| Commit message (Collapse) | Author | Age |
|---|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Create a dss_lcd_mgr_config struct instance in DPI. Fill up all the parameters
of the struct with configurations held by the panel, and the configurations
required by DPI.
Use these to write to the DISPC registers. These direct register writes would be
later replaced by a function which applies the configuration using the shadow
register programming model.
The DISPC_DIVISORo registers were written in the functions dpi_set_dispc_clk()
and dpi_set_dsi_clk(), now they just fill up the dispc_clock_info parameter in
mgr_config. They are written later in dpi_config_lcd_manager.
Signed-off-by: Archit Taneja <archit@ti.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
Create a struct dss_lcd_mgr_config which holds LCD overlay manager related
parameters. These are currently partially contained in the omap_dss_device
connected to the manager, and the rest are in the interface driver.
The parameters are directly written to the DISPC registers in the interface
drivers. These should eventually be applied at the correct time using the
shadow register programming model. This struct would help in grouping these
parameters so that they can be applied together.
Signed-off-by: Archit Taneja <archit@ti.com>
|
| |
|
|
|
|
|
|
|
|
|
| |
dipsc_mgr_set_clock div has an int return type to report errors or success.
The function doesn't really check for errors and always returns 0. Change
the return type to void.
Checking for the correct DISPC clock divider ranges will be done when a DSS2
user does a manager apply. This support will be added later.
Signed-off-by: Archit Taneja <archit@ti.com>
|
| |
|
|
|
|
|
|
|
|
| |
The dispc_disable_outputs() function currently disables all LCD managers except
LCD3. This patch adds disabling functionality for LCD3 manager thereby
maintaining consistency of Display Subsystem for in case Display Controller is
reset when LCD3 manager is in use.
Signed-off-by: Chandrabhanu Mahapatra <cmahapatra@ti.com>
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
|
| |\
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This series does the following things:
- Remove passive matrix LCD support: There are no panel drivers with
passive matrix LCD drivers in DSS2. There are no passive matrix panels
even available to test with DSS. Since no one is using passive matrix
panels, stop trying to support it. It cleans up the DSS driver.
- Add some new fields to omap_video_timings: There were some standard
panel timing fields missing from omap_video_timings. Namely
Hsync/Vsync/DE levels and interlace. Add these to omap_video_timings
to align it more with xorg modeline. Add some other OMAP DSS specific
fields to omap_video_timings.
- Remove some hacks done because omap_video_timings didn't have the
above fields.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
For DSI operation in videomode, DISPC logic levels for the signals HSYNC, VSYNC
and DE need to be specified to DSI via the fields VP_HSYNC_POL, VP_VSYNC_POL and
VP_DE_POL in DSI_CTRL registers.
This information is completely internal to DSS as logic levels for the above
signals hold no meaning on the DSI bus. Hence a DSI panel driver should be
totally oblivious of these fields.
Fix the logic levels/polarities in the DISPC and DSI registers to a default
value. This is done by overriding these fields in omap_video_timings struct
filled by the panel driver for DISPC, and use the equivalent default values
when programming DSI_CTRL registers. Also, remove the redundant polarity related
fields in omap_dss_dsi_videomode_data.
Signed-off-by: Archit Taneja <archit@ti.com>
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The hdmi CEA and VESA timings were represented by the struct hdmi_video_timings,
omap_video_timings couldn't be used as it didn't contain the fields hsync/vsync
polarities and interlaced/progressive information.
Remove hdmi_video_timings, and use omap_video_timings instead.
Cc: Mythri P K <mythripk@ti.com>
Signed-off-by: Archit Taneja <archit@ti.com>
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Use the interlace field in omap_video_timings to configure/retrieve
corresponding fb mode flags in fb_var_screeninfo and fb_videomode.
The interlace field maps with the fb mode flags FB_VMODE_INTERLACED and
FB_VMODE_NONINTERLACED.
Signed-off-by: Archit Taneja <archit@ti.com>
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
dispc_ovl_setup()
Currently the interlace parameter passed to dispc_ovl_setup() is configured by
checking the display type, and set to true if the display type is VENC.
This isn't correct as other panels can take interlaced content too. The
omap_video_timings struct in manager's private data contains the info whether
the panel is in interlaced mode or not.
Signed-off-by: Archit Taneja <archit@ti.com>
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Add a parameter called interlace which tells whether the timings are in
interlaced or progressive mode. This aligns the omap_video_timings struct with
the Xorg modeline configuration.
It also removes the hack needed to write to divide the manager height by 2 if
the connected interface is VENC.
Signed-off-by: Archit Taneja <archit@ti.com>
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
omap_panel_config contains fields which are finally written to DISPC_POL_FREQo
registers. These are now held by omap_video_timings and are set when the manager
timings are applied.
Remove the omap_panel_config enum, and remove all it's references from panel or
interface drivers.
Signed-off-by: Archit Taneja <archit@ti.com>
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Use the newly added fields in omap_video_timings(hsync, vsync and data_enable
logic levels and data, hsync and vsync latching related info) to
configure/retrieve corresponding sync flags in fb_var_screeninfo and
fb_videomode.
Out of the new fields, hsync_level and vsync_level can be mapped to the fb sync
flags FB_SYNC_HOR_HIGH_ACT and FB_SYNC_VERT_HIGH_ACT.
When converting fb mode to omap_video_timings, the fields which don't have an
equivalent parameter in fb are kept as the original values if the panel driver
has a get_timings op, else they are set to default values.
Signed-off-by: Archit Taneja <archit@ti.com>
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
dispc_mgr_set_pol_freq() configures the fields in the register DISPC_POL_FREQo.
All these fields have been moved to omap_video_timings struct, and are now
programmed in dispc_mgr_set_lcd_timings(). These will be configured when timings
are applied via dss_mgr_set_timings().
Remove dispc_mgr_set_pol_freq() and it's calls from the interface drivers.
Signed-off-by: Archit Taneja <archit@ti.com>
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Hsync, Vsync, Data enable enable logic levels and latching info of Data lanes,
Hsync and Vsync signals(with respect to pixel clock) are newly added parameters
in omap_video_timings.
Program these in dispc_mgr_set_lcd_timings. These will be configured when the
manager's timings are set via dss_mgr_set_timings().
Signed-off-by: Archit Taneja <archit@ti.com>
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
timings sysfs file
The display sysfs file for viewing/storing display timings is something which
will be deprecated. The new omap_video_timings fields (hsync_level, vsync_level
and others) are not configurable or viewable via this sysfs file.
This prevents the need to make the input more configurable to take the new
fields and at the same time work without these fields for backward
compatibility.
In display_timings_store, the omap_video_timings struct used to set the timings
is initialized to the existing panel timings so that the new fields are taken in
correctly. The other fields are taken from the user as before.
Signed-off-by: Archit Taneja <archit@ti.com>
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Some panel timing related fields are contained in omap_panel_config in the form
of flags. The fields are:
- Hsync logic level
- Vsync logic level
- Data driven on rising/falling edge of pixel clock
- Output enable/Data enable logic level
- HSYNC/VSYNC driven on rising/falling edge of pixel clock
Out of these parameters, Hsync and Vsync logic levels are a part of the timings
in the Xorg modeline configuration. So it makes sense to move the to
omap_video_timings. The rest aren't a part of modeline, but it still makes
sense to move these since they are related to panel timings.
These fields stored in omap_panel_config in dssdev are configured for LCD
panels, and the corresponding LCD managers in the DISPC_POL_FREQo registers.
Add the above fields in omap_video_timings. Represent their state via new enums.
Add these parameters to the omap_video_timings instances in the panel drivers.
Keep the corresponding IVS, IHS, IPC, IEO, RF and ONOFF flags in
omap_panel_config for now. The struct will be removed later.
Signed-off-by: Archit Taneja <archit@ti.com>
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Remove configuration of Ac-bias pins
Ac-bias pins need to be configured only for passive matrix displays. Remove
acbi and acb fields in omap_dss_device and their configuration in panel
drivers. Don't program these fields in DISP_POL_FREQo register any more.
The panel driver for sharp-ls037v7dw01, and the panel config for
Innolux AT070TN8 in generic dpi panel driver set acb to a non zero value. This
is most likely carried over from the old omapfb driver which supported passive
matrix displays.
Cc: Thomas Weber <weber@corscience.de>
Signed-off-by: Archit Taneja <archit@ti.com>
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Remove omap_lcd_display_type enum
The enum omap_lcd_display_type is used to configure the lcd display type in
DISPC. Remove this enum and always set display type to TFT by creating function
dss_mgr_set_lcd_type_tft().
Signed-off-by: Archit Taneja <archit@ti.com>
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Remove OMAP_DSS_LCD_TFT as a omap_panel_config flag.
We don't support passive matrix displays any more. Remove this flag from all the
panel drivers.
Force the display_type to OMAP_DSS_LCD_DISPLAY_TFT in the interface drivers.
Signed-off-by: Archit Taneja <archit@ti.com>
|
| |/
|
|
|
|
|
|
|
|
|
|
|
| |
Remove clock constraints related to passive matrix displays.
There is a constraint (pcd_min should be 3) for passive matrix displays. Remove
this constraint in clock divider calculations as we won't support passive
matrix displays any more.
This cleans up the functions which calculate the clock dividers with DSI's PLL
or DSS_FCLK as the clock source.
Signed-off-by: Archit Taneja <archit@ti.com>
|
| |
|
|
|
|
|
|
|
| |
State change of HDMI PHY could potentially take many millisecs, we can do
better by protecting things in hdmi_set_phy_pwr() with a mutex rather than
a spin_lock_irqsave.
Signed-off-by: Jassi Brar <jaswinder.singh@linaro.org>
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
|
| |
|
|
|
|
|
|
| |
It is simpler to read the current status from a register as compared
to maintaining a state variable to hold the information.
Signed-off-by: Jassi Brar <jaswinder.singh@linaro.org>
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
|
| |
|
|
|
|
|
|
| |
DISPC functions have been modified to provide clock and register dumps and debug
support for the LCD3 manager.
Signed-off-by: Chandrabhanu Mahapatra <cmahapatra@ti.com>
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The support for LCD3 manager has been added into the manager module. LCD3 panel
has registers as DISPC_CONTROL3 and DISPC_CONFIG3 just like those in LCD and
LCD2 panels. These registers control the Display Controller (DISPC) module for
LCD3 output. The three LCDs support Display Serial Interface (DSI), Remote Frame
Buffer Interface (RFBI) and Parallel CMOS Output Interface (DPI). These LCDs can
be connected through parallel output interface using DISPC and RFBI or DPI. For
serial interface DSS uses DSI.
The LCD3 panel, just like LCD and LCD2 panels, has a clock switch in DSS_CTRL
register which has been enabled. The clock switch chooses between DSS_CLK and
DPLL_DSI1_C_CLK1 as source for LCD3_CLK. New IRQs as DISPC_IRQ_VSYNC3,
DISPC_IRQ_FRAMEDONE3, DISPC_IRQ_ACBIAS_COUNT_STAT3 and DISPC_IRQ_SYNC_LOST3 have
been added specific to the new manager.
Signed-off-by: Chandrabhanu Mahapatra <cmahapatra@ti.com>
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
|
| |
|
|
|
|
|
|
|
|
|
| |
OMAP5 Display Subsystem (DSS) architecture comes with a additional LCD3 channel
with its own dedicated overlay manager. The current patch adds LCD3 channel and
basic register support for LCD3 channel. It adds register addresses for various
Display Controller (DISPC) registers like DISPC_DEFAULT_COLOR, DISPC_TIMING_H,
DISPC_DIVISORo, etc.
Signed-off-by: Chandrabhanu Mahapatra <cmahapatra@ti.com>
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
The current implementation of LCD channels and managers consists of a number of
if-else construct which has been replaced by a simpler interface. A constant
structure mgr_desc has been created in Display Controller (DISPC) module. The
mgr_desc contains for each channel its name, irqs and is initialized one time
with all registers and their corresponding fields to be written to enable
various features of Display Subsystem. This structure is later used by various
functions of DISPC which simplifies the further implementation of LCD channels
and its corresponding managers.
Signed-off-by: Chandrabhanu Mahapatra <cmahapatra@ti.com>
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
If runtime PM is not enabled in the kernel config, pm_runtime_get_sync()
will always return 1 and pm_runtime_put_sync() will always return
-ENOSYS. pm_runtime_get_sync() returning 1 presents no problem to the
driver, but -ENOSYS from pm_runtime_put_sync() causes the driver to
print a warning.
One option would be to ignore errors returned by pm_runtime_put_sync()
totally, as they only say that the call was unable to put the hardware
into suspend mode.
However, I chose to ignore the returned -ENOSYS explicitly, and print a
warning for other errors, as I think we should get notified if the HW
failed to go to suspend properly.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Cc: Jassi Brar <jaswinder.singh@linaro.org>
Cc: Grazvydas Ignotas <notasas@gmail.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The current way how omapdss handles system suspend and resume is that
omapdss device (a platform device, which is not part of the device
hierarchy of the DSS HW devices, like DISPC and DSI, or panels.) uses
the suspend and resume callbacks from platform_driver to handle system
suspend. It does this by disabling all enabled panels on suspend, and
resuming the previously disabled panels on resume.
This presents a few problems.
One is that as omapdss device is not related to the panel devices or the
DSS HW devices, there's no ordering in the suspend process. This means
that suspend could be first ran for DSS HW devices and panels, and only
then for omapdss device. Currently this is not a problem, as DSS HW
devices and panels do not handle suspend.
Another, more pressing problem, is that when suspending or resuming, the
runtime PM functions return -EACCES as runtime PM is disabled during
system suspend. This causes the driver to print warnings, and operations
to fail as they think that they failed to bring up the HW.
This patch changes the omapdss suspend handling to use PM notifiers,
which are called before suspend and after resume. This way we have a
normally functioning system when we are suspending and resuming the
panels.
This patch, I believe, creates a problem that somebody could enable or
disable a panel between PM_SUSPEND_PREPARE and the system suspend, and
similarly the other way around in resume. I choose to ignore the problem
for now, as it sounds rather unlikely, and if it happens, it's not
fatal.
In the long run the system suspend handling of omapdss and panels should
be thought out properly. The current approach feels rather hacky.
Perhaps the panel drivers should handle system suspend, or the users of
omapdss (omapfb, omapdrm) should handle system suspend.
Note that after this patch we could probably revert
0eaf9f52e94f756147dbfe1faf1f77a02378dbf9 (OMAPDSS: use sync versions of
pm_runtime_put). But as I said, this patch may be temporary, so let's
leave the sync version still in place.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reported-by: Jassi Brar <jaswinder.singh@linaro.org>
Tested-by: Jassi Brar <jaswinder.singh@linaro.org>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In preparation of OMAP moving to Common Clk Framework(CCF) change
clk_enable() and clk_disable() calls to clk_prepare_enable() and
clk_disable_unprepare() in omapdss. This can be safely done, as omapdss
never enables or disables clocks in atomic context.
Signed-off-by: Rajendra Nayak <rnayak@ti.com>
Cc: Tomi Valkeinen <tomi.valkeinen@ti.com>
Cc: <linux-fbdev@vger.kernel.org>
Cc: Paul Walmsley <paul@pwsan.com>
Cc: Mike Turquette <mturquette@linaro.org>
[tomi.valkeinen@ti.com: updated patch description]
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
|
| |
|
|
|
| |
Signed-off-by: Peter Meerwald <pmeerw@pmeerw.net>
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
|
| |
|
|
|
|
|
|
|
|
|
|
| |
We have two almost the same enums: omap_channel and
omap_dss_overlay_managers. omap_channel is used almost everywhere, and
omap_channel assigns explicit values to the enum values which are needed
for proper operation.
omap_dss_overlay_managers is only used in one place, so it's easy to
remove it, which is what this patch does.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
|
| |\
| |
| |
| | |
Merge v3.5-rc2 to get latest device tree and dynamic debug changes.
|
| | | |
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
If the privileges given to root threads (3% of allowable memory) or a
negative value of /proc/pid/oom_score_adj happen to exceed the amount of
rss of a thread, its badness score overflows as a result of commit
a7f638f999ff ("mm, oom: normalize oom scores to oom_score_adj scale only
for userspace").
Fix this by making the type signed and return 1, meaning the thread is
still eligible for kill, if the value is negative.
Reported-by: Dave Jones <davej@redhat.com>
Acked-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: David Rientjes <rientjes@google.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
| | |\
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull scheduler fixes from Ingo Molnar.
* 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
sched: Fix the relax_domain_level boot parameter
sched: Validate assumptions in sched_init_numa()
sched: Always initialize cpu-power
sched: Fix domain iteration
sched/rt: Fix lockdep annotation within find_lock_lowest_rq()
sched/numa: Load balance between remote nodes
sched/x86: Calculate booted cores after construction of sibling_mask
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
It does not get processed because sched_domain_level_max is 0 at the
time that setup_relax_domain_level() is run.
Simply accept the value as it is, as we don't know the value of
sched_domain_level_max until sched domain construction is completed.
Fix sched_relax_domain_level in cpuset. The build_sched_domain() routine calls
the set_domain_attribute() routine prior to setting the sd->level, however,
the set_domain_attribute() routine relies on the sd->level to decide whether
idle load balancing will be off/on.
Signed-off-by: Dimitri Sivanich <sivanich@sgi.com>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/20120605184436.GA15668@sgi.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Add some code to validate assumptions we're making and output
warnings if they are not.
If this trigger we want to know about it.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Alex Shi <lkml.alex@gmail.com>
Link: http://lkml.kernel.org/n/tip-6uc3wk5s9udxtdl9cnku0vtt@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Often when we run into mis-shapen topologies the balance iteration
fails to update the cpu power properly and we'll end up in /0 traps.
Always initialize the cpu-power to a semi-sane value so that we can
at least boot the machine, even if the load-balancer might not
function correctly.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/n/tip-3lbhyj25sr169ha7z3qht5na@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Weird topologies can lead to asymmetric domain setups. This needs
further consideration since these setups are typically non-minimal
too.
For now, make it work by adding an extra mask selecting which CPUs
are allowed to iterate up.
The topology that triggered it is the one from David Rientjes:
10 20 20 30
20 10 20 20
20 20 10 20
30 20 20 10
resulting in boxes that wouldn't even boot.
Reported-by: David Rientjes <rientjes@google.com>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/n/tip-3p86l9cuaqnxz7uxsojmz5rm@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Roland Dreier reported spurious, hard to trigger lockdep warnings
within the scheduler - without any real lockup.
This bit gives us the right clue:
> [89945.640512] [<ffffffff8103fa1a>] double_lock_balance+0x5a/0x90
> [89945.640568] [<ffffffff8104c546>] push_rt_task+0xc6/0x290
if you look at that code you'll find the double_lock_balance() in
question is the one in find_lock_lowest_rq() [yay for inlining].
Now find_lock_lowest_rq() has a bug.. it fails to use
double_unlock_balance() in one exit path, if this results in a retry in
push_rt_task() we'll call double_lock_balance() again, at which point
we'll run into said lockdep confusion.
Reported-by: Roland Dreier <roland@kernel.org>
Acked-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1337282386.4281.77.camel@twins
Signed-off-by: Ingo Molnar <mingo@kernel.org>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Commit cb83b629b ("sched/numa: Rewrite the CONFIG_NUMA sched
domain support") removed the NODE sched domain and started checking
if the node distance in SLIT table is farther than REMOTE_DISTANCE,
if so, it will lose the load balance chance at exec/fork/wake_affine
points.
But actually, even the node distance is farther than REMOTE_DISTANCE.
Modern CPUs also has QPI like connections, which ensures that memory
access is not too slow between nodes. So the above change in behavior
on NUMA machine causes a performance regression on various benchmarks:
hackbench, tbench, netperf, oltp, etc.
This patch will recover the scheduler behavior to old mode on all my
Intel platforms: NHM EP/EX, WSM EP, SNB EP/EP4S, and thus fixes the
perfromance regressions. (all of them just have 2 kinds distance, 10, 21)
Signed-off-by: Alex Shi <alex.shi@intel.com>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1338965571-9812-1-git-send-email-alex.shi@intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Commit 316ad248307fb ("sched/x86: Rewrite set_cpu_sibling_map()")
broke the booted_cores accounting.
The problem is that the booted_cores accounting needs all the
sibling links set up. So restore the second loop and add a comment as
to why its needed.
On qemu booted with -smp sockets=1,cores=2,threads=2;
Before:
$ grep cores /proc/cpuinfo
cpu cores : 2
cpu cores : 1
cpu cores : 4
cpu cores : 3
With the patch:
$ grep cores /proc/cpuinfo
cpu cores : 2
cpu cores : 2
cpu cores : 2
cpu cores : 2
Reported-by: Prarit Bhargava <prarit@redhat.com>
Reported-by: Borislav Petkov <bp@amd64.org>
Signed-off-by: Kamalesh Babulal <kamalesh@linux.vnet.ibm.com>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/20120531073738.GH7511@linux.vnet.ibm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Fix lots of new kernel-doc warnings in kernel/sched/fair.c:
Warning(kernel/sched/fair.c:3625): No description found for parameter 'env'
Warning(kernel/sched/fair.c:3625): Excess function parameter 'sd' description in 'update_sg_lb_stats'
Warning(kernel/sched/fair.c:3735): No description found for parameter 'env'
Warning(kernel/sched/fair.c:3735): Excess function parameter 'sd' description in 'update_sd_pick_busiest'
Warning(kernel/sched/fair.c:3735): Excess function parameter 'this_cpu' description in 'update_sd_pick_busiest'
.. more warnings
Signed-off-by: Randy Dunlap <rdunlap@xenotime.net>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
This reverts commit 9e612a008fa7fe493a473454def56aa321479495.
It incorrectly finds VGA connectors where none are attached, apparently
not noticing that nothing replied to the EDID queries, and happily using
the default EDID modes that have nothing to do with actual hardware.
That in turn then causes X to fall down to the lowest common
denominator, which is usually the default 1024x768 mode that is in the
default EDID and pretty much anything supports).
I'd suggest that if not relying on the HDP pin, the code should at least
check whether it gets valid EDID data back, rather than just assume
there's something on the VGA connector.
Cc: Dave Airlie <airlied@linux.ie>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
| | |\ \
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4
Pull ext4 bug fixes from Theodore Ts'o:
"This update contains two bug fixes, both destined for the stable tree.
Perhaps the most important is one which fixes ext4 when used with file
systems originally formatted for use with ext3, but then later
converted to take advantage of ext4."
* tag 'ext4_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4:
ext4: don't set i_flags in EXT4_IOC_SETFLAGS
ext4: fix the free blocks calculation for ext3 file systems w/ uninit_bg
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Commit 7990696 uses the ext4_{set,clear}_inode_flags() functions to
change the i_flags automatically but fails to remove the error setting
of i_flags. So we still have the problem of trashing state flags.
Fix this by removing the assignment.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Cc: stable@kernel.org
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Ext3 filesystems that are converted to use as many ext4 file system
features as possible will enable uninit_bg to speed up e2fsck times.
These file systems will have a native ext3 layout of inode tables and
block allocation bitmaps (as opposed to ext4's flex_bg layout).
Unfortunately, in these cases, when first allocating a block in an
uninitialized block group, ext4 would incorrectly calculate the number
of free blocks in that block group, and then errorneously report that
the file system was corrupt:
EXT4-fs error (device vdd): ext4_mb_generate_buddy:741: group 30, 32254 clusters in bitmap, 32258 in gd
This problem can be reproduced via:
mke2fs -q -t ext4 -O ^flex_bg /dev/vdd 5g
mount -t ext4 /dev/vdd /mnt
fallocate -l 4600m /mnt/test
The problem was caused by a bone headed mistake in the check to see if a
particular metadata block was part of the block group.
Many thanks to Kees Cook for finding and bisecting the buggy commit
which introduced this bug (commit fd034a84e1, present since v3.2).
Reported-by: Sander Eikelenboom <linux@eikelenboom.it>
Reported-by: Kees Cook <keescook@chromium.org>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Tested-by: Kees Cook <keescook@chromium.org>
Cc: stable@kernel.org
|
| | |\ \ \
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
git://git.kernel.org/pub/scm/linux/kernel/git/paulus/powerpc
Pull powerpc fixes from Paul Mackerras:
"Two small fixes for powerpc:
- a fix for a regression since 3.2 that causes 4-second (or longer)
pauses
- a fix for a potential oops when loading kernel modules on 32-bit
embedded systems."
* 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/paulus/powerpc:
powerpc: Fix kernel panic during kernel module load
powerpc/time: Sanity check of decrementer expiration is necessary
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
This fixes a problem which can causes kernel oopses while loading
a kernel module.
According to the PowerPC EABI specification, GPR r11 is assigned
the dedicated function to point to the previous stack frame.
In the powerpc-specific kernel module loader, do_plt_call()
(in arch/powerpc/kernel/module_32.c), GPR r11 is also used
to generate trampoline code.
This combination crashes the kernel, in the case where the compiler
chooses to use a helper function for saving GPRs on entry, and the
module loader has placed the .init.text section far away from the
.text section, meaning that it has to generate a trampoline for
functions in the .init.text section to call the GPR save helper.
Because the trampoline trashes r11, references to the stack frame
using r11 can cause an oops.
The fix just uses GPR r12 instead of GPR r11 for generating the
trampoline code. According to the statements from Freescale, this is
safe from an EABI perspective.
I've tested the fix for kernel 2.6.33 on MPC8541.
Cc: stable@vger.kernel.org
Signed-off-by: Steffen Rumler <steffen.rumler.ext@nsn.com>
[paulus@samba.org: reworded the description]
Signed-off-by: Paul Mackerras <paulus@samba.org>
|
| | | |/ /
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
This reverts 68568add2c ("powerpc/time: Remove unnecessary sanity check
of decrementer expiration"). We do need to check whether we have reached
the expiration time of the next event, because we sometimes get an early
decrementer interrupt, most notably when we set the decrementer to 1 in
arch_irq_work_raise(). The effect of not having the sanity check is that
if timer_interrupt() gets called early, we leave the decrementer set to
its maximum value, which means we then don't get any more decrementer
interrupts for about 4 seconds (or longer, depending on timebase
frequency). I saw these pauses as a consequence of getting a stray
hypervisor decrementer interrupt left over from exiting a KVM guest.
This isn't quite a straight revert because of changes to the surrounding
code, but it restores the same algorithm as was previously used.
Cc: stable@vger.kernel.org
Acked-by: Anton Blanchard <anton@samba.org>
Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
|
