summaryrefslogtreecommitdiff
path: root/drivers
AgeCommit message (Collapse)AuthorFilesLines
2014-11-18drm/exynos: Use unsigned long for possible_crtcsSean Paul5-5/+5
Change all instances of possible_crtcs in the exynos drm driver to be unsigned long. This matches the type used in the drm layer. Signed-off-by: Sean Paul <seanpaul@chromium.org> Reviewed-by: Tomasz Figa <t.figa@samsung.com> Signed-off-by: Inki Dae <inki.dae@samsung.com>
2014-11-18drm/exynos: Don't keep dpms state in encoderSean Paul1-17/+0
This patch removes the dpms state tracking in encoder. This state is at best confusing and at worst incorrect since the display drivers can turn on and off without propagating the value. Signed-off-by: Sean Paul <seanpaul@chromium.org> Reviewed-by: Tomasz Figa <t.figa@samsung.com> Signed-off-by: Inki Dae <inki.dae@samsung.com>
2014-11-18drm/exynos: Rename display_op power_on to dpmsSean Paul7-41/+10
This patch renames the display_op power_on to dpms to accurately reflect what the function does. The side-effect of this patch is that the new hdmi dpms callback is now invoked twice in the dpms path. This is safe and will be dealt with when the exynos_drm shim goes away. Signed-off-by: Sean Paul <seanpaul@chromium.org> Reviewed-by: Tomasz Figa <t.figa@samsung.com> Signed-off-by: Inki Dae <inki.dae@samsung.com>
2014-11-18drm/exynos: Remove dpms link between encoder/connectorSean Paul3-88/+8
This patch removes the call from encoder dpms into connector dpms (which will then call back into encoder dpms through the helper function). The callback is likely to keep connector->dpms in the right state when initiating dpms from crtc or encoder, but this isn't the right way to do it. This patch is the first step towards rationalizing power management in the exynos drm driver. Signed-off-by: Sean Paul <seanpaul@chromium.org> Reviewed-by: Tomasz Figa <t.figa@samsung.com> Signed-off-by: Inki Dae <inki.dae@samsung.com>
2014-11-18drm/exynos: Remove apply manager callbackSean Paul7-43/+8
This patch removes the apply() manager callback in favor of putting the relevant commits in the individual drivers. This will mitigate some of the difference between the suspend/resume path and the dpms path Signed-off-by: Sean Paul <seanpaul@chromium.org> Reviewed-by: Tomasz Figa <t.figa@samsung.com> Signed-off-by: Inki Dae <inki.dae@samsung.com>
2014-11-18drm/exynos: Pass exynos_drm_manager in manager ops instead of devSean Paul6-184/+202
This patch changes the manager ops callbacks from accepting the subdrv device pointer to taking a pointer to the manager. This will allow us to move closer to decoupling manager/display from subdrv, and subsequently decoupling the crtc/plane from the encoder. Signed-off-by: Sean Paul <seanpaul@chromium.org> Signed-off-by: Inki Dae <inki.dae@samsung.com>
2014-11-18drm/exynos: hdmi: Implement initialize op for hdmiSean Paul4-188/+219
This patch implements the initialize callback in the hdmi and mixer manager. This allows us to get rid of drm_dev in the drm_hdmi level and track it in the mixer and hdmi drivers. This is one of the things holding back the complete removal of the drm_hdmi layer. Signed-off-by: Sean Paul <seanpaul@chromium.org> Signed-off-by: Inki Dae <inki.dae@samsung.com>
2014-11-18drm/exynos: Use manager_op initialize in fimdSean Paul1-4/+15
This patch implements the intitialize manager op in fimd. This will allow us to keep track of drm_dev in context instead of using subdev, which in turn makes it easier to remove subdev from fimd. Signed-off-by: Sean Paul <seanpaul@chromium.org> Signed-off-by: Inki Dae <inki.dae@samsung.com>
2014-11-18drm/exynos: Add an initialize function to manager and displaySean Paul2-0/+26
This patch adds an initialize function to the manager and display operations. This allows them to keep track of drm_device in their local context, as well as adds an initialization hook right after the encoder is created. Signed-off-by: Sean Paul <seanpaul@chromium.org> Reviewed-by: Tomasz Figa <t.figa@samsung.com> Signed-off-by: Inki Dae <inki.dae@samsung.com>
2014-11-18drm/exynos: Merge overlay_ops into manager_opsSean Paul6-82/+55
This patch merges overlay_ops into manager_ops. In all cases, overlay_ops is implemented in the same place as manager ops, it doesn't serve a functional purpose, and doesn't make things more clear. Signed-off-by: Sean Paul <seanpaul@chromium.org> Signed-off-by: Inki Dae <inki.dae@samsung.com>
2014-11-18drm/exynos: Remove useless slab.h includeStephane Marchesin1-1/+0
Signed-off-by: Stephane Marchesin <marcheu@chromium.org> Signed-off-by: Sean Paul <seanpaul@chromium.org> Signed-off-by: Inki Dae <inki.dae@samsung.com>
2014-11-18drm/exynos: add phy settings for RB resolutionsShirish S1-0/+45
This patch adds support for the below mentioned pixel clocks in Exynos5250. Without them, following display modes won¡¯t be supported: 71 MHz - 1280x800@60Hz RB 73.25 MHz - 800x600@120Hz RB 88.75 MHz - 1440x900@60Hz RB 115.5 MHz - 1024x768@120Hz RB 119 MHz - 1680x1050@60Hz RB Signed-off-by: Shirish S <s.shirish@samsung.com> Reviewed-by: Tomasz Figa <t.figa@samsung.com> Signed-off-by: Inki Dae <inki.dae@samsung.com>
2014-11-18misc: modem_if: fix section mismatch from tizen_modem.cMyungJoo Ham1-1/+1
A function, modem_probe, had __init while it was used by a struct that is not __init. Change-Id: Ib1571c89b4bfcebc58b2669b669bd4606d1922fa Signed-off-by: MyungJoo Ham <myungjoo.ham@samsung.com>
2014-11-18extcon: Move OF helper function to extcon core and change function nameChanwoo Choi6-66/+44
This patch move simply OF helper function to extcon core and change function name as following: - of_extcon_get_extcon_dev() -> extcon_get_edev_by_phandle() Change-Id: I37fd827f945e3d713cfe15607689eab6792007b1 Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com> Acked-by: Felipe Balbi <balbi@ti.com>
2014-11-18extcon: of: Remove unnecessary function call by using the name of device_nodeChanwoo Choi1-11/+3
This patch remove unnecessary function call in of_extcon_get_extcon_dev() by using the name of device_node structure. Change-Id: Ifae59614f355d1ad1beef9aab1c497840f22a727 Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
2014-11-18misc: modem_if: regress to old tizen modem driverChanho Park38-15222/+861
Tizen modem utility and xmm626x modem uses old style modem I/F. I brought the driver from linux-3.0 kernel[1] because the driver should be matched with telephony and modem-util package. Kamil brought the driver from linux-3.4-exynos, but the driver didn't match the current telephony daemon and modem util. Data communication didn't work with the driver. So, I decided to regress to old driver. Tizen Modem Team doesn't have any plan to update latest version of modem package. To maintain consistency of the interface, I picked the driver from linux-3.0 tree. [1] : https://review.tizen.org/gerrit/gitweb?p=kernel%2Flinux-3.0.git;a=summary Change-Id: I6123f4dc149dcc38e1ae339f950f94fdc0c033f9 Signed-off-by: Chanho Park <chanho61.park@samsung.com>
2014-11-18phy: exynos4212-usb: Use refcount to power device PHY for HSIC0Tomasz Figa1-0/+10
HSIC0 requires device PHY to be powered up and the driver handles that already, but there is no reference counting of device PHY power state, so unplugging USB cable powers it down, breaking modem IF operation. This patch fixes the issue by adding handling of reference counting of device PHY to HSIC0 power-down/-up. Signed-off-by: Tomasz Figa <t.figa@samsung.com> Change-Id: I14b1825fec3cd9c1b2652b89f275a50c73359ad6
2014-11-18misc: modem_if: Use pinctrl to properly configure pinsTomasz Figa2-7/+24
Configuring an interrupt pin as output/input manually using gpio_direction_*() breaks interrupt settings and must not be done. This patch adds necessary code to enable pull-down for interrupt pin using pinctrl API instead of setting the pin to low output to fix issues with missing modem interrupts. Signed-off-by: Tomasz Figa <t.figa@samsung.com> Change-Id: Ie56d84ace1941cef3199fea99d7eb907bb9b5c03
2014-11-18mfd: max77686: Add suspend/resume callback function for alarm wake-up.Jonghwa Lee1-0/+38
While system is resuming, Max77686's Alarm IRQ should be handled after resuming of i2c controller. It disables IRQ in suspend callback funtion and enables it in resume callback to postpone handling. Change-Id: I69f5af2f8b63748853d0ba8a67f0cbb661e1ffcd Signed-off-by: Jonghwa Lee <jonghwa3.lee@samsung.com>
2014-11-18media: vb2-dma-contig: add a hack to speed up cache managementMarek Szyprowski1-2/+27
In most cases video processing is done with coherent (non-cachable) userspace mappings. In such case for USERPTR mode cache management can be skipped. This patch introduces code for such optimisation. It is called a hack, because it is not possible in generic way to detect if the given userspace mapping is coherent on not. Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com> Change-id: I86aa292455371f1b18eb3f6b765234dc04ff937b
2014-11-18s5p-mfc: Remove unused variableSylwester Nawrocki1-1/+0
This fixes following compilation warning: drivers/media/platform/s5p-mfc/s5p_mfc_dec.c: In function ‘s5p_mfc_buf_prepare’: drivers/media/platform/s5p-mfc/s5p_mfc_dec.c:1045:15: warning: unused variable ‘index’ [-Wunused-variable] Change-Id: Id6c0c264c7169871df116ba6496b2e08e2638844 Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
2014-11-18misc: modem_if: initialize gpio_link_enable if invalid gpioChanho Park1-1/+2
This patch fixes invalid gpio value caused by 47c53475 commit. The gpio_link_enable should be initialized to 0 if not provided Change-Id: Ie22a2a60a2b713a3be9f141a5dc5c71d4adeb667 Signed-off-by: Chanho Park <chanho61.park@samsung.com>
2014-11-18modem_if: Change hardcoded link enable gpio valueBeomho Seo1-7/+4
Currently, link enable gpio is hardecoded to zero. Zero equal gpa0-0 value. It is value for uart rxd line. If link enable gpio is unused, it is don't have to call gpio request. This patch change to set line enable gpio value optionally, and remove unused variable 'ret'. Change-Id: I4095d57d76362722abd5c5fd92ef529772680d97 Signed-off-by: Beomho Seo <beomho.seo@samsung.com>
2014-11-18drm/exynos: add dmabuf sync support for kms frameworkInki Dae4-21/+218
Change-Id: I5937869e5fa2ec2a41e864494eac0077ccdbdb51 Signed-off-by: Inki Dae <inki.dae@samsung.com>
2014-11-18media: s5p_mfc: fix doublly free issueZhaowei Yuan1-6/+2
When video_register_device() returns failure, vfd will be freed at once and dev->vfl_dec will be freed after label "err_dec_reg", but the two pointers point to the same area, so it was doublly freed here. Since video_device_release() will be called by video_unregister_device() indirectly, we should remove the two other calls after labels err_enc_reg and err_dec_reg. Change-Id: I5ee8f83eeb47443e2f772914127514eab996c347 Signed-off-by: Zhaowei Yuan <zhaowei.yuan@samsung.com>
2014-11-18cpufreq:LAB: Change method of boost state preservingLukasz Majewski1-4/+5
It is not necessary to change the boost state when LAB governor is entered, since LAB will change it according to its own politics. Only enter state is preserved. When leaving the LAB, only when required, work is scheduled to restore boost initial state. Change-Id: I6323f3c0011fe54a33d70c9ad0f9da5360b4a735 Signed-off-by: Lukasz Majewski <l.majewski@samsung.com>
2014-11-18cpufreq:governor: Add serialization to the cpufreq_governor_dbs() functionLukasz Majewski1-2/+12
It is necessary to serialize access to cpufreq_governor_dbs() function, since it can be accessed from different, not protected by any mutex paths like sysfs boost attribute or LAB governor internals. Change-Id: Id7b62db6ca0b7c28f5e8c6286aec312d3d0c971e Signed-off-by: Lukasz Majewski <l.majewski@samsung.com>
2014-11-18media: s5p_mfc: remove the code checking dev->plat_devZhaowei Yuan1-5/+0
We should remove the code checking dev->plat_dev since we can ensure the pointer pdev can not be NULL. Change-Id: Ibdc44403068ee4462e414d6e84757b8a4c2b512c Signed-off-by: Zhaowei Yuan <zhaowei.yuan@samsung.com>
2014-11-18media: s5p-mfc: add to set clock rateDonghwa Lee1-0/+13
from: Seung-Woo Kim <sw0312.kim@samsung.com> MFC needs 200MHz for sclk_mfc clock to work properly. The clock rate setting was missed, so this patch adds it. Change-Id: Ica696a5fda2babe81e885945fa5affd0b09ff5ba Signed-off-by: Donghwa Lee <dh09.lee@samsung.com> Signed-off-by: Seung-Woo Kim <sw0312.kim@samsung.com>
2014-11-18clocksource: exynos_mct: Change exynos4_mct_tick_clear return type to voidKrzysztof Kozlowski1-6/+2
Return value of exynos4_mct_tick_clear() was never checked so it can be safely changed to void. Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com> Change-Id: I14f872e244434002005c532adf8afc97ef77cea5
2014-11-18clocksource: exynos_mct: Fix stall after CPU hotpluggingKrzysztof Kozlowski1-13/+20
Fix stall after hotplugging CPU1. The stall was a result of starting the CPU1 local timer not in L1 timer but in L0 (which is used by CPU0). Stall information: [ 530.045259] INFO: rcu_preempt detected stalls on CPUs/tasks: [ 530.045618] 1: (6 GPs behind) idle=6d0/0/0 softirq=369/369 [ 530.050987] (detected by 0, t=6589 jiffies, g=33, c=32, q=0) [ 530.056721] Task dump for CPU 1: [ 530.059928] swapper/1 R running 0 0 1 0x00001000 [ 530.066377] [<c0524e14>] (__schedule+0x414/0x9b4) from [<c00b6610>] (rcu_idle_enter+0x18/0x38) [ 530.074955] [<c00b6610>] (rcu_idle_enter+0x18/0x38) from [<c0079a18>] (cpu_startup_entry+0x60/0x3bc) [ 530.084069] [<c0079a18>] (cpu_startup_entry+0x60/0x3bc) from [<c0517d34>] (secondary_start_kernel+0x164/0x1a0) [ 530.094029] [<c0517d34>] (secondary_start_kernel+0x164/0x1a0) from [<40517244>] (0x40517244) The timers for CPU1 were missed: [ 591.668436] cpu: 1 [ 591.670430] clock 0: [ 591.672691] .base: c0ab7750 [ 591.676160] .index: 0 [ 591.679025] .resolution: 1 nsecs [ 591.682404] .get_time: ktime_get [ 591.685970] .offset: 0 nsecs [ 591.689349] active timers: [ 591.692045] #0: <dfb51f40>, hrtimer_wakeup, S:01 [ 591.696759] # expires at 454687834257-454687884257 nsecs [in -136770537232 to -136770487232 nsecs] And the event_handler for next event was wrong: [ 591.917120] Tick Device: mode: 1 [ 591.920676] Per CPU device: 0 [ 591.923621] Clock Event Device: mct_tick0 [ 591.927623] max_delta_ns: 178956969027 [ 591.931613] min_delta_ns: 1249 [ 591.934913] mult: 51539608 [ 591.938557] shift: 32 [ 591.941681] mode: 3 [ 591.944724] next_event: 595025000000 nsecs [ 591.949227] set_next_event: exynos4_tick_set_next_event [ 591.954522] set_mode: exynos4_tick_set_mode [ 591.959296] event_handler: hrtimer_interrupt [ 591.963730] retries: 0 [ 591.966761] [ 591.968245] Tick Device: mode: 0 [ 591.971801] Per CPU device: 1 [ 591.974746] Clock Event Device: mct_tick1 [ 591.978750] max_delta_ns: 178956969027 [ 591.982739] min_delta_ns: 1249 [ 591.986037] mult: 51539608 [ 591.989681] shift: 32 [ 591.992806] mode: 3 [ 591.995848] next_event: 453685000000 nsecs [ 592.000353] set_next_event: exynos4_tick_set_next_event [ 592.005648] set_mode: exynos4_tick_set_mode [ 592.010421] event_handler: tick_handle_periodic [ 592.015115] retries: 0 [ 592.018145] After turning off the CPU1, the MCT L1 local timer was disabled but the interrupt was not cleared. Turning on the CPU1 enabled the IRQ with setup_irq() but, before setting affinity to CPU1, the pending L1 timer interrupt was processed by CPU0 in exynos4_mct_tick_isr(). The ISR then called event handler which set up the next timer event for current CPU (CPU0). Therefore the MCT L1 timer wasn't actually started. Fix the stall by: 1. Setting next timer event not on current CPU but on the CPU indicated by cpumask in 'clock_event_device'. 2. Clearing the timer interrupt upon stopping the local timer. The patch also moves around the call to exynos4_mct_tick_stop() but this is done only for the code readability as it is not essential for the fix. Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com> Change-Id: I3a1a23e2b970661b5f7c60fc633a7545aa80ed5e
2014-11-18cpufreq: serialize calls to __cpufreq_governor()Viresh Kumar1-2/+8
We can't take a big lock around __cpufreq_governor() as this causes recursive locking for some cases. But calls to this routine must be serialized for every policy. Otherwise we can see some unpredictable events. For example, consider following scenario: __cpufreq_remove_dev() __cpufreq_governor(policy, CPUFREQ_GOV_STOP); policy->governor->governor(policy, CPUFREQ_GOV_STOP); cpufreq_governor_dbs() case CPUFREQ_GOV_STOP: mutex_destroy(&cpu_cdbs->timer_mutex) cpu_cdbs->cur_policy = NULL; <PREEMPT> store() __cpufreq_set_policy() __cpufreq_governor(policy, CPUFREQ_GOV_LIMITS); policy->governor->governor(policy, CPUFREQ_GOV_LIMITS); case CPUFREQ_GOV_LIMITS: mutex_lock(&cpu_cdbs->timer_mutex); <-- Warning (destroyed mutex) if (policy->max < cpu_cdbs->cur_policy->cur) <- cur_policy == NULL And so store() will eventually result in a crash if cur_policy is NULL at this point. Introduce an additional variable which would guarantee serialization here. Change-Id: Ibae767cbd9c25c7598b39d1405fa3d98d2125101 Reported-by: Stephen Boyd <sboyd@codeaurora.org> Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
2014-11-18cpufreq: Fix serialization of frequency transitionsViresh Kumar1-3/+4
Commit 7c30ed ("cpufreq: make sure frequency transitions are serialized") interacts poorly with systems that have a single core freqency for all cores. On such systems we have a single policy for all cores with several CPUs. When we do a frequency transition the governor calls the pre and post change notifiers which causes cpufreq_notify_transition() per CPU. Since the policy is the same for all of them all CPUs after the first and the warnings added are generated by checking a per-policy flag the warnings will be triggered for all cores after the first. Fix this by allowing notifier to be called for n times. Where n is the number of cpus in policy->cpus. Change-Id: I5712dde7f992644f9c3ddc8313151f80bea0d877 Reported-and-tested-by: Mark Brown <broonie@linaro.org> Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
2014-11-18cpufreq: make sure frequency transitions are serializedViresh Kumar1-0/+14
Whenever we are changing frequency of a cpu, we are calling PRECHANGE and POSTCHANGE notifiers. They must be serialized. i.e. PRECHANGE or POSTCHANGE shouldn't be called twice contiguously. This can happen due to bugs in users of __cpufreq_driver_target() or actual cpufreq drivers who are sending these notifiers. This patch adds some protection against this. Now, we keep track of the last transaction and see if something went wrong. Change-Id: I0f5465bd515c431ae2d3711d065f70aacec7e978 Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
2014-11-18cpufreq: remove unnecessary cpufreq_cpu_{get|put}() callsViresh Kumar1-17/+2
struct cpufreq_policy is already passed as argument to some routines like: __cpufreq_driver_getavg() and so we don't really need to do cpufreq_cpu_get() before and cpufreq_cpu_put() in them to get a policy structure. Remove them. Change-Id: I6a9ff8ed483a4f4faacc2ea047d93354dccdb0b6 Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
2014-11-18modem: sipc4: Chagne the manner of recieving data for FMT,RFS type deviceJonghwa Lee1-1/+25
When packet arrives, link device call iodev's helper function to recieve packets. The way of recieving data of IPC_FMT and IPC_RFS type iodevs differs from IPC_RAW and IPC_MULTI_RAW. This patch adds specified method of recieving data for FMT, RFS typed. This modification references TIZEN 2.2 kernel. Change-Id: I01efa7678bbabfbd1011ceba42571fc221313c4d Signed-off-by: Jonghwa Lee <jonghwa3.lee@samsung.com>
2014-11-18drm/exynos: remove DRIVER_HAVE_IRQ featureJoonyoung Shim3-22/+1
Exynos drm driver cannot support DRIVER_HAVE_IRQ feature because it uses driver specific one instead of routine of drm framework to install/uninstall irq handler. Change-Id: I5796d7113cbc4283cbb41591384aaa69011818d4 Signed-off-by: Joonyoung Shim <jy0922.shim@samsung.com>
2014-11-18zram: remove zram->lock in read path and change it with mutexMinchan Kim2-12/+9
Finally, we separated zram->lock dependency from 32bit stat/ table handling so there is no reason to use rw_semaphore between read and write path so this patch removes the lock from read path totally and changes rw_semaphore with mutex. So, we could do old: read-read: OK read-write: NO write-write: NO Now: read-read: OK read-write: OK write-write: NO The below data proves mixed workload performs well 11 times and there is also enhance on write-write path because current rw-semaphore doesn't support SPIN_ON_OWNER. It's side effect but anyway good thing for us. Write-related tests perform better (from 61% to 1058%) but read path has good/bad(from -2.22% to 1.45%) but they are all marginal within stddev. CPU 12 iozone -t -T -l 12 -u 12 -r 16K -s 60M -I +Z -V 0 ==Initial write ==Initial write records: 10 records: 10 avg: 516189.16 avg: 839907.96 std: 22486.53 (4.36%) std: 47902.17 (5.70%) max: 546970.60 max: 909910.35 min: 481131.54 min: 751148.38 ==Rewrite ==Rewrite records: 10 records: 10 avg: 509527.98 avg: 1050156.37 std: 45799.94 (8.99%) std: 40695.44 (3.88%) max: 611574.27 max: 1111929.26 min: 443679.95 min: 980409.62 ==Read ==Read records: 10 records: 10 avg: 4408624.17 avg: 4472546.76 std: 281152.61 (6.38%) std: 163662.78 (3.66%) max: 4867888.66 max: 4727351.03 min: 4058347.69 min: 4126520.88 ==Re-read ==Re-read records: 10 records: 10 avg: 4462147.53 avg: 4363257.75 std: 283546.11 (6.35%) std: 247292.63 (5.67%) max: 4912894.44 max: 4677241.75 min: 4131386.50 min: 4035235.84 ==Reverse Read ==Reverse Read records: 10 records: 10 avg: 4565865.97 avg: 4485818.08 std: 313395.63 (6.86%) std: 248470.10 (5.54%) max: 5232749.16 max: 4789749.94 min: 4185809.62 min: 3963081.34 ==Stride read ==Stride read records: 10 records: 10 avg: 4515981.80 avg: 4418806.01 std: 211192.32 (4.68%) std: 212837.97 (4.82%) max: 4889287.28 max: 4686967.22 min: 4210362.00 min: 4083041.84 ==Random read ==Random read records: 10 records: 10 avg: 4410525.23 avg: 4387093.18 std: 236693.22 (5.37%) std: 235285.23 (5.36%) max: 4713698.47 max: 4669760.62 min: 4057163.62 min: 3952002.16 ==Mixed workload ==Mixed workload records: 10 records: 10 avg: 243234.25 avg: 2818677.27 std: 28505.07 (11.72%) std: 195569.70 (6.94%) max: 288905.23 max: 3126478.11 min: 212473.16 min: 2484150.69 ==Random write ==Random write records: 10 records: 10 avg: 555887.07 avg: 1053057.79 std: 70841.98 (12.74%) std: 35195.36 (3.34%) max: 683188.28 max: 1096125.73 min: 437299.57 min: 992481.93 ==Pwrite ==Pwrite records: 10 records: 10 avg: 501745.93 avg: 810363.09 std: 16373.54 (3.26%) std: 19245.01 (2.37%) max: 518724.52 max: 833359.70 min: 464208.73 min: 765501.87 ==Pread ==Pread records: 10 records: 10 avg: 4539894.60 avg: 4457680.58 std: 197094.66 (4.34%) std: 188965.60 (4.24%) max: 4877170.38 max: 4689905.53 min: 4226326.03 min: 4095739.72 Change-Id: I7d2299149ce6982d76caaaadb936b7385cbee515 Signed-off-by: Minchan Kim <minchan@kernel.org> Cc: Nitin Gupta <ngupta@vflare.org> Tested-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com> Cc: Jerome Marchand <jmarchan@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-11-18zram: remove workqueue for freeing removed pending slotMinchan Kim2-58/+6
Commit a0c516cbfc74 ("zram: don't grab mutex in zram_slot_free_noity") introduced free request pending code to avoid scheduling by mutex under spinlock and it was a mess which made code lenghty and increased overhead. Now, we don't need zram->lock any more to free slot so this patch reverts it and then, tb_lock should protect it. Change-Id: I3429e568bab78c197da3fc5cbd5afb9355bf7d21 Signed-off-by: Minchan Kim <minchan@kernel.org> Cc: Nitin Gupta <ngupta@vflare.org> Tested-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com> Cc: Jerome Marchand <jmarchan@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-11-18zram: introduce zram->tb_lockMinchan Kim2-6/+23
Currently, the zram table is protected by zram->lock but it's rather coarse-grained lock and it makes hard for scalibility. Let's use own rwlock instead of depending on zram->lock. This patch adds new locking so obviously, it would make slow but this patch is just prepartion for removing coarse-grained rw_semaphore(ie, zram->lock) which is hurdle about zram scalability. Final patch in this patchset series will remove the lock from read-path and change rw_semaphore with mutex in write path. With bonus, we could drop pending slot free mess in next patch. Change-Id: If5456f871bc6b0d6ee1f8218fde3f5a13d261c8b Signed-off-by: Minchan Kim <minchan@kernel.org> Cc: Nitin Gupta <ngupta@vflare.org> Tested-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com> Cc: Jerome Marchand <jmarchan@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-11-18zram: use atomic operation for statMinchan Kim2-20/+16
Some of fields in zram->stats are protected by zram->lock which is rather coarse-grained so let's use atomic operation without explict locking. This patch is ready for removing dependency of zram->lock in read path which is very coarse-grained rw_semaphore. Of course, this patch adds new atomic operation so it might make slow but my 12CPU test couldn't spot any regression. All gain/lose is marginal within stddev. iozone -t -T -l 12 -u 12 -r 16K -s 60M -I +Z -V 0 ==Initial write ==Initial write records: 50 records: 50 avg: 412875.17 avg: 415638.23 std: 38543.12 (9.34%) std: 36601.11 (8.81%) max: 521262.03 max: 502976.72 min: 343263.13 min: 351389.12 ==Rewrite ==Rewrite records: 50 records: 50 avg: 416640.34 avg: 397914.33 std: 60798.92 (14.59%) std: 46150.42 (11.60%) max: 543057.07 max: 522669.17 min: 304071.67 min: 316588.77 ==Read ==Read records: 50 records: 50 avg: 4147338.63 avg: 4070736.51 std: 179333.25 (4.32%) std: 223499.89 (5.49%) max: 4459295.28 max: 4539514.44 min: 3753057.53 min: 3444686.31 ==Re-read ==Re-read records: 50 records: 50 avg: 4096706.71 avg: 4117218.57 std: 229735.04 (5.61%) std: 171676.25 (4.17%) max: 4430012.09 max: 4459263.94 min: 2987217.80 min: 3666904.28 ==Reverse Read ==Reverse Read records: 50 records: 50 avg: 4062763.83 avg: 4078508.32 std: 186208.46 (4.58%) std: 172684.34 (4.23%) max: 4401358.78 max: 4424757.22 min: 3381625.00 min: 3679359.94 ==Stride read ==Stride read records: 50 records: 50 avg: 4094933.49 avg: 4082170.22 std: 185710.52 (4.54%) std: 196346.68 (4.81%) max: 4478241.25 max: 4460060.97 min: 3732593.23 min: 3584125.78 ==Random read ==Random read records: 50 records: 50 avg: 4031070.04 avg: 4074847.49 std: 192065.51 (4.76%) std: 206911.33 (5.08%) max: 4356931.16 max: 4399442.56 min: 3481619.62 min: 3548372.44 ==Mixed workload ==Mixed workload records: 50 records: 50 avg: 149925.73 avg: 149675.54 std: 7701.26 (5.14%) std: 6902.09 (4.61%) max: 191301.56 max: 175162.05 min: 133566.28 min: 137762.87 ==Random write ==Random write records: 50 records: 50 avg: 404050.11 avg: 393021.47 std: 58887.57 (14.57%) std: 42813.70 (10.89%) max: 601798.09 max: 524533.43 min: 325176.99 min: 313255.34 ==Pwrite ==Pwrite records: 50 records: 50 avg: 411217.70 avg: 411237.96 std: 43114.99 (10.48%) std: 33136.29 (8.06%) max: 530766.79 max: 471899.76 min: 320786.84 min: 317906.94 ==Pread ==Pread records: 50 records: 50 avg: 4154908.65 avg: 4087121.92 std: 151272.08 (3.64%) std: 219505.04 (5.37%) max: 4459478.12 max: 4435857.38 min: 3730512.41 min: 3101101.67 Change-Id: Ib0d538597fbc4a2037b0464f8d62fb73fa0b0c24 Signed-off-by: Minchan Kim <minchan@kernel.org> Cc: Nitin Gupta <ngupta@vflare.org> Tested-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com> Cc: Jerome Marchand <jmarchan@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-11-18zram: remove unnecessary freeMinchan Kim1-8/+0
Commit a0c516cbfc74 ("zram: don't grab mutex in zram_slot_free_noity") introduced pending zram slot free in zram's write path in case of missing slot free by memory allocation failure in zram_slot_free_notify but it is not necessary because we have already freed the slot right before overwriting. Change-Id: I5048bce2ca8c377d9539f0397a04bddc5f5a5e92 Signed-off-by: Minchan Kim <minchan@kernel.org> Cc: Nitin Gupta <ngupta@vflare.org> Cc: Jerome Marchand <jmarchan@redhat.com> Tested-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-11-18zram: delay pending free request in read pathMinchan Kim1-1/+0
Sergey reported we don't need to handle pending free request every I/O so that this patch removes it in read path while we remain it in write path. Let's consider below example. Swap subsystem ask to zram "A" block free by swap_slot_free_notify but zram had been pended it without real freeing. Swap subsystem allocates "A" block for new data but request pended for a long time just handled and zram blindly free new data on the "A" block. :( That's why we couldn't remove handle pending free request right before zram-write. Change-Id: Ib4409bfad7b1ae263e2708c74875c322da72c7b3 Signed-off-by: Minchan Kim <minchan@kernel.org> Reported-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com> Tested-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com> Cc: Nitin Gupta <ngupta@vflare.org> Cc: Jerome Marchand <jmarchan@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-11-18zram: fix race between reset and flushing pending workMinchan Kim1-2/+2
Dan and Sergey reported that there is a racy between reset and flushing of pending work so that it could make oops by freeing zram->meta in reset while zram_slot_free can access zram->meta if new request is adding during the race window. This patch moves flush after taking init_lock so it prevents new request so that it closes the race. Change-Id: Ibc09001d1ad4a4ef852d661384259b53f0f9c19b Signed-off-by: Minchan Kim <minchan@kernel.org> Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Cc: Nitin Gupta <ngupta@vflare.org> Cc: Jerome Marchand <jmarchan@redhat.com> Tested-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-11-18zram: add copyrightMinchan Kim2-0/+2
Add my copyright to the zram source code which I maintain. Change-Id: I8816064aa958c9304c53fae0972e011060cc2bcc Signed-off-by: Minchan Kim <minchan@kernel.org> Cc: Nitin Gupta <ngupta@vflare.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-11-18zram: remove old private project commentMinchan Kim3-3/+0
Remove the old private compcache project address so upcoming patches should be sent to LKML because we Linux kernel community will take care. Change-Id: Ia5bf208791c8fa6e96161fd9fb842d6829f14698 Signed-off-by: Minchan Kim <minchan@kernel.org> Cc: Nitin Gupta <ngupta@vflare.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-11-18zram: promote zram from stagingMinchan Kim9-80/+3
Zram has lived in staging for a LONG LONG time and have been fixed/improved by many contributors so code is clean and stable now. Of course, there are lots of product using zram in real practice. The major TV companys have used zram as swap since two years ago and recently our production team released android smart phone with zram which is used as swap, too and recently Android Kitkat start to use zram for small memory smart phone. And there was a report Google released their ChromeOS with zram, too and cyanogenmod have been used zram long time ago. And I heard some disto have used zram block device for tmpfs. In addition, I saw many report from many other peoples. For example, Lubuntu start to use it. The benefit of zram is very clear. With my experience, one of the benefit was to remove jitter of video application with backgroud memory pressure. It would be effect of efficient memory usage by compression but more issue is whether swap is there or not in the system. Recent mobile platforms have used JAVA so there are many anonymous pages. But embedded system normally are reluctant to use eMMC or SDCard as swap because there is wear-leveling and latency issues so if we do not use swap, it means we can't reclaim anoymous pages and at last, we could encounter OOM kill. :( Although we have real storage as swap, it was a problem, too. Because it sometime ends up making system very unresponsible caused by slow swap storage performance. Quote from Luigi on Google "Since Chrome OS was mentioned: the main reason why we don't use swap to a disk (rotating or SSD) is because it doesn't degrade gracefully and leads to a bad interactive experience. Generally we prefer to manage RAM at a higher level, by transparently killing and restarting processes. But we noticed that zram is fast enough to be competitive with the latter, and it lets us make more efficient use of the available RAM. " and he announced. http://www.spinics.net/lists/linux-mm/msg57717.html Other uses case is to use zram for block device. Zram is block device so anyone can format the block device and mount on it so some guys on the internet start zram as /var/tmp. http://forums.gentoo.org/viewtopic-t-838198-start-0.html Let's promote zram and enhance/maintain it instead of removing. Signed-off-by: Minchan Kim <minchan@kernel.org> Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Acked-by: Nitin Gupta <ngupta@vflare.org> Acked-by: Pekka Enberg <penberg@kernel.org> Cc: Bob Liu <bob.liu@oracle.com> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Hugh Dickins <hughd@google.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Luigi Semenzato <semenzato@google.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Rik van Riel <riel@redhat.com> Cc: Seth Jennings <sjenning@linux.vnet.ibm.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Conflicts: drivers/block/Makefile Change-Id: I368f76a5368fffaacbf349cfd78f79cba5da0a0d
2014-11-18zsmalloc: move it under mmMinchan Kim7-1189/+1
This patch moves zsmalloc under mm directory. Before that, description will explain why we have needed custom allocator. Zsmalloc is a new slab-based memory allocator for storing compressed pages. It is designed for low fragmentation and high allocation success rate on large object, but <= PAGE_SIZE allocations. zsmalloc differs from the kernel slab allocator in two primary ways to achieve these design goals. zsmalloc never requires high order page allocations to back slabs, or "size classes" in zsmalloc terms. Instead it allows multiple single-order pages to be stitched together into a "zspage" which backs the slab. This allows for higher allocation success rate under memory pressure. Also, zsmalloc allows objects to span page boundaries within the zspage. This allows for lower fragmentation than could be had with the kernel slab allocator for objects between PAGE_SIZE/2 and PAGE_SIZE. With the kernel slab allocator, if a page compresses to 60% of it original size, the memory savings gained through compression is lost in fragmentation because another object of the same size can't be stored in the leftover space. This ability to span pages results in zsmalloc allocations not being directly addressable by the user. The user is given an non-dereferencable handle in response to an allocation request. That handle must be mapped, using zs_map_object(), which returns a pointer to the mapped region that can be used. The mapping is necessary since the object data may reside in two different noncontigious pages. The zsmalloc fulfills the allocation needs for zram perfectly [sjenning@linux.vnet.ibm.com: borrow Seth's quote] Signed-off-by: Minchan Kim <minchan@kernel.org> Acked-by: Nitin Gupta <ngupta@vflare.org> Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Cc: Bob Liu <bob.liu@oracle.com> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Hugh Dickins <hughd@google.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Luigi Semenzato <semenzato@google.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Pekka Enberg <penberg@kernel.org> Cc: Rik van Riel <riel@redhat.com> Cc: Seth Jennings <sjenning@linux.vnet.ibm.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Conflicts: mm/Kconfig Change-Id: I57dad090a3c48db4a67c88e6fa20a4bdbb82d984
2014-11-18zsmalloc: add more commentNitin Cupta2-11/+64
This patch adds lots of comments and it will help others to review and enhance. Change-Id: I743596bf18e8acf1082c21437c2caef5f15aad71 Signed-off-by: Seth Jennings <sjenning@linux.vnet.ibm.com> Signed-off-by: Nitin Gupta <ngupta@vflare.org> Signed-off-by: Minchan Kim <minchan@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-11-18zsmalloc: add Kconfig for enabling page table methodMinchan Kim2-15/+17
Zsmalloc has two methods 1) copy-based and 2) pte based to access objects that span two pages. You can see history why we supported two approach from [1]. But it was bad choice that adding hard coding to select arch which want to use pte based method because there are lots of SoC in an architecure and they can have different cache size, CPU speed and so on so it would be better to expose it to user as selectable Kconfig option like Andrew Morton suggested. [1] https://lkml.org/lkml/2012/7/11/58 Change-Id: Ieedde9cfac0a7d9bbcb3d5d5b36318efd41132eb Acked-by: Nitin Gupta <ngupta@vflare.org> Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Signed-off-by: Minchan Kim <minchan@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
generated by cgit v1.2.3 (git 2.25.1) at 2024-11-05 20:45:16 +0000