aboutsummaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAge
...
| * ahci_sunxi: Use msleep instead of mdelayHans de Goede2014-02-23
| | | | | | | | | | | | | | | | ahci_sunxi_phy_init is called from the probe and resume code paths, and sleeping is safe in both, so use msleep instead of mdelay. Signed-off-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Tejun Heo <tj@kernel.org>
| * ahci_platform: Drop unused ahci_platform_data membersHans de Goede2014-02-22
| | | | | | | | | | | | | | | | | | These members are not used anywhere, and in the future we want ahci_platform_data to go away entirely so there is no reason to keep these around. Signed-off-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Tejun Heo <tj@kernel.org>
| * ahci_platform: Drop support for imx53-ahci platform device typeHans de Goede2014-02-22
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Since the 3.13 release the ahci_imx driver has proper devicetree enabled support for ahci on imx53 and that is used instead of the old board file created imx53-ahci platform device. Note this patch also complete drops the id-table, an id-table is not needed for a single id platform driver, the name field in the driver struct suffices. And the code already has an explicit "MODULE_ALIAS("platform:ahci");" so the id-table is not needed for that either. Cc: Marek Vasut <marex@denx.de> Signed-off-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Tejun Heo <tj@kernel.org>
| * ahci_platform: Drop support for ahci-strict platform device typeHans de Goede2014-02-22
| | | | | | | | | | | | | | | | I've done a grep over the entire kernel tree and nothing is using this (anymore?). Signed-off-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Tejun Heo <tj@kernel.org>
| * ata: ahci_platform: runtime resume the device before useRoger Quadros2014-02-22
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | On OMAP platforms the device needs to be runtime resumed before it can be accessed. The OMAP HWMOD framework takes care of enabling the module and its resources based on the device's runtime PM state. In this patch we runtime resume during .probe() and runtime suspend after .remove(). We also update the runtime PM state during .resume(). CC: Balaji T K <balajitk@ti.com> Signed-off-by: Roger Quadros <rogerq@ti.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Tejun Heo <tj@kernel.org>
| * ata: ahci_platform: Manage SATA PHYRoger Quadros2014-02-22
| | | | | | | | | | | | | | | | | | | | | | | | | | Some platforms have a PHY hooked up to the SATA controller. The PHY needs to be initialized and powered up for SATA to work. We do that using the PHY framework. tj: Minor comment formatting updates. CC: Balaji T K <balajitk@ti.com> Signed-off-by: Roger Quadros <rogerq@ti.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Tejun Heo<tj@kernel.org>
| * ata: ahci_platform: Update DT compatible listRoger Quadros2014-02-22
| | | | | | | | | | | | | | | | | | | | The ahci_platform driver supports "snps,dwc-ahci". Add this to the DT binding information. Signed-off-by: Roger Quadros <rogerq@ti.com> Reviewed-by: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Tejun Heo <tj@kernel.org>
| * ata: ahci_platform: Add DT compatible for Synopsis DWC AHCI controllerRoger Quadros2014-02-22
| | | | | | | | | | | | | | | | | | | | Add compatible string "snps,dwc-ahci", which should be used for Synopsis Designware SATA cores. e.g. on TI OMAP5 and DRA7 platforms. Signed-off-by: Roger Quadros <rogerq@ti.com> Reviewed-by: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Tejun Heo <tj@kernel.org>
| * ahci-imx: Port to library-ised ahci_platformHans de Goede2014-02-22
| | | | | | | | | | | | | | | | | | | | | | | | | | This avoids the ugliness of creating a nested platform device from probe. While moving it around anyways, move the mk6q phy init code from probe to imx_sata_enable, as the phy needs to be re-initialized on resume too, otherwise the drive won't be recognized after resume. Tested on a wandboard i.mx6 quad. Signed-off-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Tejun Heo <tj@kernel.org>
| * ARM: sunxi: Add support for Allwinner SUNXi SoCs sata to ahci_platformOlliver Schinagl2014-02-22
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This patch adds support for the ahci sata controler found on Allwinner A10 and A20 SoCs to the ahci_platform driver. Orignally written by Olliver Schinagl using the approach of having a platform device which probe method creates a new child platform device which gets driven by ahci_platform.c, as done by ahci_imx.c . Refactored by Hans de Goede to add most of the non sunxi specific functionality to ahci_platform.c and use a platform_data pointer from of_device_id for the sunxi specific bits. Signed-off-by: Olliver Schinagl <oliver@schinagl.nl> Signed-off-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Tejun Heo <tj@kernel.org>
| * ahci-platform: "Library-ise" suspend / resume functionalityHans de Goede2014-02-22
| | | | | | | | | | | | | | | | | | | | | | Split suspend / resume code into host suspend / resume functionality and resource enable / disabling phases, and export the new suspend_ / resume_host functions. tj: Minor comment formatting updates. Signed-off-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Tejun Heo <tj@kernel.org>
| * ahci-platform: "Library-ise" ahci_probe functionalityHans de Goede2014-02-22
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | ahci_probe consists of 3 steps: 1) Get resources (get mmio, clks, regulator) 2) Enable resources, handled by ahci_platform_enable_resouces 3) The more or less standard ahci-host controller init sequence This commit refactors step 1 and 3 into separate functions, so the platform drivers for AHCI implementations which need a specific order in step 2, and / or need to do some custom register poking at some time, can re-use ahci-platform.c code without needing to copy and paste it. Note that ahci_platform_init_host's prototype takes the 3 non function members of ahci_platform_data as arguments, the idea is that drivers using the new exported utility functions will not use ahci_platform_data at all, and hopefully in the future ahci_platform_data can go away entirely. tj: Minor comment formatting updates. Signed-off-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Tejun Heo <tj@kernel.org>
| * ahci-platform: Add enable_ / disable_resources helper functionsHans de Goede2014-02-22
| | | | | | | | | | | | | | tj: Minor comment formatting updates. Signed-off-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Tejun Heo <tj@kernel.org>
| * ahci-platform: Add support for an optional regulator for sata-target powerHans de Goede2014-02-22
| | | | | | | | | | Signed-off-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Tejun Heo <tj@kernel.org>
| * ahci-platform: Add support for devices with more then 1 clockHans de Goede2014-02-22
| | | | | | | | | | | | | | | | | | | | The allwinner-sun4i AHCI controller needs 2 clocks to be enabled and the imx AHCI controller needs 3 clocks to be enabled. tj: Minor comment formatting updates. Signed-off-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Tejun Heo <tj@kernel.org>
| * libahci: Allow drivers to override start_engineHans de Goede2014-02-22
| | | | | | | | | | | | | | | | | | | | | | | | | | | | Allwinner A10 and A20 ARM SoCs have an AHCI sata controller which needs a special register to be poked before starting the DMA engine. This register gets reset on an ahci_stop_engine call, so there is no other place then ahci_start_engine where this poking can be done. This commit allows drivers to override ahci_start_engine behavior for use by the Allwinner AHCI driver (and potentially other drivers in the future). Signed-off-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Tejun Heo <tj@kernel.org>
| * ata: libahci: replace obsolete simple_strtoul() with kstrtouint()Daeseok Youn2014-02-19
| | | | | | | | | | Signed-off-by: Daeseok Youn <daeseok.youn@gmail.com> Signed-off-by: Tejun Heo <tj@kernel.org>
| * ata: libahci: make ahci_pmp_retry_softreset() as staticDaeseok Youn2014-02-19
| | | | | | | | | | | | | | | | | | | | sparse says: drivers/ata/libahci.c:1390:5: warning: symbol 'ahci_pmp_retry_softreset' was not declared. Should it be static? Signed-off-by: Daeseok Youn <daeseok.youn@gmail.com> Signed-off-by: Tejun Heo <tj@kernel.org>
| * ata: CONFIG_ATA is libataRandy Dunlap2014-02-18
| | | | | | | | | | | | | | | | | | | | | | Let users know that CONFIG_ATA is the kconfig symbol for libata, since libata is mentioned in documentation and messages several times. Also correct a grammar typo. Signed-off-by: Randy Dunlap <rdunlap@infradead.org> Signed-off-by: Tejun Heo <tj@kernel.org>
| * pata_imx: Propagate the real error code on platform_get_irq() failureFabio Estevam2014-02-18
| | | | | | | | | | | | | | | | | | | | | | | | No need to return a 'fake' return value on platform_get_irq() failure. Just return the error code itself instead. Also, change the error condition to irq < 0, so that only negative values are treated as errors. Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com> Signed-off-by: Tejun Heo <tj@kernel.org>
| * pata_imx: Use devm_ioremap_resource() to simplify codeFabio Estevam2014-02-18
| | | | | | | | | | | | | | | | | | | | | | Using devm_ioremap_resource() can lead to code simplication, as we don't need to explicitily check for error returned by platform_get_resource(). Also, no need to print an error message when devm_ioremap_resource() fails, as the OOM code code will shout loudly on such condition. Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com> Signed-off-by: Tejun Heo <tj@kernel.org>
| * ata: delete non-required instances of include <linux/init.h>Paul Gortmaker2014-02-13
| | | | | | | | | | | | | | | | | | | | | | None of these files are actually using any __init type directives and hence don't need to include <linux/init.h>. Most are just a left over from __devinit and __cpuinit removal, or simply due to code getting copied from one driver to the next. Cc: linux-ide@vger.kernel.org Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com> Signed-off-by: Tejun Heo <tj@kernel.org>
* | Merge branch 'for-3.15' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/wqLinus Torvalds2014-03-31
|\ \ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Pull workqueue changes from Tejun Heo: "PREPARE_[DELAYED_]WORK() were used to change the work function of work items without fully reinitializing it; however, this makes workqueue consider the work item as a different one from before and allows the work item to start executing before the previous instance is finished which can lead to extremely subtle issues which are painful to debug. The interface has never been popular. This pull request contains patches to remove existing usages and kill the interface. As one of the changes was routed during the last devel cycle and another depended on a pending change in nvme, for-3.15 contains a couple merge commits. In addition, interfaces which were deprecated quite a while ago - __cancel_delayed_work() and WQ_NON_REENTRANT - are removed too" * 'for-3.15' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq: workqueue: remove deprecated WQ_NON_REENTRANT workqueue: Spelling s/instensive/intensive/ workqueue: remove PREPARE_[DELAYED_]WORK() staging/fwserial: don't use PREPARE_WORK afs: don't use PREPARE_WORK nvme: don't use PREPARE_WORK usb: don't use PREPARE_DELAYED_WORK floppy: don't use PREPARE_[DELAYED_]WORK ps3-vuart: don't use PREPARE_WORK wireless/rt2x00: don't use PREPARE_WORK in rt2800usb.c workqueue: Remove deprecated __cancel_delayed_work()
| * | workqueue: remove deprecated WQ_NON_REENTRANTZhangZhen2014-03-29
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Tejun Heo has made WQ_NON_REENTRANT useless in the dbf2576e37 ("workqueue: make all workqueues non-reentrant"). So remove its usages and definition. This patch doesn't introduce any behavior changes. tj: minor description updates. Signed-off-by: ZhangZhen <zhenzhang.zhang@huawei.com> Sigend-off-by: Tejun Heo <tj@kernel.org> Acked-by: James Chapman <jchapman@katalix.com> Acked-by: Ulf Hansson <ulf.hansson@linaro.org>
| * | workqueue: Spelling s/instensive/intensive/Geert Uytterhoeven2014-03-24
| | | | | | | | | | | | | | | Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be> Signed-off-by: Tejun Heo <tj@kernel.org>
| * | workqueue: remove PREPARE_[DELAYED_]WORK()Tejun Heo2014-03-07
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Peter Hurley noticed that since a2c1c57be8d9 ("workqueue: consider work function when searching for busy work items"), a work item which gets assigned a different work function would break out of the non-reentrancy guarantee as workqueue would consider it a different work item. This is fragile and extremely subtle. PREPARE_[DELAYED_]WORK() have never been used widely and its semantics has always been somewhat iffy. If the work item is known not to be on queue when PREPARE_WORK() is called, there's no difference from using INIT_WORK(). If the work item may be queued at the time of PREPARE_WORK(), we can't really tell whether the old or new function will be executed the next time. We really don't want this level of subtlety in workqueue interface for such marginal use cases. The previous patches converted all existing users away from PREPARE_[DELAYED_]WORK(). Let's remove them. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Peter Hurley <peter@hurleysoftware.com> Link: http://lkml.kernel.org/g/1392493119-9277-1-git-send-email-peter@hurleysoftware.com
| * | staging/fwserial: don't use PREPARE_WORKTejun Heo2014-03-07
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | PREPARE_[DELAYED_]WORK() are being phased out. They have few users and a nasty surprise in terms of reentrancy guarantee as workqueue considers work items to be different if they don't have the same work function. fwtty_peer->work is multiplexed with multiple work functions. Introduce fwserial_peer_workfn() which invokes fwtty_peer->workfn and always use it as the work function and update the users to set the ->workfn field instead of overriding the work function using PREPARE_WORK(). It would probably be best to route this with other related updates through the workqueue tree. Compile tested. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Peter Hurley <peter@hurleysoftware.com>
| * | afs: don't use PREPARE_WORKTejun Heo2014-03-07
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | PREPARE_[DELAYED_]WORK() are being phased out. They have few users and a nasty surprise in terms of reentrancy guarantee as workqueue considers work items to be different if they don't have the same work function. afs_call->async_work is multiplexed with multiple work functions. Introduce afs_async_workfn() which invokes afs_call->async_workfn and always use it as the work function and update the users to set the ->async_workfn field instead of overriding the work function using PREPARE_WORK(). It would probably be best to route this with other related updates through the workqueue tree. Compile tested. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: David Howells <dhowells@redhat.com> Cc: linux-afs@lists.infradead.org
| * | nvme: don't use PREPARE_WORKTejun Heo2014-03-07
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | PREPARE_[DELAYED_]WORK() are being phased out. They have few users and a nasty surprise in terms of reentrancy guarantee as workqueue considers work items to be different if they don't have the same work function. nvme_dev->reset_work is multiplexed with multiple work functions. Introduce nvme_reset_workfn() which invokes nvme_dev->reset_workfn and always use it as the work function and update the users to set the ->reset_workfn field instead of overriding the work function using PREPARE_WORK(). It would probably be best to route this with other related updates through the workqueue tree. Compile tested. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Matthew Wilcox <willy@linux.intel.com> Cc: linux-nvme@lists.infradead.org
| * | usb: don't use PREPARE_DELAYED_WORKTejun Heo2014-03-07
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | PREPARE_[DELAYED_]WORK() are being phased out. They have few users and a nasty surprise in terms of reentrancy guarantee as workqueue considers work items to be different if they don't have the same work function. usb_hub->init_work is multiplexed with multiple work functions; however, the work item is never queued while in-flight, so we can simply use INIT_DELAYED_WORK() before each queueing. It would probably be best to route this with other related updates through the workqueue tree. Lightly tested. v2: Greg and Alan confirm that the work item is never queued while in-flight. Simply use INIT_DELAYED_WORK(). Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Alan Stern <stern@rowland.harvard.edu> Cc: linux-usb@vger.kernel.org
| * | floppy: don't use PREPARE_[DELAYED_]WORKTejun Heo2014-03-07
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | PREPARE_[DELAYED_]WORK() are being phased out. They have few users and a nasty surprise in terms of reentrancy guarantee as workqueue considers work items to be different if they don't have the same work function. floppy has been multiplexing floppy_work and fd_timer with multiple work functions. Introduce floppy_work_workfn() and fd_timer_workfn() which invoke floppy_work_fn and fd_timer_fn respectively and always use the two functions as the work functions and update the users to set floppy_work_fn and fd_timer_fn instead of overriding work functions using PREPARE_[DELAYED_]WORK(). It would probably be best to route this with other related updates through the workqueue tree. Lightly tested using qemu. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Jiri Kosina <jkosina@suse.cz>
| * | ps3-vuart: don't use PREPARE_WORKTejun Heo2014-03-07
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | PREPARE_[DELAYED_]WORK() are being phased out. They have few users and a nasty surprise in terms of reentrancy guarantee as workqueue considers work items to be different if they don't have the same work function. ps3_vuart wasn't overriding the work item with multiple work functions but was using NULL for INIT_WORK() and then single PREPARE_WORK() to set the work function. We can simply invoke INIT_WORK() with the work function and remove the PREPARE_WORK() usage. It would probably be best to route this with other related updates through the workqueue tree. Geoff: Tested on ps3 (DECR-1400). Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Geoff Levand <geoff@infradead.org> Cc: linuxppc-dev@lists.ozlabs.org Cc: cbe-oss-dev@lists.ozlabs.org
| * | wireless/rt2x00: don't use PREPARE_WORK in rt2800usb.cTejun Heo2014-03-07
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | PREPARE_[DELAYED_]WORK() are being phased out. They have few users and a nasty surprise in terms of reentrancy guarantee as workqueue considers work items to be different if they don't have the same work function. Update rt2800usb.c to use INIT_WORK() instead of PREPARE_WORK(). As the work item isn't in active use during rt2800usb_probe_hw(), this doesn't cause any behavior difference. It would probably be best to route this with other related updates through the workqueue tree. Only compile tested. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Ivo van Doorn <IvDoorn@gmail.com> Cc: Gertjan van Wingerde <gwingerde@gmail.com> Cc: Helmut Schaa <helmut.schaa@googlemail.com> Cc: linux-wireless@vger.kernel.org
| * | Merge branch 'master' into wq/for-3.15Tejun Heo2014-03-07
| |\ \ | | | | | | | | | | | | | | | | | | | | | | | | Pull 3.14-rc5 into wq/for-3.15 to receive nvme updates which the scheduled PREPARE_DELAYED_WORK() updates depend on. Signed-off-by: Tejun Heo <tj@kernel.org>
| * \ \ Merge branch 'wq/for-3.14-fixes' into wq/for-3.15Tejun Heo2014-03-07
| |\ \ \ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | To receive 70044d71d31d ("firewire: don't use PREPARE_DELAYED_WORK"). There will be further related updates in for-3.15 branch. Signed-off-by: Tejun Heo <tj@kernel.org>
| * | | | workqueue: Remove deprecated __cancel_delayed_work()Tan Xiaojun2014-02-18
| | |_|/ | |/| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | __cancel_delayed_work() was deprecated by 136b5721d75a ("workqueue: deprecate __cancel_delayed_work()") as cancel_delayed_work() was updated so that it could be used from all contexts. Enough time has passed since the deprecation. Let's remove it. tj: description update Signed-off-by: Tan Xiaojun <tanxiaojun@huawei.com> Signed-off-by: Tejun Heo <tj@kernel.org>
* | | | Merge branch 'for-3.15' of ↵Linus Torvalds2014-03-31
|\ \ \ \ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | git://git.kernel.org/pub/scm/linux/kernel/git/tj/percpu Pull percpu changes from Tejun Heo: "The percpu allocation is now popular enough for the extremely naive range allocator to cause scalability issues. The existing allocator linearly scanned the allocation map on both alloc and free without making use of hint or anything. Al reimplemented the range allocator so that it can use binary search instead of linear scan during free and alloc path uses simple hinting to avoid scanning in common cases. Combined, the new allocator resolves the scalability issue percpu allocator was showing during container benchmark workload" * 'for-3.15' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/percpu: percpu: renew the max_contig if we merge the head and previous block percpu: allocation size should be even percpu: speed alloc_pcpu_area() up percpu: store offsets instead of lengths in ->map[] perpcu: fold pcpu_split_block() into the only caller
| * | | | percpu: renew the max_contig if we merge the head and previous blockJianyu Zhan2014-03-29
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | During pcpu_alloc_area(), we might merge the current head with the previous block. Since we have calculated the max_contig using the size of previous block before we skip it, and now we update the size of previous block, so we should renew the max_contig. Signed-off-by: Jianyu Zhan <nasa4836@gmail.com> Signed-off-by: Tejun Heo <tj@kernel.org>
| * | | | percpu: allocation size should be evenViro2014-03-17
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | 723ad1d90b56 ("percpu: store offsets instead of lengths in ->map[]") updated percpu area allocator to use the lowest bit, instead of sign, to signify whether the area is occupied and forced min align to 2; unfortunately, it forgot to force the allocation size to be even causing malfunctions for the very rare odd-sized allocations. Always force the allocations to be even sized. tj: Wrote patch description. Original-patch-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Tejun Heo <tj@kernel.org>
| * | | | percpu: speed alloc_pcpu_area() upAl Viro2014-03-07
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | If we know that first N areas are all in use, we can obviously skip them when searching for a free one. And that kind of hint is very easy to maintain. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Tejun Heo <tj@kernel.org>
| * | | | percpu: store offsets instead of lengths in ->map[]Al Viro2014-03-07
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Current code keeps +-length for each area in chunk->map[]. It has several unpleasant consequences: * even if we know that first 50 areas are all in use, allocation still needs to go through all those areas just to sum their sizes, just to get the offset of free one. * freeing needs to find the array entry refering to the area in question; again, the need to sum the sizes until we reach the offset we are interested in. Note that offsets are monotonous, so simple binary search would do here. New data representation: array of <offset,in-use flag> pairs. Each pair is represented by one int - we use offset|1 for <offset, in use> and offset for <offset, free> (we make sure that all offsets are even). In the end we put a sentry entry - <total size, in use>. The first entry is <0, flag>; it would be possible to store together the flag for Nth area and offset for N+1st, but that leads to much hairier code. In other words, where the old variant would have 4, -8, -4, 4, -12, 100 (4 bytes free, 8 in use, 4 in use, 4 free, 12 in use, 100 free) we store <0,0>, <4,1>, <12,1>, <16,0>, <20,1>, <32,0>, <132,1> i.e. 0, 5, 13, 16, 21, 32, 133 This commit switches to new data representation and takes care of a couple of low-hanging fruits in free_pcpu_area() - one is the switch to binary search, another is not doing two memmove() when one would do. Speeding the alloc side up (by keeping track of how many areas in the beginning are known to be all in use) also becomes possible - that'll be done in the next commit. Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Tejun Heo <tj@kernel.org>
| * | | | perpcu: fold pcpu_split_block() into the only callerAl Viro2014-03-07
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | ... and simplify the results a bit. Makes the next step easier to deal with - we will be changing the data representation for chunk->map[] and it's easier to do if the code in question is not split between pcpu_alloc_area() and pcpu_split_block(). Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Tejun Heo <tj@kernel.org>
* | | | | Merge tag 'arm64-upstream' of ↵Linus Torvalds2014-03-31
|\ \ \ \ \ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux Pull ARM64 updates from Catalin Marinas: - KGDB support for arm64 - PCI I/O space extended to 16M (in preparation of PCIe support patches) - Dropping ZONE_DMA32 in favour of ZONE_DMA (we only need one for the time being), together with swiotlb late initialisation to correctly setup the bounce buffer - DMA API cache maintenance support (not all ARMv8 platforms have hardware cache coherency) - Crypto extensions advertising via ELF_HWCAP2 for compat user space - Perf support for dwarf unwinding in compat mode - asm/tlb.h converted to the generic mmu_gather code - asm-generic rwsem implementation - Code clean-up * tag 'arm64-upstream' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux: (42 commits) arm64: Remove pgprot_dmacoherent() arm64: Support DMA_ATTR_WRITE_COMBINE arm64: Implement custom mmap functions for dma mapping arm64: Fix __range_ok macro arm64: Fix duplicated Kconfig entries arm64: mm: Route pmd thp functions through pte equivalents arm64: rwsem: use asm-generic rwsem implementation asm-generic: rwsem: de-PPCify rwsem.h arm64: enable generic CPU feature modalias matching for this architecture arm64: smp: make local symbol static arm64: debug: make local symbols static ARM64: perf: support dwarf unwinding in compat mode ARM64: perf: add support for frame pointer unwinding in compat mode ARM64: perf: add support for perf registers API arm64: Add boot time configuration of Intermediate Physical Address size arm64: Do not synchronise I and D caches for special ptes arm64: Make DMA coherent and strongly ordered mappings not executable arm64: barriers: add dmb barrier arm64: topology: Implement basic CPU topology support arm64: advertise ARMv8 extensions to 32-bit compat ELF binaries ...
| * | | | | arm64: Remove pgprot_dmacoherent()Catalin Marinas2014-03-24
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Since this macro is identical to pgprot_writecombine() and is only used in a single place, remove it completely to avoid confusion. On ARMv7+ processors, the coherent DMA mapping must be Normal NonCacheable (a.k.a. writecombine) to avoid mismatched hardware attribute aliases (with the kernel linear mapping as Normal Cacheable). Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
| * | | | | arm64: Support DMA_ATTR_WRITE_COMBINELaura Abbott2014-03-24
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | DMA_ATTR_WRITE_COMBINE is currently ignored. Set the pgprot appropriately for non coherent opperations. Signed-off-by: Laura Abbott <lauraa@codeaurora.org> Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
| * | | | | arm64: Implement custom mmap functions for dma mappingLaura Abbott2014-03-24
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The current dma_ops do not specify an mmap function so maping falls back to the default implementation. There are at least two issues with using the default implementation: 1) The pgprot is always pgprot_noncached (strongly ordered) memory even with coherent operations 2) dma_common_mmap calls virt_to_page on the remapped non-coherent address which leads to invalid memory being mapped. Fix both these issue by implementing a custom mmap function which correctly accounts for remapped addresses and sets vm_pg_prot appropriately. Signed-off-by: Laura Abbott <lauraa@codeaurora.org> [catalin.marinas@arm.com: replaced "arm64_" with "__" prefix for consistency] Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
| * | | | | arm64: Fix __range_ok macroChristopher Covington2014-03-20
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Without this, the following scenario is incorrectly determined to be invalid. addr 0x7f_ffffe000 size 8192 addr_limit 0x80_00000000 This behavior was observed while trying to vmsplice the stack as part of a CRIU dump of a process on a system started with the norandmaps kernel parameter. Signed-off-by: Christopher Covington <cov@codeaurora.org> Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
| * | | | | arm64: Fix duplicated Kconfig entriesMark Brown2014-03-19
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Probably due to rebasing over the lengthy time it took to get the patch merged commit addea9ef055b (cpufreq: enable ARM drivers on arm64) added a duplicate Power management options section. Add CPUfreq to the CPU power management section and remove a duplicate include of the main power section. Signed-off-by: Mark Brown <broonie@linaro.org> Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
| * | | | | arm64: mm: Route pmd thp functions through pte equivalentsSteve Capper2014-03-14
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Rather than have separate hugetlb and transparent huge page pmd manipulation functions, re-wire our thp functions to simply call the pte equivalents. This allows THP to take advantage of the new PTE_WRITE logic introduced in: c2c93e5 arm64: mm: Introduce PTE_WRITE To represent splitting THPs we use the PTE_SPECIAL bit as this is not used for pmds. Signed-off-by: Steve Capper <steve.capper@linaro.org> Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
| * | | | | arm64: rwsem: use asm-generic rwsem implementationWill Deacon2014-03-14
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | asm-generic offers an atomic-add based rwsem implementation, which can avoid the need for heavier, spinlock-based synchronisation on the fast path. This patch makes use of the optimised implementation for arm64 CPUs. Signed-off-by: Will Deacon <will.deacon@arm.com> Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
generated by cgit v1.2.2 (git 2.25.0) at 2024-09-24 20:25:51 -0400