aboutsummaryrefslogtreecommitdiffstats
path: root/arch
Commit message (Collapse)AuthorAge
* davinci: Initial support for Omapl138-HawkboardVictor Rodriguez2010-09-24
| | | | | | | | | | | | | This patch adds initial support for the Hawkboard-L138 system It is under the machine name "omapl138_hawkboard". This system is based on the da850 davinci CPU architecture. Information on these system may be found at http://www.hawkboard.org. Basic support for the UART console is included in this patch. It's tested with latest Angstrom File Systems like ramdisk from http://alturl.com/imb45. Signed-off-by: Victor Rodriguez <victor.rodriguez@sasken.com> Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
* davinci: MityDSP-L138/MityARM-1808 read MAC address from I2C PromMichael Williamson2010-09-24
| | | | | | | | | For the MityDSP-L138/MityARM-1808 SOMS, read the factory assigned MAC address from the onboard I2C EPROM and assign it to the emac device during platform initialization. Signed-off-by: Michael Williamson <michael.williamson@criticallink.com> Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
* davinci: add tnetv107x touchscreen platform deviceCyril Chemparathy2010-09-24
| | | | | | | | This patch adds a platform device definition for tnetv107x's touchscreen controller. Signed-off-by: Cyril Chemparathy <cyril@ti.com> Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
* davinci: add keypad config for tnetv107x evm boardCyril Chemparathy2010-09-24
| | | | | | | | This patch adds evm board specific keymap definitions and controller configuration data for on-chip keypad controller on tnetv107x silicon. Signed-off-by: Cyril Chemparathy <cyril@ti.com> Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
* davinci: add tnetv107x keypad platform deviceCyril Chemparathy2010-09-24
| | | | | | | | This patch adds a platform device definition for tnetv107x's keypad controller. Signed-off-by: Cyril Chemparathy <cyril@ti.com> Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
* omap: cleanup unused davinci mdio arch codeCyril Chemparathy2010-09-24
| | | | | | | | | | This patch removes davinci architecture code that has now been rendered useless by the previous patches in the MDIO separation series. Signed-off-by: Cyril Chemparathy <cyril@ti.com> Acked-by: David S. Miller <davem@davemloft.net> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
* davinci: cleanup mdio arch code and switch to phy_idCyril Chemparathy2010-09-24
| | | | | | | | | | | | | This patch removes davinci architecture code that has now been rendered useless by the previous patches in the MDIO separation series. In addition, the earlier phy_mask definitions have been replaced with corresponding phy_id definitions. Signed-off-by: Cyril Chemparathy <cyril@ti.com> Tested-by: Michael Williamson <michael.williamson@criticallink.com> Tested-by: Caglar Akyuz <caglarakyuz@gmail.com> Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
* omap: add mdio platform devicesCyril Chemparathy2010-09-24
| | | | | | | | | | | | | This patch adds mdio platform devices on SoCs that have the necessary hardware. Clock lookup entries (aliases) have also been added, so that the MDIO and EMAC drivers can independently enable/disable a shared underlying clock. Further, the EMAC MMR region has been split down into separate MDIO and EMAC regions. Signed-off-by: Cyril Chemparathy <cyril@ti.com> Acked-by: David S. Miller <davem@davemloft.net> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
* davinci: add mdio platform devicesCyril Chemparathy2010-09-24
| | | | | | | | | | | | | | This patch adds mdio platform devices on SoCs that have the necessary hardware. Clock lookup entries (aliases) have also been added, so that the MDIO and EMAC drivers can independently enable/disable a shared underlying clock. Further, the EMAC MMR region has been split down into separate MDIO and EMAC regions. Signed-off-by: Cyril Chemparathy <cyril@ti.com> Acked-by: David S. Miller <davem@davemloft.net> Tested-by: Michael Williamson <michael.williamson@criticallink.com> Tested-by: Caglar Akyuz <caglarakyuz@gmail.com> Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
* davinci: Add I2C0 devices to MityDSP-L138/MityARM-1808 platformMichael Williamson2010-09-24
| | | | | | | | | | This patch adds the i2c0 bus and attached devices to the MityDSP-L138 and MityARM-1808 davinci SoM. Included is a TPS65023 voltage regulator needed for power management and a small 24c02 EPROM that contains factory configuration data. Signed-off-by: Michael Williamson <michael.williamson@criticallink.com> Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
* davinci: add idcode for tnetv107x rev 1.1/1.2Cyril Chemparathy2010-09-24
| | | | | | | | This patch adds an entry into the idcode table for tnetv107x silicon revision 1.1 and 1.2 devices. Signed-off-by: Cyril Chemparathy <cyril@ti.com> Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
* davinci: Initial support for MityDSP-L138/MityARM-1808Michael Williamson2010-09-24
| | | | | | | | | | | | | This patch adds initial support for the MityDSP-L138 and MityDSP-1808 system on Module (SOM) under the machine name "mityomapl138". These SOMs are based on the da850 davinci CPU architecture. Information on these SOMs may be found at http://www.mitydsp.com. Basic support for the console UART, NAND, and EMAC (MII interface) is included in this patch. Signed-off-by: Michael Williamson <michael.williamson@criticallink.com> Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
* davinci: Add CONFIG_REGULATOR_DUMMY to DA8XX defconfig file.Michael Williamson2010-09-24
| | | | | | | | | In order to support reference DA8XX machines not providing a voltage regulator control for the core voltage, the REGULATOR_DUMMY option is required. Signed-off-by: Michael Williamson <michael.williamson@criticallink.com> Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
* davinci: Add machine checks to DA8XX serial console init routinesMichael Williamson2010-09-24
| | | | | | | | | | This patch adds machine checks in the serial console init routines for the DA8XX EVM boards. This is needed because there are other DA8XX based machines that use a different UART/tty as the console and may be included in a common kernel build. Signed-off-by: Michael Williamson <michael.williamson@criticallink.com> Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
* davinci: dm6467t evm: setup NAND flash timingSekhar Nori2010-09-24
| | | | | | | | | | | | | | | Setup NAND flash timing on DM6467T EVM. Without the timing setup, the NAND flash on DM6467T RevC EVM reports a number of random bad blocks because of read errors. Also, with this, copying a 100M file on RevB EVM takes ~35 sec against 1 minute 30 seconds earlier. Signed-off-by: Sekhar Nori <nsekhar@ti.com> Acked-by: Kevin Hilman <khilman@deeprootsystems.com> Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
* davinci: am18x/da850/omap-l138 evm: setup NAND flash timingSekhar Nori2010-09-24
| | | | | | | | | | | | | | | | | | Setup the NAND flash timings for DA850 EVM Before configuring the timing values, throughput calculation using dd command yielded 469 kB/s write and 966 kB/s read speed. After the timing configuration, the throughput was measured to be 2.4 MB/s write and 5 MB/s read. [Mukul Bhatnagar: actual calculation of timing values from the NAND datasheet] Signed-off-by: Sekhar Nori <nsekhar@ti.com> Cc: Mukul Bhatnagar <mbhatnagar@ti.com> Acked-by: Kevin Hilman <khilman@deeprootsystems.com> Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
* davinci: am17x/da830/omap-l137 evm: setup NAND flash timingSudhakar Rajashekhara2010-09-24
| | | | | | | | | | | | | | | | | | | Setup the NAND flash timings for DA830 EVM. Before configuring the timing values, throughput calculation using dd command yielded 477 kB/s write and 970 kB/s read speed. After the timing configuration, the throughput was measured to be 2.5 MB/s write and 5.1 MB/s read. [Mukul Bhatnagar: actual calculation of timing values from the NAND datasheet] Signed-off-by: Sudhakar Rajashekhara <sudhakar.raj@ti.com> Cc: Mukul Bhatnagar <mbhatnagar@ti.com> Signed-off-by: Sekhar Nori <nsekhar@ti.com> Acked-by: Kevin Hilman <khilman@deeprootsystems.com> Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
* davinci: dm644x evm: setup NAND flash timingSekhar Nori2010-09-24
| | | | | | | | | | | | | | | | | | The DM644x EVM nand flash timing was earlier being done as a special case in the NAND driver itself. With the NAND driver now capable of progamming the AEMIF interface using timing data passed from the platform, the timing values are being moved into their rightful place in the EVM specific board file. The values being programmed match what was being done earlier and thus do not represent any change in performance/functionality. Signed-off-by: Sekhar Nori <nsekhar@ti.com> Acked-by: Kevin Hilman <khilman@deeprootsystems.com> Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
* nand: davinci: add support for timing configurationSekhar Nori2010-09-24
| | | | | | | | | | | | | | | | | | | | | This patch modifies the DaVinci NAND driver to use the new AEMIF timing setup API to configure the NAND access timings. Earlier, AEMIF configuration was being done as a special case for DM644x board, but now more boards emerge which have capability to boot for other media (SPI flash, NOR flash) and have the kernel access NAND flash. This means that kernel cannot always depend on the bootloader to setup the NAND. Also, on platforms such as da850/omap-l138, the aemif input frequency changes as cpu frequency changes; necessiating re-calculation of timimg values as part of cpufreq transtitions. This patch forms the basis for adding that support. Signed-off-by: Sekhar Nori <nsekhar@ti.com> Acked-by: Kevin Hilman <khilman@deeprootsystems.com> Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com>
* davinci: add support for aemif timing configurationSekhar Nori2010-09-24
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | This patch adds support to configure the AEMIF interface with supplied timing values. Since this capability is useful both from NOR and NAND flashes, it is provided as a new interface and in a file of its own. AEMIF timing configuration is required in cases: 1) Where the AEMIF clock rate can change at runtime (a side affect of cpu frequency change). 2) Where U-Boot does not support NAND/NOR but supports other media like SPI Flash or MMC/SD and thus does not care about setting up the AEMIF timing for kernel to use. 3) Where U-Boot just hasn't configured the timing values and cannot be upgraded because the box is already in the field. Since there is now a header file for AEMIF interface, the common (non-NAND specific) defines for AEMIF registers have been moved from nand.h into the newly created aemif.h Signed-off-by: Sekhar Nori <nsekhar@ti.com> Acked-by: Kevin Hilman <khilman@deeprootsystems.com> Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
* davinci: dm365 evm: use EDMAQ_3 as the audio DMA queueSekhar Nori2010-09-24
| | | | | | | | | | | | | | By default the audio driver uses EDMAQ_0 as the DMA queue, but on DM365 this queue is specially designed for video transfers with a large fifo size. Having both audio and video transfers on the same queue leads to noise on the audio side. This patch changes the audio queue number for DM365 to EDMAQ_3. Signed-off-by: Sekhar Nori <nsekhar@ti.com> Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
* DA850: Add MMCSD1 resources, platform device and convenience registration ↵Juha Kuikka2010-09-24
| | | | | | | | | function Add resources, platform device and convenience registration function for DA850's second MMC/SD controller (MMCSD1). Signed-off-by: Juha Kuikka <juha.kuikka@elektrobit.com> Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
* DA850: Split MMCSD clock into two to support both MMCSD peripheralsJuha Kuikka2010-09-24
| | | | | | | | Split mmcsd_clk into mmcsd0_clk and mmcsd1_clk and add davinci_mmc.1 in preparation for adding support for MMCSD1 peripheral in DA850. Signed-off-by: Juha Kuikka <juha.kuikka@elektrobit.com> Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
* DA850: Add LPSC id for MMCSD1 peripheralJuha Kuikka2010-09-24
| | | | | | | Add LPSC id for DA850's MMCSD1 peripheral. Signed-off-by: Juha Kuikka <juha.kuikka@elektrobit.com> Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
* arm: mach-davinci: check irq2ctlr() resultKulikov Vasiliy2010-09-24
| | | | | | | | | | | | | | | | | | | | | | | | | | | If irq2ctlr() fails return IRQ_NONE. Also as it can fail make 'ctlr' signed. The semantic patch that finds this problem (many false-positive results): (http://coccinelle.lip6.fr/) // <smpl> @ r1 @ identifier f; @@ int f(...) { ... } @@ identifier r1.f; type T; unsigned T x; @@ *x = f(...) ... *x > 0 Signed-off-by: Kulikov Vasiliy <segooon@gmail.com> Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
* DA850: move MII/RMII pin lists to the board fileSergei Shtylyov2010-09-24
| | | | | | | | | | | | The CPGMAC pin list in da850.c was incorrectly split into two MII/RMII mode specific pin lists, while what pin group is used is a function of how the board is wired. Copy the pin lists to board-da850-evm.c, renaming them accordingly, and merge the two lists in da850.c into one, da850_cpgmac_pins[], representing the CPGMAC module as a whole... Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com> Tested-by: Ben Gardiner <bengardiner@nanometrics.ca> Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
* DA850: move NAND/NOR pin lists to the board fileSergei Shtylyov2010-09-24
| | | | | | | | | | | | | | | The NAND/NOR flash pin lists (da850_nand_pins/da850_nor_pins) are purely board specific and as such shouldn't be in da850.c -- copy them to board-da850-evm.c, renaming to da850_evm_nand_pins/da850_evm_nor_pins respectively, and merge the two lists in da850.c into one, representing the EMIF 2.5 module as a whole, just like we have it in da830.c... While at it, remove the '__init' modifier from da850_evm_setup_nor_nand() as this function is called from non '__init' code... Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com> Tested-by: Ben Gardiner <bengardiner@nanometrics.ca> Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
* Davinci: tnetv107x: retain psc reg base after initCyril Chemparathy2010-09-24
| | | | | | | | | This is a bugfix for the original tnetv107x submission series. The psc_regs base array was being discarded post-init, and this was causing a crash during post-init clock enable/disable. Signed-off-by: Cyril Chemparathy <cyril@ti.com> Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
* davinci: am18x/da850/omap-l138: keep async clock constant with cpufreqSekhar Nori2010-09-24
| | | | | | | | | Keep PLL0 SYSCLK3 at a constant rate of 100MHz. This enables the AEMIF timing to remain valid even as the PLL0 output is changed by cpufreq driver to save power. Signed-off-by: Sekhar Nori <nsekhar@ti.com> Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
* davinci: cpufreq: add support for keeping an additional clock constantSekhar Nori2010-09-24
| | | | | | | | | | | | | | | | On OMAP-L138 SoC, some of the sysclks need not be at a fixed ratio to CPU clock and can be kept at a relatively constant rate by adjusting the PLLDIVn ratio even as cpufreq goes ahead and changes the CPU clock. This feature can be used to keep the EMIFA (PLL0 SYSCLK3) clock at a constant rate so that the EMIF timings need not be re-programmed whenever the CPU frequency changes. This patch adds the required suppport to cpufreq driver. Signed-off-by: Sekhar Nori <nsekhar@ti.com> Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
* davinci: clock: add support for setting sysclk rateSekhar Nori2010-09-24
| | | | | | | | | | | | | | | Setting sysclk rate will be useful in cases where the sysclk is not at a fixed ratio to the PLL output but can asynchronously be changed. This support forms the basis of attempt to keep the AEMIF clock constant on OMAP-L138 even as PLL0 output changes as ARM clock is changed to save power. This patch has been tested on OMAP-L138. Signed-off-by: Sekhar Nori <nsekhar@ti.com> Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
* davinci: dm365: disable pulldowns for all MMC/SD1 pins.Sekhar Nori2010-09-24
| | | | | | | | | | | | | This patch disables internal pulldowns for all MMC/SD1 pins. Presently only MMCSD1_CMD pin's pull down is disabled, but with this some MMC/SD cards do not get detected on MMC/SD1 slot of the EVM. The problem was reproducible with SanDisk 4GB SDHC card. Reported-by: Stephane Bovagne <s-bovagne@ti.com> Signed-off-by: Sekhar Nori <nsekhar@ti.com> Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
* davinci: am18x/da850/omap-l138: use 'NOM' voltage defined in datasheet as ↵Sekhar Nori2010-09-24
| | | | | | | | | | | | | | | | | | min voltage For each DA850 OPP, the normal ('NOM') voltage defined in the tecnical reference manual (TRM) is actually the minimum voltage the frequency is supported at. The minimum ('MIN') voltage defined in TRM is meant to take care of voltage fluctuations and the device should not be run at this voltage for extended periods of time. Fix the OPP definitions to define the cvdd_min as the normal voltage defined in the datasheet. Signed-off-by: Sekhar Nori <nsekhar@ti.com> Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
* davinci: introduce support for AM1x ARM9 microprocessorsSekhar Nori2010-09-24
| | | | | | | | | | | | | | | | | | | | | | | | | The Sitara AM17x SoCs from TI are an OMAP-L137 pin-to-pin compatible ARM9 microprocessor offering from TI. The Sitara AM18x SoCs from TI are an OMAP-L138 pin-to-pin compatible ARM9 microprocessor offering from TI. More information about these processors available at: www.ti.com/am1x Because of their compatibiliy with OMAP-L1x, the kernel support for OMAP-L1x is fully relevant to AM1x processors. This patch updates the Kconfig prompt and help text to include the AM1x part names to help users select configurations required for these parts easily. Also, the hardware information that shows up in /proc/cpuinfo is updated to show applicability of the respective OMAP-L1x EVMs for AM1x parts. Signed-off-by: Sekhar Nori <nsekhar@ti.com> Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
* davinci: clean up inconsistent usage of spaces in KconfigSekhar Nori2010-09-24
| | | | | | | | | | | | | In arch/arm/mach-davinci/Kconfig, some of the configuration items are indented with multiple spaces instead of tabs. Also, in couple of places, two spaces are used in the middle of help text where one should do. This patch fixes both issues. Signed-off-by: Sekhar Nori <nsekhar@ti.com> Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
* davinci: cpufreq: bailout on regulator errorsSekhar Nori2010-09-24
| | | | | | | | | | | | | | | | | | | | | | Current cpufreq code does not consider errors that can occur while changing voltage. Code to increase CPU frequency goes ahead even in the case the regulator has failed to increase the voltage. This leads to hard error since lower voltages cannot support increased frequency. Prevent this by not increasing frequency in case increasing voltage is not successful. Also, do not lower the voltage if changing the cpu frequency has failed for some reason. Note that we do not return error on failure to decrease voltage as that is not a hard error. Build fix for non-cpufreq kernels by Caglar Akyuz. Signed-off-by: Sekhar Nori <nsekhar@ti.com> Cc: Caglar Akyuz <caglar@bilkon-kontrol.com.tr> Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
* davinci: clock: make 'disable unused clocks' printk debug onlyKevin Hilman2010-09-24
| | | | | | | The long list of clocks being disabled on boot is noisy and not needed for standard boots. Make this a debug printk instead. Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
* davinci: da850 evm: sparse cleanup: make file local variables staticSekhar Nori2010-09-24
| | | | | | | | | | | | | | | Without this cleanup, sparse checker reports warnings of the type: CHECK arch/arm/mach-davinci/board-da850-evm.c arch/arm/mach-davinci/board-da850-evm.c:112:22: warning: symbol 'da850_evm_nandflash_partition' was not declared. Should it be static? The nand flash partitions and regulator supplies are used within the EVM file and so should have been static This patch has been boot tested on DA830 and DA850 EVMs. Signed-off-by: Sekhar Nori <nsekhar@ti.com> Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
* frv: double syscall restarts, syscall restart in sigreturn()Al Viro2010-09-20
| | | | | | | | | | | We need to make sure that only the first do_signal() to be handled on the way out syscall will bother with syscall restarts; additionally, the check on the "signal has user handler" path had been wrong - compare with restart prevention in sigreturn()... Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
* frv: handling of restart into restart_syscall is fsckedAl Viro2010-09-20
| | | | | | | | | do_signal() should place the syscall number in gr7, not gr8 when handling ERESTART_WOULDBLOCK. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
* frv: avoid infinite loop of SIGSEGV deliveryAl Viro2010-09-20
| | | | | | | | | | | Use force_sigsegv() rather than force_sig(SIGSEGV, ...) as the former resets the SEGV handler pointer which will kill the process, rather than leaving it open to an infinite loop if the SEGV handler itself caused a SEGV signal. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
* frv: fix address verification holes in setup_frame/setup_rt_frameAl Viro2010-09-20
| | | | | | | | | | | | | | a) sa_handler might be maliciously set to point to kernel memory; blindly dereferencing it in FDPIC case is a Bad Idea(tm). b) I'm not sure you need that set_fs(USER_DS) there at all, but if you do, you'd better do it *before* checking the frame you've decided to use with access_ok(), lest sigaltstack() becomes a convenient roothole. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
* frv: restart_block.fn needs to be reset on sigreturnAl Viro2010-09-20
| | | | | | | | | Reset restart_block.fn on executing a sigreturn such that any currently pending system call restarts will be forced to return -EINTR. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
* Merge branch 'for-linus' of ↵Linus Torvalds2010-09-19
|\ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | git://git.kernel.org/pub/scm/linux/kernel/git/mattst88/alpha-2.6 * 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mattst88/alpha-2.6: alpha: deal with multiple simultaneously pending signals alpha: fix a 14 years old bug in sigreturn tracing alpha: unb0rk sigsuspend() and rt_sigsuspend() alpha: belated ERESTART_RESTARTBLOCK race fix alpha: Shift perf event pending work earlier in timer interrupt alpha: wire up fanotify and prlimit64 syscalls alpha: kill big kernel lock alpha: fix build breakage in asm/cacheflush.h alpha: remove unnecessary cast from void* in assignment. alpha: Use static const char * const where possible
| * alpha: deal with multiple simultaneously pending signalsAl Viro2010-09-18
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Unlike the other targets, alpha sets _one_ sigframe and buggers off until the next syscall/interrupt, even if more signals are pending. It leads to quite a few unpleasant inconsistencies, starting with SIGSEGV potentially arriving not where it should and including e.g. mess with sigsuspend(); consider two pending signals blocked until sigsuspend() unblocks them. We pick the first one; then, if we are hit by interrupt while in the handler, we process the second one as well. If we are not, and if no syscalls had been made, we get out of the first handler and leave the second signal pending; normally sigreturn() would've picked it anyway, but here it starts with restoring the original mask and voila - the second signal is blocked again. On everything else we get both delivered consistently. It's actually easy to fix; the only thing to watch out for is prevention of double syscall restart. Fortunately, the idea I've nicked from arm fix by rmk works just fine... Testcase demonstrating the behaviour in question; on alpha we get one or both flags set (usually one), on everything else both are always set. #include <signal.h> #include <stdio.h> int had1, had2; void f1(int sig) { had1 = 1; } void f2(int sig) { had2 = 1; } main() { sigset_t set1, set2; sigemptyset(&set1); sigemptyset(&set2); sigaddset(&set2, 1); sigaddset(&set2, 2); signal(1, f1); signal(2, f2); sigprocmask(SIG_SETMASK, &set2, NULL); raise(1); raise(2); sigsuspend(&set1); printf("had1:%d had2:%d\n", had1, had2); } Tested-by: Michael Cree <mcree@orcon.net.nz> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Matt Turner <mattst88@gmail.com>
| * alpha: fix a 14 years old bug in sigreturn tracingAl Viro2010-09-18
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The way sigreturn() is implemented on alpha breaks PTRACE_SYSCALL, all way back to 1.3.95 when alpha has grown PTRACE_SYSCALL support. What happens is direct return to ret_from_syscall, in order to bypass mangling of a3 (error indicator) and prevent other mutilations of registers (e.g. by syscall restart). That's fine, but... the entire TIF_SYSCALL_TRACE codepath is kept separate on alpha and post-syscall stopping/notifying the tracer is after the syscall. And the normal path we are forcibly switching to doesn't have it. So we end up with *one* stop in traced sigreturn() vs. two in other syscalls. And yes, strace is visibly broken by that; try to strace the following #include <signal.h> #include <stdio.h> void f(int sig) {} main() { signal(SIGHUP, f); raise(SIGHUP); write(1, "eeeek\n", 6); } and watch the show. The close(1) = 405 in the end of strace output is coming from return value of write() (6 == __NR_close on alpha) and syscall number of exit_group() (__NR_exit_group == 405 there). The fix is fairly simple - the only thing we end up missing is the call of syscall_trace() and we can tell whether we'd been called from the SYSCALL_TRACE path by checking ra value. Since we are setting the switch_stack up (that's what sys_sigreturn() does), we have the right environment for calling syscall_trace() - just before we call undo_switch_stack() and return. Since undo_switch_stack() will overwrite s0 anyway, we can use it to store the result of "has it been called from SYSCALL_TRACE path?" check. The same thing applies in rt_sigreturn(). Tested-by: Michael Cree <mcree@orcon.net.nz> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Matt Turner <mattst88@gmail.com>
| * alpha: unb0rk sigsuspend() and rt_sigsuspend()Al Viro2010-09-18
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Old code used to set regs->r0 and regs->r19 to force the right return value. Leaving that after switch to ERESTARTNOHAND was a Bad Idea(tm), since now that screws the restart - if we hit the case when get_signal_to_deliver() returns 0, we will step back to syscall insn, with v0 set to EINTR and a3 to 1. The latter won't matter, since EINTR is 4, aka __NR_write. Testcase: #include <signal.h> #define _GNU_SOURCE #include <unistd.h> #include <sys/syscall.h> main() { sigset_t mask; sigemptyset(&mask); sigaddset(&mask, SIGCONT); sigprocmask(SIG_SETMASK, &mask, NULL); kill(0, SIGCONT); syscall(__NR_sigsuspend, 1, "b0rken\n", 7); } results on alpha in immediate message to stdout... Fix is obvious; moreover, since we don't need regs anymore, we can switch to normal prototypes for these guys and lose the wrappers. Even better, rt_sigsuspend() is identical to generic version in kernel/signal.c now. Tested-by: Michael Cree <mcree@orcon.net.nz> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Matt Turner <mattst88@gmail.com>
| * alpha: belated ERESTART_RESTARTBLOCK race fixAl Viro2010-09-18
| | | | | | | | | | | | | | | | | | same thing as had been done on other targets back in 2003 - move setting ->restart_block.fn into {rt_,}sigreturn(). Tested-by: Michael Cree <mcree@orcon.net.nz> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Matt Turner <mattst88@gmail.com>
| * alpha: Shift perf event pending work earlier in timer interruptMichael Cree2010-09-18
| | | | | | | | | | | | | | | | | | | | | | Pending work from the performance event subsystem is executed in the timer interrupt. This patch shifts the call to perf_event_do_pending() before the call to update_process_times() as the latter may call back into the perf event subsystem and it is prudent to have the pending work executed first. Signed-off-by: Michael Cree <mcree@orcon.net.nz> Signed-off-by: Matt Turner <mattst88@gmail.com>
| * alpha: wire up fanotify and prlimit64 syscallsMikael Pettersson2010-09-18
| | | | | | | | | | | | | | | | | | | | | | The 2.6.36-rc kernel added three new system calls: fanotify_init, fanotify_mark, and prlimit64. This patch wires them up on Alpha. Built and booted on an XP900. Untested beyond that. Signed-off-by: Mikael Pettersson <mikpe@it.uu.se> Signed-off-by: Matt Turner <mattst88@gmail.com>
generated by cgit v1.2.2 (git 2.25.0) at 2024-10-03 01:32:13 -0400