PVR driver
commit
f0caf5295ff37b0f4ecb39b6b12f9f61b83f5eec
Author: Charles Johnson <charles.f.johnson@intel.com>
Date: Tue Oct 11 14:01:37 2011 +0300
GFX: Fix display init by moving mdfld_dsi_pkg_sender_init() earlier
The DSI-LVDS bridge driver code moved the mdfld_dsi_pkg_sender_init() call
later in mdfld_dsi_output_init(). This breaks display init for command mode
panels, so move it back.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
23bb3b4c739813c0e461d0c00143b8ccb0bdfd12
Author: Jani Nikula <jani.nikula@intel.com>
Date: Tue Oct 18 11:06:37 2011 +0300
gfx: pvr: drop Moorestown specific files
Remove unused Moorestown specific files in PVR.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
989d469cf7e30d79933c46ac981e8eaea159d9e3
Author: Jani Nikula <jani.nikula@intel.com>
Date: Tue Oct 18 10:49:23 2011 +0300
gfx: display: drop support for Medfield A0
Remove hacky workarounds for an old board revision.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
895910f95ab83852ad7b9ea48e73d5abd151ec11
Author: Jani Nikula <jani.nikula@intel.com>
Date: Tue Oct 18 10:29:28 2011 +0300
gfx: pvr: drop #ifdef CONFIG_X86_MRST
Moorestown is not supported. Remove code paths depending on
CONFIG_X86_MRST=y.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
5412b558670e148323100475a466d54f2968718e
Author: Jani Nikula <jani.nikula@intel.com>
Date: Tue Oct 18 10:28:03 2011 +0300
gfx: display: drop #ifdefs on CONFIG_X86_MRST
Moorestown is not supported. Remove code paths depending on
CONFIG_X86_MRST=y.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
1654d116601bc06ee8c0576b0ebcc6aa3e5430d3
Author: Jani Nikula <jani.nikula@intel.com>
Date: Tue Oct 18 10:13:19 2011 +0300
gfx: display: drop IS_FW_UPDATED macro
Yes, let's assume we really have updated the firmware. Remove IS_FW_UPDATED
macro, unconditionally defined as 1, and regard it as 1.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
eeb911cf7a55c36df4c1f3acb7cdf17d703217e7
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Oct 17 17:53:18 2011 +0300
gfx: display: fix a number of compiler warnings
Fix some compiler warnings.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
b204c62730c2ddb8a89f231928476a598b1b8121
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Oct 17 17:33:48 2011 +0300
gfx: imgv: drop Lincroft topaz support
Lincroft is not relevant for Medfield. Remove Lincroft topaz support.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
c6b25a0bc70b5627f03ba335797c04412dbd5662
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Oct 17 17:02:11 2011 +0300
gfx: display: clean up psb_drv.h a bit
The psb_drv.h has accumulated excess cruft. Tidy it up a bit, removing
unused code.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
8e7642d52cfbf2e0a25c4088dc80aa7783a909f3
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Oct 17 16:46:09 2011 +0300
gfx: display: drop unused i2c module
Remove unused code left behind in earlier cleanups.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
39de15266755794e0defa94d3639478bc5d98114
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Oct 17 16:37:47 2011 +0300
gfx: display: drop unused modes support
Remove unused code left behind in earlier cleanups.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
a7c57928751bcaf932d0d56dc19701c651d5ad54
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Oct 17 14:33:49 2011 +0300
gfx: display: drop unused Pyrenees panel support
Remove unused and bitrotting panel driver.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
5e1855228771a00d5cc44f850720c16f57456a69
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Oct 17 14:23:14 2011 +0300
gfx: display: drop unused BIOS support files
These BIOS routines are not used in Medfield. Remove them.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
46b2bed9b7e628763751f2d19e871ab5dad9b96d
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Oct 17 14:20:36 2011 +0300
gfx: display: drop the rest of the unused LVDS code
Keep purging LVDS related code not relevant in Medfield.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
676b6125308e5cee924a124a81b796b73f0a0526
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Oct 17 14:12:12 2011 +0300
gfx: display: drop unused SDVO support
SDVO is not supported in Medfield. Remove the code.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
f32caf1bb703c5bd2f769dad960626f2deb4f62c
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Oct 17 14:08:59 2011 +0300
gfx: display: drop unused Aava DSI display driver
Remove unused and bitrotting display driver.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
e45e1c13d1de3ff7542bf4725154a834af4c0b78
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Oct 17 14:06:29 2011 +0300
gfx: display: drop unused Moorestown DSI files
Remove Moorestown specific DSI support code.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
34903fb8888caec18249d1c38660676d34f42fbb
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Oct 17 13:59:20 2011 +0300
gfx: display: clean up unused backlight functionality
Remove backlight related code not relevant in Medfield.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
193157347041dbc678b4875d15c80b0b624152b5
Author: Jani Nikula <jani.nikula@intel.com>
Date: Fri Oct 14 14:46:54 2011 +0300
gfx: display: remove unused watchdog functionality
Remove watchdog related functionality that is not used on Medfield.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
6cce56805947da0bcd6e9a4d585e3b55b6c7cacc
Author: Jani Nikula <jani.nikula@intel.com>
Date: Fri Oct 14 14:39:10 2011 +0300
gfx: display: remove unused Poulsbo lid related functionality
Drop Poulsbo specific code related to lid.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
14e00f79b5ba7f94628dfefa26bc0f6396f0e302
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Oct 17 13:53:36 2011 +0300
gfx: display: clean up display and LVDS support, dropping unused code
Focusing on Medfield support only allows us to drop unused and bitrotting
code.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
40c24db86884cc1804ca4d30f8b57515e33d50c4
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Oct 17 12:31:23 2011 +0300
gfx: topaz and msvdx are always enabled in Medfield
Topaz and msvdx are present in Medfield. Remove IS_TOPAZ() and IS_MSVDX()
macros and consider them always true.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
7b2b1ae58b2ce07ea2f43fc0f25ef822b52cbb69
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Oct 17 12:27:43 2011 +0300
gfx: drop non-Medfield support
Drop support for anything other than Medfield. Remove IS_MDFLD() macro and
consider it always true.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
00d0cfba8f7fafc2f9f901896eb8d928698f14ce
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Oct 17 12:06:25 2011 +0300
gfx: drop Moorestown support
Drop Moorestown support. Remove IS_MRST() macro and consider it always
false.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
5be8a9e6524f5eb9caf9aacb2010fca52280f1bb
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Oct 17 11:34:44 2011 +0300
gfx: display: drop Moorestown MM support
Drop Moorestown MM support. Remove IS_MRST_MM() macro and consider it
always false.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
ec418c2d3342dce4c66cbbc14bbe08f1338dddd0
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Oct 17 11:32:30 2011 +0300
gfx: display: drop Oaktrail support
Drop Oaktrail support. Remove IS_OAKTRAIL() macro and consider it always
false.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
d3b4ea082fdc2d9a6c27f9457a0e09d8dad6b9fe
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Oct 17 11:30:55 2011 +0300
gfx: display: drop non-MID support
Drop non-MID support. Remove IS_MID() macro and consider it always true.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
2d12c777b382a51c9273b79ebe48d586299085de
Author: Jani Nikula <jani.nikula@intel.com>
Date: Fri Oct 14 14:08:32 2011 +0300
gfx: display: drop Poulsbo support
Drop Poulsbo support. Remove IS_POULSBO() macro and consider it always
false.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
8ac4007de566b708e1daed763c1b703bb28de1e9
Author: Jani Nikula <jani.nikula@intel.com>
Date: Fri Oct 14 12:29:57 2011 +0300
gfx: display: Initialize DSI-LVDS bridge chip only when it's used
Fix modular build, but does not allow the module to be unloaded.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
74a69ce53506de2d7970b497f89440553bfbe3ba
Author: Jani Nikula <jani.nikula@intel.com>
Date: Thu Oct 13 17:43:00 2011 +0300
Revert "graphics: Fix build as module"
This reverts commit "graphics: Fix build as module", which oopses on
graphics driver probe.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
ede699dc31f726bc06d65f09d17e485e0e39cabc
Author: Alan Cox <alan@linux.intel.com>
Date: Thu Oct 13 10:41:23 2011 +0100
gfx: drop Moorestown support
First step at dropping Moorestown support: don't build it. This is part of
a larger commit in AC tree filtered to drivers/staging/mrst.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
379b9b51688ed33cd2385954cfd61350ad7d6756
Author: Imre Deak <imre.deak@intel.com>
Date: Thu Oct 13 14:36:19 2011 +0100
gfx: pvr: change SECURE_FD_EXPORT compile option into a run-time check
At the moment we need to support user space libraries compiled both with
and without this option on. Check which library is in use and change the
functionality accordingly.
Note that this option is only a poor workaround for a problem in user
space. Once we have a proper fix there, we can remove the corresponding
code here.
Also move -DSUPPORT_MEMINFO_IDS to where similar defines are specified
and remove -DPVR_SECURE_HANDLES which is already specified elsewhere.
Signed-off-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
commit
8cfeef14fa2f72e68ba7c87f67ced2ad8776a918
Author: Imre Deak <imre.deak@intel.com>
Date: Thu Oct 13 14:36:18 2011 +0100
gfx: pvr: prepare for build options that can mismatch
Needed by the next patch.
We need a place for options bits that user space can pass to the driver.
These bits - unlike the current compile option bits - will not be
compared to the driver's compile option bits, the same driver binary will
support user space libraries both with these bits on or off.
Also sanitize some of the debug/error prints.
Signed-off-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
commit
4862da2e2dcf43226b7b66d8eee4c7e840144ae4
Author: Imre Deak <imre.deak@intel.com>
Date: Thu Oct 13 14:36:18 2011 +0100
gfx: pvr: add pvr_get_sgx_dev_info
Factor out this functionality to its own function and use it in the
spot where it was inlined earlier.
Needed by the next patch.
Signed-off-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
commit
c2db37c90938424be4b9493f2b2615d17f37a140
Author: Pauli Nieminen <pauli.nieminen@linux.intel.com>
Date: Thu Oct 13 14:36:18 2011 +0100
gfx: pvr: Add dummy ioctl entry for a new UMG ioctl
To have same ABI with HC userspace we need stub ioctl entry for a new
UMG ioctl. The ioctl is to query current buffer index from swapchain.
But we have depracated swapchain support in kernel so no point
implementing the ioctl.
Signed-off-by: Pauli Nieminen <pauli.nieminen@linux.intel.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
commit
62fd2f750283b44a0ec529ec6d32889ed0e9fab6
Author: Imre Deak <imre.deak@intel.com>
Date: Thu Oct 13 14:36:17 2011 +0100
gfx: pvr: check also the PVR specific retval of IOCTLs
In addition to the standard IOCTL error return, the PVR driver can
return a custom error code in it's IOCTL out structure. Check for this
too.
Signed-off-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
commit
16a1c431f86110e8d3d19bcc1b57ecb9afb7b708
Author: Imre Deak <imre.deak@intel.com>
Date: Wed Oct 12 19:22:47 2011 +0100
gfx: pvr: add white list based firmware version check
Do this to support multiple SGX firmware versions. Until now there had
to be an exact match between firmware and kernel driver versions, which
made maintenance of the kernel driver difficult. This change is the
beginning of a series of changes that relaxes the user
space/kernel/firmware dependency as much as possible and allows us to
have a single kernel driver for all supported configurations.
Also do some error/debug printing sanitizing on the way.
Signed-off-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
commit
39897129184c47ea32ef87917c88a7d7bb3165cd
Author: Imre Deak <imre.deak@intel.com>
Date: Wed Oct 12 19:22:46 2011 +0100
gfx: pvr: check sizes of IOCTL parameter structures
This sanity check can be done easier using the IOC_* macros for
well-written IOCTLs. In case of PVR IOCTLs we have to do them manually,
since each input/output IOCTL parameter is accessed through an extra
redirection from a fixed IOCTL structure that is the same for every PVR
IOCTL.
Signed-off-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
commit
50ee3670f22df7525b83dfe163c3586afcbf79b5
Author: Imre Deak <imre.deak@intel.com>
Date: Wed Oct 12 19:22:46 2011 +0100
gfx: pvr: add convenience variable
Needed by the upcoming patch.
Also sanitize a debug printing on the way.
Signed-off-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
commit
96b148ad595dd104c7a2f0d06d0fae6cd0e090c3
Author: Imre Deak <imre.deak@intel.com>
Date: Wed Oct 12 19:22:45 2011 +0100
gfx: pvr: fix IOCTL ABI w.r.t SGX kick IOCTL
Recent PVR user space library versions extended the parameter structure
of this IOCTL effectively changing the ABI. Fix the structure
accordingly.
bLastInScene is supposed to optimize on some of the cache flushes before
starting a rendering operation. It's still under investigation whether
it's safe to do this though. Until we figure this out, ignore this
parameter and perform cache flushes as before.
SGX_MAX_SRC_SYNCS is the maximum number of source buffers per render
operation for which buffers we can do a dependency check.
Signed-off-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
commit
30fee8cb6bd1ef947c80aa927a8b4bdfffa674d5
Author: Imre Deak <imre.deak@intel.com>
Date: Wed Oct 12 19:22:45 2011 +0100
gfx: pvr: fix IOCTL ABI w.r.t MEMINFO_IDS, SECURE_FD_EXPORT, SECURE_HANDLES
Recent PVR user space library versions extended some of the IOCTL
parameter structures, effectively changing the ABI. These extensions
are optionally built based on the above Kconfig options. They are
in fact only workarounds for problems in those library versions, so
after we fixed those problems in the libraries properly we shoud remove
the support for them in the driver.
For other library versions not depending on these features, enabling
them won't have any functional difference. The net result is that we
remove the ABI dependency and can use both types of library versions
with the same driver binary.
Signed-off-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
commit
5e809df345a679f2d99bf79cfbba041ab33bed99
Author: Pauli Nieminen <pauli.nieminen@linux.intel.com>
Date: Fri Oct 7 15:36:19 2011 +0300
gfx: Use drm mode page flipping also in medfield
The page flip was originally developed and tested using mrst prototype
but same code works correctly in medfield too. To enable flip support
driver has to add the flip function to medfield crtc function pointers.
Signed-off-by: Pauli Nieminen <pauli.nieminen@linux.intel.com>
commit
dd97cf8808b283727a0b7114268559db43536c16
Author: Alan Cox <alan@linux.intel.com>
Date: Mon Oct 3 12:46:53 2011 +0100
graphics: Fix build as module
One module_init only
Signed-off-by: Alan Cox <alan@linux.intel.com>
commit
122715042e4127e8fa27fb745ae0ae405e9a9fef
Author: Ander Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com>
Date: Fri Sep 30 12:52:44 2011 +0100
pvr: drv: implement drm page flip ioctl
Signed-off-by: Ander Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com>
Signed-off-by: Pauli Nieminen <pauli.nieminen@intel.com>
commit
d6171d71d2839559253d9bf0c33bcba4271f73f8
Author: Ander Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com>
Date: Fri Sep 30 12:52:43 2011 +0100
pvr: srvkm: add a callback based sync mechanism
Users of this mechanism can register a callback that will be called
when the specified sync object's completed ops turn to be greater or
equal to the current pending values. If the specified object is already
in sync, the callback will be called immediately. The flags control
whether read, write or both types of operations should be considered.
Signed-off-by: Ander Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com>
Signed-off-by: Pauli Nieminen <pauli.nieminen@intel.com>
commit
a881d79a4ef3d994de797e6910cd3f24e3b06b28
Author: Jani Nikula <jani.nikula@intel.com>
Date: Tue Sep 20 16:54:00 2011 +0300
GFX: Add tc358764xbg/65xbg MIPI DSI-LVDS bridge chip support
Add preliminary support for Toshiba tc358764xbg/65xbg MIPI DSI-LVDS bridge
chip. Set panelid=TC35876X on the kernel command line to enable support.
ISSUES/TODO:
* The I2C device and GPIO line numbers should be defined in firmware,
rather than hard coded in the driver. Consequently, this implementation
only works on dv0.9.
* The driver exposes functions that are called from here and there in the
display controller driver, even before the driver has been initialized.
* The implementation touches DSI clock registers, among other things,
directly without locking. It seems to usually work by accident.
* Many of the problems in the implementation (like the above) come from the
fact that the whole display controller driver lacks proper infrastructure
for pretty much everything, including easily adding new display
drivers. It's not trivial to fix this when adding a new driver.
* Fix checkpatch/sparse errors.
This patch is based on earlier work by (at least) Srinivas Pandruvada, Paul
Drews, Ke Ai, and Michael Demeter.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
6e9f8dd97fc07f19c47f3d4160ce442aecbac830
Author: Jani Nikula <jani.nikula@intel.com>
Date: Fri Sep 23 17:16:28 2011 +0300
GFX: Redefine register addresses with pipe parameter
Replace MIPIA_SOME_REG + reg_offset constructs with MIPI_SOME_REG(pipe)
macros to simplify error prone code and to improve readability.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
7e5a33bdbd3297c75b69bfd2a802a8da582f3629
Author: Jani Nikula <jani.nikula@intel.com>
Date: Fri Sep 23 10:40:15 2011 +0300
GFX: Remove duplicate DSI controller init code
DSI controller init functions dsi_controller_init(),
dsi_controller_dpi_init(), and dsi_controller_dbi_init() are mostly just
duplicates of the corresponding functions with mdfld_ prefix. Handle the
minor differences, and remove code duplication.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
2fa11c603a4c95edcedc6333310e5184cba8da9c
Author: Jani Nikula <jani.nikula@intel.com>
Date: Tue Sep 6 16:39:53 2011 +0300
GFX: Remove redundant and unused ioctl definitions
Remove redundant, unused, and commented out ioctl definitions.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
f24b654d4a226490a9ba753ae722014da78acae6
Author: Jani Nikula <jani.nikula@intel.com>
Date: Tue Sep 6 14:45:33 2011 +0300
GFX: Remove another useless DSI command send wrapper
Use mdfld_dsi_send_mcs_short() directly rather than through a wrapper.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
b901ccfe425f89686f8b1149cdf21f2912c39c21
Author: Jani Nikula <jani.nikula@intel.com>
Date: Tue Sep 6 13:07:13 2011 +0300
GFX: Replace the BIT<N> macros with the generic BIT(N) from bitops.h
Replace BIT<N> macros with generic BIT(N) macros from bitops.h. The bulk of
the change was done using sed:
sed -i 's/\([^A-Za-z0-9_]*\)\(BIT\)\([0-9][0-9]*\)\([^A-Za-z_]*\)/\1\2(\3)\4/g' *.c *.h
Then manually dropped the #defines from psb_intel_reg.h.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
4088df73c3ca8d050db82b19c20e6733ec34cd74
Author: Jani Nikula <jani.nikula@intel.com>
Date: Tue Sep 6 12:28:12 2011 +0300
GFX: Make display config mode getter functions static
The display config mode getter functions are used through function pointers
within struct panel_funcs, so they can be static.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
1bfa41a52df2f3e7540afd8bab26c6dce3aa5a43
Author: Jani Nikula <jani.nikula@intel.com>
Date: Tue Sep 6 12:26:10 2011 +0300
DRV: Remove unused display header files pyr_vid.h and tmd_cmd.h
The display header files pyr_vid.h and tmd_cmd.h contain only declarations
for functions that don't even exist. Remove the files.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
5eb3e79030f0b0e528af5c46848cbd69a49da7ec
Author: Jani Nikula <jani.nikula@intel.com>
Date: Tue Sep 6 12:12:45 2011 +0300
GFX: Remove useless unreachable BIT macros
Remove useless #defines within #if 0.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
b7ab02155911f5fd8221282a30301ad802bf37f4
Author: Jani Nikula <jani.nikula@intel.com>
Date: Tue Sep 6 12:01:37 2011 +0300
PVR: Drop support for kernel versions 2.6.34 and older
There is absolutely no guarantee that the driver would work with kernel
versions 2.6.34 or older, even with the conditional compilation in place,
and the further modifications that will be made to the driver won't take
older kernels into account anyway. Thus explicitly drop support for 2.6.34
and earlier instead of pretending.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
ed996159001324249dc71a12e7bd60202e21014d
Author: Jani Nikula <jani.nikula@intel.com>
Date: Tue Sep 6 11:50:09 2011 +0300
GFX: Drop support for kernel versions 2.6.34 and older
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
3cc85dd6288c5b7bd4a8f6e05ab99c346850a330
Author: Jani Nikula <jani.nikula@intel.com>
Date: Tue Sep 6 11:41:18 2011 +0300
GFX: Drop support for kernel versions 2.6.34 and older
There is absolutely no guarantee that the driver would work with kernel
versions 2.6.34 or older, even with the conditional compilation in place,
and the further modifications that will be made to the driver won't take
older kernels into account anyway. Thus explicitly drop support for 2.6.34
and earlier instead of pretending.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
018d062de792c070e2d0bea469421ae8b6a4dd78
Author: Jani Nikula <jani.nikula@intel.com>
Date: Tue Sep 6 10:56:18 2011 +0300
GFX: Remove unused dsi pkg structs
The mdfld_dsi_{gen_short_pkg,gen_long_pkg,dcs_pkg,pkg} structs are no
longer used and can be removed.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
fad1f6978e0138126e41fe183ba25c02a806fb90
Author: Pauli Nieminen <pauli.nieminen@intel.com>
Date: Wed Sep 7 18:00:27 2011 +0300
pvr: Free export handle only after memory is freed
Export handles are currently freed when the exporter frees it local
mapping of the memory. Next export may (is very likely) to reuse the
handle for different memory.
While process that received the export handle wants to avoid expensive
memory mapping when it already has mapped a matching handle. If the
memory behind a handle changes suddenly process receiving process will
continue use old memory.
Problem can be seen in screen that random client would lose content
because it renders to old memory while display server shows the new
memory in screen.
The fix is to delay the free of the handle until the memory is freed.
Signed-off-by: Pauli Nieminen <pauli.nieminen@intel.com>
Reviewed-by: Imre Deak <imre.deak@intel.com>
commit
c3c6c7df9ac604855c99a9d05a7b8bef8495e264
Author: Imre Deak <imre.deak@intel.com>
Date: Thu Sep 1 18:01:28 2011 +0300
mrst: pvr: add debugfs support
Signed-off-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Pauli Nieminen <pauli.nieminen@intel.com>
commit
bc6de1ede5c35870d1113fe0ef6c90b0f1e8f473
Author: Imre Deak <imre.deak@intel.com>
Date: Thu Sep 1 18:01:27 2011 +0300
mrst: pvr: add tracing for flip command
Signed-off-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Pauli Nieminen <pauli.nieminen@intel.com>
commit
193f1a94f45e820085bc3402b2c72f6ee103b57e
Author: Imre Deak <imre.deak@intel.com>
Date: Thu Sep 1 18:01:26 2011 +0300
mrst: pvr: add tracing to SGX queryblits command
Signed-off-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Pauli Nieminen <pauli.nieminen@intel.com>
commit
eea730158b268d0df9a3d96e3c95887b6393a9c0
Author: Imre Deak <imre.deak@intel.com>
Date: Thu Sep 1 18:01:25 2011 +0300
mrst: pvr: add tracing to the SGX kick and transfer commands
Signed-off-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Pauli Nieminen <pauli.nieminen@intel.com>
commit
ea703813ba51bedf059ff4a182a845f2235d6b55
Author: Imre Deak <imre.deak@intel.com>
Date: Thu Sep 1 18:01:24 2011 +0300
mrst: pvr: pass proc info to various IOCTL handlers
Needed by the next patch adding tracing to these commands.
Signed-off-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Pauli Nieminen <pauli.nieminen@intel.com>
commit
71249e1b27c430b522cd4919af413bc136356d91
Author: Imre Deak <imre.deak@intel.com>
Date: Thu Sep 1 18:01:23 2011 +0300
mrst: pvr: add command tracing support
Add a lightweight tracer to track commands submitted by clients. This
can help for example to debug dead-lock situations where command
synchronization counters form a circular dependency.
Signed-off-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Pauli Nieminen <pauli.nieminen@intel.com>
commit
447184e302d7efc0577fc16b181203bbb2f6e7f8
Author: Imre Deak <imre.deak@intel.com>
Date: Thu Sep 1 18:01:22 2011 +0300
mrst: pvr: include missing header files
Signed-off-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Pauli Nieminen <pauli.nieminen@intel.com>
commit
129c16270bb4c2190d2a15c77fb61aec984a5454
Author: Imre Deak <imre.deak@intel.com>
Date: Thu Sep 1 18:01:21 2011 +0300
mrst: pvr: get proc name during process attach time
This will be needed by the upcoming patches where we need a cheap way
to get to the current process name.
Signed-off-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Pauli Nieminen <pauli.nieminen@intel.com>
commit
77bf123bd4a87cb6355133037c25601200b2fc60
Author: Jani Nikula <jani.nikula@intel.com>
Date: Fri Aug 26 14:56:27 2011 +0300
GFX: Replace out-of-range udelay() with mdelay()
Fix build error:
ERROR: "__bad_udelay" [drivers/staging/mrst/medfield/medfield_gfx.ko] undefined!
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
f7fdba671596913b63fd4f11f1b88c71425c50ae
Author: Jani Nikula <jani.nikula@intel.com>
Date: Wed Aug 24 12:49:53 2011 +0300
GFX/PVR: Clean up function and variable declarations
Make more functions and variables static, and remove a few unused
variables. Move most extern declarations scattered around inline in .c
files to .h files, and add proper #includes. This helps tracking the
dependencies between modules, and also fixes a number of sparse warnings.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
ae6010f7d2694bde629955de56c74e8c4287e1ce
Author: Jani Nikula <jani.nikula@intel.com>
Date: Tue Aug 23 16:08:18 2011 +0300
PVR: Delete unused header file sysirq.h
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
f7d2693a454cc6d393d1e0691fce060d84b065b0
Author: Jani Nikula <jani.nikula@intel.com>
Date: Tue Aug 23 16:07:20 2011 +0300
GFX: Remove duplicate declarations for psb_irq.c functions
Declare functions in psb_irq.c only in psb_irq.h, not also psb_drv.h.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
38051b54f4093d0fac3f188770feb1006203cd85
Author: Jani Nikula <jani.nikula@intel.com>
Date: Tue Aug 23 15:21:37 2011 +0300
GFX: Delete unused files psb_intel_dsi2.c and psb_setup.c
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
2c8cfd18094037e23698e20d2059de14f4b55cc0
Author: Jani Nikula <jani.nikula@intel.com>
Date: Tue Aug 23 10:12:42 2011 +0300
GFX: Reserve mdfld_dsi_send_dcs() for mem write command only
mdfld_dsi_send_dcs() is mostly about handling mem write as a special case,
passing other commands to other functions. Only use the function for mem
write.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
7f2e39a3b86a73bc7b0125a891d360658da41d10
Author: Jani Nikula <jani.nikula@intel.com>
Date: Tue Aug 23 09:37:17 2011 +0300
GFX: Pass DSI command data as data and length instead of struct
Pass DSI command data as data and length, and go on with the mdfld_dsi_pkg
struct removal.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
f57208f9d8094b1cc6fd248588d8a9fbad4e0288
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Aug 22 18:06:20 2011 +0300
GFX: Clean up the low level DSI sender functions send_{short,long}_pkg()
Also keep removing the mdfld_dsi_pkg struct from DSI sending.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
84e21960c8fa8d3ba499f0394c8df754db4ee0c2
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Aug 22 17:04:09 2011 +0300
GFX: pkg_type is really the MIPI DSI data_type, rename it accordingly
The data_type is used in registers, but the definitions come directly from
MIPI DSI specs. Rename the enums accordingly. Also drop the unused and
non-existent DCS packet type.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
2f95aa967f0a96ce54943d6bdffc73acb9ebd128
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Aug 22 11:52:14 2011 +0300
GFX: Make transmission mode a bool parameter to simplify code
Make transmission mode a bool parameter to simplify code. This is also in
preparation for further simplifications later.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
f0d40649f1e595ee6017e8b38aa36b15d057e107
Author: Jani Nikula <jani.nikula@intel.com>
Date: Fri Aug 19 13:55:58 2011 +0300
GFX: Remove queued DSI command sending
The queued DSI command sending was overly complex, buggy (for example
keeping pointers to past stack frames or deallocated data), limited (number
of commands in queue, data size in command), and it was only used in a
couple of places. Simply remove it, and send all DSI commands immediately.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
a4b4942508f634fd8bd13165d46d4d6db142ad4a
Author: Jani Nikula <jani.nikula@intel.com>
Date: Fri Aug 19 12:43:22 2011 +0300
GFX: Hide dsi_pkg_sender internal structs and enums
Move some structs and enums from .h to .c. This makes later refactoring of
mdfld_dsi_pkg_sender.c easier.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
5c44bbec1a1768791a740c7733b3e4689c2bff7e
Author: Jani Nikula <jani.nikula@intel.com>
Date: Thu Aug 18 14:25:52 2011 +0300
GFX: Remove a useless abstraction layer in DSI command sending
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
e666aef44b106bae04fb65f7e068665bb3079882
Author: Jani Nikula <jani.nikula@intel.com>
Date: Thu Aug 18 12:58:03 2011 +0300
GFX: Remove an abstraction layer for some DSI commands
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
ef071cdccaafeed42229af9e6265c7867e84ffe2
Author: Jani Nikula <jani.nikula@intel.com>
Date: Thu Aug 18 12:38:32 2011 +0300
GFX: Refactor: Make DSI HS/LP a parameter rather than an abstraction layer
This is a start towards a more sensible DSI command interface.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
affbb07de18e381824e722c1794e07c7acb3c8cf
Author: Jani Nikula <jani.nikula@intel.com>
Date: Wed Aug 17 16:27:19 2011 +0300
GFX: Fix a number of sparse warnings
Fix issues found by sparse: Make functions and variables static where
possible. Use NULL rather than 0 for pointers. Use __iomem to reduce
address space warnings. Remove some unused variables and functions. Fix
locking imbalance. Use unsigned types for 1-bit bitfields.
Note: Only sparse warnings are fixed in the changed code, which means this
will produce checkpatch warnings.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
2d70898399fbf8cf515a2a93bd46c69265a5ff84
Author: Jani Nikula <jani.nikula@intel.com>
Date: Wed Aug 17 16:30:41 2011 +0300
GFX: Do not include .c file from another
Simply append psb_intel_display2.c to psb_intel_display.c rather than
include it.
Note: This patch generates lots of checkpatch warnings, because the
included file is inserted as-is, without modification.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
f1c8ab0bfbeff46fe6a1cea3c5afb7f25265d8db
Author: Jani Nikula <jani.nikula@intel.com>
Date: Tue Aug 16 13:07:41 2011 +0300
GFX: Quick fix to silence build warnings about unused procfs functions
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
9e82f9fcbf6f9fb508c2c090c5a49813da5a015e
Author: Jani Nikula <jani.nikula@intel.com>
Date: Tue Aug 16 13:04:06 2011 +0300
GFX: Quick fix to silence build warnings about unused variables
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
5408ea71f8a7bdae8d03089d48986746dd57b272
Author: Jani Nikula <jani.nikula@intel.com>
Date: Tue Aug 16 13:01:21 2011 +0300
GFX: Fix build warning about unused backlight related function and variables
Add #ifdef CONFIG_BACKLIGHT_CLASS_DEVICE around the function and variables.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
abfe9e60cf2acc0cde07dcbe9229219c90f6113a
Author: Jani Nikula <jani.nikula@intel.com>
Date: Tue Aug 16 12:55:25 2011 +0300
GFX: Fix printk format warnings
Reduce build noise by fixing printk format warnings.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
6906b5a242fe460827f0b43f0e270d8461ac02dd
Author: Jani Nikula <jani@nikula.org>
Date: Tue Aug 16 10:56:18 2011 +0300
PVR: Fix PVR build so that it doesn't get rebuilt every time
Use symlinks instead of copies of the source files, and preserve them using
.PRECIOUS to 1) build the driver so that it doesn't get rebuilt every time,
and 2) produce useful paths in the build log that point to the actual
source files through the symlinks. Add the intermediate symlinks to
.gitignore.
This is a slight improvement to the original hack of building more than one
kernel module from the same sources with different config options and
defines. It should be fixed later, but this makes life easier now.
Also make the Medfield and Moorestown Makefiles as similar to each other as
possible, to make later unification easier.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
bb926aae8fbb0cfb98f387e504337ac3c01147d0
Author: Pauli Nieminen <pauli.nieminen@intel.com>
Date: Tue Aug 16 14:02:40 2011 +0300
pvr: Remove timer hack from vblank handler
The timer hack isn't required after userspace fixes to flip
implementation.
Signed-off-by: Pauli Nieminen <pauli.nieminen@intel.com>
commit
6ccfe7bacbffc1900add826fb1f4efa0bc40702a
Author: Pauli Nieminen <pauli.nieminen@intel.com>
Date: Thu Aug 11 09:31:21 2011 +0300
pvr: Fix device poll implementation
Driver always signaling that there is data available misleads display
server to call blocking read. Blocked display server makes whole
system unuseable.
drm_poll checks if there is queued events to userspace. Replacing
psb_poll with drm_poll fixes the poll implementation.
Signed-off-by: Pauli Nieminen <pauli.nieminen@intel.com>
commit
5fef5630c4fddd4c775c91c0ba11c468af001cc2
Author: Alan Cox <alan@linux.intel.com>
Date: Fri Aug 19 14:05:32 2011 +0100
gfx: Fix build failure of graphics when compiling for MRST
The graphics code tries to use RAR headers which are not where it expects. We
don't use the RAR feature or care about it for the this GFX driver so throw
it out.
Signed-off-by: Alan Cox <alan@linux.intel.com>
commit
803ff08d23f4f0164f24f664c23ce772724f8722
Author: Pauli Nieminen <pauli.nieminen@intel.com>
Date: Fri Aug 19 14:05:17 2011 +0100
Backlight type was missing from properties which caused warning in
register.
Reported-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Signed-off-by: Pauli Nieminen <pauli.nieminen@intel.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
commit
ce064800f88e9731c1d1333fbccc4a05e8e4fc68
Author: Pauli Nieminen <pauli.nieminen@intel.com>
Date: Fri Aug 19 14:03:20 2011 +0100
pvr: Fix out of array access
PVR driver defines its own connector and encoder types for display which
aren't suppported by drm code. When converting the type to string kernel
would crash or output random junk because conversion table doesn't have
entry for MIPI.
[ 19.455764] BUG: unable to handle kernel paging request at
65646f6d
[ 19.455779] IP: [<
c127e0c9>] strnlen+0x9/0x20
[ 19.455802] *pde =
00000000
[ 19.455812] Oops: 0000 [#1] PREEMPT SMP
[ 19.455825] Modules linked in: medfield_gfx(C+) fbcon font bitblit
softcursor ttm drm_kms_helper drm fb fbdev
+cfbcopyarea cfbimgblt cfbfillrect
[ 19.455860]
[ 19.455871] Pid: 129, comm: modprobe Tainted: G C 3.0.0+ #30
[ 19.455884] EIP: 0060:[<
c127e0c9>] EFLAGS:
00010297 CPU: 0
[ 19.455895] EIP is at strnlen+0x9/0x20
[ 19.455903] EAX:
65646f6d EBX:
f86ccd60 ECX:
65646f6d EDX:
fffffffe
[ 19.455912] ESI:
f86ccd80 EDI:
0000ffff EBP:
f5b596fc ESP:
f5b596fc
[ 19.455922] DS: 007b ES: 007b FS: 00d8 GS: 0033 SS: 0068
[ 19.455932] Process modprobe (pid: 129, ti=
f5b58000 task=
f6312be0 task.ti=
f5b58000)
[ 19.455940] Stack:
[ 19.455945]
f5b5971c c127bb3f f999f67b f86c8b25 65646f6d f5b59784 f86ccd60 00000002
[ 19.455965]
f5b59760 c127cad6 00070100 0006ffff 0006ffff 00070190 00060010 0006000c
[ 19.455985]
00000020 f86ccd60 f86ccd80 f86c8b27 ff0a0004 ffffffff 0000000b f63b44f8
[ 19.456005] Call Trace:
[ 19.456021] [<
c127bb3f>] string.isra.4+0x2f/0xb0
[ 19.456133] [<
f999f67b>] ? mdfld_dsi_gen_fifo_ready+0x3b/0xb0 [medfield_gfx]
[ 19.456151] [<
c127cad6>] vsnprintf+0x1e6/0x3a0
[ 19.456166] [<
c127ccea>] snprintf+0x1a/0x20
[ 19.456210] [<
f86bb287>] drm_get_encoder_name+0x37/0x40 [drm]
[ 19.456231] [<
f876fac5>] drm_crtc_helper_set_mode+0x405/0x470 [drm_kms_helper]
MIPI (or DSI) is protocol specification on top of LVDS serial bus. That
makes it resonable to call MIPI connectors and encoders LVDS.
NOT FOR UPSTREAM
Signed-off-by: Pauli Nieminen <pauli.nieminen@intel.com>
Reported-and-Tested-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
commit
e6a9e72108f5444b754d68dd60328e0a8740ab36
Author: Pauli Nieminen <pauli.nieminen@intel.com>
Date: Fri Jul 8 11:04:57 2011 +0300
configs: Update medfield config to have touch&keypad support
Signed-off-by: Pauli Nieminen <pauli.nieminen@intel.com>
commit
27a326770da336624834f101ff889421dae1b007
Author: Jani Nikula <jani.nikula@intel.com>
Date: Tue Jul 5 11:19:08 2011 +0300
mrst: Add drivers/staging/mrst/README documenting the forward port
Document where the driver is coming from and what was done to make it work
against linux-3.0-mid-ref.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
c8355336d838d0e19135783dbf0bb60504bcbd14
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Jul 4 17:46:02 2011 +0300
mrst: Add arch/x86/configs/medfield_pvr_defconfig
Add a known good defconfig.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Pauli Nieminen <pauli.nieminen@intel.com>
commit
c5a69015cba9486adab91a987d13ed8698db8368
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Jul 4 15:14:23 2011 +0300
drm: Protect drm/drm_os_linux.h inclusion with #ifdef
Fix problems caused by including both drm/drmP.h and drm/drm_os_linux.h.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
353c55ac88cf9525fea65d797723c5538967915c
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Jul 4 14:51:43 2011 +0300
drv: Add DRM_MODE_CONNECTOR_MIPI definition
REVISIT: This definition and value 15 are from the linux-3.0-mid-ref
tree. psb_intel_dsi.c also has a commented out definition with value
13. The definition should be added to include/drm/drm_mode.h eventually.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
37a03b495ca98bee1adcb391e47a411d5aa82e77
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Jul 4 14:44:07 2011 +0300
drv: Remove psb_validate_kernel_buffer() and val_seq
The function does not compile, is not used, and is the only user of
val_seq, after ttm_eu_reserve_buffers() has switched to handling the seq
internally.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
11ef8154687d734ba0d4d4238c6101e826de1b9e
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Jul 4 14:26:47 2011 +0300
drv: Use bdev->fence_lock for locking
The lock field has been dropped from struct ttm_buffer_object. Use the
bdev->fence_lock of the struct for locking instead.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
d19a6a3dc015ff165f79a286fd5ba416d992f81e
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Jul 4 14:15:36 2011 +0300
drv: Add force parameter to struct drm_connector_funcs .detect functions
An additional force parameter has been added to the .detect function in
struct drm_connector_funcs. Add the parameter to the functions
dsi_connector_detect, mdfld_dsi_connector_detect, mdfld_hdmi_detect,
mrst_dsi_detect, psb_intel_lvds_detect, and psb_intel_sdvo_detect.
FIXME: Start actually using the force parameter.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
52ebe11a7cf8c0ca37bef03c0cc411d63bfcfc84
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Jul 4 14:12:54 2011 +0300
drv: Add start parameter to psb_intel_crtc_gamma_set()
The start parameter has been added to struct drm_crtc_funcs .gamma_set
function, so add it to psb_intel_crtc_gamma_set() also.
REVISIT: Start using the start and size parameters in the function.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
b8a0cd91217e0395ca402e3a6a4c34ae2d233494
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Jul 4 14:01:04 2011 +0300
drv: Remove initializations that have been removed from struct drm_driver
REVISIT: The procfs stuff should be ported to sysfs/debugfs.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
d82d658906ca20cff1758f776f414fb48897fcc9
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Jul 4 13:57:28 2011 +0300
drv: Separate the drm driver and pci driver structs
The pci_driver struct is no longer part of struct drm_driver.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
266d1b7ace965b1203f7de738568fc325d1dd77f
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Jul 4 13:11:27 2011 +0300
drv: Replace acquire/release_console_sem with console_lock/unlock
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
d3301724bfd035c91d5c561d1f38c2753835526f
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Jul 4 13:03:54 2011 +0300
drv: Remove #include <linux/i2c-id.h>
The linux/i2c-id.h file has been removed.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
4613ef2bd9389b10779cdde1b2239f26d727604c
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Jul 4 13:23:14 2011 +0300
drv: Remove an informational message
The code block would not compile, because struct ttm_mem_type_manager no
longer has manager field. Instead of resolving the issue, just remove the
code block because it only prints an informational message if "GATT was not
clean after VT switch".
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
1546b44a62af1eb1dd72f14be2736a049b5101cc
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Jul 4 13:20:45 2011 +0300
HACK: drv: Do not use BKL for drm_ioctl() locking
The BKL no longer exists, do not use it. This is a hack because the BKL is
not replaced with any other locking.
REVISIT: Need to check the locking.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
55218c31dfe2a3e597df144ab4007f5fae473928
Author: Pauli Nieminen <pauli.nieminen@intel.com>
Date: Mon Jul 4 22:16:37 2011 +0300
pvr: Add bo manager functions to custom memory managers
FIXME: Check if pvr need custom memory manager like nouveau
Signed-off-by: Pauli Nieminen <pauli.nieminen@intel.com>
commit
b8ed80804204bf769d8b8fe6f4681f92e56aea38
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Jul 4 15:46:56 2011 +0300
pvr: Replace ttm_global_reference with drm_global_reference
struct ttm_global_reference has been dropped in favor of struct
drm_global_reference.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
47818d739cf33acc766dd7af7a9eeb20f7a3c6aa
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Jul 4 15:17:54 2011 +0300
pvr: Use bdev->fence_lock for locking
The lock field has been dropped from struct ttm_buffer_object. Use the
bdev->fence_lock of the struct for locking instead.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
5f42477b7df47850047008bd2848f775dbb1fee3
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Jul 4 15:36:52 2011 +0300
pvr: Add dma_addrs parameter to drm_psb_tbe_populate
The dma_addrs parameter has been added to struct ttm_backend_func .populate
function. Add it to drm_psb_tbe_populate also.
REVISIT: Is there a need to use the dma_addrs parameter?
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
838f8a6f87ba90aa3489a0ebd83c06bdc4a7db9c
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Jul 4 15:34:01 2011 +0300
pvr: Use the start field of struct ttm_mem_reg
The mm_node field of struct ttm_mem_reg has become a void*, and can no
longer be referenced.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
e81fb9c9f888ff477034b6dd63f18c5bb386d5d1
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Jul 4 13:12:13 2011 +0300
pvr: Replace acquire/release_console_sem with console_lock/unlock
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
b23657906cf1d8cd4c083d69021f9d69169b59ab
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Jul 4 12:06:03 2011 +0300
pvr: Use DEFINE_SPINLOCK() instead of SPIN_LOCK_UNLOCKED
SPIN_LOCK_UNLOCKED has been removed.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
581bb8e42869ebce5e33e13262cb5e54d22a830d
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Jul 4 11:59:22 2011 +0300
pvr: Fix calls to pci_restore_state()
Fix calls to pci_restore_state(), which no longer returns a value.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
38896d85369f49b2f8ca106d97708947df925bb0
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Jul 4 11:48:23 2011 +0300
pvr: Remove references to AUTOCONF_INCLUDED and linux/config.h
AUTOCONF_INCLUDED is no longer defined, and linux/config.h no longer
exists.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
fa229e9d85048d19a3b2de19b66d3a87eab8a2ca
Author: Jani Nikula <jani.nikula@intel.com>
Date: Mon Jul 4 11:47:35 2011 +0300
mrst: Remove local versions of drm_global.{h,c}
The mrst fork looks incorrect.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Pauli Nieminen <pauli.nieminen@intel.com>
commit
f173452bed6a7384d258b11658ec55afe87fb71c
Author: Jani Nikula <jani.nikula@intel.com>
Date: Fri Jul 1 12:43:00 2011 +0300
staging: Add mrst graphics driver to Kconfig and Makefile
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
commit
0aae9776ad86ab77dfe67cf549db78667a46ee62
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Wed Jun 22 13:17:04 2011 -0700
GFX: Graphics Driver build release 5.3.0.0047
Weekly Graphics driver release.
Please use usermode package 0047 with this build.
Signed-off-by: Hitesh K. Patel <hitesh.k.patel@intel.com>
commit
de023b4f5dbbc6bdd5a4184c561727b2b2ac2df4
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Fri Jun 17 10:00:52 2011 -0700
GFX: Enable Display Island for DPST
When enabling DPST by default we need to make sure that
display island is enabled to get DPST interrupt generated.
Signed-off-by: Hitesh K. Patel <hitesh.k.patel@intel.com>
commit
35cc686f6a56cb1ee959a9ff68d557f9d19ac5a9
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Thu Jun 16 12:31:04 2011 -0700
GFX: Allow multiple registers dump
Allow multiple registers to be dumped from procsys.
Signed-off-by: Hitesh K. Patel <hitesh.k.patel@intel.com>
commit
f5b3a4c10c0a385471b8b65bede0509681a06626
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Fri Jun 10 11:42:42 2011 -0700
GFX: Fix invalid return when creat_proc_init fails
when init failed for any of the procfs creation the failure
would be returned. Now it just makes sure that error is printed
and continued.
Signed-off-by: Hitesh K. Patel <hitesh.k.patel@intel.com>
commit
75ede86757ffc57c2503b615b4172160b7284364
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Mon Jun 20 15:50:19 2011 -0700
kernel rebase kernel-adaptation-mrst-2.6.37.6-102.1
commit
094854b8292dba86e57248f686efe35b8de425cd
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Fri Jun 17 10:00:52 2011 -0700
GFX: Enable Display Island for DPST
When enabling DPST by default we need to make sure that
display island is enabled to get DPST interrupt generated.
Signed-off-by: Hitesh K. Patel <hitesh.k.patel@intel.com>
commit
c92e91c54ab7d6d5e5d4c1df1156fef4088c8943
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Thu Jun 16 12:31:04 2011 -0700
GFX: Allow multiple registers dump
Allow multiple registers to be dumped from procsys.
Signed-off-by: Hitesh K. Patel <hitesh.k.patel@intel.com>
commit
a52a45c44aa5a807f8733abbdd8cf55c16b676f1
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Fri Jun 10 11:42:42 2011 -0700
GFX: Fix invalid return when creat_proc_init fails
when init failed for any of the procfs creation the failure
would be returned. Now it just makes sure that error is printed
and continued.
Signed-off-by: Hitesh K. Patel <hitesh.k.patel@intel.com>
commit
d25fe71cd95a6fe761315e882605b0e74f466eec
Author: Pauli Nieminen <pauli.nieminen@intel.com>
Date: Thu Jun 9 11:03:20 2011 +0300
moorestown/pvr: Enable memory debug only in debug build
Memory debug flags have huge runtime footprint which can be seen
as one of larges CPU users in system. In my QT/demos/declarative/snake
profile cost was about 15% CPU time.
Actual test case is GPU limited so optimization doesn't give fps boost
but it maybe meaningfull optimization for applciation startup and
shutdown times.
Also fixing incorrectly typed DEBUG_LINUX_MEM_AREAS to actually enable
the feature.
Signed-off-by: Pauli Nieminen <pauli.nieminen@intel.com>
commit
0d13a11049f84fa86d8029cb69baead764ccfa59
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Thu Jun 9 09:02:49 2011 -0700
GFX: Graphics Driver build release 5.3.0.0046
Weekly Graphics driver release.
Please use usermode package 0046 with this build.
Signed-off-by: Hitesh K. Patel <hitesh.k.patel@intel.com>
commit
afb5c438c374397df4d3136d1a7c24850daa0f5f
Author: Ren, Zhaohan <zhaohan.ren@intel.com>
Date: Wed May 11 15:56:48 2011 +0800
Update psb_drm.h
commit
63ae47109df731c117a61ce523cf5473764f2f69
Author: Daemon Anastas <daemonx.g.anastas@intel.com>
Date: Sat May 28 15:08:46 2011 -0700
revert back to a 19ms display DMA delay for SwapChain buffers.
commit
96f710fdc9c4244562239af87571376c959598b7
Author: Daemon Anastas <daemonx.g.anastas@intel.com>
Date: Wed Jun 1 11:09:03 2011 -0700
Change DMA delay for when we have a swapchain to 3ms. This works well and pushs the image out without a delay.
commit
e9205fb80d61a280812e96507e61ff29c538524a
Author: Daemon Anastas <daemonx.g.anastas@intel.com>
Date: Tue May 31 18:05:55 2011 -0700
Implement a different Display Buffer DMA delay from the VSYNC interrupt. This delay for the 60hz display refresh rate is 18ms for a QT rendered display and 11ms for a GLES2 rendered display.
commit
4106ee3e3a5d348868f5bcf7d3d3a8b6e2e83755
Author: Peter Zick <peterz@peterz-VirtualBox>
Date: Tue May 31 13:10:15 2011 -0700
Remove printk debug message from ospm_suspend_display.
commit
81022bb3d5da8696084fbc85949b4f35bfca78b1
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Thu May 19 16:11:37 2011 -0700
GFX: Graphics Driver build release 5.3.0.0045
Weekly Graphics driver release.
Please use usermode package 0045 with this build.
Signed-off-by: Hitesh K. Patel <hitesh.k.patel@intel.com>
commit
967dbe161d67b94f95bf7f2149ed07f1906a5803
Author: Peter Zick <peterx.zick@intel.com>
Date: Fri May 27 09:24:05 2011 -0700
GFX: Support graphics & Display island power gating.
added to support the Moorestown configuration of the
graphics and display islands power gating.
Signed-off-by: Peter Zick <peterx.zick@intel.com>
commit
14693c1032a9abb43c42d8d6c3240549c52a58b0
Author: Daemon Anastas <daemonx.g.anastas@intel.com>
Date: Fri May 20 16:04:58 2011 -0700
GFX: Change the jiffies use to proper function
Use msec_to_jiffies() to insure an 11ms delay for the
Display Buffer DMA timer. Also set the display buffer DMA delay
to 18ms. This is to optimized for QT/MeeGo-ux.
Signed-off-by: Daemon Anastas <daemonx.g.anastas@intel.com>
commit
ce09e906f985b0894f614943cacf533839871b83
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Fri May 27 15:42:48 2011 -0700
kernel rebase to kernel-adaptation-mrst-2.6.37.6-54.2
commit
48a951768628047cd317603268c9410f62be9cb4
Author: Peter Zick <peterx.zick@intel.com>
Date: Fri May 27 09:24:05 2011 -0700
Changes to psb_powermgmt.c added to support the Moorestown configuration of the graphics and display islands power gating.
commit
d631df0a967b002114ae0da22c885d97b24451fe
Author: Daemon Anastas <daemonx.g.anastas@intel.com>
Date: Wed May 25 12:52:32 2011 -0700
Set the display buffer DMA delay offset to 18ms. This is optimized for QT/MEEGO-UX.
commit
f85694cb7d2a8d9d6e06857359a173af43639170
Author: Daemon Anastas <daemonx.g.anastas@intel.com>
Date: Fri May 20 16:04:58 2011 -0700
Use msec_to_jiffies() to insure an 11ms delay for the Display Buffer DMA timer.
commit
56c41edf7ee26b2f4f4aeb48b869e781253f149d
Author: Li Zeng <li.zeng@intel.com>
Date: Thu May 12 13:16:57 2011 +0800
ved: Decode hw sequence value is initilise to 1
BZ: 1848
After reboot, Decode hw sequence and sw sequence are both 0, so sync buffer always success even if it is not really ready.
HW sequence value is initilise to 1 to avoid such issue
Change-Id: I8886a67d70b412e147e3b99453442bf57490dc7d
Signed-off-by: Li Zeng <li.zeng@intel.com>
commit
d705c82d06bee78d767379bdf3913eaabdfc6725
Author: Austin Yuan <shengquan.yuan@gmail.com>
Date: Tue May 17 13:27:52 2011 +0800
Video: optimized udelay usage
1) From Liu, Haiyang: there is a call path psb_setup_fw() -> psb_wait_for_register()
-> udelay(1000), which is not C0 friendly, reduce it to 1
2) export three runtime options for further optimization/tuning
-Export option "cpu_relax" to replace all "udelay" to "cpu_relax" at runtime
-Export option "udelay_multiplier" to multiple the udealy usec with a value (increase)
-Export option "udelay_divider" to divide the udelay usec with a value (decrease)
Signed-off-by: Austin Yuan <shengquan.yuan@gmail.com>
commit
09c31a9d4a4fb604ad7b5e798765bf70c5960bc7
Author: Daemon Anastas <daemonx.g.anastas@intel.com>
Date: Wed May 11 13:19:43 2011 -0700
Update the delay for DMA'ing the display buffer from VSync to 11ms (from 15ms). This is to account for the display frequency change of 48.8hz to 60hz.
commit
1ee4119468c1949cb6fbf931c5bdc24c83a5ba4d
Author: Ren, Zhaohan <zhaohan.ren@intel.com>
Date: Wed May 11 15:56:48 2011 +0800
Update psb_drm.h
commit
0e03d0f66fcec154a24b65e25726e13052d3fe4f
Author: Elaine Wang <elaine.wang@intel.com>
Date: Wed May 4 15:42:29 2011 +0800
Video:Fix Topaz D0i3 hang issue.
BZ: 1478
After enalbe Topaz D0i3 manually by "echo 0x2 > /sys/module/medfield_gfx/parameters/topaz_pmpolicy", video encoding performance is very low and the en
coded clip is corruptted. See error message "TOPAZ timeout (probable lockup) detected" in terminal.
This patch fix several potential problems in Topaz save and restore MTX state functions. Save the content of Topaz communication registers before power down Topaz and restore it after resume. Before power down Topaz, poll MMU control register until make sure there is no pending memory request. Reset MVEA twice according to IMG's advice. After transfer firmware to Topaz, kick NULL command to Topaz core 0 and core 1 and wait for its completion. If polling the writeback of NULL commands times out, read back 4 words from command FIFO to recover the state of it.
Inserting NULL commands would cause extra delay of resuming Topaz and reading back 4 words from command FIFO doesn't guarantee successfuly recovery everytime, but it makes video encoding much more stable while enter and exit D0i3. It will be removed after IMG provide a total fix in Topaz firmware.
Signed-off-by: Elaine Wang <elaine.wang@intel.com>
commit
75a8b0604ff47effc895a5279965c2b5fc06a6db
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Thu May 5 09:09:17 2011 -0700
Driver Package release 0043
commit
813601f6e4561dcea1bf914f02917b5e4ff57842
Author: Ren Zhaohan <zhaohan.ren@intel.com>
Date: Thu Apr 28 09:42:53 2011 -0400
GFX: Removed buffer 4 bytes alignment
This restriction isn't needed. This will allow creating buffer of any size.
commit
97f81d82513bda7f05593dd900ef009b4afbd8dd
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Tue May 3 14:06:11 2011 -0700
GFX: add platform detect support
refactor the code that checking MM system and Oaktrail platform
put the dmi match code in psb_platform_detect(), add helper macros
like IS_MRST_MM and IS_OAKTRAIL
Signed-off-by: Li Peng <peng.li@intel.com>
commit
1017a004a954a3a416a0e39d7466ed7c4443f01b
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Tue May 3 13:56:09 2011 -0700
GFX: use raw_edid on mrst tablet
currently the modeline used on mrst tablet is retrieved from GCT
table in which the vertical refresh rate is about 49 HZ, but it isn't
seem to be the best modeline because the EDID probed from LVDS panel
recommending 60 HZ vrefresh rate.
Also the h/w design on mrst tablet requires i2c0 as the channel for
EDID probe, fixing this by adding DMI_MATCH code for mrst tablet.
Another issue is that the EDID checksum on mrst tablet is bad, which block
driver doing further modeline parse, this maybe a problem of LVDS panel,
add workaround to use raw_edid on mrst tablet.
Signed-off-by: Li Peng <peng.li@intel.com>
commit
02968b64eba3ce5ac3c6a9d2bc5de3ae93814d45
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Wed Apr 20 17:01:53 2011 -0700
GFX: Graphics Driver build release 5.3.0.0042
Weekly Graphics driver release.
Please use usermode package 0042 with this build.
Signed-off-by: Hitesh K. Patel <hitesh.k.patel@intel.com>
commit
1f64690d7a4e0d4699a3db5d587e5d37e2c70772
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Wed Apr 20 14:33:06 2011 -0700
GFX-Display: Enabled S0i3 for TMD panel.
Now the system could enter and exit S0i3 for several times, with TMD/TPO panel
attached on MDFLD, by the steps:
1. Enable S0i3:
a) echo 12000 > /proc/sys/vm/dirty_writeback_centisecs
b) echo s0i3 > /sys/module/mid_pmu/parameters/s0ix
c) echo 1 >/sys/devices/platform/max3110/adaptive
2. xset dpms force off (or "xset s $TIMEOUT $INTERVAL")
Also fixed the "Fabrice Error" issue when in S0i3 state.
Signed-off-by: Austin Hu <austin.hu@intel.com>
commit
77c3db295d9977b37c9cc4804b39d311ee074eb2
Author: Yu(Alex) Dai <yu.dai@intel.com>
Date: Mon Mar 21 12:09:57 2011 -0700
GFX: Fix memory release issue when X is crashed.
This is an improvement for previous Fix.
Now we add 3-level sanity check for PVRDeviceNode, BMContext and BMHeap.
Signed-off-by: Yu(Alex) Dai <yu.dai@intel.com>
commit
2fd2837ee86bd4edf5b3a27803adaa8bbc928fa5
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Tue Apr 5 14:39:34 2011 -0700
GFX: Removed SUPPORT_EGL_EXTENSIONS
Removing Some EGL Extension support.
Signed-off-by: Hitesh K. Patel <hitesh.k.patel@intel.com>
commit
a0588a08a98795698b9ecc3e52b5429ec04fdacf
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Wed Apr 20 15:35:55 2011 -0700
GFX: Cleanup video codying style
Cleaning up Video driver coding style to match OpenSource
Coding style.
Signed-off-by: Austin Yuan <shengquan.yuan@gmail.com>
commit
465cad2312c487a284e1689f6533813dafc72048
Author: Daemon Anastas <daemonx.g.anastas@intel.com>
Date: Sun Apr 10 12:49:06 2011 -0700
GFX: Fix tearing issue on Moorestown (Bugzilla 5503)
Fix tearing on Moorestown due to race condition of Display moving
buffer to screen on vysnc and driver swapping buffer on vsync.
This is fixed here by using a timer to delay the driver flipping
the display buffer. This change is made to not effect Medfield.
Signed-off-by: Daemon Anastas <daemonx.g.anastas@intel.com>
commit
801029054b4de50d58c4b17abfc68fb270c5574d
Author: Elaine Wang <elaine.wang@intel.com>
Date: Wed Apr 13 15:23:56 2011 +0800
GFX: Fix device firmware loading issue
Fix a code defect that cause loading device firmware repeatedly
in different video encoding sessions.
Use Topaz context handle instead of variable topaz_fw_loaded to check
if any video encoding process is running. When driver removes Topaz
context, mark current MTX data as invalid to prohibit retoring MTX
data wrongly at the beginning of next encoding session.
Signed-off-by: Elaine Wang <elaine.wang@intel.com>
commit
ba38a2fed5bfa6ab988e83f8632ba4dee6eb76c2
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Wed Apr 13 14:34:59 2011 -0700
GFX: Fix to avoid infinite loop in ospm_power_island_up/down.
- Removed redundant code to power on/off GL3 (PSB_PWRGT_GFX_MASK_B0)
- Added counters with debug prints to detect infinite loops
- Added code to handle ISP on/off to be used by camera driver.
- invalidate GL3 before suspend.
- Add in OSPM_ISP_ISLAND check in ospm_power_island_up.
Signed-off-by: Rajesh Poornachandran <rajesh.poornachandran@intel.com>
commit
45e5aa0e53c403df69d4ef6de7927aa1433b7194
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Tue May 3 15:40:04 2011 -0700
kernel rebase kernel-adaptation-mrst-2.6.37.6-38.2
commit
d099ff36cb3385cd139e0ba5882128b68df67c5a
Author: Ren Zhaohan <zhaohan.ren@intel.com>
Date: Thu Apr 28 09:42:53 2011 -0400
Removed buffer 4 bytes alignment restriction to allow creating buffer of any size.
commit
aa1ef8d15b18c0175dc0564601e842e682d5e130
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Fri Apr 22 11:24:37 2011 -0700
GFX: Read EDID from i2c Adaptor #0 for MM
MM Systems have EDID hooked up to i2c adaptor 0.
Detect the MM system and read the EDID.
Signed-off-by: Hitesh K. Patel <hitesh.k.patel@intel.com>
commit
0954d9c2eaf4f1d581a13e37ac2f3353c703d1d0
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Wed Apr 20 17:01:53 2011 -0700
driver release 0042
commit
993db7289b98c5a055ff1f3d3c18d0e3d3b27721
Author: Austin Hu <austin.hu@intel.com>
Date: Mon Apr 18 17:16:23 2011 +0800
GFX-Display: Enabled D0i3 for the Display sub-system of Acer Tablet.
Enable it by checking the current DC status in run-time runtime_suspend()/runtime_idle()
callbacks, rather than checking DSR status which isn't supported by Acer Tablet's LVDS panel.
Enable/Disable run-time PM with pm_runtime_put_noidle()/pm_runtime_get_noresume(), according
to the common run-time PM changes since kernel 37.
And here is the /sys/kernel/debug/mrst_pmu diff before and after DPMS off:
Before:
GFX: [BLOCKS s0i3]
0000:00:02.0 8086/4102 pvrsrvkm PCI-D0 D0 10 0 0 6 0
After:
GFX:
0000:00:02.0 8086/4102 pvrsrvkm PCI-D3hot D0 10 0 0 8 0
Signed-off-by: Austin Hu <austin.hu@intel.com>
commit
df8c26f576b0fbdef4e5be7540ca38475c4eeee2
Author: Austin Yuan <shengquan.yuan@gmail.com>
Date: Fri Apr 15 10:27:56 2011 +0800
Cleanup video codying style
Signed-off-by: Austin Yuan <shengquan.yuan@gmail.com>
commit
1b0ce19f3ad640e10ec5e95bbbcb690ac40aeba4
Author: Daemon Anastas <daemonx.g.anastas@intel.com>
Date: Sun Apr 10 12:49:06 2011 -0700
GFX: Fix tearing issue on Moorestown (Bugzilla 5503)
Fix tearing on Moorestown due to race condition of Display moving
buffer to screen on vysnc and driver swapping buffer on vsync.
This is fixed here by using a timer to delay the driver flipping
the display buffer. This change is made to not effect Medfield.
Signed-off-by: Daemon Anastas <daemonx.g.anastas@intel.com>
commit
3efd32effa12928feb16eb5641262e5b175ac763
Author: Elaine Wang <elaine.wang@intel.com>
Date: Wed Apr 13 15:23:56 2011 +0800
GFX: Fix device firmware loading issue
Fix a code defect that cause loading device firmware repeatedly
in different video encoding sessions.
Use Topaz context handle instead of variable topaz_fw_loaded to check
if any video encoding process is running. When driver removes Topaz
context, mark current MTX data as invalid to prohibit retoring MTX
data wrongly at the beginning of next encoding session.
Signed-off-by: Elaine Wang <elaine.wang@intel.com>
commit
bf9cf397c15cdefb37e5d19dc04b194c77c5909e
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Tue Apr 5 14:39:34 2011 -0700
GFX: Removed SUPPORT_EGL_EXTENSIONS
Removing Some EGL Extension support.
Signed-off-by: Hitesh K. Patel <hitesh.k.patel@intel.com>
commit
1e90ac03aa57a1b6a2dcf3fb2d958150f5223de0
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Wed Apr 13 14:34:59 2011 -0700
GFX: Fix to avoid infinite loop in ospm_power_island_up/down.
- Removed redundant code to power on/off GL3 (PSB_PWRGT_GFX_MASK_B0)
- Added counters with debug prints to detect infinite loops
- Added code to handle ISP on/off to be used by camera driver.
- invalidate GL3 before suspend.
- Add in OSPM_ISP_ISLAND check in ospm_power_island_up.
Signed-off-by: Rajesh Poornachandran <rajesh.poornachandran@intel.com>
commit
efe1620f8beb9e7c3c9b2c22792c2fc891fbbf3c
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Thu Apr 14 11:24:46 2011 -0700
Finally rebase to kernel-adaptation-mrst-2.6.37.6-26.1
commit
6ce9100dfde3e3e68764a5ccba9c2bbb590aede7
Author: Daemon Anastas <daemonx.g.anastas@intel.com>
Date: Sun Apr 10 12:49:06 2011 -0700
GFX: Fix tearing issue on Moorestown (Bugzilla 5503)
Fix tearing on Moorestown due to race condition of Display moving
buffer to screen on vysnc and driver swapping buffer on vsync.
This is fixed here by using a timer to delay the driver flipping
the display buffer. This change is made to not effect Medfield.
Signed-off-by: Daemon Anastas <daemonx.g.anastas@intel.com>
commit
b0a6a0a3cd76de8b32dcc341b36ad220a37d8498
Author: Elaine Wang <elaine.wang@intel.com>
Date: Wed Apr 13 15:23:56 2011 +0800
GFX: Fix device firmware loading issue
Fix a code defect that cause loading device firmware repeatedly
in different video encoding sessions.
Use Topaz context handle instead of variable topaz_fw_loaded to check
if any video encoding process is running. When driver removes Topaz
context, mark current MTX data as invalid to prohibit retoring MTX
data wrongly at the beginning of next encoding session.
Signed-off-by: Elaine Wang <elaine.wang@intel.com>
commit
f5541df0bf689ac4488b2b2ab5cfb18120023fd5
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Tue Apr 5 14:39:34 2011 -0700
GFX: Removed SUPPORT_EGL_EXTENSIONS
Removing Some EGL Extension support.
Signed-off-by: Hitesh K. Patel <hitesh.k.patel@intel.com>
commit
5fb7d6107bf4a2b3903d45e7431fbe7b5c242d50
Author: Li Peng <peng.li@intel.com>
Date: Thu Mar 31 11:22:33 2011 +0800
mrst gfx: Oaktrail HDMI enabling patch
[v3] more bugfixes in mode setting test
[v2] adding audio support.
[v1] video and i2c implementation.
Signed-off-by: Li Peng <peng.li@intel.com>
Signed-off-by: Li Peng <peng.li@linux.intel.com>
commit
38d8c9c05c5477d907f9a8f1b6f9dfe615d4b58f
Author: Li Peng <peng.li@intel.com>
Date: Thu Mar 31 11:16:47 2011 +0800
mrst gfx: change pipe number for oaktrail
oaktrail platform (a variant of mrst) has two pipes, change
pipe number to two.
Signed-off-by: Li Peng <peng.li@intel.com>
Signed-off-by: Li Peng <peng.li@linux.intel.com>
commit
b88f9845e65b9dd0f2f7450795e1085918d2ade5
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Wed Apr 13 14:34:59 2011 -0700
GFX: Fix to avoid infinite loop in ospm_power_island_up/down.
- Removed redundant code to power on/off GL3 (PSB_PWRGT_GFX_MASK_B0)
- Added counters with debug prints to detect infinite loops
- Added code to handle ISP on/off to be used by camera driver.
- invalidate GL3 before suspend.
- Add in OSPM_ISP_ISLAND check in ospm_power_island_up.
Signed-off-by: Rajesh Poornachandran <rajesh.poornachandran@intel.com>
commit
fa347620e74ec7aa76e96506514c51ec23ef7cfd
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Wed Apr 13 16:34:21 2011 -0700
rebase kernel-adaptation-mrst-2.6.37.6-26.1
commit
fe87b34dd827d2970dc0bb5e0eaf5b103ae9ca02
Author: Daemon Anastas <daemonx.g.anastas@intel.com>
Date: Sun Apr 10 12:49:06 2011 -0700
GFX: Fix tearing issue on Moorestown (Bugzilla 5503)
Fix tearing on Moorestown due to race condition of Display moving
buffer to screen on vysnc and driver swapping buffer on vsync.
This is fixed here by using a timer to delay the driver flipping
the display buffer. This change is made to not effect Medfield.
Signed-off-by: Daemon Anastas <daemonx.g.anastas@intel.com>
commit
5f44890f95b1e28e441c7c7963e6706e7f11a881
Author: Elaine Wang <elaine.wang@intel.com>
Date: Wed Apr 13 15:23:56 2011 +0800
GFX: Fix device firmware loading issue
Fix a code defect that cause loading device firmware repeatedly
in different video encoding sessions.
Use Topaz context handle instead of variable topaz_fw_loaded to check
if any video encoding process is running. When driver removes Topaz
context, mark current MTX data as invalid to prohibit retoring MTX
data wrongly at the beginning of next encoding session.
Signed-off-by: Elaine Wang <elaine.wang@intel.com>
commit
83429d2111104b378299458df9e24d373aeb0489
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Tue Apr 5 14:39:34 2011 -0700
GFX: Removed SUPPORT_EGL_EXTENSIONS
Removing Some EGL Extension support.
Signed-off-by: Hitesh K. Patel <hitesh.k.patel@intel.com>
commit
65b6c93017bf226f79be269d4847a509de7ed055
Author: Li Peng <peng.li@intel.com>
Date: Thu Mar 31 11:22:33 2011 +0800
mrst gfx: Oaktrail HDMI enabling patch
[v3] more bugfixes in mode setting test
[v2] adding audio support.
[v1] video and i2c implementation.
Signed-off-by: Li Peng <peng.li@intel.com>
Signed-off-by: Li Peng <peng.li@linux.intel.com>
commit
dd770ce3d548ae3c463ca8411e59a3de37796017
Author: Li Peng <peng.li@intel.com>
Date: Thu Mar 31 11:16:47 2011 +0800
mrst gfx: change pipe number for oaktrail
oaktrail platform (a variant of mrst) has two pipes, change
pipe number to two.
Signed-off-by: Li Peng <peng.li@intel.com>
Signed-off-by: Li Peng <peng.li@linux.intel.com>
commit
4f213a9d02f689d9fe00102b6db71f8ac994a8b5
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Wed Apr 13 14:34:59 2011 -0700
GFX: Fix to avoid infinite loop in ospm_power_island_up/down.
- Removed redundant code to power on/off GL3 (PSB_PWRGT_GFX_MASK_B0)
- Added counters with debug prints to detect infinite loops
- Added code to handle ISP on/off to be used by camera driver.
- invalidate GL3 before suspend.
- Add in OSPM_ISP_ISLAND check in ospm_power_island_up.
Signed-off-by: Rajesh Poornachandran <rajesh.poornachandran@intel.com>
commit
cc5b6236d07d1ffd5eb9202625baad9d1d0f4247
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Wed Apr 13 15:46:21 2011 -0700
git rebase kernel-adaptation-mrst-2.6.37.6-26.1
commit
56486202dfc76a2544a52bcf53d936824100d68b
Author: Daemon Anastas <daemonx.g.anastas@intel.com>
Date: Tue Apr 12 19:07:24 2011 -0700
check a spinlock and timer is not pending before setting timer (bugzilla 5503)
commit
2a6398ae1f523df0ba99c1dcafdd7a8a201d515a
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Thu Apr 7 17:27:17 2011 -0700
fix a typo
commit
063f2d34545728e7fc0df8e9b35217d94270bb1d
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Thu Apr 7 16:43:19 2011 -0700
GFX: Fix OSPM__ALL_ISLAND define
OSPM_ALL_ISLAND is defined as 0x2f it needs to be 0x3f
Signed-off-by: Hitesh K. Patel <hitesh.k.patel@intel.com>
commit
a022158016906934e28fc5e43cab076bfc040e6c
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Thu Apr 7 11:06:31 2011 -0700
GFX: Fix to avoid infinite loop in ospm_power_island_up/down.
- Removed redundant code to power on/off GL3 (PSB_PWRGT_GFX_MASK_B0)
- Added counters with debug prints to detect infinite loops
- Added code to handle ISP on/off to be used by camera driver.
- invalidate GL3 before suspend.
- Add in OSPM_ISP_ISLAND check in ospm_power_island_up.
Signed-off-by: Rajesh Poornachandran <rajesh.poornachandran@intel.com>
commit
c929a97c9e94eb687f015936f7309fdbeb015e8d
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Tue Apr 5 14:39:34 2011 -0700
GFX: Removed SUPPORT_EGL_EXTENSIONS
Removing Some EGL Extension support.
Signed-off-by: Hitesh K. Patel <hitesh.k.patel@intel.com>
commit
39e1a780478c305e511ba6d1d51b7a29f9a0c7fb
Author: Li Peng <peng.li@intel.com>
Date: Thu Mar 31 11:22:33 2011 +0800
mrst gfx: Oaktrail HDMI enabling patch
[v3] more bugfixes in mode setting test
[v2] adding audio support.
[v1] video and i2c implementation.
Signed-off-by: Li Peng <peng.li@intel.com>
Signed-off-by: Li Peng <peng.li@linux.intel.com>
commit
524e0ef2b155c944eb911b72589433377175c09c
Author: Li Peng <peng.li@intel.com>
Date: Thu Mar 31 11:16:47 2011 +0800
mrst gfx: change pipe number for oaktrail
oaktrail platform (a variant of mrst) has two pipes, change
pipe number to two.
Signed-off-by: Li Peng <peng.li@intel.com>
Signed-off-by: Li Peng <peng.li@linux.intel.com>
commit
822f7c7b0906ac49cbc975dda57ed378aa085741
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Thu Mar 31 12:24:58 2011 -0700
GFX: Graphics Driver build release 5.3.0.0041
Weekly Graphics driver release.
Please use usermode package 0041 with this build.
Signed-off-by: Hitesh K. Patel <hitesh.k.patel@intel.com>
commit
f2e96e8f02812a0db19aadb8507f655a9948ac35
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Thu Mar 31 12:14:22 2011 -0700
GFX: Optimize and Disable GL3
Use the bottom page of KERN_DATA_HEAP for GL3_USSE_WR_ADDR
Disable GL3.
Signed-off-by: Hitesh K. Patel <hitesh.k.patel@intel.com>
commit
dea7a3147923567ffe8bb9ffa5345b995ca5fa85
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Thu Mar 31 11:56:29 2011 -0700
GFX: Fix for Runtime PM
Runtime PM counters were not correctly incremneted/decremented
since we were always decrementing even in fail condition but not
incrementing in failed condition.
Signed-off-by: Hitesh K. Patel <hitesh.k.patel@intel.com>
commit
80bdde90d17087d071d4d33111f11cd7f22b8fc4
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Wed Mar 30 17:10:55 2011 -0700
GFX: Add PMIC Backlight controls
On Some morrestown platform the backlight is attached to PMIC. This patch will...
1. Enable Brighness level adjustment through PMIC Backlight controller
2. Turn off Backlight completely when LVDS panel is off
Signed-off-by: Hitesh K. Patel <hitesh.k.patel@intel.com>
commit
7399372ed507dfb318debe28a5644b4ee026f020
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Wed Mar 30 17:10:15 2011 -0700
revert PMIC backlight source
commit
454de637c7019a24da0d03efd169b0b9b0b70940
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Mon Mar 28 15:10:59 2011 -0700
GFX: Add PMIC Backlight controls
On Some morrestown platform the backlight is attached to PMIC. This patch will...
1. Enable Brighness level adjustment through PMIC Backlight controller
2. Turn off Backlight completely when LVDS panel is off
Signed-off-by: Hitesh K. Patel <hitesh.k.patel@intel.com>
commit
cb94ee7a31a2a276c120b6edea9f01fa7d1bdb32
Author: Austin Hu <austin.hu@intel.com>
Date: Thu Mar 24 15:43:00 2011 +0800
GFX: Fixed a memory leak issue in the DRM_PSB_GETPAGEADDRS IOCTL.
Fix Memory leak in TTM-GLUE.
Signed-off-by: Austin Hu <austin.hu@intel.com>
commit
9397a7c2e7d74e40c9a17bcf8989782b2181ce05
Author: Austin Yuan <shengquan.yuan@gmail.com>
Date: Thu Mar 24 13:51:18 2011 +0800
GFX: ospm_power_island_up/down: use loop counter to avoid infinite dead loop
Attemp to fix 3577740:UI freeze during video playback due to events/0 over CPU consumption
The rootcause is ospm_power_island_up/down run into dead loop
Signed-off-by: Austin Yuan <shengquan.yuan@gmail.com>
commit
c0a076ee8d51e0e28b651801f45e62f563264769
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Wed Mar 23 14:32:10 2011 -0700
GFX: Graphics Driver build release 5.3.0.0040
Weekly Graphics driver release.
Please use usermode package 0040 with this build.
Signed-off-by: Hitesh K. Patel <hitesh.k.patel@intel.com>
commit
e9687aef40806d9577559ee0f2519a5a7e4f7f0a
Author: Hitesh K. Patel <hitesh.k.patel@intel.com>
Date: Mon Mar 28 14:55:47 2011 -0700
kernel rebase 2.6.37.57.1
commit
cdb8d38531e2fbfb36762dd4cfa25516bd37e75e
Author: Yu(Alex) Dai <yu.dai@intel.com>
Date: Thu Mar 3 21:30:46 2011 +0800
Enlarge GENERAL_MAPPING heap to make sure more apps can run on SGX535 platform simultaneously.
The GENERAL_MAPPING heap is created as shared heap for multi-contexts
to fix one HW issue (BRN_23410). The heap size is set to 128M, which
is not enough for high memory usage under MeeGo Tablet 1280x600 mode.
Now set it to 512M to satisfy UI memory request.
Signed-off-by: Yu(Alex) Dai <yu.dai@intel.com>
commit
b7561fedff1a11f785d2e2e2abbeeb3577f6cffb
Author: Jackie Li <yaodong.li@intel.com>
Date: Thu Mar 3 21:30:08 2011 +0800
GFX-Display: Fixed glsync tearing issue on MIPI panel.
Tearing happens while running glsync -s b on both MIPI panel with vsync enabled.
The root cause of this issue is that 3D blit happens during write_mem_start command transmission.
Add a fixing to wait for DBI FIFO empty before schedule 3D blit to system frame buffer.
TE interrupt of second MIPI panel was enabled since we need to send out write_mem_start command only when TE interrupt happens.
Known issue:
In dual MIPI clone mode, two MIPI panels share the same system frame buffer and TE interrupts may NOT happen at the same time,
so if one of the TE interrupt handler woke up 3D blit to frame buffer, there would be tearing on the other MIPI panel
if it is sending write_mem_start command
Signed-off-by: Jackie Li <yaodong.li@intel.com>
commit
9ef6a38fe330be33f74983b63c406a2fca52bdbc
Author: Justin Dou <Justin.Dou@intel.com>
Date: Thu Mar 3 21:14:54 2011 +0800
Added the kernel-mid.config for mrst kernel build and updated our gfx driver to 0038 version
The kernel-mid.config is now aligned with configs/kernel-adaptation-mrst.config which was used for handset image build.
Signed-off-by: Justin Dou <Justin.Dou@intel.com>
commit
f8aadafe3d6f4c4db5a5cb698b7df5e957c6517a
Author: Justin Dou <Justin.Dou@intel.com>
Date: Thu Mar 3 20:15:22 2011 +0800
Create the MRST kernel branch with kernel-adaptation-mrst-2.6.37-13.2.src.rpm
This is the up-to-date kernel from http://download.meego.com/live/Trunk/standard/src/ which integrated our 0037 gfx driver.
Signed-off-by: Justin Dou <Justin.Dou@intel.com>
usb/penwell_otg: set charging current limitation to 100mA by default for SDP case
BZ: 38168
This is in order to workaround one issue in low battery level case,
set charging current to 100mA once SDP is detected. Late setting
may cause system shutdown due to low power state.
Change-Id: I829e3283b32a858b446e7287e72179e65825516b
Signed-off-by: Wu, Hao <hao.wu@intel.com>
Reviewed-on: http://android.intel.com:8080/53444
Reviewed-by: Mukherjee, DevapriyaX <devapriyax.mukherjee@intel.com>
Reviewed-by: Pallala, Ramakrishna <ramakrishna.pallala@intel.com>
Reviewed-by: Tang, Richard <richard.tang@intel.com>
Reviewed-by: Jena, TapanX <tapanx.jena@intel.com>
Reviewed-by: Meng, Zhe <zhe.meng@intel.com>
Tested-by: Meng, Zhe <zhe.meng@intel.com>
Tested-by: Kallappa Manjanna, MadhukumarX <madhukumarx.kallappa.manjanna@intel.com>
Reviewed-by: buildbot <buildbot@intel.com>
Tested-by: buildbot <buildbot@intel.com>
atomisp: [mt9e013]set frame_valid flag in work queue call back.
BZ: 41629
root cause:
gain control was designed below;
First: control the filter at SOF via work queue.
Next: respond the gain setup request via ioctl from user space.
However work queue service delayed when cpu is heavily-loaded.
This situation happens at 1080 recording.
The work queue call back delayed 12 msec from enqueued at worst.
In this case, control order is broken like below;
First: respond the gain setup request.
Second: control the filter.
In the result, gain is setupped twice during one frame.
solution:
set frame_valid flag in the work queue call back instead receipt SOF signal.
then the gain setup request is suspended when work queue service delayed.
Change-Id: I4827b589591e010cb4fbefb3cbf8700f1224e5ec
Signed-off-by: Yoshio Wada <yoshio.wada@intel.com>
Reviewed-on: http://android.intel.com:8080/52711
Reviewed-by: Kruger, Jozef <jozef.kruger@intel.com>
Reviewed-by: Kantola, Lasse <lasse.kantola@intel.com>
Reviewed-by: Koskinen, Ilkka <ilkka.koskinen@intel.com>
Reviewed-by: Toivonen, Tuukka <tuukka.toivonen@intel.com>
Reviewed-by: Koski, Anttu <anttu.koski@intel.com>
Tested-by: Koski, Anttu <anttu.koski@intel.com>
Reviewed-by: buildbot <buildbot@intel.com>
Tested-by: buildbot <buildbot@intel.com>
GFX-Display: Added Post2 support for frame buffer flipping.
BZ: 33412
This is 1/3 patches to solve Bug 33412.
Since post() path only supports update display plane surface in
kernel driver, we cannot configure other display plane settings
dynamically from user space.
This patch eanbled frame buffer flipping in Post2() path, and it
enhanced kernel driver to be able to re-config all display plane
registers, so that HWC can adjust display plane configs such as
plane format, z order dynamically.
This patch also increased the number of buffer of display class
so that we can multiple buffers to kernel driver in one flip cmd.
Change-Id: Id9b6bfcd0c59a2fbcb6338dc20dbbfd219ce5e9a
Signed-off-by: Jackie Li <yaodong.li@intel.com>
Signed-off-by: Lei Zhang <lei.zhang@intel.com>
Reviewed-on: http://android.intel.com:8080/51149
Reviewed-by: Xu, Randy <randy.xu@intel.com>
Tested-by: Xu, Randy <randy.xu@intel.com>
Reviewed-by: buildbot <buildbot@intel.com>
Tested-by: buildbot <buildbot@intel.com>
GFX-Display: fix mistake in /proc/display_status.
BZ: 42140
system will hange ,when using /proc/display_status.
it is related to buf overflow ,which been not protected well.
one example:
echo "a b000 b100">/proc/display_status
cat /proc/display_status
Change-Id: I4695e6b8a9a31faaee6e437bbd0ae3e5d5b7900f
Signed-off-by: Ke Ai <ke.ai@intel.com>
Reviewed-on: http://android.intel.com:8080/53129
Reviewed-by: Xu, Randy <randy.xu@intel.com>
Tested-by: Tong, BoX <box.tong@intel.com>
Reviewed-by: buildbot <buildbot@intel.com>
Tested-by: buildbot <buildbot@intel.com>
GFX-Display: detach overlay rw ioctrl, gtt map/umap ioctrl from drm global lock.
BZ: 38361
1. Gtt map/unmap ioctrl don't need hold drm global lock. There is already a mm lock
for protecting gtt mm resource.
2. overlay wait vblank function should not block on mode_config lock, that will cause
actual wait time exceed a vblank period.
3. Add a mutex to protect overlay rw ioctrl instead of drm global lock. Because this
ioctrl is only used by overlay, it is not necessary to lock other display registers.
Change-Id: Ic9f98fb5bcb28925740133a4fb0c0aa3679b25ac
Signed-off-by: Jason Hu <jason.hu@intel.com>
Reviewed-on: http://android.intel.com:8080/52606
Reviewed-by: Ding, Haitao <haitao.ding@intel.com>
Tested-by: Ding, Haitao <haitao.ding@intel.com>
Reviewed-by: buildbot <buildbot@intel.com>
Tested-by: buildbot <buildbot@intel.com>
[GFX-3D]Release last flip buffer after a while even if there is no new frame coming
BZ:42074
Release last flip buffer after a while even if there is no new frame coming,
as looks like user space try to lock last frame without submitting a new frame first
and cause dead lock then
Signed-off-by: Yun Tu(Mark Tu) <yun.tu@intel.com>
Change-Id: Ifa76783985418cd2011546c01169c996baf25901
Reviewed-on: http://android.intel.com:8080/52876
Reviewed-by: Xu, Randy <randy.xu@intel.com>
Tested-by: Xu, Randy <randy.xu@intel.com>
Tested-by: Tong, BoX <box.tong@intel.com>
Reviewed-by: buildbot <buildbot@intel.com>
Tested-by: buildbot <buildbot@intel.com>
[PORT FROM R2][REVERTME] cpu: trigger a BUG() when boot cpu1 failed when being woke up from S3.
BZ: 36367
From yanmin:
It's hard to reproduce this issue and we don't know where CPU 1 is
running when it happens. A potential patch is to call BUG() to trigger
a panic when CPU 0 couldn't boot CPU 1, so system would reboot by
software WDT and wouldn't consume power unconsciously.
Change-Id: I32624a6875b435655dc0fe7e3f647dbbeb6d793a
Orig-Change-Id: I700068ab629cf60061ac24315733aeedb5710b86
Signed-off-by: Yanmin Zhang <yanmin.zhang@intel.com>
Signed-off-by: ShuoX Liu <shuox.liu@intel.com>
Reviewed-on: http://android.intel.com:8080/52325
Reviewed-by: Popescu, CatalinX <catalinx.popescu@intel.com>
Reviewed-by: Hogander, Jouni <jouni.hogander@intel.com>
Reviewed-by: Mansoor, Illyas <illyas.mansoor@intel.com>
Tested-by: Martin, LoicX <loicx.martin@intel.com>
Reviewed-by: buildbot <buildbot@intel.com>
Tested-by: buildbot <buildbot@intel.com>
GFX-Display: Fix Out Of Memory issue when playing video while rotating the device with suspend and resume
BZ: 39131
Add definitaion of SUPPORT_LARGE_GENERAL_HEAP in kernel/driver/staging/mrst/medfield/Makefile
to enlarge the GeneralHeap in gpu, and ensure there is enough video memory for video playing
while rotate the screen with suspend and resume.
Change-Id: If7653c1068a01f22dca77d2342555222dc0765cf
Signed-off-by: Lingyun Zhu <lingyun.zhu@intel.com>
Reviewed-on: http://android.intel.com:8080/52413
Reviewed-by: Tong, BoX <box.tong@intel.com>
Tested-by: Tong, BoX <box.tong@intel.com>
Reviewed-by: Xu, Randy <randy.xu@intel.com>
Tested-by: Xu, Randy <randy.xu@intel.com>
Reviewed-by: buildbot <buildbot@intel.com>
Tested-by: buildbot <buildbot@intel.com>
smb347_charger: Add a small delay after OTG-disable.
BZ: 39775
When OTG disabled and charging enabled, send the uevent after a
short delay, as Status Register C is not updated immediately.
Change-Id: I24160a7a5f9350aa50ca5dbad6af5b0c23a4188b
Signed-off-by: Pavan Kumar S <pavan.kumar.s@intel.com>
Reviewed-on: http://android.intel.com:8080/51924
Reviewed-by: Tc, Jenny <jenny.tc@intel.com>
Reviewed-by: Jena, TapanX <tapanx.jena@intel.com>
Tested-by: Kallappa Manjanna, MadhukumarX <madhukumarx.kallappa.manjanna@intel.com>
Reviewed-by: buildbot <buildbot@intel.com>
Tested-by: buildbot <buildbot@intel.com>
atomisp: flash-light appears on the screen.
BZ: 36446
keep the timing to write the itg and gain value into mt9e013 register.
- use SOF irq
- implement the gain delay filter in the sensor driver
Solution :
- atomisp
- libcamera
- libmfldadvci
- bionic
Change-Id: Iaba0da1906af9895af2b6f9054d3601d3455c044
Signed-off-by: Yoshio Wada <yoshio.wada@intel.com>
Reviewed-on: http://android.intel.com:8080/51348
Reviewed-by: Kantola, Lasse <lasse.kantola@intel.com>
Reviewed-by: Kruger, Jozef <jozef.kruger@intel.com>
Reviewed-by: Koski, Anttu <anttu.koski@intel.com>
Tested-by: Koski, Anttu <anttu.koski@intel.com>
Reviewed-by: Koskinen, Ilkka <ilkka.koskinen@intel.com>
Reviewed-by: buildbot <buildbot@intel.com>
Tested-by: buildbot <buildbot@intel.com>
GFX-Display: resolved the flashing issue when booting up SC1.
BZ: 32042
Fixed by:
1. Not doing mode setting, DPMS operation and suspending/resuming before driver
loadling completes, as IFWI turns on the MIPI panel when launching system;
2. Replacing mdfld_dsi_dbi_update_power() with mdfld_auo_dsi_dbi_set_power(),
to avoid potential incorrect DCS command sequence.
3. Making DBI panel update_fb() work when the frame buffer device is used to
update screen in droidboot mode.
4. Also fixed the droidboot displaying nothing issue caused by patch #47785,
by sending "write_mem_start" DSI command in TE interrupt handler in droidboot
mode where there isn't page flip request.
This issue is from Enzo BZ #570.
Change-Id: I9c52ef7e4e28ec6504be67843121293fbc2a6cd8
Signed-off-by: Jackie Li <yaodong.li@intel.com>
Signed-off-by: Austin Hu <austin.hu@intel.com>
Signed-off-by: Ke Ai <ke.ai@intel.com>
Reviewed-on: http://android.intel.com:8080/51962
Reviewed-by: Tong, BoX <box.tong@intel.com>
Tested-by: Tong, BoX <box.tong@intel.com>
Reviewed-by: Xu, Randy <randy.xu@intel.com>
Reviewed-by: buildbot <buildbot@intel.com>
Tested-by: buildbot <buildbot@intel.com>
projects / kernel / kernel-mfld-blackbay.git / log