_size,
				     u64 search_start, u64 search_end,
				     u64 hint_byte, struct btrfs_key *ins,
				     u64 exclude_start, u64 exclude_nr,
				     int data)
{
	int ret = 0;
	struct btrfs_root *root = orig_root->fs_info->extent_root;
	struct btrfs_free_cluster *last_ptr = NULL;
	struct btrfs_block_group_cache *block_group = NULL;
	int empty_cluster = 2 * 1024 * 1024;
	int allowed_chunk_alloc = 0;
	struct btrfs_space_info *space_info;
	int last_ptr_loop = 0;
	int loop = 0;

	WARN_ON(num_bytes < root->sectorsize);
	btrfs_set_key_type(ins, BTRFS_EXTENT_ITEM_KEY);
	ins->objectid = 0;
	ins->offset = 0;

	space_info = __find_space_info(root->fs_info, data);

	if (orig_root->ref_cows || empty_size)
		allowed_chunk_alloc = 1;

	if (data & BTRFS_BLOCK_GROUP_METADATA) {
		last_ptr = &root->fs_info->meta_alloc_cluster;
		if (!btrfs_test_opt(root, SSD))
			empty_cluster = 64 * 1024;
	}

	if ((data & BTRFS_BLOCK_GROUP_DATA) && btrfs_test_opt(root, SSD)) {
		last_ptr = &root->fs_info->data_alloc_cluster;
	}

	if (last_ptr) {
		spin_lock(&last_ptr->lock);
		if (last_ptr->block_group)
			hint_byte = last_ptr->window_start;
		spin_unlock(&last_ptr->lock);
	}

	search_start = max(search_start, first_logical_byte(root, 0));
	search_start = max(search_start, hint_byte);

	if (!last_ptr) {
		empty_cluster = 0;
		loop = 1;
	}

	if (search_start == hint_byte) {
		block_group = btrfs_lookup_block_group(root->fs_info,
						       search_start);
		if (block_group && block_group_bits(block_group, data)) {
			down_read(&space_info->groups_sem);
			if (list_empty(&block_group->list) ||
			    block_group->ro) {
				/*
				 * someone is removing this block group,
				 * we can't jump into the have_block_group
				 * target because our list pointers are not
				 * valid
				 */
				btrfs_put_block_group(block_group);
				up_read(&space_info->groups_sem);
			} else
				goto have_block_group;
		} else if (block_group) {
			btrfs_put_block_group(block_group);
		}
	}

search:
	down_read(&space_info->groups_sem);
	list_for_each_entry(block_group, &space_info->block_groups, list) {
		u64 offset;

		atomic_inc(&block_group->count);
		search_start = block_group->key.objectid;

have_block_group:
		if (unlikely(!block_group->cached)) {
			mutex_lock(&block_group->cache_mutex);
			ret = cache_block_group(root, block_group);
			mutex_unlock(&block_group->cache_mutex);
			if (ret) {
				btrfs_put_block_group(block_group);
				break;
			}
		}

		if (unlikely(block_group->ro))
			goto loop;

		if (last_ptr) {
			/*
			 * the refill lock keeps out other
			 * people trying to start a new cluster
			 */
			spin_lock(&last_ptr->refill_lock);
			if (last_ptr->block_group &&
			    (last_ptr->block_group->ro ||
			    !block_group_bits(last_ptr->block_group, data))) {
				offset = 0;
				goto refill_cluster;
			}

			offset = btrfs_alloc_from_cluster(block_group, last_ptr,
						 num_bytes, search_start);
			if (offset) {
				/* we have a block, we're done */
				spin_unlock(&last_ptr->refill_lock);
				goto checks;
			}

			spin_lock(&last_ptr->lock);
			/*
			 * whoops, this cluster doesn't actually point to
			 * this block group.  Get a ref on the block
			 * group is does point to and try again
			 */
			if (!last_ptr_loop && last_ptr->block_group &&
			    last_ptr->block_group != block_group) {

				btrfs_put_block_group(block_group);
				block_group = last_ptr->block_group;
				atomic_inc(&block_group->count);
				spin_unlock(&last_ptr->lock);
				spin_unlock(&last_ptr->refill_lock);

				last_ptr_loop = 1;
				search_start = block_group->key.objectid;
				/*
				 * we know this block group is properly
				 * in the list because
				 * btrfs_remove_block_group, drops the
				 * cluster before it removes the block
				 * group from the list
				 */
				goto have_block_group;
			}
			spin_unlock(&last_ptr->lock);
refill_cluster:
			/*
			 * this cluster didn't work out, free it and
			 * start over
			 */
			btrfs_return_cluster_to_free_space(NULL, last_ptr);

			last_ptr_loop = 0;

			/* allocate a cluster in this block group */
			ret = btrfs_find_space_cluster(trans, root,
					       block_group, last_ptr,
					       offset, num_bytes,
					       empty_cluster + empty_size);
			if (ret == 0) {
				/*
				 * now pull our allocation out of this
				 * cluster
				 */
				offset = btrfs_alloc_from_cluster(block_group,
						  last_ptr, num_bytes,
						  search_start);
				if (offset) {
					/* we found one, proceed */
					spin_unlock(&last_ptr->refill_lock);
					goto checks;
				}
			}
			/*
			 * at this point we either didn't find a cluster
			 * or we weren't able to allocate a block from our
			 * cluster.  Free the cluster we've been trying
			 * to use, and go to the next block group
			 */
			if (loop < 2) {
				btrfs_return_cluster_to_free_space(NULL,
								   last_ptr);
				spin_unlock(&last_ptr->refill_lock);
				goto loop;
			}
			spin_unlock(&last_ptr->refill_lock);
		}

		offset = btrfs_find_space_for_alloc(block_group, search_start,
						    num_bytes, empty_size);
		if (!offset)
			goto loop;
checks:
		search_start = stripe_align(root, offset);

		/* move on to the next group */
		if (search_start + num_bytes >= search_end) {
			btrfs_add_free_space(block_group, offset, num_bytes);
			goto loop;
		}

		/* move on to the next group */
		if (search_start + num_bytes >
		    block_group->key.objectid + block_group->key.offset) {
			btrfs_add_free_space(block_group, offset, num_bytes);
			goto loop;
		}

		if (exclude_nr > 0 &&
		    (search_start + num_bytes > exclude_start &&
		     search_start < exclude_start + exclude_nr)) {
			search_start = exclude_start + exclude_nr;

			btrfs_add_free_space(block_group, offset, num_bytes);
			/*
			 * if search_start is still in this block group
			 * then we just re-search this block group
			 */
			if (search_start >= block_group->key.objectid &&
			    search_start < (block_group->key.objectid +
					    block_group->key.offset))
				goto have_block_group;
			goto loop;
		}

		ins->objectid = search_start;
		ins->offset = num_bytes;

		if (offset < search_start)
			btrfs_add_free_space(block_group, offset,
					     search_start - offset);
		BUG_ON(offset > search_start);

		/* we are all good, lets return */
		break;
loop:
		btrfs_put_block_group(block_group);
	}
	up_read(&space_info->groups_sem);

	/* loop == 0, try to find a clustered alloc in every block group
	 * loop == 1, try again after forcing a chunk allocation
	 * loop == 2, set empty_size and empty_cluster to 0 and try again
	 */
	if (!ins->objectid && loop < 3 &&
	    (empty_size || empty_cluster || allowed_chunk_alloc)) {
		if (loop >= 2) {
			empty_size = 0;
			empty_cluster = 0;
		}

		if (allowed_chunk_alloc) {
			ret = do_chunk_alloc(trans, root, num_bytes +
					     2 * 1024 * 1024, data, 1);
			allowed_chunk_alloc = 0;
		} else {
			space_info->force_alloc = 1;
		}

		if (loop < 3) {
			loop++;
			goto search;
		}
		ret = -ENOSPC;
	} else if (!ins->objectid) {
		ret = -ENOSPC;
	}

	/* we found what we needed */
	if (ins->objectid) {
		if (!(data & BTRFS_BLOCK_GROUP_DATA))
			trans->block_group = block_group->key.objectid;

		btrfs_put_block_group(block_group);
		ret = 0;
	}

	return ret;
}

static void dump_space_info(struct btrfs_space_info *info, u64 bytes)
{
	struct btrfs_block_group_cache *cache;

	printk(KERN_INFO "space_info has %llu free, is %sfull\n",
	       (unsigned long long)(info->total_bytes - info->bytes_used -
				    info->bytes_pinned - info->bytes_reserved),
	       (info->full) ? "" : "not ");
	printk(KERN_INFO "space_info total=%llu, pinned=%llu, delalloc=%llu,"
	       " may_use=%llu, used=%llu\n",
	       (unsigned long long)info->total_bytes,
	       (unsigned long long)info->bytes_pinned,
	       (unsigned long long)info->bytes_delalloc,
	       (unsigned long long)info->bytes_may_use,
	       (unsigned long long)info->bytes_used);

	down_read(&info->groups_sem);
	list_for_each_entry(cache, &info->block_groups, list) {
		spin_lock(&cache->lock);
		printk(KERN_INFO "block group %llu has %llu bytes, %llu used "
		       "%llu pinned %llu reserved\n",
		       (unsigned long long)cache->key.objectid,
		       (unsigned long long)cache->key.offset,
		       (unsigned long long)btrfs_block_group_used(&cache->item),
		       (unsigned long long)cache->pinned,
		       (unsigned long long)cache->reserved);
		btrfs_dump_free_space(cache, bytes);
		spin_unlock(&cache->lock);
	}
	up_read(&info->groups_sem);
}

static int __btrfs_reserve_extent(struct btrfs_trans_handle *trans,
				  struct btrfs_root *root,
				  u64 num_bytes, u64 min_alloc_size,
				  u64 empty_size, u64 hint_byte,
				  u64 search_end, struct btrfs_key *ins,
				  u64 data)
{
	int ret;
	u64 search_start = 0;
	struct btrfs_fs_info *info = root->fs_info;

	data = btrfs_get_alloc_profile(root, data);
again:
	/*
	 * the only place that sets empty_size is btrfs_realloc_node, which
	 * is not called recursively on allocations
	 */
	if (empty_size || root->ref_cows) {
		if (!(data & BTRFS_BLOCK_GROUP_METADATA)) {
			ret = do_chunk_alloc(trans, root->fs_info->extent_root,
				     2 * 1024 * 1024,
				     BTRFS_BLOCK_GROUP_METADATA |
				     (info->metadata_alloc_profile &
				      info->avail_metadata_alloc_bits), 0);
		}
		ret = do_chunk_alloc(trans, root->fs_info->extent_root,
				     num_bytes + 2 * 1024 * 1024, data, 0);
	}

	WARN_ON(num_bytes < root->sectorsize);
	ret = find_free_extent(trans, root, num_bytes, empty_size,
			       search_start, search_end, hint_byte, ins,
			       trans->alloc_exclude_start,
			       trans->alloc_exclude_nr, data);

	if (ret == -ENOSPC && num_bytes > min_alloc_size) {
		num_bytes = num_bytes >> 1;
		num_bytes = num_bytes & ~(root->sectorsize - 1);
		num_bytes = max(num_bytes, min_alloc_size);
		do_chunk_alloc(trans, root->fs_info->extent_root,
			       num_bytes, data, 1);
		goto again;
	}
	if (ret) {
		struct btrfs_space_info *sinfo;

		sinfo = __find_space_info(root->fs_info, data);
		printk(KERN_ERR "btrfs allocation failed flags %llu, "
		       "wanted %llu\n", (unsigned long long)data,
		       (unsigned long long)num_bytes);
		dump_space_info(sinfo, num_bytes);
		BUG();
	}

	return ret;
}

int btrfs_free_reserved_extent(struct btrfs_root *root, u64 start, u64 len)
{
	struct btrfs_block_group_cache *cache;
	int ret = 0;

	cache = btrfs_lookup_block_group(root->fs_info, start);
	if (!cache) {
		printk(KERN_ERR "Unable to find block group for %llu\n",
		       (unsigned long long)start);
		return -ENOSPC;
	}

	ret = btrfs_discard_extent(root, start, len);

	btrfs_add_free_space(cache, start, len);
	btrfs_put_block_group(cache);
	update_reserved_extents(root, start, len, 0);

	return ret;
}

int btrfs_reserve_extent(struct btrfs_trans_handle *trans,
				  struct btrfs_root *root,
				  u64 num_bytes, u64 min_alloc_size,
				  u64 empty_size, u64 hint_byte,
				  u64 search_end, struct btrfs_key *ins,
				  u64 data)
{
	int ret;
	ret = __btrfs_reserve_extent(trans, root, num_bytes, min_alloc_size,
				     empty_size, hint_byte, search_end, ins,
				     data);
	update_reserved_extents(root, ins->objectid, ins->offset, 1);
	return ret;
}

static int alloc_reserved_file_extent(struct btrfs_trans_handle *trans,
				      struct btrfs_root *root,
				      u64 parent, u64 root_objectid,
				      u64 flags, u64 owner, u64 offset,
				      struct btrfs_key *ins, int ref_mod)
{
	int ret;
	struct btrfs_fs_info *fs_info = root->fs_info;
	struct btrfs_extent_item *extent_item;
	struct btrfs_extent_inline_ref *iref;
	struct btrfs_path *path;
	struct extent_buffer *leaf;
	int type;
	u32 size;

	if (parent > 0)
		type = BTRFS_SHARED_DATA_REF_KEY;
	else
		type = BTRFS_EXTENT_DATA_REF_KEY;

	size = sizeof(*extent_item) + btrfs_extent_inline_ref_size(type);

	path = btrfs_alloc_path();
	BUG_ON(!path);

	path->leave_spinning = 1;
	ret = btrfs_insert_empty_item(trans, fs_info->extent_root, path,
				      ins, size);
	BUG_ON(ret);

	leaf = path->nodes[0];
	extent_item = btrfs_item_ptr(leaf, path->slots[0],
				     struct btrfs_extent_item);
	btrfs_set_extent_refs(leaf, extent_item, ref_mod);
	btrfs_set_extent_generation(leaf, extent_item, trans->transid);
	btrfs_set_extent_flags(leaf, extent_item,
			       flags | BTRFS_EXTENT_FLAG_DATA);

	iref = (struct btrfs_extent_inline_ref *)(extent_item + 1);
	btrfs_set_extent_inline_ref_type(leaf, iref, type);
	if (parent > 0) {
		struct btrfs_shared_data_ref *ref;
		ref = (struct btrfs_shared_data_ref *)(iref + 1);
		btrfs_set_extent_inline_ref_offset(leaf, iref, parent);
		btrfs_set_shared_data_ref_count(leaf, ref, ref_mod);
	} else {
		struct btrfs_extent_data_ref *ref;
		ref = (struct btrfs_extent_data_ref *)(&iref->offset);
		btrfs_set_extent_data_ref_root(leaf, ref, root_objectid);
		btrfs_set_extent_data_ref_objectid(leaf, ref, owner);
		btrfs_set_extent_data_ref_offset(leaf, ref, offset);
		btrfs_set_extent_data_ref_count(leaf, ref, ref_mod);
	}

	btrfs_mark_buffer_dirty(path->nodes[0]);
	btrfs_free_path(path);

	ret = update_block_group(trans, root, ins->objectid, ins->offset,
				 1, 0);
	if (ret) {
		printk(KERN_ERR "btrfs update block group failed for %llu "
		       "%llu\n", (unsigned long long)ins->objectid,
		       (unsigned long long)ins->offset);
		BUG();
	}
	return ret;
}

static int alloc_reserved_tree_block(struct btrfs_trans_handle *trans,
				     struct btrfs_root *root,
				     u64 parent, u64 root_objectid,
				     u64 flags, struct btrfs_disk_key *key,
				     int level, struct btrfs_key *ins)
{
	int ret;
	struct btrfs_fs_info *fs_info = root->fs_info;
	struct btrfs_extent_item *extent_item;
	struct btrfs_tree_block_info *block_info;
	struct btrfs_extent_inline_ref *iref;
	struct btrfs_path *path;
	struct extent_buffer *leaf;
	u32 size = sizeof(*extent_item) + sizeof(*block_info) + sizeof(*iref);

	path = btrfs_alloc_path();
	BUG_ON(!path);

	path->leave_spinning = 1;
	ret = btrfs_insert_empty_item(trans, fs_info->extent_root, path,
				      ins, size);
	BUG_ON(ret);

	leaf = path->nodes[0];
	extent_item = btrfs_item_ptr(leaf, path->slots[0],
				     struct btrfs_extent_item);
	btrfs_set_extent_refs(leaf, extent_item, 1);
	btrfs_set_extent_generation(leaf, extent_item, trans->transid);
	btrfs_set_extent_flags(leaf, extent_item,
			       flags | BTRFS_EXTENT_FLAG_TREE_BLOCK);
	block_info = (struct btrfs_tree_block_info *)(extent_item + 1);

	btrfs_set_tree_block_key(leaf, block_info, key);
	btrfs_set_tree_block_level(leaf, block_info, level);

	iref = (struct btrfs_extent_inline_ref *)(block_info + 1);
	if (parent > 0) {
		BUG_ON(!(flags & BTRFS_BLOCK_FLAG_FULL_BACKREF));
		btrfs_set_extent_inline_ref_type(leaf, iref,
						 BTRFS_SHARED_BLOCK_REF_KEY);
		btrfs_set_extent_inline_ref_offset(leaf, iref, parent);
	} else {
		btrfs_set_extent_inline_ref_type(leaf, iref,
						 BTRFS_TREE_BLOCK_REF_KEY);
		btrfs_set_extent_inline_ref_offset(leaf, iref, root_objectid);
	}

	btrfs_mark_buffer_dirty(leaf);
	btrfs_free_path(path);

	ret = update_block_group(trans, root, ins->objectid, ins->offset,
				 1, 0);
	if (ret) {
		printk(KERN_ERR "btrfs update block group failed for %llu "
		       "%llu\n", (unsigned long long)ins->objectid,
		       (unsigned long long)ins->offset);
		BUG();
	}
	return ret;
}

int btrfs_alloc_reserved_file_extent(struct btrfs_trans_handle *trans,
				     struct btrfs_root *root,
				     u64 root_objectid, u64 owner,
				     u64 offset, struct btrfs_key *ins)
{
	int ret;

	BUG_ON(root_objectid == BTRFS_TREE_LOG_OBJECTID);

	ret = btrfs_add_delayed_data_ref(trans, ins->objectid, ins->offset,
					 0, root_objectid, owner, offset,
					 BTRFS_ADD_DELAYED_EXTENT, NULL);
	return ret;
}

/*
 * this is used by the tree logging recovery code.  It records that
 * an extent has been allocated and makes sure to clear the free
 * space cache bits as well
 */
int btrfs_alloc_logged_file_extent(struct btrfs_trans_handle *trans,
				   struct btrfs_root *root,
				   u64 root_objectid, u64 owner, u64 offset,
				   struct btrfs_key *ins)
{
	int ret;
	struct btrfs_block_group_cache *block_group;

	block_group = btrfs_lookup_block_group(root->fs_info, ins->objectid);
	mutex_lock(&block_group->cache_mutex);
	cache_block_group(root, block_group);
	mutex_unlock(&block_group->cache_mutex);

	ret = btrfs_remove_free_space(block_group, ins->objectid,
				      ins->offset);
	BUG_ON(ret);
	btrfs_put_block_group(block_group);
	ret = alloc_reserved_file_extent(trans, root, 0, root_objectid,
					 0, owner, offset, ins, 1);
	return ret;
}

/*
 * finds a free extent and does all the dirty work required for allocation
 * returns the key for the extent through ins, and a tree buffer for
 * the first block of the extent through buf.
 *
 * returns 0 if everything worked, non-zero otherwise.
 */
static int alloc_tree_block(struct btrfs_trans_handle *trans,
			    struct btrfs_root *root,
			    u64 num_bytes, u64 parent, u64 root_objectid,
			    struct btrfs_disk_key *key, int level,
			    u64 empty_size, u64 hint_byte, u64 search_end,
			    struct btrfs_key *ins)
{
	int ret;
	u64 flags = 0;

	ret = __btrfs_reserve_extent(trans, root, num_bytes, num_bytes,
				     empty_size, hint_byte, search_end,
				     ins, 0);
	BUG_ON(ret);

	if (root_objectid == BTRFS_TREE_RELOC_OBJECTID) {
		if (parent == 0)
			parent = ins->objectid;
		flags |= BTRFS_BLOCK_FLAG_FULL_BACKREF;
	} else
		BUG_ON(parent > 0);

	update_reserved_extents(root, ins->objectid, ins->offset, 1);
	if (root_objectid != BTRFS_TREE_LOG_OBJECTID) {
		struct btrfs_delayed_extent_op *extent_op;
		extent_op = kmalloc(sizeof(*extent_op), GFP_NOFS);
		BUG_ON(!extent_op);
		if (key)
			memcpy(&extent_op->key, key, sizeof(extent_op->key));
		else
			memset(&extent_op->key, 0, sizeof(extent_op->key));
		extent_op->flags_to_set = flags;
		extent_op->update_key = 1;
		extent_op->update_flags = 1;
		extent_op->is_data = 0;

		ret = btrfs_add_delayed_tree_ref(trans, ins->objectid,
					ins->offset, parent, root_objectid,
					level, BTRFS_ADD_DELAYED_EXTENT,
					extent_op);
		BUG_ON(ret);
	}
	return ret;
}

struct extent_buffer *btrfs_init_new_buffer(struct btrfs_trans_handle *trans,
					    struct btrfs_root *root,
					    u64 bytenr, u32 blocksize,
					    int level)
{
	struct extent_buffer *buf;

	buf = btrfs_find_create_tree_block(root, bytenr, blocksize);
	if (!buf)
		return ERR_PTR(-ENOMEM);
	btrfs_set_header_generation(buf, trans->transid);
	btrfs_set_buffer_lockdep_class(buf, level);
	btrfs_tree_lock(buf);
	clean_tree_block(trans, root, buf);

	btrfs_set_lock_blocking(buf);
	btrfs_set_buffer_uptodate(buf);

	if (root->root_key.objectid == BTRFS_TREE_LOG_OBJECTID) {
		set_extent_dirty(&root->dirty_log_pages, buf->start,
			 buf->start + buf->len - 1, GFP_NOFS);
	} else {
		set_extent_dirty(&trans->transaction->dirty_pages, buf->start,
			 buf->start + buf->len - 1, GFP_NOFS);
	}
	trans->blocks_used++;
	/* this returns a buffer locked for blocking */
	return buf;
}

/*
 * helper function to allocate a block for a given tree
 * returns the tree buffer or NULL.
 */
struct extent_buffer *btrfs_alloc_free_block(struct btrfs_trans_handle *trans,
					struct btrfs_root *root, u32 blocksize,
					u64 parent, u64 root_objectid,
					struct btrfs_disk_key *key, int level,
					u64 hint, u64 empty_size)
{
	struct btrfs_key ins;
	int ret;
	struct extent_buffer *buf;

	ret = alloc_tree_block(trans, root, blocksize, parent, root_objectid,
			       key, level, empty_size, hint, (u64)-1, &ins);
	if (ret) {
		BUG_ON(ret > 0);
		return ERR_PTR(ret);
	}

	buf = btrfs_init_new_buffer(trans, root, ins.objectid,
				    blocksize, level);
	return buf;
}

#if 0
int btrfs_drop_leaf_ref(struct btrfs_trans_handle *trans,
			struct btrfs_root *root, struct extent_buffer *leaf)
{
	u64 disk_bytenr;
	u64 num_bytes;
	struct btrfs_key key;
	struct btrfs_file_extent_item *fi;
	u32 nritems;
	int i;
	int ret;

	BUG_ON(!btrfs_is_leaf(leaf));
	nritems = btrfs_header_nritems(leaf);

	for (i = 0; i < nritems; i++) {
		cond_resched();
		btrfs_item_key_to_cpu(leaf, &key, i);

		/* only extents have references, skip everything else */
		if (btrfs_key_type(&key) != BTRFS_EXTENT_DATA_KEY)
			continue;

		fi = btrfs_item_ptr(leaf, i, struct btrfs_file_extent_item);

		/* inline extents live in the btree, they don't have refs */
		if (btrfs_file_extent_type(leaf, fi) ==
		    BTRFS_FILE_EXTENT_INLINE)
			continue;

		disk_bytenr = btrfs_file_extent_disk_bytenr(leaf, fi);

		/* holes don't have refs */
		if (disk_bytenr == 0)
			continue;

		num_bytes = btrfs_file_extent_disk_num_bytes(leaf, fi);
		ret = btrfs_free_extent(trans, root, disk_bytenr, num_bytes,
					leaf->start, 0, key.objectid, 0);
		BUG_ON(ret);
	}
	return 0;
}

static noinline int cache_drop_leaf_ref(struct btrfs_trans_handle *trans,
					struct btrfs_root *root,
					struct btrfs_leaf_ref *ref)
{
	int i;
	int ret;
	struct btrfs_extent_info *info;
	struct refsort *sorted;

	if (ref->nritems == 0)
		return 0;

	sorted = kmalloc(sizeof(*sorted) * ref->nritems, GFP_NOFS);
	for (i = 0; i < ref->nritems; i++) {
		sorted[i].bytenr = ref->extents[i].bytenr;
		sorted[i].slot = i;
	}
	sort(sorted, ref->nritems, sizeof(struct refsort), refsort_cmp, NULL);

	/*
	 * the items in the ref were sorted when the ref was inserted
	 * into the ref cache, so this is already in order
	 */
	for (i = 0; i < ref->nritems; i++) {
		info = ref->extents + sorted[i].slot;
		ret = btrfs_free_extent(trans, root, info->bytenr,
					  info->num_bytes, ref->bytenr,
					  ref->owner, ref->generation,
					  info->objectid, 0);

		atomic_inc(&root->fs_info->throttle_gen);
		wake_up(&root->fs_info->transaction_throttle);
		cond_resched();

		BUG_ON(ret);
		info++;
	}

	kfree(sorted);
	return 0;
}


static int drop_snap_lookup_refcount(struct btrfs_trans_handle *trans,
				     struct btrfs_root *root, u64 start,
				     u64 len, u32 *refs)
{
	int ret;

	ret = btrfs_lookup_extent_refs(trans, root, start, len, refs);
	BUG_ON(ret);

#if 0 /* some debugging code in case we see problems here */
	/* if the refs count is one, it won't get increased again.  But
	 * if the ref count is > 1, someone may be decreasing it at
	 * the same time we are.
	 */
	if (*refs != 1) {
		struct extent_buffer *eb = NULL;
		eb = btrfs_find_create_tree_block(root, start, len);
		if (eb)
			btrfs_tree_lock(eb);

		mutex_lock(&root->fs_info->alloc_mutex);
		ret = lookup_extent_ref(NULL, root, start, len, refs);
		BUG_ON(ret);
		mutex_unlock(&root->fs_info->alloc_mutex);

		if (eb) {
			btrfs_tree_unlock(eb);
			free_extent_buffer(eb);
		}
		if (*refs == 1) {
			printk(KERN_ERR "btrfs block %llu went down to one "
			       "during drop_snap\n", (unsigned long long)start);
		}

	}
#endif

	cond_resched();
	return ret;
}


/*
 * this is used while deleting old snapshots, and it drops the refs
 * on a whole subtree starting from a level 1 node.
 *
 * The idea is to sort all the leaf pointers, and then drop the
 * ref on all the leaves in order.  Most of the time the leaves
 * will have ref cache entries, so no leaf IOs will be required to
 * find the extents they have references on.
 *
 * For each leaf, any references it has are also dropped in order
 *
 * This ends up dropping the references in something close to optimal
 * order for reading and modifying the extent allocation tree.
 */
static noinline int drop_level_one_refs(struct btrfs_trans_handle *trans,
					struct btrfs_root *root,
					struct btrfs_path *path)
{
	u64 bytenr;
	u64 root_owner;
	u64 root_gen;
	struct extent_buffer *eb = path->nodes[1];
	struct extent_buffer *leaf;
	struct btrfs_leaf_ref *ref;
	struct refsort *sorted = NULL;
	int nritems = btrfs_header_nritems(eb);
	int ret;
	int i;
	int refi = 0;
	int slot = path->slots[1];
	u32 blocksize = btrfs_level_size(root, 0);
	u32 refs;

	if (nritems == 0)
		goto out;

	root_owner = btrfs_header_owner(eb);
	root_gen = btrfs_header_generation(eb);
	sorted = kmalloc(sizeof(*sorted) * nritems, GFP_NOFS);

	/*
	 * step one, sort all the leaf pointers so we don't scribble
	 * randomly into the extent allocation tree
	 */
	for (i = slot; i < nritems; i++) {
		sorted[refi].bytenr = btrfs_node_blockptr(eb, i);
		sorted[refi].slot = i;
		refi++;
	}

	/*
	 * nritems won't be zero, but if we're picking up drop_snapshot
	 * after a crash, slot might be > 0, so double check things
	 * just in case.
	 */
	if (refi == 0)
		goto out;

	sort(sorted, refi, sizeof(struct refsort), refsort_cmp, NULL);

	/*
	 * the first loop frees everything the leaves point to
	 */
	for (i = 0; i < refi; i++) {
		u64 ptr_gen;

		bytenr = sorted[i].bytenr;

		/*
		 * check the reference count on this leaf.  If it is > 1
		 * we just decrement it below and don't update any
		 * of the refs the leaf points to.
		 */
		ret = drop_snap_lookup_refcount(trans, root, bytenr,
						blocksize, &refs);
		BUG_ON(ret);
		if (refs != 1)
			continue;

		ptr_gen = btrfs_node_ptr_generation(eb, sorted[i].slot);

		/*
		 * the leaf only had one reference, which means the
		 * only thing pointing to this leaf is the snapshot
		 * we're deleting.  It isn't possible for the reference
		 * count to increase again later
		 *
		 * The reference cache is checked for the leaf,
		 * and if found we'll be able to drop any refs held by
		 * the leaf without needing to read it in.
		 */
		ref = btrfs_lookup_leaf_ref(root, bytenr);
		if (ref && ref->generation != ptr_gen) {
			btrfs_free_leaf_ref(root, ref);
			ref = NULL;
		}
		if (ref) {
			ret = cache_drop_leaf_ref(trans, root, ref);
			BUG_ON(ret);
			btrfs_remove_leaf_ref(root, ref);
			btrfs_free_leaf_ref(root, ref);
		} else {
			/*
			 * the leaf wasn't in the reference cache, so
			 * we have to read it.
			 */
			leaf = read_tree_block(root, bytenr, blocksize,
					       ptr_gen);
			ret = btrfs_drop_leaf_ref(trans, root, leaf);
			BUG_ON(ret);
			free_extent_buffer(leaf);
		}
		atomic_inc(&root->fs_info->throttle_gen);
		wake_up(&root->fs_info->transaction_throttle);
		cond_resched();
	}

	/*
	 * run through the loop again to free the refs on the leaves.
	 * This is faster than doing it in the loop above because
	 * the leaves are likely to be clustered together.  We end up
	 * working in nice chunks on the extent allocation tree.
	 */
	for (i = 0; i < refi; i++) {
		bytenr = sorted[i].bytenr;
		ret = btrfs_free_extent(trans, root, bytenr,
					blocksize, eb->start,
					root_owner, root_gen, 0, 1);
		BUG_ON(ret);

		atomic_inc(&root->fs_info->throttle_gen);
		wake_up(&root->fs_info->transaction_throttle);
		cond_resched();
	}
out:
	kfree(sorted);

	/*
	 * update the path to show we've processed the entire level 1
	 * node.  This will get saved into the root's drop_snapshot_progress
	 * field so these drops are not repeated again if this transaction
	 * commits.
	 */
	path->slots[1] = nritems;
	return 0;
}

/*
 * helper function for drop_snapshot, this walks down the tree dropping ref
 * counts as it goes.
 */
static noinline int walk_down_tree(struct btrfs_trans_handle *trans,
				   struct btrfs_root *root,
				   struct btrfs_path *path, int *level)
{
	u64 root_owner;
	u64 root_gen;
	u64 bytenr;
	u64 ptr_gen;
	struct extent_buffer *next;
	struct extent_buffer *cur;
	struct extent_buffer *parent;
	u32 blocksize;
	int ret;
	u32 refs;

	WARN_ON(*level < 0);
	WARN_ON(*level >= BTRFS_MAX_LEVEL);
	ret = drop_snap_lookup_refcount(trans, root, path->nodes[*level]->start,
				path->nodes[*level]->len, &refs);
	BUG_ON(ret);
	if (refs > 1)
		goto out;

	/*
	 * walk down to the last node level and free all the leaves
	 */
	while (*level >= 0) {
		WARN_ON(*level < 0);
		WARN_ON(*level >= BTRFS_MAX_LEVEL);
		cur = path->nodes[*level];

		if (btrfs_header_level(cur) != *level)
			WARN_ON(1);

		if (path->slots[*level] >=
		    btrfs_header_nritems(cur))
			break;

		/* the new code goes down to level 1 and does all the
		 * leaves pointed to that node in bulk.  So, this check
		 * for level 0 will always be false.
		 *
		 * But, the disk format allows the drop_snapshot_progress
		 * field in the root to leave things in a state where
		 * a leaf will need cleaning up here.  If someone crashes
		 * with the old code and then boots with the new code,
		 * we might find a leaf here.
		 */
		if (*level == 0) {
			ret = btrfs_drop_leaf_ref(trans, root, cur);
			BUG_ON(ret);
			break;
		}

		/*
		 * once we get to level one, process the whole node
		 * at once, including everything below it.
		 */
		if (*level == 1) {
			ret = drop_level_one_refs(trans, root, path);
			BUG_ON(ret);
			break;
		}

		bytenr = btrfs_node_blockptr(cur, path->slots[*level]);
		ptr_gen = btrfs_node_ptr_generation(cur, path->slots[*level]);
		blocksize = btrfs_level_size(root, *level - 1);

		ret = drop_snap_lookup_refcount(trans, root, bytenr,
						blocksize, &refs);
		BUG_ON(ret);

		/*
		 * if there is more than one reference, we don't need
		 * to read that node to drop any references it has.  We
		 * just drop the ref we hold on that node and move on to the
		 * next slot in this level.
		 */
		if (refs != 1) {
			parent = path->nodes[*level];
			root_owner = btrfs_header_owner(parent);
			root_gen = btrfs_header_generation(parent);
			path->slots[*level]++;

			ret = btrfs_free_extent(trans, root, bytenr,
						blocksize, parent->start,
						root_owner, root_gen,
						*level - 1, 1);
			BUG_ON(ret);

			atomic_inc(&root->fs_info->throttle_gen);
			wake_up(&root->fs_info->transaction_throttle);
			cond_resched();

			continue;
		}

		/*
		 * we need to keep freeing things in the next level down.
		 * read the block and loop around to process it
		 */
		next = read_tree_block(root, bytenr, blocksize, ptr_gen);
		WARN_ON(*level <= 0);
		if (path->nodes[*level-1])
			free_extent_buffer(path->nodes[*level-1]);
		path->nodes[*level-1] = next;
		*level = btrfs_header_level(next);
		path->slots[*level] = 0;
		cond_resched();
	}
out:
	WARN_ON(*level < 0);
	WARN_ON(*level >= BTRFS_MAX_LEVEL);

	if (path->nodes[*level] == root->node) {
		parent = path->nodes[*level];
		bytenr = path->nodes[*level]->start;
	} else {
		parent = path->nodes[*level + 1];
		bytenr = btrfs_node_blockptr(parent, path->slots[*level + 1]);
	}

	blocksize = btrfs_level_size(root, *level);
	root_owner = btrfs_header_owner(parent);
	root_gen = btrfs_header_generation(parent);

	/*
	 * cleanup and free the reference on the last node
	 * we processed
	 */
	ret = btrfs_free_extent(trans, root, bytenr, blocksize,
				  parent->start, root_owner, root_gen,
				  *level, 1);
	free_extent_buffer(path->nodes[*level]);
	path->nodes[*level] = NULL;

	*level += 1;
	BUG_ON(ret);

	cond_resched();
	return 0;
}
#endif

struct walk_control {
	u64 refs[BTRFS_MAX_LEVEL];
	u64 flags[BTRFS_MAX_LEVEL];
	struct btrfs_key update_progress;
	int stage;
	int level;
	int shared_level;
	int update_ref;
	int keep_locks;
};

#define DROP_REFERENCE	1
#define UPDATE_BACKREF	2

/*
 * hepler to process tree block while walking down the tree.
 *
 * when wc->stage == DROP_REFERENCE, this function checks
 * reference count of the block. if the block is shared and
 * we need update back refs for the subtree rooted at the
 * block, this function changes wc->stage to UPDATE_BACKREF
 *
 * when wc->stage == UPDATE_BACKREF, this function updates
 * back refs for pointers in the block.
 *
 * NOTE: return value 1 means we should stop walking down.
 */
static noinline int walk_down_proc(struct btrfs_trans_handle *trans,
				   struct btrfs_root *root,
				   struct btrfs_path *path,
				   struct walk_control *wc)
{
	int level = wc->level;
	struct extent_buffer *eb = path->nodes[level];
	struct btrfs_key key;
	u64 flag = BTRFS_BLOCK_FLAG_FULL_BACKREF;
	int ret;

	if (wc->stage == UPDATE_BACKREF &&
	    btrfs_header_owner(eb) != root->root_key.objectid)
		return 1;

	/*
	 * when reference count of tree block is 1, it won't increase
	 * again. once full backref flag is set, we never clear it.
	 */
	if ((wc->stage == DROP_REFERENCE && wc->refs[level] != 1) ||
	    (wc->stage == UPDATE_BACKREF && !(wc->flags[level] & flag))) {
		BUG_ON(!path->locks[level]);
		ret = btrfs_lookup_extent_info(trans, root,
					       eb->start, eb->len,
					       &wc->refs[level],
					       &wc->flags[level]);
		BUG_ON(ret);
		BUG_ON(wc->refs[level] == 0);
	}

	if (wc->stage == DROP_REFERENCE &&
	    wc->update_ref && wc->refs[level] > 1) {
		BUG_ON(eb == root->node);
		BUG_ON(path->slots[level] > 0);
		if (level == 0)
			btrfs_item_key_to_cpu(eb, &key, path->slots[level]);
		else
			btrfs_node_key_to_cpu(eb, &key, path->slots[level]);
		if (btrfs_header_owner(eb) == root->root_key.objectid &&
		    btrfs_comp_cpu_keys(&key, &wc->update_progress) >= 0) {
			wc->stage = UPDATE_BACKREF;
			wc->shared_level = level;
		}
	}

	if (wc->stage == DROP_REFERENCE) {
		if (wc->refs[level] > 1)
			return 1;

		if (path->locks[level] && !wc->keep_locks) {
			btrfs_tree_unlock(eb);
			path->locks[level] = 0;
		}
		return 0;
	}

	/* wc->stage == UPDATE_BACKREF */
	if (!(wc->flags[level] & flag)) {
		BUG_ON(!path->locks[level]);
		ret = btrfs_inc_ref(trans, root, eb, 1);
		BUG_ON(ret);
		ret = btrfs_dec_ref(trans, root, eb, 0);
		BUG_ON(ret);
		ret = btrfs_set_disk_extent_flags(trans, root, eb->start,
						  eb->len, flag, 0);
		BUG_ON(ret);
		wc->flags[level] |= flag;
	}

	/*
	 * the block is shared by multiple trees, so it's not good to
	 * keep the tree lock
	 */
	if (path->locks[level] && level > 0) {
		btrfs_tree_unlock(eb);
		path->locks[level] = 0;
	}
	return 0;
}

/*
 * hepler to process tree block while walking up the tree.
 *
 * when wc->stage == DROP_REFERENCE, this function drops
 * reference count on the block.
 *
 * when wc->stage == UPDATE_BACKREF, this function changes
 * wc->stage back to DROP_REFERENCE if we changed wc->stage
 * to UPDATE_BACKREF previously while processing the block.
 *
 * NOTE: return value 1 means we should stop walking up.
 */
static noinline int walk_up_proc(struct btrfs_trans_handle *trans,
				 struct btrfs_root *root,
				 struct btrfs_path *path,
				 struct walk_control *wc)
{
	int ret = 0;
	int level = wc->level;
	struct extent_buffer *eb = path->nodes[level];
	u64 parent = 0;

	if (wc->stage == UPDATE_BACKREF) {
		BUG_ON(wc->shared_level < level);
		if (level < wc->shared_level)
			goto out;

		BUG_ON(wc->refs[level] <= 1);
		ret = find_next_key(path, level + 1, &wc->update_progress);
		if (ret > 0)
			wc->update_ref = 0;

		wc->stage = DROP_REFERENCE;
		wc->shared_level = -1;
		path->slots[level] = 0;

		/*
		 * check reference count again if the block isn't locked.
		 * we should start walking down the tree again if reference
		 * count is one.
		 */
		if (!path->locks[level]) {
			BUG_ON(level == 0);
			btrfs_tree_lock(eb);
			btrfs_set_lock_blocking(eb);
			path->locks[level] = 1;

			ret = btrfs_lookup_extent_info(trans, root,
						       eb->start, eb->len,
						       &wc->refs[level],
						       &wc->flags[level]);
			BUG_ON(ret);
			BUG_ON(wc->refs[level] == 0);
			if (wc->refs[level] == 1) {
				btrfs_tree_unlock(eb);
				path->locks[level] = 0;
				return 1;
			}
		} else {
			BUG_ON(level != 0);
		}
	}

	/* wc->stage == DROP_REFERENCE */
	BUG_ON(wc->refs[level] > 1 && !path->locks[level]);

	if (wc->refs[level] == 1) {
		if (level == 0) {
			if (wc->flags[level] & BTRFS_BLOCK_FLAG_FULL_BACKREF)
				ret = btrfs_dec_ref(trans, root, eb, 1);
			else
				ret = btrfs_dec_ref(trans, root, eb, 0);
			BUG_ON(ret);
		}
		/* make block locked assertion in clean_tree_block happy */
		if (!path->locks[level] &&
		    btrfs_header_generation(eb) == trans->transid) {
			btrfs_tree_lock(eb);
			btrfs_set_lock_blocking(eb);
			path->locks[level] = 1;
		}
		clean_tree_block(trans, root, eb);
	}

	if (eb == root->node) {
		if (wc->flags[level] & BTRFS_BLOCK_FLAG_FULL_BACKREF)
			parent = eb->start;
		else
			BUG_ON(root->root_key.objectid !=
			       btrfs_header_owner(eb));
	} else {
		if (wc->flags[level + 1] & BTRFS_BLOCK_FLAG_FULL_BACKREF)
			parent = path->nodes[level + 1]->start;
		else
			BUG_ON(root->root_key.objectid !=
			       btrfs_header_owner(path->nodes[level + 1]));
	}

	ret = btrfs_free_extent(trans, root, eb->start, eb->len, parent,
				root->root_key.objectid, level, 0);
	BUG_ON(ret);
out:
	wc->refs[level] = 0;
	wc->flags[level] = 0;
	return ret;
}

static noinline int walk_down_tree(struct btrfs_trans_handle *trans,
				   struct btrfs_root *root,
				   struct btrfs_path *path,
				   struct walk_control *wc)
{
	struct extent_buffer *next;
	struct extent_buffer *cur;
	u64 bytenr;
	u64 ptr_gen;
	u32 blocksize;
	int level = wc->level;
	int ret;

	while (level >= 0) {
		cur = path->nodes[level];
		BUG_ON(path->slots[level] >= btrfs_header_nritems(cur));

		ret = walk_down_proc(trans, root, path, wc);
		if (ret > 0)
			break;

		if (level == 0)
			break;

		bytenr = btrfs_node_blockptr(cur, path->slots[level]);
		blocksize = btrfs_level_size(root, level - 1);
		ptr_gen = btrfs_node_ptr_generation(cur, path->slots[level]);

		next = read_tree_block(root, bytenr, blocksize, ptr_gen);
		btrfs_tree_lock(next);
		btrfs_set_lock_blocking(next);

		level--;
		BUG_ON(level != btrfs_header_level(next));
		path->nodes[level] = next;
		path->slots[level] = 0;
		path->locks[level] = 1;
		wc->level = level;
	}
	return 0;
}

static noinline int walk_up_tree(struct btrfs_trans_handle *trans,
				 struct btrfs_root *root,
				 struct btrfs_path *path,
				 struct walk_control *wc, int max_level)
{
	int level = wc->level;
	int ret;

	path->slots[level] = btrfs_header_nritems(path->nodes[level]);
	while (level < max_level && path->nodes[level]) {
		wc->level = level;
		if (path->slots[level] + 1 <
		    btrfs_header_nritems(path->nodes[level])) {
			path->slots[level]++;
			return 0;
		} else {
			ret = walk_up_proc(trans, root, path, wc);
			if (ret > 0)
				return 0;

			if (path->locks[level]) {
				btrfs_tree_unlock(path->nodes[level]);
				path->locks[level] = 0;
			}
			free_extent_buffer(path->nodes[level]);
			path->nodes[level] = NULL;
			level++;
		}
	}
	return 1;
}

/*
 * drop a subvolume tree.
 *
 * this function traverses the tree freeing any blocks that only
 * referenced by the tree.
 *
 * when a shared tree block is found. this function decreases its
 * reference count by one. if update_ref is true, this function
 * also make sure backrefs for the shared block and all lower level
 * blocks are properly updated.
 */
int btrfs_drop_snapshot(struct btrfs_root *root, int update_ref)
{
	struct btrfs_path *path;
	struct btrfs_trans_handle *trans;
	struct btrfs_root *tree_root = root->fs_info->tree_root;
	struct btrfs_root_item *root_item = &root->root_item;
	struct walk_control *wc;
	struct btrfs_key key;
	int err = 0;
	int ret;
	int level;

	path = btrfs_alloc_path();
	BUG_ON(!path);

	wc = kzalloc(sizeof(*wc), GFP_NOFS);
	BUG_ON(!wc);

	trans = btrfs_start_transaction(tree_root, 1);

	if (btrfs_disk_key_objectid(&root_item->drop_progress) == 0) {
		level = btrfs_header_level(root->node);
		path->nodes[level] = btrfs_lock_root_node(root);
		btrfs_set_lock_blocking(path->nodes[level]);
		path->slots[level] = 0;
		path->locks[level] = 1;
		memset(&wc->update_progress, 0,
		       sizeof(wc->update_progress));
	} else {
		btrfs_disk_key_to_cpu(&key, &root_item->drop_progress);
		memcpy(&wc->update_progress, &key,
		       sizeof(wc->update_progress));

		level = root_item->drop_level;
		BUG_ON(level == 0);
		path->lowest_level = level;
		ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
		path->lowest_level = 0;
		if (ret < 0) {
			err = ret;
			goto out;
		}
		btrfs_node_key_to_cpu(path->nodes[level], &key,
				      path->slots[level]);
		WARN_ON(memcmp(&key, &wc->update_progress, sizeof(key)));

		/*
		 * unlock our path, this is safe because only this
		 * function is allowed to delete this snapshot
		 */
		btrfs_unlock_up_safe(path, 0);

		level = btrfs_header_level(root->node);
		while (1) {
			btrfs_tree_lock(path->nodes[level]);
			btrfs_set_lock_blocking(path->nodes[level]);

			ret = btrfs_lookup_extent_info(trans, root,
						path->nodes[level]->start,
						path->nodes[level]->len,
						&wc->refs[level],
						&wc->flags[level]);
			BUG_ON(ret);
			BUG_ON(wc->refs[level] == 0);

			if (level == root_item->drop_level)
				break;

			btrfs_tree_unlock(path->nodes[level]);
			WARN_ON(wc->refs[level] != 1);
			level--;
		}
	}

	wc->level = level;
	wc->shared_level = -1;
	wc->stage = DROP_REFERENCE;
	wc->update_ref = update_ref;
	wc->keep_locks = 0;

	while (1) {
		ret = walk_down_tree(trans, root, path, wc);
		if (ret < 0) {
			err = ret;
			break;
		}

		ret = walk_up_tree(trans, root, path, wc, BTRFS_MAX_LEVEL);
		if (ret < 0) {
			err = ret;
			break;
		}

		if (ret > 0) {
			BUG_ON(wc->stage != DROP_REFERENCE);
			break;
		}

		if (wc->stage == DROP_REFERENCE) {
			level = wc->level;
			btrfs_node_key(path->nodes[level],
				       &root_item->drop_progress,
				       path->slots[level]);
			root_item->drop_level = level;
		}

		BUG_ON(wc->level == 0);
		if (trans->transaction->in_commit ||
		    trans->transaction->delayed_refs.flushing) {
			ret = btrfs_update_root(trans, tree_root,
						&root->root_key,
						root_item);
			BUG_ON(ret);

			btrfs_end_transaction(trans, tree_root);
			trans = btrfs_start_transaction(tree_root, 1);
		} else {
			unsigned long update;
			update = trans->delayed_ref_updates;
			trans->delayed_ref_updates = 0;
			if (update)
				btrfs_run_delayed_refs(trans, tree_root,
						       update);
		}
	}
	btrfs_release_path(root, path);
	BUG_ON(err);

	ret = btrfs_del_root(trans, tree_root, &root->root_key);
	BUG_ON(ret);

	free_extent_buffer(root->node);
	free_extent_buffer(root->commit_root);
	kfree(root);
out:
	btrfs_end_transaction(trans, tree_root);
	kfree(wc);
	btrfs_free_path(path);
	return err;
}

/*
 * drop subtree rooted at tree block 'node'.
 *
 * NOTE: this function will unlock and release tree block 'node'
 */
int btrfs_drop_subtree(struct btrfs_trans_handle *trans,
			struct btrfs_root *root,
			struct extent_buffer *node,
			struct extent_buffer *parent)
{
	struct btrfs_path *path;
	struct walk_control *wc;
	int level;
	int parent_level;
	int ret = 0;
	int wret;

	BUG_ON(root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID);

	path = btrfs_alloc_path();
	BUG_ON(!path);

	wc = kzalloc(sizeof(*wc), GFP_NOFS);
	BUG_ON(!wc);

	btrfs_assert_tree_locked(parent);
	parent_level = btrfs_header_level(parent);
	extent_buffer_get(parent);
	path->nodes[parent_level] = parent;
	path->slots[parent_level] = btrfs_header_nritems(parent);

	btrfs_assert_tree_locked(node);
	level = btrfs_header_level(node);
	path->nodes[level] = node;
	path->slots[level] = 0;
	path->locks[level] = 1;

	wc->refs[parent_level] = 1;
	wc->flags[parent_level] = BTRFS_BLOCK_FLAG_FULL_BACKREF;
	wc->level = level;
	wc->shared_level = -1;
	wc->stage = DROP_REFERENCE;
	wc->update_ref = 0;
	wc->keep_locks = 1;

	while (1) {
		wret = walk_down_tree(trans, root, path, wc);
		if (wret < 0) {
			ret = wret;
			break;
		}

		wret = walk_up_tree(trans, root, path, wc, parent_level);
		if (wret < 0)
			ret = wret;
		if (wret != 0)
			break;
	}

	kfree(wc);
	btrfs_free_path(path);
	return ret;
}

#if 0
static unsigned long calc_ra(unsigned long start, unsigned long last,
			     unsigned long nr)
{
	return min(last, start + nr - 1);
}

static noinline int relocate_inode_pages(struct inode *inode, u64 start,
					 u64 len)
{
	u64 page_start;
	u64 page_end;
	unsigned long first_index;
	unsigned long last_index;
	unsigned long i;
	struct page *page;
	struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
	struct file_ra_state *ra;
	struct btrfs_ordered_extent *ordered;
	unsigned int total_read = 0;
	unsigned int total_dirty = 0;
	int ret = 0;

	ra = kzalloc(sizeof(*ra), GFP_NOFS);

	mutex_lock(&inode->i_mutex);
	first_index = start >> PAGE_CACHE_SHIFT;
	last_index = (start + len - 1) >> PAGE_CACHE_SHIFT;

	/* make sure the dirty trick played by the caller work */
	ret = invalidate_inode_pages2_range(inode->i_mapping,
					    first_index, last_index);
	if (ret)
		goto out_unlock;

	file_ra_state_init(ra, inode->i_mapping);

	for (i = first_index ; i <= last_index; i++) {
		if (total_read % ra->ra_pages == 0) {
			btrfs_force_ra(inode->i_mapping, ra, NULL, i,
				       calc_ra(i, last_index, ra->ra_pages));
		}
		total_read++;
again:
		if (((u64)i << PAGE_CACHE_SHIFT) > i_size_read(inode))
			BUG_ON(1);
		page = grab_cache_page(inode->i_mapping, i);
		if (!page) {
			ret = -ENOMEM;
			goto out_unlock;
		}
		if (!PageUptodate(page)) {
			btrfs_readpage(NULL, page);
			lock_page(page);
			if (!PageUptodate(page)) {
				unlock_page(page);
				page_cache_release(page);
				ret = -EIO;
				goto out_unlock;
			}
		}
		wait_on_page_writeback(page);

		page_start = (u64)page->index << PAGE_CACHE_SHIFT;
		page_end = page_start + PAGE_CACHE_SIZE - 1;
		lock_extent(io_tree, page_start, page_end, GFP_NOFS);

		ordered = btrfs_lookup_ordered_extent(inode, page_start);
		if (ordered) {
			unlock_extent(io_tree, page_start, page_end, GFP_NOFS);
			unlock_page(page);
			page_cache_release(page);
			btrfs_start_ordered_extent(inode, ordered, 1);
			btrfs_put_ordered_extent(ordered);
			goto again;
		}
		set_page_extent_mapped(page);

		if (i == first_index)
			set_extent_bits(io_tree, page_start, page_end,
					EXTENT_BOUNDARY, GFP_NOFS);
		btrfs_set_extent_delalloc(inode, page_start, page_end);

		set_page_dirty(page);
		total_dirty++;

		unlock_extent(io_tree, page_start, page_end, GFP_NOFS);
		unlock_page(page);
		page_cache_release(page);
	}

out_unlock:
	kfree(ra);
	mutex_unlock(&inode->i_mutex);
	balance_dirty_pages_ratelimited_nr(inode->i_mapping, total_dirty);
	return ret;
}

static noinline int relocate_data_extent(struct inode *reloc_inode,
					 struct btrfs_key *extent_key,
					 u64 offset)
{
	struct btrfs_root *root = BTRFS_I(reloc_inode)->root;
	struct extent_map_tree *em_tree = &BTRFS_I(reloc_inode)->extent_tree;
	struct extent_map *em;
	u64 start = extent_key->objectid - offset;
	u64 end = start + extent_key->offset - 1;

	em = alloc_extent_map(GFP_NOFS);
	BUG_ON(!em || IS_ERR(em));

	em->start = start;
	em->len = extent_key->offset;
	em->block_len = extent_key->offset;
	em->block_start = extent_key->objectid;
	em->bdev = root->fs_info->fs_devices->latest_bdev;
	set_bit(EXTENT_FLAG_PINNED, &em->flags);

	/* setup extent map to cheat btrfs_readpage */
	lock_extent(&BTRFS_I(reloc_inode)->io_tree, start, end, GFP_NOFS);
	while (1) {
		int ret;
		spin_lock(&em_tree->lock);
		ret = add_extent_mapping(em_tree, em);
		spin_unlock(&em_tree->lock);
		if (ret != -EEXIST) {
			free_extent_map(em);
			break;
		}
		btrfs_drop_extent_cache(reloc_inode, start, end, 0);
	}
	unlock_extent(&BTRFS_I(reloc_inode)->io_tree, start, end, GFP_NOFS);

	return relocate_inode_pages(reloc_inode, start, extent_key->offset);
}

struct btrfs_ref_path {
	u64 extent_start;
	u64 nodes[BTRFS_MAX_LEVEL];
	u64 root_objectid;
	u64 root_generation;
	u64 owner_objectid;
	u32 num_refs;
	int lowest_level;
	int current_level;
	int shared_level;

	struct btrfs_key node_keys[BTRFS_MAX_LEVEL];
	u64 new_nodes[BTRFS_MAX_LEVEL];
};

struct disk_extent {
	u64 ram_bytes;
	u64 disk_bytenr;
	u64 disk_num_bytes;
	u64 offset;
	u64 num_bytes;
	u8 compression;
	u8 encryption;
	u16 other_encoding;
};

static int is_cowonly_root(u64 root_objectid)
{
	if (root_objectid == BTRFS_ROOT_TREE_OBJECTID ||
	    root_objectid == BTRFS_EXTENT_TREE_OBJECTID ||
	    root_objectid == BTRFS_CHUNK_TREE_OBJECTID ||
	    root_objectid == BTRFS_DEV_TREE_OBJECTID ||
	    root_objectid == BTRFS_TREE_LOG_OBJECTID ||
	    root_objectid == BTRFS_CSUM_TREE_OBJECTID)
		return 1;
	return 0;
}

static noinline int __next_ref_path(struct btrfs_trans_handle *trans,
				    struct btrfs_root *extent_root,
				    struct btrfs_ref_path *ref_path,
				    int first_time)
{
	struct extent_buffer *leaf;
	struct btrfs_path *path;
	struct btrfs_extent_ref *ref;
	struct btrfs_key key;
	struct btrfs_key found_key;
	u64 bytenr;
	u32 nritems;
	int level;
	int ret = 1;

	path = btrfs_alloc_path();
	if (!path)
		return -ENOMEM;

	if (first_time) {
		ref_path->lowest_level = -1;
		ref_path->current_level = -1;
		ref_path->shared_level = -1;
		goto walk_up;
	}
walk_down:
	level = ref_path->current_level - 1;
	while (level >= -1) {
		u64 parent;
		if (level < ref_path->lowest_level)
			break;

		if (level >= 0)
			bytenr = ref_path->nodes[level];
		else
			bytenr = ref_path->extent_start;
		BUG_ON(bytenr == 0);

		parent = ref_path->nodes[level + 1];
		ref_path->nodes[level + 1] = 0;
		ref_path->current_level = level;
		BUG_ON(parent == 0);

		key.objectid = bytenr;
		key.offset = parent + 1;
		key.type = BTRFS_EXTENT_REF_KEY;

		ret = btrfs_search_slot(trans, extent_root, &key, path, 0, 0);
		if (ret < 0)
			goto out;
		BUG_ON(ret == 0);

		leaf = path->nodes[0];
		nritems = btrfs_header_nritems(leaf);
		if (path->slots[0] >= nritems) {
			ret = btrfs_next_leaf(extent_root, path);
			if (ret < 0)
				goto out;
			if (ret > 0)
				goto next;
			leaf = path->nodes[0];
		}

		btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
		if (found_key.objectid == bytenr &&
		    found_key.type == BTRFS_EXTENT_REF_KEY) {
			if (level < ref_path->shared_level)
				ref_path->shared_level = level;
			goto found;
		}
next:
		level--;
		btrfs_release_path(extent_root, path);
		cond_resched();
	}
	/* reached lowest level */
	ret = 1;
	goto out;
walk_up:
	level = ref_path->current_level;
	while (level < BTRFS_MAX_LEVEL - 1) {
		u64 ref_objectid;

		if (level >= 0)
			bytenr = ref_path->nodes[level];
		else
			bytenr = ref_path->extent_start;

		BUG_ON(bytenr == 0);

		key.objectid = bytenr;
		key.offset = 0;
		key.type = BTRFS_EXTENT_REF_KEY;

		ret = btrfs_search_slot(trans, extent_root, &key, path, 0, 0);
		if (ret < 0)
			goto out;

		leaf = path->nodes[0];
		nritems = btrfs_header_nritems(leaf);
		if (path->slots[0] >= nritems) {
			ret = btrfs_next_leaf(extent_root, path);
			if (ret < 0)
				goto out;
			if (ret > 0) {
				/* the extent was freed by someone */
				if (ref_path->lowest_level == level)
					goto out;
				btrfs_release_path(extent_root, path);
				goto walk_down;
			}
			leaf = path->nodes[0];
		}

		btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
		if (found_key.objectid != bytenr ||
				found_key.type != BTRFS_EXTENT_REF_KEY) {
			/* the extent was freed by someone */
			if (ref_path->lowest_level == level) {
				ret = 1;
				goto out;
			}
			btrfs_release_path(extent_root, path);
			goto walk_down;
		}
found:
		ref = btrfs_item_ptr(leaf, path->slots[0],
				struct btrfs_extent_ref);
		ref_objectid = btrfs_ref_objectid(leaf, ref);
		if (ref_objectid < BTRFS_FIRST_FREE_OBJECTID) {
			if (first_time) {
				level = (int)ref_objectid;
				BUG_ON(level >= BTRFS_MAX_LEVEL);
				ref_path->lowest_level = level;
				ref_path->current_level = level;
				ref_path->nodes[level] = bytenr;
			} else {
				WARN_ON(ref_objectid != level);
			}
		} else {
			WARN_ON(level != -1);
		}
		first_time = 0;

		if (ref_path->lowest_level == level) {
			ref_path->owner_objectid = ref_objectid;
			ref_path->num_refs = btrfs_ref_num_refs(leaf, ref);
		}

		/*
		 * the block is tree root or the block isn't in reference
		 * counted tree.
		 */
		if (found_key.objectid == found_key.offset ||
		    is_cowonly_root(btrfs_ref_root(leaf, ref))) {
			ref_path->root_objectid = btrfs_ref_root(leaf, ref);
			ref_path->root_generation =
				btrfs_ref_generation(leaf, ref);
			if (level < 0) {
				/* special reference from the tree log */
				ref_path->nodes[0] = found_key.offset;
				ref_path->current_level = 0;
			}
			ret = 0;
			goto out;
		}

		level++;
		BUG_ON(ref_path->nodes[level] != 0);
		ref_path->nodes[level] = found_key.offset;
		ref_path->current_level = level;

		/*
		 * the reference was created in the running transaction,
		 * no need to continue walking up.
		 */
		if (btrfs_ref_generation(leaf, ref) == trans->transid) {
			ref_path->root_objectid = btrfs_ref_root(leaf, ref);
			ref_path->root_generation =
				btrfs_ref_generation(leaf, ref);
			ret = 0;
			goto out;
		}

		btrfs_release_path(extent_root, path);
		cond_resched();
	}
	/* reached max tree level, but no tree root found. */
	BUG();
out:
	btrfs_free_path(path);
	return ret;
}

static int btrfs_first_ref_path(struct btrfs_trans_handle *trans,
				struct btrfs_root *extent_root,
				struct btrfs_ref_path *ref_path,
				u64 extent_start)
{
	memset(ref_path, 0, sizeof(*ref_path));
	ref_path->extent_start = extent_start;

	return __next_ref_path(trans, extent_root, ref_path, 1);
}

static int btrfs_next_ref_path(struct btrfs_trans_handle *trans,
			       struct btrfs_root *extent_root,
			       struct btrfs_ref_path *ref_path)
{
	return __next_ref_path(trans, extent_root, ref_path, 0);
}

static noinline int get_new_locations(struct inode *reloc_inode,
				      struct btrfs_key *extent_key,
				      u64 offset, int no_fragment,
				      struct disk_extent **extents,
				      int *nr_extents)
{
	struct btrfs_root *root = BTRFS_I(reloc_inode)->root;
	struct btrfs_path *path;
	struct btrfs_file_extent_item *fi;
	struct extent_buffer *leaf;
	struct disk_extent *exts = *extents;
	struct btrfs_key found_key;
	u64 cur_pos;
	u64 last_byte;
	u32 nritems;
	int nr = 0;
	int max = *nr_extents;
	int ret;

	WARN_ON(!no_fragment && *extents);
	if (!exts) {
		max = 1;
		exts = kmalloc(sizeof(*exts) * max, GFP_NOFS);
		if (!exts)
			return -ENOMEM;
	}

	path = btrfs_alloc_path();
	BUG_ON(!path);

	cur_pos = extent_key->objectid - offset;
	last_byte = extent_key->objectid + extent_key->offset;
	ret = btrfs_lookup_file_extent(NULL, root, path, reloc_inode->i_ino,
				       cur_pos, 0);
	if (ret < 0)
		goto out;
	if (ret > 0) {
		ret = -ENOENT;
		goto out;
	}

	while (1) {
		leaf = path->nodes[0];
		nritems = btrfs_header_nritems(leaf);
		if (path->slots[0] >= nritems) {
			ret = btrfs_next_leaf(root, path);
			if (ret < 0)
				goto out;
			if (ret > 0)
				break;
			leaf = path->nodes[0];
		}

		btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
		if (found_key.offset != cur_pos ||
		    found_key.type != BTRFS_EXTENT_DATA_KEY ||
		    found_key.objectid != reloc_inode->i_ino)
			break;

		fi = btrfs_item_ptr(leaf, path->slots[0],
				    struct btrfs_file_extent_item);
		if (btrfs_file_extent_type(leaf, fi) !=
		    BTRFS_FILE_EXTENT_REG ||
		    btrfs_file_extent_disk_bytenr(leaf, fi) == 0)
			break;

		if (nr == max) {
			struct disk_extent *old = exts;
			max *= 2;
			exts = kzalloc(sizeof(*exts) * max, GFP_NOFS);
			memcpy(exts, old, sizeof(*exts) * nr);
			if (old != *extents)
				kfree(old);
		}

		exts[nr].disk_bytenr =
			btrfs_file_extent_disk_bytenr(leaf, fi);
		exts[nr].disk_num_bytes =
			btrfs_file_extent_disk_num_bytes(leaf, fi);
		exts[nr].offset = btrfs_file_extent_offset(leaf, fi);
		exts[nr].num_bytes = btrfs_file_extent_num_bytes(leaf, fi);
		exts[nr].ram_bytes = btrfs_file_extent_ram_bytes(leaf, fi);
		exts[nr].compression = btrfs_file_extent_compression(leaf, fi);
		exts[nr].encryption = btrfs_file_extent_encryption(leaf, fi);
		exts[nr].other_encoding = btrfs_file_extent_other_encoding(leaf,
									   fi);
		BUG_ON(exts[nr].offset > 0);
		BUG_ON(exts[nr].compression || exts[nr].encryption);
		BUG_ON(exts[nr].num_bytes != exts[nr].disk_num_bytes);

		cur_pos += exts[nr].num_bytes;
		nr++;

		if (cur_pos + offset >= last_byte)
			break;

		if (no_fragment) {
			ret = 1;
			goto out;
		}
		path->slots[0]++;
	}

	BUG_ON(cur_pos + offset > last_byte);
	if (cur_pos + offset < last_byte) {
		ret = -ENOENT;
		goto out;
	}
	ret = 0;
out:
	btrfs_free_path(path);
	if (ret) {
		if (exts != *extents)
			kfree(exts);
	} else {
		*extents = exts;
		*nr_extents = nr;
	}
	return ret;
}

static noinline int replace_one_extent(struct btrfs_trans_handle *trans,
					struct btrfs_root *root,
					struct btrfs_path *path,
					struct btrfs_key *extent_key,
					struct btrfs_key *leaf_key,
					struct btrfs_ref_path *ref_path,
					struct disk_extent *new_extents,
					int nr_extents)
{
	struct extent_buffer *leaf;
	struct btrfs_file_extent_item *fi;
	struct inode *inode = NULL;
	struct btrfs_key key;
	u64 lock_start = 0;
	u64 lock_end = 0;
	u64 num_bytes;
	u64 ext_offset;
	u64 search_end = (u64)-1;
	u32 nritems;
	int nr_scaned = 0;
	int extent_locked = 0;
	int extent_type;
	int ret;

	memcpy(&key, leaf_key, sizeof(key));
	if (ref_path->owner_objectid != BTRFS_MULTIPLE_OBJECTIDS) {
		if (key.objectid < ref_path->owner_objectid ||
		    (key.objectid == ref_path->owner_objectid &&
		     key.type < BTRFS_EXTENT_DATA_KEY)) {
			key.objectid = ref_path->owner_objectid;
			key.type = BTRFS_EXTENT_DATA_KEY;
			key.offset = 0;
		}
	}

	while (1) {
		ret = btrfs_search_slot(trans, root, &key, path, 0, 1);
		if (ret < 0)
			goto out;

		leaf = path->nodes[0];
		nritems = btrfs_header_nritems(leaf);
next:
		if (extent_locked && ret > 0) {
			/*
			 * the file extent item was modified by someone
			 * before the extent got locked.
			 */
			unlock_extent(&BTRFS_I(inode)->io_tree, lock_start,
				      lock_end, GFP_NOFS);
			extent_locked = 0;
		}

		if (path->slots[0] >= nritems) {
			if (++nr_scaned > 2)
				break;

			BUG_ON(extent_locked);
			ret = btrfs_next_leaf(root, path);
			if (ret < 0)
				goto out;
			if (ret > 0)
				break;
			leaf = path->nodes[0];
			nritems = btrfs_header_nritems(leaf);
		}

		btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);

		if (ref_path->owner_objectid != BTRFS_MULTIPLE_OBJECTIDS) {
			if ((key.objectid > ref_path->owner_objectid) ||
			    (key.objectid == ref_path->owner_objectid &&
			     key.type > BTRFS_EXTENT_DATA_KEY) ||
			    key.offset >= search_end)
				break;
		}

		if (inode && key.objectid != inode->i_ino) {
			BUG_ON(extent_locked);
			btrfs_release_path(root, path);
			mutex_unlock(&inode->i_mutex);
			iput(inode);
			inode = NULL;
			continue;
		}

		if (key.type != BTRFS_EXTENT_DATA_KEY) {
			path->slots[0]++;
			ret = 1;
			goto next;
		}
		fi = btrfs_item_ptr(leaf, path->slots[0],
				    struct btrfs_file_extent_item);
		extent_type = btrfs_file_extent_type(leaf, fi);
		if ((extent_type != BTRFS_FILE_EXTENT_REG &&
		     extent_type != BTRFS_FILE_EXTENT_PREALLOC) ||
		    (btrfs_file_extent_disk_bytenr(leaf, fi) !=
		     extent_key->objectid)) {
			path->slots[0]++;
			ret = 1;
			goto next;
		}

		num_bytes = btrfs_file_extent_num_bytes(leaf, fi);
		ext_offset = btrfs_file_extent_offset(leaf, fi);

		if (search_end == (u64)-1) {
			search_end = key.offset - ext_offset +
				btrfs_file_extent_ram_bytes(leaf, fi);
		}

		if (!extent_locked) {
			lock_start = key.offset;
			lock_end = lock_start + num_bytes - 1;
		} else {
			if (lock_start > key.offset ||
			    lock_end + 1 < key.offset + num_bytes) {
				unlock_extent(&BTRFS_I(inode)->io_tree,
					      lock_start, lock_end, GFP_NOFS);
				extent_locked = 0;
			}
		}

		if (!inode) {
			btrfs_release_path(root, path);

			inode = btrfs_iget_locked(root->fs_info->sb,
						  key.objectid, root);
			if (inode->i_state & I_NEW) {
				BTRFS_I(inode)->root = root;
				BTRFS_I(inode)->location.objectid =
					key.objectid;
				BTRFS_I(inode)->location.type =
					BTRFS_INODE_ITEM_KEY;
				BTRFS_I(inode)->location.offset = 0;
				btrfs_read_locked_inode(inode);
				unlock_new_inode(inode);
			}
			/*
			 * some code call btrfs_commit_transaction while
			 * holding the i_mutex, so we can't use mutex_lock
			 * here.
			 */
			if (is_bad_inode(inode) ||
			    !mutex_trylock(&inode->i_mutex)) {
				iput(inode);
				inode = NULL;
				key.offset = (u64)-1;
				goto skip;
			}
		}

		if (!extent_locked) {
			struct btrfs_ordered_extent *ordered;

			btrfs_release_path(root, path);

			lock_extent(&BTRFS_I(inode)->io_tree, lock_start,
				    lock_end, GFP_NOFS);
			ordered = btrfs_lookup_first_ordered_extent(inode,
								    lock_end);
			if (ordered &&
			    ordered->file_offset <= lock_end &&
			    ordered->file_offset + ordered->len > lock_start) {
				unlock_extent(&BTRFS_I(inode)->io_tree,
					      lock_start, lock_end, GFP_NOFS);
				btrfs_start_ordered_extent(inode, ordered, 1);
				btrfs_put_ordered_extent(ordered);
				key.offset += num_bytes;
				goto skip;
			}
			if (ordered)
				btrfs_put_ordered_extent(ordered);

			extent_locked = 1;
			continue;
		}

		if (nr_extents == 1) {
			/* update extent pointer in place */
			btrfs_set_file_extent_disk_bytenr(leaf, fi,
						new_extents[0].disk_bytenr);
			btrfs_set_file_extent_disk_num_bytes(leaf, fi,
						new_extents[0].disk_num_bytes);
			btrfs_mark_buffer_dirty(leaf);

			btrfs_drop_extent_cache(inode, key.offset,
						key.offset + num_bytes - 1, 0);

			ret = btrfs_inc_extent_ref(trans, root,
						new_extents[0].disk_bytenr,
						new_extents[0].disk_num_bytes,
						leaf->start,
						root->root_key.objectid,
						trans->transid,
						key.objectid);
			BUG_ON(ret);

			ret = btrfs_free_extent(trans, root,
						extent_key->objectid,
						extent_key->offset,
						leaf->start,
						btrfs_header_owner(leaf),
						btrfs_header_generation(leaf),
						key.objectid, 0);
			BUG_ON(ret);

			btrfs_release_path(root, path);
			key.offset += num_bytes;
		} else {
			BUG_ON(1);
#if 0
			u64 alloc_hint;
			u64 extent_len;
			int i;
			/*
			 * drop old extent pointer at first, then insert the
			 * new pointers one bye one
			 */
			btrfs_release_path(root, path);
			ret = btrfs_drop_extents(trans, root, inode, key.offset,
						 key.offset + num_bytes,
						 key.offset, &alloc_hint);
			BUG_ON(ret);

			for (i = 0; i < nr_extents; i++) {
				if (ext_offset >= new_extents[i].num_bytes) {
					ext_offset -= new_extents[i].num_bytes;
					continue;
				}
				extent_len = min(new_extents[i].num_bytes -
						 ext_offset, num_bytes);

				ret = btrfs_insert_empty_item(trans, root,
							      path, &key,
							      sizeof(*fi));
				BUG_ON(ret);

				leaf = path->nodes[0];
				fi = btrfs_item_ptr(leaf, path->slots[0],
						struct btrfs_file_extent_item);
				btrfs_set_file_extent_generation(leaf, fi,
							trans->transid);
				btrfs_set_file_extent_type(leaf, fi,
							BTRFS_FILE_EXTENT_REG);
				btrfs_set_file_extent_disk_bytenr(leaf, fi,
						new_extents[i].disk_bytenr);
				btrfs_set_file_extent_disk_num_bytes(leaf, fi,
						new_extents[i].disk_num_bytes);
				btrfs_set_file_extent_ram_bytes(leaf, fi,
						new_extents[i].ram_bytes);

				btrfs_set_file_extent_compression(leaf, fi,
						new_extents[i].compression);
				btrfs_set_file_extent_encryption(leaf, fi,
						new_extents[i].encryption);
				btrfs_set_file_extent_other_encoding(leaf, fi,
						new_extents[i].other_encoding);

				btrfs_set_file_extent_num_bytes(leaf, fi,
							extent_len);
				ext_offset += new_extents[i].offset;
				btrfs_set_file_extent_offset(leaf, fi,
							ext_offset);
				btrfs_mark_buffer_dirty(leaf);

				btrfs_drop_extent_cache(inode, key.offset,
						key.offset + extent_len - 1, 0);

				ret = btrfs_inc_extent_ref(trans, root,
						new_extents[i].disk_bytenr,
						new_extents[i].disk_num_bytes,
						leaf->start,
						root->root_key.objectid,
						trans->transid, key.objectid);
				BUG_ON(ret);
				btrfs_release_path(root, path);

				inode_add_bytes(inode, extent_len);

				ext_offset = 0;
				num_bytes -= extent_len;
				key.offset += extent_len;

				if (num_bytes == 0)
					break;
			}
			BUG_ON(i >= nr_extents);
#endif
		}

		if (extent_locked) {
			unlock_extent(&BTRFS_I(inode)->io_tree, lock_start,
				      lock_end, GFP_NOFS);
			extent_locked = 0;
		}
skip:
		if (ref_path->owner_objectid != BTRFS_MULTIPLE_OBJECTIDS &&
		    key.offset >= search_end)
			break;

		cond_resched();
	}
	ret = 0;
out:
	btrfs_release_path(root, path);
	if (inode) {
		mutex_unlock(&inode->i_mutex);
		if (extent_locked) {
			unlock_extent(&BTRFS_I(inode)->io_tree, lock_start,
				      lock_end, GFP_NOFS);
		}
		iput(inode);
	}
	return ret;
}

int btrfs_reloc_tree_cache_ref(struct btrfs_trans_handle *trans,
			       struct btrfs_root *root,
			       struct extent_buffer *buf, u64 orig_start)
{
	int level;
	int ret;

	BUG_ON(btrfs_header_generation(buf) != trans->transid);
	BUG_ON(root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID);

	level = btrfs_header_level(buf);
	if (level == 0) {
		struct btrfs_leaf_ref *ref;
		struct btrfs_leaf_ref *orig_ref;

		orig_ref = btrfs_lookup_leaf_ref(root, orig_start);
		if (!orig_ref)
			return -ENOENT;

		ref = btrfs_alloc_leaf_ref(root, orig_ref->nritems);
		if (!ref) {
			btrfs_free_leaf_ref(root, orig_ref);
			return -ENOMEM;
		}

		ref->nritems = orig_ref->nritems;
		memcpy(ref->extents, orig_ref->extents,
			sizeof(ref->extents[0]) * ref->nritems);

		btrfs_free_leaf_ref(root, orig_ref);

		ref->root_gen = trans->transid;
		ref->bytenr = buf->start;
		ref->owner = btrfs_header_owner(buf);
		ref->generation = btrfs_header_generation(buf);

		ret = btrfs_add_leaf_ref(root, ref, 0);
		WARN_ON(ret);
		btrfs_free_leaf_ref(root, ref);
	}
	return 0;
}

static noinline int invalidate_extent_cache(struct btrfs_root *root,
					struct extent_buffer *leaf,
					struct btrfs_block_group_cache *group,
					struct btrfs_root *target_root)
{
	struct btrfs_key key;
	struct inode *inode = NULL;
	struct btrfs_file_extent_item *fi;
	u64 num_bytes;
	u64 skip_objectid = 0;
	u32 nritems;
	u32 i;

	nritems = btrfs_header_nritems(leaf);
	for (i = 0; i < nritems; i++) {
		btrfs_item_key_to_cpu(leaf, &key, i);
		if (key.objectid == skip_objectid ||
		    key.type != BTRFS_EXTENT_DATA_KEY)
			continue;
		fi = btrfs_item_ptr(leaf, i, struct btrfs_file_extent_item);
		if (btrfs_file_extent_type(leaf, fi) ==
		    BTRFS_FILE_EXTENT_INLINE)
			continue;
		if (btrfs_file_extent_disk_bytenr(leaf, fi) == 0)
			continue;
		if (!inode || inode->i_ino != key.objectid) {
			iput(inode);
			inode = btrfs_ilookup(target_root->fs_info->sb,
					      key.objectid, target_root, 1);
		}
		if (!inode) {
			skip_objectid = key.objectid;
			continue;
		}
		num_bytes = btrfs_file_extent_num_bytes(leaf, fi);

		lock_extent(&BTRFS_I(inode)->io_tree, key.offset,
			    key.offset + num_bytes - 1, GFP_NOFS);
		btrfs_drop_extent_cache(inode, key.offset,
					key.offset + num_bytes - 1, 1);
		unlock_extent(&BTRFS_I(inode)->io_tree, key.offset,
			      key.offset + num_bytes - 1, GFP_NOFS);
		cond_resched();
	}
	iput(inode);
	return 0;
}

static noinline int replace_extents_in_leaf(struct btrfs_trans_handle *trans,
					struct btrfs_root *root,
					struct extent_buffer *leaf,
					struct btrfs_block_group_cache *group,
					struct inode *reloc_inode)
{
	struct btrfs_key key;
	struct btrfs_key extent_key;
	struct btrfs_file_extent_item *fi;
	struct btrfs_leaf_ref *ref;
	struct disk_extent *new_extent;
	u64 bytenr;
	u64 num_bytes;
	u32 nritems;
	u32 i;
	int ext_index;
	int nr_extent;
	int ret;

	new_extent = kmalloc(sizeof(*new_extent), GFP_NOFS);
	BUG_ON(!new_extent);

	ref = btrfs_lookup_leaf_ref(root, leaf->start);
	BUG_ON(!ref);

	ext_index = -1;
	nritems = btrfs_header_nritems(leaf);
	for (i = 0; i < nritems; i++) {
		btrfs_item_key_to_cpu(leaf, &key, i);
		if (btrfs_key_type(&key) != BTRFS_EXTENT_DATA_KEY)
			continue;
		fi = btrfs_item_ptr(leaf, i, struct btrfs_file_extent_item);
		if (btrfs_file_extent_type(leaf, fi) ==
		    BTRFS_FILE_EXTENT_INLINE)
			continue;
		bytenr = btrfs_file_extent_disk_bytenr(leaf, fi);
		num_bytes = btrfs_file_extent_disk_num_bytes(leaf, fi);
		if (bytenr == 0)
			continue;

		ext_index++;
		if (bytenr >= group->key.objectid + group->key.offset ||
		    bytenr + num_bytes <= group->key.objectid)
			continue;

		extent_key.objectid = bytenr;
		extent_key.offset = num_bytes;
		extent_key.type = BTRFS_EXTENT_ITEM_KEY;
		nr_extent = 1;
		ret = get_new_locations(reloc_inode, &extent_key,
					group->key.objectid, 1,
					&new_extent, &nr_extent);
		if (ret > 0)
			continue;
		BUG_ON(ret < 0);

		BUG_ON(ref->extents[ext_index].bytenr != bytenr);
		BUG_ON(ref->extents[ext_index].num_bytes != num_bytes);
		ref->extents[ext_index].bytenr = new_extent->disk_bytenr;
		ref->extents[ext_index].num_bytes = new_extent->disk_num_bytes;

		btrfs_set_file_extent_disk_bytenr(leaf, fi,
						new_extent->disk_bytenr);
		btrfs_set_file_extent_disk_num_bytes(leaf, fi,
						new_extent->disk_num_bytes);
		btrfs_mark_buffer_dirty(leaf);

		ret = btrfs_inc_extent_ref(trans, root,
					new_extent->disk_bytenr,
					new_extent->disk_num_bytes,
					leaf->start,
					root->root_key.objectid,
					trans->transid, key.objectid);
		BUG_ON(ret);

		ret = btrfs_free_extent(trans, root,
					bytenr, num_bytes, leaf->start,
					btrfs_header_owner(leaf),
					btrfs_header_generation(leaf),
					key.objectid, 0);
		BUG_ON(ret);
		cond_resched();
	}
	kfree(new_extent);
	BUG_ON(ext_index + 1 != ref->nritems);
	btrfs_free_leaf_ref(root, ref);
	return 0;
}

int btrfs_free_reloc_root(struct btrfs_trans_handle *trans,
			  struct btrfs_root *root)
{
	struct btrfs_root *reloc_root;
	int ret;

	if (root->reloc_root) {
		reloc_root = root->reloc_root;
		root->reloc_root = NULL;
		list_add(&reloc_root->dead_list,
			 &root->fs_info->dead_reloc_roots);

		btrfs_set_root_bytenr(&reloc_root->root_item,
				      reloc_root->node->start);
		btrfs_set_root_level(&root->root_item,
				     btrfs_header_level(reloc_root->node));
		memset(&reloc_root->root_item.drop_progress, 0,
			sizeof(struct btrfs_disk_key));
		reloc_root->root_item.drop_level = 0;

		ret = btrfs_update_root(trans, root->fs_info->tree_root,
					&reloc_root->root_key,
					&reloc_root->root_item);
		BUG_ON(ret);
	}
	return 0;
}

int btrfs_drop_dead_reloc_roots(struct btrfs_root *root)
{
	struct btrfs_trans_handle *trans;
	struct btrfs_root *reloc_root;
	struct btrfs_root *prev_root = NULL;
	struct list_head dead_roots;
	int ret;
	unsigned long nr;

	INIT_LIST_HEAD(&dead_roots);
	list_splice_init(&root->fs_info->dead_reloc_roots, &dead_roots);

	while (!list_empty(&dead_roots)) {
		reloc_root = list_entry(dead_roots.prev,
					struct btrfs_root, dead_list);
		list_del_init(&reloc_root->dead_list);

		BUG_ON(reloc_root->commit_root != NULL);
		while (1) {
			trans = btrfs_join_transaction(root, 1);
			BUG_ON(!trans);

			mutex_lock(&root->fs_info->drop_mutex);
			ret = btrfs_drop_snapshot(trans, reloc_root);
			if (ret != -EAGAIN)
				break;
			mutex_unlock(&root->fs_info->drop_mutex);

			nr = trans->blocks_used;
			ret = btrfs_end_transaction(trans, root);
			BUG_ON(ret);
			btrfs_btree_balance_dirty(root, nr);
		}

		free_extent_buffer(reloc_root->node);

		ret = btrfs_del_root(trans, root->fs_info->tree_root,
				     &reloc_root->root_key);
		BUG_ON(ret);
		mutex_unlock(&root->fs_info->drop_mutex);

		nr = trans->blocks_used;
		ret = btrfs_end_transaction(trans, root);
		BUG_ON(ret);
		btrfs_btree_balance_dirty(root, nr);

		kfree(prev_root);
		prev_root = reloc_root;
	}
	if (prev_root) {
		btrfs_remove_leaf_refs(prev_root, (u64)-1, 0);
		kfree(prev_root);
	}
	return 0;
}

int btrfs_add_dead_reloc_root(struct btrfs_root *root)
{
	list_add(&root->dead_list, &root->fs_info->dead_reloc_roots);
	return 0;
}

int btrfs_cleanup_reloc_trees(struct btrfs_root *root)
{
	struct btrfs_root *reloc_root;
	struct btrfs_trans_handle *trans;
	struct btrfs_key location;
	int found;
	int ret;

	mutex_lock(&root->fs_info->tree_reloc_mutex);
	ret = btrfs_find_dead_roots(root, BTRFS_TREE_RELOC_OBJECTID, NULL);
	BUG_ON(ret);
	found = !list_empty(&root->fs_info->dead_reloc_roots);
	mutex_unlock(&root->fs_info->tree_reloc_mutex);

	if (found) {
		trans = btrfs_start_transaction(root, 1);
		BUG_ON(!trans);
		ret = btrfs_commit_transaction(trans, root);
		BUG_ON(ret);
	}

	location.objectid = BTRFS_DATA_RELOC_TREE_OBJECTID;
	location.offset = (u64)-1;
	location.type = BTRFS_ROOT_ITEM_KEY;

	reloc_root = btrfs_read_fs_root_no_name(root->fs_info, &location);
	BUG_ON(!reloc_root);
	btrfs_orphan_cleanup(reloc_root);
	return 0;
}

static noinline int init_reloc_tree(struct btrfs_trans_handle *trans,
				    struct btrfs_root *root)
{
	struct btrfs_root *reloc_root;
	struct extent_buffer *eb;
	struct btrfs_root_item *root_item;
	struct btrfs_key root_key;
	int ret;

	BUG_ON(!root->ref_cows);
	if (root->reloc_root)
		return 0;

	root_item = kmalloc(sizeof(*root_item), GFP_NOFS);
	BUG_ON(!root_item);

	ret = btrfs_copy_root(trans, root, root->commit_root,
			      &eb, BTRFS_TREE_RELOC_OBJECTID);
	BUG_ON(ret);

	root_key.objectid = BTRFS_TREE_RELOC_OBJECTID;
	root_key.offset = root->root_key.objectid;
	root_key.type = BTRFS_ROOT_ITEM_KEY;

	memcpy(root_item, &root->root_item, sizeof(root_item));
	btrfs_set_root_refs(root_item, 0);
	btrfs_set_root_bytenr(root_item, eb->start);
	btrfs_set_root_level(root_item, btrfs_header_level(eb));
	btrfs_set_root_generation(root_item, trans->transid);

	btrfs_tree_unlock(eb);
	free_extent_buffer(eb);

	ret = btrfs_insert_root(trans, root->fs_info->tree_root,
				&root_key, root_item);
	BUG_ON(ret);
	kfree(root_item);

	reloc_root = btrfs_read_fs_root_no_radix(root->fs_info->tree_root,
						 &root_key);
	BUG_ON(!reloc_root);
	reloc_root->last_trans = trans->transid;
	reloc_root->commit_root = NULL;
	reloc_root->ref_tree = &root->fs_info->reloc_ref_tree;

	root->reloc_root = reloc_root;
	return 0;
}

/*
 * Core function of space balance.
 *
 * The idea is using reloc trees to relocate tree blocks in reference
 * counted roots. There is one reloc tree for each subvol, and all
 * reloc trees share same root key objectid. Reloc trees are snapshots
 * of the latest committed roots of subvols (root->commit_root).
 *
 * To relocate a tree block referenced by a subvol, there are two steps.
 * COW the block through subvol's reloc tree, then update block pointer
 * in the subvol to point to the new block. Since all reloc trees share
 * same root key objectid, doing special handing for tree blocks owned
 * by them is easy. Once a tree block has been COWed in one reloc tree,
 * we can use the resulting new block directly when the same block is
 * required to COW again through other reloc trees. By this way, relocated
 * tree blocks are shared between reloc trees, so they are also shared
 * between subvols.
 */
static noinline int relocate_one_path(struct btrfs_trans_handle *trans,
				      struct btrfs_root *root,
				      struct btrfs_path *path,
				      struct btrfs_key *first_key,
				      struct btrfs_ref_path *ref_path,
				      struct btrfs_block_group_cache *group,
				      struct inode *reloc_inode)
{
	struct btrfs_root *reloc_root;
	struct extent_buffer *eb = NULL;
	struct btrfs_key *keys;
	u64 *nodes;
	int level;
	int shared_level;
	int lowest_level = 0;
	int ret;

	if (ref_path->owner_objectid < BTRFS_FIRST_FREE_OBJECTID)
		lowest_level = ref_path->owner_objectid;

	if (!root->ref_cows) {
		path->lowest_level = lowest_level;
		ret = btrfs_search_slot(trans, root, first_key, path, 0, 1);
		BUG_ON(ret < 0);
		path->lowest_level = 0;
		btrfs_release_path(root, path);
		return 0;
	}

	mutex_lock(&root->fs_info->tree_reloc_mutex);
	ret = init_reloc_tree(trans, root);
	BUG_ON(ret);
	reloc_root = root->reloc_root;

	shared_level = ref_path->shared_level;
	ref_path->shared_level = BTRFS_MAX_LEVEL - 1;

	keys = ref_path->node_keys;
	nodes = ref_path->new_nodes;
	memset(&keys[shared_level + 1], 0,
	       sizeof(*keys) * (BTRFS_MAX_LEVEL - shared_level - 1));
	memset(&nodes[shared_level + 1], 0,
	       sizeof(*nodes) * (BTRFS_MAX_LEVEL - shared_level - 1));

	if (nodes[lowest_level] == 0) {
		path->lowest_level = lowest_level;
		ret = btrfs_search_slot(trans, reloc_root, first_key, path,
					0, 1);
		BUG_ON(ret);
		for (level = lowest_level; level < BTRFS_MAX_LEVEL; level++) {
			eb = path->nodes[level];
			if (!eb || eb == reloc_root->node)
				break;
			nodes[level] = eb->start;
			if (level == 0)
				btrfs_item_key_to_cpu(eb, &keys[level], 0);
			else
				btrfs_node_key_to_cpu(eb, &keys[level], 0);
		}
		if (nodes[0] &&
		    ref_path->owner_objectid >= BTRFS_FIRST_FREE_OBJECTID) {
			eb = path->nodes[0];
			ret = replace_extents_in_leaf(trans, reloc_root, eb,
						      group, reloc_inode);
			BUG_ON(ret);
		}
		btrfs_release_path(reloc_root, path);
	} else {
		ret = btrfs_merge_path(trans, reloc_root, keys, nodes,
				       lowest_level);
		BUG_ON(ret);
	}

	/*
	 * replace tree blocks in the fs tree with tree blocks in
	 * the reloc tree.
	 */
	ret = btrfs_merge_path(trans, root, keys, nodes, lowest_level);
	BUG_ON(ret < 0);

	if (ref_path->owner_objectid >= BTRFS_FIRST_FREE_OBJECTID) {
		ret = btrfs_search_slot(trans, reloc_root, first_key, path,
					0, 0);
		BUG_ON(ret);
		extent_buffer_get(path->nodes[0]);
		eb = path->nodes[0];
		btrfs_release_path(reloc_root, path);
		ret = invalidate_extent_cache(reloc_root, eb, group, root);
		BUG_ON(ret);
		free_extent_buffer(eb);
	}

	mutex_unlock(&root->fs_info->tree_reloc_mutex);
	path->lowest_level = 0;
	return 0;
}

static noinline int relocate_tree_block(struct btrfs_trans_handle *trans,
					struct btrfs_root *root,
					struct btrfs_path *path,
					struct btrfs_key *first_key,
					struct btrfs_ref_path *ref_path)
{
	int ret;

	ret = relocate_one_path(trans, root, path, first_key,
				ref_path, NULL, NULL);
	BUG_ON(ret);

	return 0;
}

static noinline int del_extent_zero(struct btrfs_trans_handle *trans,
				    struct btrfs_root *extent_root,
				    struct btrfs_path *path,
				    struct btrfs_key *extent_key)
{
	int ret;

	ret = btrfs_search_slot(trans, extent_root, extent_key, path, -1, 1);
	if (ret)
		goto out;
	ret = btrfs_del_item(trans, extent_root, path);
out:
	btrfs_release_path(extent_root, path);
	return ret;
}

static noinline struct btrfs_root *read_ref_root(struct btrfs_fs_info *fs_info,
						struct btrfs_ref_path *ref_path)
{
	struct btrfs_key root_key;

	root_key.objectid = ref_path->root_objectid;
	root_key.type = BTRFS_ROOT_ITEM_KEY;
	if (is_cowonly_root(ref_path->root_objectid))
		root_key.offset = 0;
	else
		root_key.offset = (u64)-1;

	return btrfs_read_fs_root_no_name(fs_info, &root_key);
}

static noinline int relocate_one_extent(struct btrfs_root *extent_root,
					struct btrfs_path *path,
					struct btrfs_key *extent_key,
					struct btrfs_block_group_cache *group,
					struct inode *reloc_inode, int pass)
{
	struct btrfs_trans_handle *trans;
	struct btrfs_root *found_root;
	struct btrfs_ref_path *ref_path = NULL;
	struct disk_extent *new_extents = NULL;
	int nr_extents = 0;
	int loops;
	int ret;
	int level;
	struct btrfs_key first_key;
	u64 prev_block = 0;


	trans = btrfs_start_transaction(extent_root, 1);
	BUG_ON(!trans);

	if (extent_key->objectid == 0) {
		ret = del_extent_zero(trans, extent_root, path, extent_key);
		goto out;
	}

	ref_path = kmalloc(sizeof(*ref_path), GFP_NOFS);
	if (!ref_path) {
		ret = -ENOMEM;
		goto out;
	}

	for (loops = 0; ; loops++) {
		if (loops == 0) {
			ret = btrfs_first_ref_path(trans, extent_root, ref_path,
						   extent_key->objectid);
		} else {
			ret = btrfs_next_ref_path(trans, extent_root, ref_path);
		}
		if (ret < 0)
			goto out;
		if (ret > 0)
			break;

		if (ref_path->root_objectid == BTRFS_TREE_LOG_OBJECTID ||
		    ref_path->root_objectid == BTRFS_TREE_RELOC_OBJECTID)
			continue;

		found_root = read_ref_root(extent_root->fs_info, ref_path);
		BUG_ON(!found_root);
		/*
		 * for reference counted tree, only process reference paths
		 * rooted at the latest committed root.
		 */
		if (found_root->ref_cows &&
		    ref_path->root_generation != found_root->root_key.offset)
			continue;

		if (ref_path->owner_objectid >= BTRFS_FIRST_FREE_OBJECTID) {
			if (pass == 0) {
				/*
				 * copy data extents to new locations
				 */
				u64 group_start = group->key.objectid;
				ret = relocate_data_extent(reloc_inode,
							   extent_key,
							   group_start);
				if (ret < 0)
					goto out;
				break;
			}
			level = 0;
		} else {
			level = ref_path->owner_objectid;
		}

		if (prev_block != ref_path->nodes[level]) {
			struct extent_buffer *eb;
			u64 block_start = ref_path->nodes[level];
			u64 block_size = btrfs_level_size(found_root, level);

			eb = read_tree_block(found_root, block_start,
					     block_size, 0);
			btrfs_tree_lock(eb);
			BUG_ON(level != btrfs_header_level(eb));

			if (level == 0)
				btrfs_item_key_to_cpu(eb, &first_key, 0);
			else
				btrfs_node_key_to_cpu(eb, &first_key, 0);

			btrfs_tree_unlock(eb);
			free_extent_buffer(eb);
			prev_block = block_start;
		}

		mutex_lock(&extent_root->fs_info->trans_mutex);
		btrfs_record_root_in_trans(found_root);
		mutex_unlock(&extent_root->fs_info->trans_mutex);
		if (ref_path->owner_objectid >= BTRFS_FIRST_FREE_OBJECTID) {
			/*
			 * try to update data extent references while
			 * keeping metadata shared between snapshots.
			 */
			if (pass == 1) {
				ret = relocate_one_path(trans, found_root,
						path, &first_key, ref_path,
						group, reloc_inode);
				if (ret < 0)
					goto out;
				continue;
			}
			/*
			 * use fallback method to process the remaining
			 * references.
			 */
			if (!new_extents) {
				u64 group_start = group->key.objectid;
				new_extents = kmalloc(sizeof(*new_extents),
						      GFP_NOFS);
				nr_extents = 1;
				ret = get_new_locations(reloc_inode,
							extent_key,
							group_start, 1,
							&new_extents,
							&nr_extents);
				if (ret)
					goto out;
			}
			ret = replace_one_extent(trans, found_root,
						path, extent_key,
						&first_key, ref_path,
						new_extents, nr_extents);
		} else {
			ret = relocate_tree_block(trans, found_root, path,
						  &first_key, ref_path);
		}
		if (ret < 0)
			goto out;
	}
	ret = 0;
out:
	btrfs_end_transaction(trans, extent_root);
	kfree(new_extents);
	kfree(ref_path);
	return ret;
}
#endif

static u64 update_block_group_flags(struct btrfs_root *root, u64 flags)
{
	u64 num_devices;
	u64 stripped = BTRFS_BLOCK_GROUP_RAID0 |
		BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_RAID10;

	num_devices = root->fs_info->fs_devices->rw_devices;
	if (num_devices == 1) {
		stripped |= BTRFS_BLOCK_GROUP_DUP;
		stripped = flags & ~stripped;

		/* turn raid0 into single device chunks */
		if (flags & BTRFS_BLOCK_GROUP_RAID0)
			return stripped;

		/* turn mirroring into duplication */
		if (flags & (BTRFS_BLOCK_GROUP_RAID1 |
			     BTRFS_BLOCK_GROUP_RAID10))
			return stripped | BTRFS_BLOCK_GROUP_DUP;
		return flags;
	} else {
		/* they already had raid on here, just return */
		if (flags & stripped)
			return flags;

		stripped |= BTRFS_BLOCK_GROUP_DUP;
		stripped = flags & ~stripped;

		/* switch duplicated blocks with raid1 */
		if (flags & BTRFS_BLOCK_GROUP_DUP)
			return stripped | BTRFS_BLOCK_GROUP_RAID1;

		/* turn single device chunks into raid0 */
		return stripped | BTRFS_BLOCK_GROUP_RAID0;
	}
	return flags;
}

static int __alloc_chunk_for_shrink(struct btrfs_root *root,
		     struct btrfs_block_group_cache *shrink_block_group,
		     int force)
{
	struct btrfs_trans_handle *trans;
	u64 new_alloc_flags;
	u64 calc;

	spin_lock(&shrink_block_group->lock);
	if (btrfs_block_group_used(&shrink_block_group->item) +
	    shrink_block_group->reserved > 0) {
		spin_unlock(&shrink_block_group->lock);

		trans = btrfs_start_transaction(root, 1);
		spin_lock(&shrink_block_group->lock);

		new_alloc_flags = update_block_group_flags(root,
						   shrink_block_group->flags);
		if (new_alloc_flags != shrink_block_group->flags) {
			calc =
			     btrfs_block_group_used(&shrink_block_group->item);
		} else {
			calc = shrink_block_group->key.offset;
		}
		spin_unlock(&shrink_block_group->lock);

		do_chunk_alloc(trans, root->fs_info->extent_root,
			       calc + 2 * 1024 * 1024, new_alloc_flags, force);

		btrfs_end_transaction(trans, root);
	} else
		spin_unlock(&shrink_block_group->lock);
	return 0;
}


int btrfs_prepare_block_group_relocation(struct btrfs_root *root,
					 struct btrfs_block_group_cache *group)

{
	__alloc_chunk_for_shrink(root, group, 1);
	set_block_group_readonly(group);
	return 0;
}

#if 0
static int __insert_orphan_inode(struct btrfs_trans_handle *trans,
				 struct btrfs_root *root,
				 u64 objectid, u64 size)
{
	struct btrfs_path *path;
	struct btrfs_inode_item *item;
	struct extent_buffer *leaf;
	int ret;

	path = btrfs_alloc_path();
	if (!path)
		return -ENOMEM;

	path->leave_spinning = 1;
	ret = btrfs_insert_empty_inode(trans, root, path, objectid);
	if (ret)
		goto out;

	leaf = path->nodes[0];
	item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_inode_item);
	memset_extent_buffer(leaf, 0, (unsigned long)item, sizeof(*item));
	btrfs_set_inode_generation(leaf, item, 1);
	btrfs_set_inode_size(leaf, item, size);
	btrfs_set_inode_mode(leaf, item, S_IFREG | 0600);
	btrfs_set_inode_flags(leaf, item, BTRFS_INODE_NOCOMPRESS);
	btrfs_mark_buffer_dirty(leaf);
	btrfs_release_path(root, path);
out:
	btrfs_free_path(path);
	return ret;
}

static noinline struct inode *create_reloc_inode(struct btrfs_fs_info *fs_info,
					struct btrfs_block_group_cache *group)
{
	struct inode *inode = NULL;
	struct btrfs_trans_handle *trans;
	struct btrfs_root *root;
	struct btrfs_key root_key;
	u64 objectid = BTRFS_FIRST_FREE_OBJECTID;
	int err = 0;

	root_key.objectid = BTRFS_DATA_RELOC_TREE_OBJECTID;
	root_key.type = BTRFS_ROOT_ITEM_KEY;
	root_key.offset = (u64)-1;
	root = btrfs_read_fs_root_no_name(fs_info, &root_key);
	if (IS_ERR(root))
		return ERR_CAST(root);

	trans = btrfs_start_transaction(root, 1);
	BUG_ON(!trans);

	err = btrfs_find_free_objectid(trans, root, objectid, &objectid);
	if (err)
		goto out;

	err = __insert_orphan_inode(trans, root, objectid, group->key.offset);
	BUG_ON(err);

	err = btrfs_insert_file_extent(trans, root, objectid, 0, 0, 0,
				       group->key.offset, 0, group->key.offset,
				       0, 0, 0);
	BUG_ON(err);

	inode = btrfs_iget_locked(root->fs_info->sb, objectid, root);
	if (inode->i_state & I_NEW) {
		BTRFS_I(inode)->root = root;
		BTRFS_I(inode)->location.objectid = objectid;
		BTRFS_I(inode)->location.type = BTRFS_INODE_ITEM_KEY;
		BTRFS_I(inode)->location.offset = 0;
		btrfs_read_locked_inode(inode);
		unlock_new_inode(inode);
		BUG_ON(is_bad_inode(inode));
	} else {
		BUG_ON(1);
	}
	BTRFS_I(inode)->index_cnt = group->key.objectid;

	err = btrfs_orphan_add(trans, inode);
out:
	btrfs_end_transaction(trans, root);
	if (err) {
		if (inode)
			iput(inode);
		inode = ERR_PTR(err);
	}
	return inode;
}

int btrfs_reloc_clone_csums(struct inode *inode, u64 file_pos, u64 len)
{

	struct btrfs_ordered_sum *sums;
	struct btrfs_sector_sum *sector_sum;
	struct btrfs_ordered_extent *ordered;
	struct btrfs_root *root = BTRFS_I(inode)->root;
	struct list_head list;
	size_t offset;
	int ret;
	u64 disk_bytenr;

	INIT_LIST_HEAD(&list);

	ordered = btrfs_lookup_ordered_extent(inode, file_pos);
	BUG_ON(ordered->file_offset != file_pos || ordered->len != len);

	disk_bytenr = file_pos + BTRFS_I(inode)->index_cnt;
	ret = btrfs_lookup_csums_range(root->fs_info->csum_root, disk_bytenr,
				       disk_bytenr + len - 1, &list);

	while (!list_empty(&list)) {
		sums = list_entry(list.next, struct btrfs_ordered_sum, list);
		list_del_init(&sums->list);

		sector_sum = sums->sums;
		sums->bytenr = ordered->start;

		offset = 0;
		while (offset < sums->len) {
			sector_sum->bytenr += ordered->start - disk_bytenr;
			sector_sum++;
			offset += root->sectorsize;
		}

		btrfs_add_ordered_sum(inode, ordered, sums);
	}
	btrfs_put_ordered_extent(ordered);
	return 0;
}

int btrfs_relocate_block_group(struct btrfs_root *root, u64 group_start)
{
	struct btrfs_trans_handle *trans;
	struct btrfs_path *path;
	struct btrfs_fs_info *info = root->fs_info;
	struct extent_buffer *leaf;
	struct inode *reloc_inode;
	struct btrfs_block_group_cache *block_group;
	struct btrfs_key key;
	u64 skipped;
	u64 cur_byte;
	u64 total_found;
	u32 nritems;
	int ret;
	int progress;
	int pass = 0;

	root = root->fs_info->extent_root;

	block_group = btrfs_lookup_block_group(info, group_start);
	BUG_ON(!block_group);

	printk(KERN_INFO "btrfs relocating block group %llu flags %llu\n",
	       (unsigned long long)block_group->key.objectid,
	       (unsigned long long)block_group->flags);

	path = btrfs_alloc_path();
	BUG_ON(!path);

	reloc_inode = create_reloc_inode(info, block_group);
	BUG_ON(IS_ERR(reloc_inode));

	__alloc_chunk_for_shrink(root, block_group, 1);
	set_block_group_readonly(block_group);

	btrfs_start_delalloc_inodes(info->tree_root);
	btrfs_wait_ordered_extents(info->tree_root, 0);
again:
	skipped = 0;
	total_found = 0;
	progress = 0;
	key.objectid = block_group->key.objectid;
	key.offset = 0;
	key.type = 0;
	cur_byte = key.objectid;

	trans = btrfs_start_transaction(info->tree_root, 1);
	btrfs_commit_transaction(trans, info->tree_root);

	mutex_lock(&root->fs_info->cleaner_mutex);
	btrfs_clean_old_snapshots(info->tree_root);
	btrfs_remove_leaf_refs(info->tree_root, (u64)-1, 1);
	mutex_unlock(&root->fs_info->cleaner_mutex);

	trans = btrfs_start_transaction(info->tree_root, 1);
	btrfs_commit_transaction(trans, info->tree_root);

	while (1) {
		ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
		if (ret < 0)
			goto out;
next:
		leaf = path->nodes[0];
		nritems = btrfs_header_nritems(leaf);
		if (path->slots[0] >= nritems) {
			ret = btrfs_next_leaf(root, path);
			if (ret < 0)
				goto out;
			if (ret == 1) {
				ret = 0;
				break;
			}
			leaf = path->nodes[0];
			nritems = btrfs_header_nritems(leaf);
		}

		btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);

		if (key.objectid >= block_group->key.objectid +
		    block_group->key.offset)
			break;

		if (progress && need_resched()) {
			btrfs_release_path(root, path);
			cond_resched();
			progress = 0;
			continue;
		}
		progress = 1;

		if (btrfs_key_type(&key) != BTRFS_EXTENT_ITEM_KEY ||
		    key.objectid + key.offset <= cur_byte) {
			path->slots[0]++;
			goto next;
		}

		total_found++;
		cur_byte = key.objectid + key.offset;
		btrfs_release_path(root, path);

		__alloc_chunk_for_shrink(root, block_group, 0);
		ret = relocate_one_extent(root, path, &key, block_group,
					  reloc_inode, pass);
		BUG_ON(ret < 0);
		if (ret > 0)
			skipped++;

		key.objectid = cur_byte;
		key.type = 0;
		key.offset = 0;
	}

	btrfs_release_path(root, path);

	if (pass == 0) {
		btrfs_wait_ordered_range(reloc_inode, 0, (u64)-1);
		invalidate_mapping_pages(reloc_inode->i_mapping, 0, -1);
	}

	if (total_found > 0) {
		printk(KERN_INFO "btrfs found %llu extents in pass %d\n",
		       (unsigned long long)total_found, pass);
		pass++;
		if (total_found == skipped && pass > 2) {
			iput(reloc_inode);
			reloc_inode = create_reloc_inode(info, block_group);
			pass = 0;
		}
		goto again;
	}

	/* delete reloc_inode */
	iput(reloc_inode);

	/* unpin extents in this range */
	trans = btrfs_start_transaction(info->tree_root, 1);
	btrfs_commit_transaction(trans, info->tree_root);

	spin_lock(&block_group->lock);
	WARN_ON(block_group->pinned > 0);
	WARN_ON(block_group->reserved > 0);
	WARN_ON(btrfs_block_group_used(&block_group->item) > 0);
	spin_unlock(&block_group->lock);
	btrfs_put_block_group(block_group);
	ret = 0;
out:
	btrfs_free_path(path);
	return ret;
}
#endif

static int find_first_block_group(struct btrfs_root *root,
		struct btrfs_path *path, struct btrfs_key *key)
{
	int ret = 0;
	struct btrfs_key found_key;
	struct extent_buffer *leaf;
	int slot;

	ret = btrfs_search_slot(NULL, root, key, path, 0, 0);
	if (ret < 0)
		goto out;

	while (1) {
		slot = path->slots[0];
		leaf = path->nodes[0];
		if (slot >= btrfs_header_nritems(leaf)) {
			ret = btrfs_next_leaf(root, path);
			if (ret == 0)
				continue;
			if (ret < 0)
				goto out;
			break;
		}
		btrfs_item_key_to_cpu(leaf, &found_key, slot);

		if (found_key.objectid >= key->objectid &&
		    found_key.type == BTRFS_BLOCK_GROUP_ITEM_KEY) {
			ret = 0;
			goto out;
		}
		path->slots[0]++;
	}
	ret = -ENOENT;
out:
	return ret;
}

int btrfs_free_block_groups(struct btrfs_fs_info *info)
{
	struct btrfs_block_group_cache *block_group;
	struct btrfs_space_info *space_info;
	struct rb_node *n;

	spin_lock(&info->block_group_cache_lock);
	while ((n = rb_last(&info->block_group_cache_tree)) != NULL) {
		block_group = rb_entry(n, struct btrfs_block_group_cache,
				       cache_node);
		rb_erase(&block_group->cache_node,
			 &info->block_group_cache_tree);
		spin_unlock(&info->block_group_cache_lock);

		btrfs_remove_free_space_cache(block_group);
		down_write(&block_group->space_info->groups_sem);
		list_del(&block_group->list);
		up_write(&block_group->space_info->groups_sem);

		WARN_ON(atomic_read(&block_group->count) != 1);
		kfree(block_group);

		spin_lock(&info->block_group_cache_lock);
	}
	spin_unlock(&info->block_group_cache_lock);

	/* now that all the block groups are freed, go through and
	 * free all the space_info structs.  This is only called during
	 * the final stages of unmount, and so we know nobody is
	 * using them.  We call synchronize_rcu() once before we start,
	 * just to be on the safe side.
	 */
	synchronize_rcu();

	while(!list_empty(&info->space_info)) {
		space_info = list_entry(info->space_info.next,
					struct btrfs_space_info,
					list);

		list_del(&space_info->list);
		kfree(space_info);
	}
	return 0;
}

int btrfs_read_block_groups(struct btrfs_root *root)
{
	struct btrfs_path *path;
	int ret;
	struct btrfs_block_group_cache *cache;
	struct btrfs_fs_info *info = root->fs_info;
	struct btrfs_space_info *space_info;
	struct btrfs_key key;
	struct btrfs_key found_key;
	struct extent_buffer *leaf;

	root = info->extent_root;
	key.objectid = 0;
	key.offset = 0;
	btrfs_set_key_type(&key, BTRFS_BLOCK_GROUP_ITEM_KEY);
	path = btrfs_alloc_path();
	if (!path)
		return -ENOMEM;

	while (1) {
		ret = find_first_block_group(root, path, &key);
		if (ret > 0) {
			ret = 0;
			goto error;
		}
		if (ret != 0)
			goto error;

		leaf = path->nodes[0];
		btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
		cache = kzalloc(sizeof(*cache), GFP_NOFS);
		if (!cache) {
			ret = -ENOMEM;
			break;
		}

		atomic_set(&cache->count, 1);
		spin_lock_init(&cache->lock);
		spin_lock_init(&cache->tree_lock);
		mutex_init(&cache->cache_mutex);
		INIT_LIST_HEAD(&cache->list);
		INIT_LIST_HEAD(&cache->cluster_list);
		read_extent_buffer(leaf, &cache->item,
				   btrfs_item_ptr_offset(leaf, path->slots[0]),
				   sizeof(cache->item));
		memcpy(&cache->key, &found_key, sizeof(found_key));

		key.objectid = found_key.objectid + found_key.offset;
		btrfs_release_path(root, path);
		cache->flags = btrfs_block_group_flags(&cache->item);

		ret = update_space_info(info, cache->flags, found_key.offset,
					btrfs_block_group_used(&cache->item),
					&space_info);
		BUG_ON(ret);
		cache->space_info = space_info;
		down_write(&space_info->groups_sem);
		list_add_tail(&cache->list, &space_info->block_groups);
		up_write(&space_info->groups_sem);

		ret = btrfs_add_block_group_cache(root->fs_info, cache);
		BUG_ON(ret);

		set_avail_alloc_bits(root->fs_info, cache->flags);
		if (btrfs_chunk_readonly(root, cache->key.objectid))
			set_block_group_readonly(cache);
	}
	ret = 0;
error:
	btrfs_free_path(path);
	return ret;
}

int btrfs_make_block_group(struct btrfs_trans_handle *trans,
			   struct btrfs_root *root, u64 bytes_used,
			   u64 type, u64 chunk_objectid, u64 chunk_offset,
			   u64 size)
{
	int ret;
	struct btrfs_root *extent_root;
	struct btrfs_block_group_cache *cache;

	extent_root = root->fs_info->extent_root;

	root->fs_info->last_trans_log_full_commit = trans->transid;

	cache = kzalloc(sizeof(*cache), GFP_NOFS);
	if (!cache)
		return -ENOMEM;

	cache->key.objectid = chunk_offset;
	cache->key.offset = size;
	cache->key.type = BTRFS_BLOCK_GROUP_ITEM_KEY;
	atomic_set(&cache->count, 1);
	spin_lock_init(&cache->lock);
	spin_lock_init(&cache->tree_lock);
	mutex_init(&cache->cache_mutex);
	INIT_LIST_HEAD(&cache->list);
	INIT_LIST_HEAD(&cache->cluster_list);

	btrfs_set_block_group_used(&cache->item, bytes_used);
	btrfs_set_block_group_chunk_objectid(&cache->item, chunk_objectid);
	cache->flags = type;
	btrfs_set_block_group_flags(&cache->item, type);

	ret = update_space_info(root->fs_info, cache->flags, size, bytes_used,
				&cache->space_info);
	BUG_ON(ret);
	down_write(&cache->space_info->groups_sem);
	list_add_tail(&cache->list, &cache->space_info->block_groups);
	up_write(&cache->space_info->groups_sem);

	ret = btrfs_add_block_group_cache(root->fs_info, cache);
	BUG_ON(ret);

	ret = btrfs_insert_item(trans, extent_root, &cache->key, &cache->item,
				sizeof(cache->item));
	BUG_ON(ret);

	set_avail_alloc_bits(extent_root->fs_info, type);

	return 0;
}

int btrfs_remove_block_group(struct btrfs_trans_handle *trans,
			     struct btrfs_root *root, u64 group_start)
{
	struct btrfs_path *path;
	struct btrfs_block_group_cache *block_group;
	struct btrfs_free_cluster *cluster;
	struct btrfs_key key;
	int ret;

	root = root->fs_info->extent_root;

	block_group = btrfs_lookup_block_group(root->fs_info, group_start);
	BUG_ON(!block_group);
	BUG_ON(!block_group->ro);

	memcpy(&key, &block_group->key, sizeof(key));

	/* make sure this block group isn't part of an allocation cluster */
	cluster = &root->fs_info->data_alloc_cluster;
	spin_lock(&cluster->refill_lock);
	btrfs_return_cluster_to_free_space(block_group, cluster);
	spin_unlock(&cluster->refill_lock);

	/*
	 * make sure this block group isn't part of a metadata
	 * allocation cluster
	 */
	cluster = &root->fs_info->meta_alloc_cluster;
	spin_lock(&cluster->refill_lock);
	btrfs_return_cluster_to_free_space(block_group, cluster);
	spin_unlock(&cluster->refill_lock);

	path = btrfs_alloc_path();
	BUG_ON(!path);

	spin_lock(&root->fs_info->block_group_cache_lock);
	rb_erase(&block_group->cache_node,
		 &root->fs_info->block_group_cache_tree);
	spin_unlock(&root->fs_info->block_group_cache_lock);
	btrfs_remove_free_space_cache(block_group);
	down_write(&block_group->space_info->groups_sem);
	/*
	 * we must use list_del_init so people can check to see if they
	 * are still on the list after taking the semaphore
	 */
	list_del_init(&block_group->list);
	up_write(&block_group->space_info->groups_sem);

	spin_lock(&block_group->space_info->lock);
	block_group->space_info->total_bytes -= block_group->key.offset;
	block_group->space_info->bytes_readonly -= block_group->key.offset;
	spin_unlock(&block_group->space_info->lock);
	block_group->space_info->full = 0;

	btrfs_put_block_group(block_group);
	btrfs_put_block_group(block_group);

	ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
	if (ret > 0)
		ret = -EIO;
	if (ret < 0)
		goto out;

	ret = btrfs_del_item(trans, root, path);
out:
	btrfs_free_path(path);
	return ret;
}