Age | Commit message (Collapse) | Author |
|
Some functionality used by android binary drivers
were missing.
Signed-off-by: Dominik Sliwa <dominik.sliwa@toradex.com>
Acked-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
|
|
Merge NVIDIA's latest Linux for Tegra aka L4T R21.6 Linux kernel changes
from git://nv-tegra.nvidia.com/linux-3.10.git commit:
b271e8fa67a6d9c4600274a25636cfe00fdd1b68
Signed-off-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
Acked-by: Dominik Sliwa <dominik.sliwa@toradex.com>
|
|
With gcc 7 the following compile time error occurs:
| drivers/video/tegra/dc/dp.c:1178:12: error: the omitted middle operand in ?: will always be 'true', suggest explicit middle operand [-Werror=parentheses]
| cr_done ? : ({ret = -EINVAL; goto fail; });
| ^
| drivers/video/tegra/dc/dp.c:1186:12: error: the omitted middle operand in ?: will always be 'true', suggest explicit middle operand [-Werror=parentheses]
| lt_done ? : ({ret = -EINVAL; goto fail; });
| ^
Signed-off-by: Max Krummenacher <max.krummenacher@toradex.com>
Acked-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
|
|
With gcc 7 the following compile time error occurs:
| drivers/video/tegra/dc/dsi.c:766:34: error: '*' in boolean context, suggest '&&' instead [-Werror=int-in-bool-context]
| drivers/video/tegra/dc/dsi.c:757:1:
| (platform_t_phy_ps) ? ( \
| ~~~~~~~~~~~~~~~~~~~
| kernel-source/drivers/video/tegra/dc/dsi.c:766:34:
| dsi->info.phy_timing.t_tlpx_ns * 1000,
| drivers/video/tegra/dc/dsi.c:757:2: note: in definition of macro 'SELECT_T_PHY'
| (platform_t_phy_ps) ? ( \
| ^~~~~~~~~~~~~~~~~
Signed-off-by: Max Krummenacher <max.krummenacher@toradex.com>
Acked-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
|
|
Handle the race condition between malicious fd close and
copy_to_user error, which can create use after free condition.
This is fixed by deferring the fd install, which eliminates
the race that leads to use after free condition.
Fixing Google Bug 32160775.
Bug 1835857
Change-Id: I337807e4360661beced8f9e1155c47b66607b8df
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/1248391
Reviewed-on: https://git-master.nvidia.com/r/1512958
GVS: Gerrit_Virtual_Submit
Reviewed-by: Bibek Basu <bbasu@nvidia.com>
Tested-by: Bibek Basu <bbasu@nvidia.com>
|
|
Drive current for LANE4 was not set if configured as 24bpp LVDS out.
Fix it by programming proper drive current register if using 24bpp out.
https://devtalk.nvidia.com/default/topic/1003030
Signed-off-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
Acked-by: Dominik Sliwa <dominik.sliwa@toradex.com>
|
|
In the debugfs support for reading panel registers, max payload
needs to be limited to the buff array size to avoid stack corruption.
Bug 1873360
Change-Id: Ibee7bd81027d2669297942c09b905f1dd3bb09ee
Signed-off-by: Pavan Kunapuli <pkunapuli@nvidia.com>
Signed-off-by: sakets <sakets@nvidia.com>
Reviewed-on: https://git-master/r/1507653
Reviewed-by: Automatic_Commit_Validation_User
GVS: Gerrit_Virtual_Submit
Reviewed-by: Bibek Basu <bbasu@nvidia.com>
|
|
When the NVMAP_IOC_PIN_MULT_32 and NVMAP_IOC_UNPIN_MULT_32 are
called it is possible that the op.addr is not initialized. This
can cause write to some random address thus causing corruption.
This patch fixes Google Bug 31668540
bug 1832092
Change-Id: I4d12d1a6c777131ba1fa2a753ea640861f8e82a6
Signed-off-by: Sri Krishna chowdary <schowdary@nvidia.com>
Reviewed-on: http://git-master/r/1314406
(cherry picked from commit da0c43534bb61e2e0849e297d389517d5e4ed168)
Reviewed-on: http://git-master/r/1504673
GVS: Gerrit_Virtual_Submit
Reviewed-by: Bibek Basu <bbasu@nvidia.com>
Tested-by: Bibek Basu <bbasu@nvidia.com>
Reviewed-by: Winnie Hsu <whsu@nvidia.com>
|
|
Consider the following case:
1. NVMAP_IOC_CREATE on IOVMM gives a valid fd to user space
2. user space does not call NVMAP_IOC_ALLOC.
3. user space calls a client driver IOCTL which calls dma_buf_map_attachment
4. call to dma_buf_map_attachment propagates till__nvmap_sg_table
which has heap_pgalloc as true and tries to access pages[]
which has all NULL.
5. Similarly, a dma_buf_kmap() can result in __nvmap_kmap() being called
which again results in NULL dereference if pages[] is accessed.
A valid __nvmap_sg_table should occur only when h->alloc is true.
So, add check for it.
bug 1838597
bug 1883708
Change-Id: I400d9d8a94ff1003db207fc9c252b9256d796f60
Signed-off-by: Sri Krishna chowdary <schowdary@nvidia.com>
Signed-off-by: Debarshi Dutta <ddutta@nvidia.com>
(cherry picked from commit 8244d104b7635cb0b26b651b6851498b9a84d7d6)
Reviewed-on: http://git-master/r/1489579
Reviewed-by: Automatic_Commit_Validation_User
GVS: Gerrit_Virtual_Submit
Reviewed-by: Bibek Basu <bbasu@nvidia.com>
Tested-by: Bibek Basu <bbasu@nvidia.com>
|
|
Initialized the uninitialized variables and handled return status
from nvmap_get_handle_param.
Bug 1884311
Bug 1820242
Change-Id: I2390c859d2b2af39eaff44749ca64e60920fe944
Signed-off-by: Gagan Grover <ggrover@nvidia.com>
Reviewed-on: http://git-master/r/1259560
Reviewed-on: http://git-master/r/1489707
GVS: Gerrit_Virtual_Submit
Tested-by: Sumit Gupta <sumitg@nvidia.com>
Reviewed-by: Bibek Basu <bbasu@nvidia.com>
|
|
Check all pages' parameters before reserve pages.
Bug 1883463
Bug 1831426
Bug 200247013
Manual port: http://git-psac/r/9287
(cherry picked from commit 61a05b52b8a17593e2817076b9bf59efdd9268ad)
Change-Id: I2f47c385ff8f4a9ca6bf37ee41749bd684ca1a20
Signed-off-by: Xia Yang <xiay@nvidia.com>
Signed-off-by: Sagar Kadamati <skadamati@nvidia.com>
Reviewed-on: http://git-master/r/1273326
Reviewed-on: http://git-master/r/1488769
GVS: Gerrit_Virtual_Submit
Tested-by: Sumit Gupta <sumitg@nvidia.com>
Reviewed-by: Bibek Basu <bbasu@nvidia.com>
|
|
Currently nvhost performs minimal checking for submits it passes
to hardware: The kernel does not check if job syncpoints are allocated
and the gather classes are not verified currently.
This patch adds checks for syncpoint ids and gather classes.
Adapted from 0abcbd69c4cbd0093e223b6c248fdd53c2886951.
Bug 1831406
Change-Id: Ifb9d2090009d16d0f56bc11546036167c7f72228
Signed-off-by: Mikko Perttunen <mperttunen@nvidia.com>
Reviewed-on: http://git-master/r/1242190
Reviewed-by: Winnie Hsu <whsu@nvidia.com>
Tested-by: Winnie Hsu <whsu@nvidia.com>
|
|
Channel ioctl interface is not multithreading safe and as the
common case is that we have only a single active user for an open
fd, add a mutex to force serialization of ioctl calls.
Bug 1830021
Change-Id: Ifa6595a105b913345104f216f0541c371e89efe5
(cherry picked from commit 7b24caa9a8d2ab08fe0c7be112e805e44906d956)
Signed-off-by: Gagan Grover <ggrover@nvidia.com>
Reviewed-on: http://git-master/r/1248801
Reviewed-by: Bibek Basu <bbasu@nvidia.com>
Reviewed-by: Automatic_Commit_Validation_User
GVS: Gerrit_Virtual_Submit
|
|
Fix possible use after free issue.
Bug 1814555
Bug 1884319
Change-Id: I826aa34f61d43fda5419a528697ce84ba2ce1eae
Reviewed-on: http://git-master/r/1221643
Signed-off-by: Gagan Grover <ggrover@nvidia.com>
Signed-off-by: Debarshi Dutta <ddutta@nvidia.com>
Reviewed-on: http://git-master/r/1257999
(cherry picked from commit b1647da33cff0c498ca8439a722ea1962ecf6901 in rel-24)
Reviewed-on: http://git-master/r/1461184
GVS: Gerrit_Virtual_Submit
Reviewed-by: Bibek Basu <bbasu@nvidia.com>
Tested-by: Bibek Basu <bbasu@nvidia.com>
|
|
Change kmalloc to kmalloc_array to prevent overflow issues
caused by large values supplied by user.
Based on "video: tegra: host: Fix overflow issues in allocation"
in nvhost/.
Coverity ID 27942
Bug 1856419
Change-Id: I5e96d0ec184543782dfe8814ad7e856b3b71221c
Signed-off-by: Mikko Perttunen <mperttunen@nvidia.com>
Reviewed-on: http://git-master/r/1295062
GVS: Gerrit_Virtual_Submit
Reviewed-by: Bibek Basu <bbasu@nvidia.com>
|
|
Consider the following case:
1. NVMAP_IOC_CREATE gives a valid fd to user space
2. user space calls NVMAP_IOC_ALLOC and it fails. So, all
of the handle's allocation fields are zero.
3. Subsequent dma_buf_vmap, mmap on fd leads to __nvmap_mmap
call.
4. handle is valid but h->alloc, h->carveout, h->heap_pgalloc,
h->vaddr all are 0.
5. We check for h->heap_pgalloc which is false, so proceed and
dereference h->carveout leading to NULL pointer exception.
A valid __nvmap_mmap should occur only when h->alloc is true.
So, add check for it.
bug 1837468
Change-Id: I9be9d94f9b74c25b9b588fb1a16a74e96161ceda
Signed-off-by: Sri Krishna chowdary <schowdary@nvidia.com>
Reviewed-on: http://git-master/r/1253236
(cherry picked from commit c5da78cf3d0c19f1e04501a4b3f64a5acacd0ff3)
Reviewed-on: http://git-master/r/1312264
GVS: Gerrit_Virtual_Submit
Reviewed-by: Bibek Basu <bbasu@nvidia.com>
Tested-by: Bibek Basu <bbasu@nvidia.com>
|
|
Allow to specify framebuffer videomode using kernel command line
parameters. NVIDIAs binary X driver later on picks up those settings
and start X with current mode settings, if no EDID data are available.
Reused some of the implementation from Stefan's work for modedb
support[1] on Tegra20/30
[1] http://git.toradex.com/cgit/linux-toradex.git/commit/?h=tegra-next&id=1d3625dd9903bcc59e2df56836565ebb682948c1
Signed-off-by: Bhuvanchandra DV <bhuvanchandra.dv@toradex.com>
Acked-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
|
|
Fix suspend and resume for LP1 sleep.
Signed-off-by: Dominik Sliwa <dominik.sliwa@toradex.com>
Acked-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
|
|
Validate the region specified by offset and size before performing
the operations like nvmap_prot_handle, nvmap_cache_maint and nvmap_handle_mk*.
This validation of offset and size once the values are in local variables
guarantees that even though user space changes the values in user buffers,
nvmap continues to perform operations with the contents that are validated.
Fixes Google Bug 34113000.
bug 1862379
Change-Id: Ief81887b3d94b49f3dcf4d2680d9d7b257c54092
Signed-off-by: Sri Krishna chowdary <schowdary@nvidia.com>
Signed-off-by: Bibek Basu <bbasu@nvidia.com>
Reviewed-on: http://git-master/r/1298712
(cherry picked from commit f45441da608d8015ece73d253d4bdb48863f99e2)
Reviewed-on: http://git-master/r/1310316
(cherry picked from commit 57367ab3be5f1c52dd6b885f114ae90dfce5a363)
Reviewed-on: http://git-master/r/1319910
GVS: Gerrit_Virtual_Submit
|
|
Moved fd_install() at the end of the channel_open ioctl. So, the fd
can't be used until open ioctl completes.
Bug 1832094
Change-Id: Ib33d43bf5164418a38f98677d4e3295f3d1c1450
Signed-off-by: Gagan Grover <ggrover@nvidia.com>
Reviewed-on: http://git-master/r/1248180
(cherry picked from commit e6a41d5c0049c2878543006b67b7ee2b2bbda2ab)
Reviewed-on: http://git-master/r/1249505
Reviewed-by: Winnie Hsu <whsu@nvidia.com>
Tested-by: Winnie Hsu <whsu@nvidia.com>
|
|
Check if there is at least one syncpt_incrs in each job.
Bug 1812182
Change-Id: I0bd0b2e7c4d01641c83ba729ec34390ddea81496
Reviewed-on: http://git-master/r/1221226
Signed-off-by: Gagan Grover <ggrover@nvidia.com>
Reviewed-on: http://git-master/r/1248797
GVS: Gerrit_Virtual_Submit
Reviewed-by: Arto Merilainen <amerilainen@nvidia.com>
|
|
drivers/video/tegra/dc/dc.c:116:3: error: 'can_filter' defined but not
used [-Werror=unused-const-variable=]
} can_filter[] = {
^~~~~~~~~~
Signed-off-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
Acked-by: Max Krummenacher <max.krummenacher@toradex.com>
|
|
Below addition on 32 bit architecture machines could
cause integer overflow since we will assign overflowed
value to "num_unpins"
s64 num_unpins = num_cmdbufs + num_relocs
Fix this and other calculations by explicitly typecasting
variables to u64 first
Bug 1781393
Change-Id: Ib7d9c0be4ac61dc404512b4bb0331aa20a6978bc
Signed-off-by: Deepak Nibade <dnibade@nvidia.com>
Reviewed-on: http://git-master/r/1171748
(cherry picked from commit 8f00b96c137b9c4cb43a8dbe2e153fae49524113)
Reviewed-on: http://git-master/r/1172519
(cherry picked from commit 61229625b1e19d5a93a9458f04e0cce356dbdee3)
Reviewed-on: http://git-master/r/1190218
Reviewed-by: Automatic_Commit_Validation_User
GVS: Gerrit_Virtual_Submit
Reviewed-by: Jeetesh Burman <jburman@nvidia.com>
Tested-by: Jeetesh Burman <jburman@nvidia.com>
Reviewed-by: Arto Merilainen <amerilainen@nvidia.com>
Reviewed-by: Bibek Basu <bbasu@nvidia.com>
Reviewed-by: Winnie Hsu <whsu@nvidia.com>
|
|
We allocate below without checking if num_syncpt_incrs
is valid or not
struct nvhost_ctrl_sync_fence_info pts[num_syncpt_incrs];
If UMD passes a negative value in num_syncpt_incrs, then
it is possible to corrupt the stack
Hence, first check if num_syncpt_incrs is valid (i.e.
not negative)
And then allocate the array dynamically using kzalloc
instead of allocating it on stack
Bug 1781393
Change-Id: I5389fd271149b457f63831a41c104c9814299ddf
Signed-off-by: Deepak Nibade <dnibade@nvidia.com>
Reviewed-on: http://git-master/r/1171747
(cherry picked from commit 07fb347b4060a888b19df3524f36fcf7974a79d1)
Reviewed-on: http://git-master/r/1172518
(cherry picked from commit 1db2d69b6abeb6fc9d4257db88f631d9c8aef74d)
Reviewed-on: http://git-master/r/1190211
GVS: Gerrit_Virtual_Submit
Reviewed-by: Jeetesh Burman <jburman@nvidia.com>
Tested-by: Jeetesh Burman <jburman@nvidia.com>
Reviewed-by: Arto Merilainen <amerilainen@nvidia.com>
Reviewed-by: Bibek Basu <bbasu@nvidia.com>
Reviewed-by: Winnie Hsu <whsu@nvidia.com>
|
|
pll_d2 runs at a minimum of 100MHz on T124. Update logic
to choose parent clock rate more than 100MHz.
e.g.: A mode with 32MHz pclk chooses parent clock of
96MHz with a divider of 3.0, which fails as pll_d
can't be pulled below 100MHz.
bug 1785365
Change-Id: I12400549a3ed42295ddd46adcb6493232f2d896a
Signed-off-by: Naveen Kumar S <nkumars@nvidia.com>
Reviewed-on: http://git-master/r/1184235
Reviewed-by: Automatic_Commit_Validation_User
GVS: Gerrit_Virtual_Submit
Reviewed-by: Aleksandr Frid <afrid@nvidia.com>
Reviewed-by: Aly Hirani <ahirani@nvidia.com>
Reviewed-by: Venu Byravarasu <vbyravarasu@nvidia.com>
Reviewed-by: Bibek Basu <bbasu@nvidia.com>
|
|
Patch taken from Manoj Gupta's post on NVIDIA's public embedded systems forum:
https://devtalk.nvidia.com/default/topic/822612/jetson-tk1/-issue-lvds-panel-enabled-effect-hdmi-out-image-pull-down-menu-items/post/4673174/#4673174
Signed-off-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
Acked-by: Dominik Sliwa <dominik.sliwa@toradex.com>
|
|
This fixes the following build time error in case CONFIG_TEGRA_LVDS is enabled:
drivers/video/tegra/dc/sor.h: In function 'tegra_sor_clk_enable':
drivers/video/tegra/dc/sor.h:180:2: error: implicit declaration of function 'clk_prepare_enable' [-Werror=implicit-function-declaration]
clk_prepare_enable(sor->sor_clk);
^
drivers/video/tegra/dc/sor.h: In function 'tegra_sor_clk_disable':
drivers/video/tegra/dc/sor.h:185:2: error: implicit declaration of function 'clk_disable_unprepare' [-Werror=implicit-function-declaration]
clk_disable_unprepare(sor->sor_clk);
^
In file included from drivers/video/tegra/dc/dc_priv_defs.h:26:0,
from drivers/video/tegra/dc/dc_priv.h:23,
from drivers/video/tegra/dc/lvds.c:23:
include/linux/clk.h: At top level:
include/linux/clk.h:330:19: error: static declaration of 'clk_prepare_enable' follows non-static declaration
static inline int clk_prepare_enable(struct clk *clk)
^
In file included from drivers/video/tegra/dc/lvds.h:20:0,
from drivers/video/tegra/dc/lvds.c:22:
drivers/video/tegra/dc/sor.h:180:2: note: previous implicit declaration of 'clk_prepare_enable' was here
clk_prepare_enable(sor->sor_clk);
^
In file included from drivers/video/tegra/dc/dc_priv_defs.h:26:0,
from drivers/video/tegra/dc/dc_priv.h:23,
from drivers/video/tegra/dc/lvds.c:23:
include/linux/clk.h:345:20: error: conflicting types for 'clk_disable_unprepare' [-Werror]
static inline void clk_disable_unprepare(struct clk *clk)
^
include/linux/clk.h:345:20: error: static declaration of 'clk_disable_unprepare' follows non-static declaration
In file included from drivers/video/tegra/dc/lvds.h:20:0,
from drivers/video/tegra/dc/lvds.c:22:
drivers/video/tegra/dc/sor.h:185:2: note: previous implicit declaration of 'clk_disable_unprepare' was here
clk_disable_unprepare(sor->sor_clk);
^
Final patch taken from Manoj Gupta's post on NVIDIA's public embedded systems forum:
https://devtalk.nvidia.com/default/topic/822612/jetson-tk1/-issue-lvds-panel-enabled-effect-hdmi-out-image-pull-down-menu-items/post/4663817/#4663817
Signed-off-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
Acked-by: Dominik Sliwa <dominik.sliwa@toradex.com>
|
|
In case kernel fails to open a channel (e.g. due to inability to
allocate hardware context or turn on the device), the channel open
function releases the resources that were already allocated
successfully.
However, currently the error path additionally calls the channel
release function for putting the channel pointer after the private
data structures have been freed - thereby causing use-after-free
memory usage.
This patch reworks error handling in channel open to release
channel without risking usage of already freed memory.
Bug 1763577
Change-Id: Ic7562e69f2babad653afc7a11e413701494a30b4
Signed-off-by: Arto Merilainen <amerilainen@nvidia.com>
Reviewed-on: http://git-master/r/1148081
Reviewed-by: Winnie Hsu <whsu@nvidia.com>
Tested-by: Winnie Hsu <whsu@nvidia.com>
|
|
In nvhost_ioctl_ctrl_module_regrdwr(), we copy offset
to read/write from user space but we do not have
any check on it
So it is possible for user space to add unaligned
offset and request read/write which would crash the
system
Fix this by explicitly checking alignment of the
offset passed by user space
Bug 1739935
Change-Id: Iea2a07c60500af876b732a0e9d9d08535aa53b5c
Signed-off-by: Deepak Nibade <dnibade@nvidia.com>
Reviewed-on: http://git-master/r/1029405
(cherry picked from commit 422baa09a17a6a17f4e572aa5441ca174634de0d)
Reviewed-on: http://git-master/r/1123363
Reviewed-by: Automatic_Commit_Validation_User
GVS: Gerrit_Virtual_Submit
Reviewed-by: Bibek Basu <bbasu@nvidia.com>
Tested-by: Bibek Basu <bbasu@nvidia.com>
Reviewed-by: Matthew Pedro <mapedro@nvidia.com>
|
|
Add ref count to prevent invalid vma removal from the h->vmas list
and also allow addition of a different vma which also has same
nvmap_vma_priv as vm_private_data into the h->vmas list. Both cases
are allowed in valid usage of nvmap_vma_open/nvmap_vma_close.
Bug 200164002
Change-Id: Ifc4d281dd91e1d072a9a3ee85e925040bd65a6bc
Signed-off-by: Sri Krishna chowdary <schowdary@nvidia.com>
Signed-off-by: Bryan Wu <pengw@nvidia.com>
Reviewed-on: http://git-master/r/1133708
Reviewed-by: Krishna Reddy <vdumpa@nvidia.com>
|
|
Make sure that the notifier object fits within the supplied buffer.
Bug 1739183
Change-Id: Ifbf46eddea86bedf0236851ea1c3f73e5f820beb
Signed-off-by: Konsta Holtta <kholtta@nvidia.com>
Reviewed-on: http://git-master/r/1026409
(cherry picked from commit 4086d2137e9b51137aa335fa264d924c73dea5fc)
Reviewed-on: http://git-master/r/1029074
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Matthew Pedro <mapedro@nvidia.com>
Reviewed-by: Shridhar Rasal <srasal@nvidia.com>
Reviewed-by: Bibek Basu <bbasu@nvidia.com>
Reviewed-by: Deepak Nibade <dnibade@nvidia.com>
Reviewed-by: Arto Merilainen <amerilainen@nvidia.com>
|
|
Query max vi/isp clk-rate runtime to calcuate max BW.
Remove max-bw defines.
Bug 1538490
Bug 1695435
Change-Id: I86a5c22fa3c7c9582351bbe9a95776aaea6a613d
Signed-off-by: Sudhir Vyas <svyas@nvidia.com>
Reviewed-on: http://git-master/r/461278
(cherry picked from commit bbcd86c917430ceea1603e03964296ca4e26ac3a)
Reviewed-on: http://git-master/r/825139
Reviewed-by: Automatic_Commit_Validation_User
GVS: Gerrit_Virtual_Submit
Reviewed-by: Jihoon Bang <jbang@nvidia.com>
Tested-by: Frank Shi <fshi@nvidia.com>
Reviewed-by: Matthew Pedro <mapedro@nvidia.com>
|
|
In case tegra_dc_sync_windows is interrupted by signal,
return the error to caller application
Bug 200090492
Change-Id: Id69fbe38d0abe0b3e71eb5a413db241ebcf0a0ae
Signed-off-by: Bibek Basu <bbasu@nvidia.com>
Reviewed-on: http://git-master/r/784754
GVS: Gerrit_Virtual_Submit
Reviewed-by: Winnie Hsu <whsu@nvidia.com>
|
|
Providing a small delay after writing to dc registers while
stopping dc stream helps in stabilizing the registers. This
helps in resolving the intermittent register read failure issue.
bug 200087039
Change-Id: I159d1d75aa2472b9e33bc42d890382f33def218a
Signed-off-by: Naveen Kumar S <nkumars@nvidia.com>
Reviewed-on: http://git-master/r/746062
(cherry picked from commit d29669af88735a2aeeb87b26f8794c9bcbb9f058)
Reviewed-on: http://git-master/r/756015
Reviewed-by: Pankaj Dabade <pdabade@nvidia.com>
Reviewed-by: Automatic_Commit_Validation_User
GVS: Gerrit_Virtual_Submit
Reviewed-by: Venkat Moganty <vmoganty@nvidia.com>
Reviewed-by: Matthew Pedro <mapedro@nvidia.com>
|
|
Tune nvhost_podgov scaling algo params for
vic03 frequency scaling.
Bug 1640539
Change-Id: Id5583b5cd60d6b4449470d8c3df1e5d06bc4aedb
Signed-off-by: Arun Kannan <akannan@nvidia.com>
Reviewed-on: http://git-master/r/741438
Reviewed-by: Automatic_Commit_Validation_User
GVS: Gerrit_Virtual_Submit
Reviewed-by: Kamal Balagopalan <kbalagopalan@nvidia.com>
Reviewed-by: Ming Wong <miwong@nvidia.com>
Reviewed-by: Bibek Basu <bbasu@nvidia.com>
Reviewed-by: Matthew Pedro <mapedro@nvidia.com>
|
|
While programming DC controller it is decided at runtime
if a particular mode can be satisfied. This check relies
on the parent pixel clock settings. This patch adds same
check in the mode filter function for HDMI, eliminating
entries from modelist which will not be supported.
Bug 200080781
Change-Id: I6e39249e419bc83d549e561267a57ad25d4652ad
Signed-off-by: Pankaj Dabade <pdabade@nvidia.com>
Reviewed-on: http://git-master/r/741686
Reviewed-by: Automatic_Commit_Validation_User
GVS: Gerrit_Virtual_Submit
Reviewed-by: Miguel Angel Vico Moya <mvicomoya@nvidia.com>
Reviewed-by: Venkat Moganty <vmoganty@nvidia.com>
|
|
The HDMI state machine registers an IRQ handler before all the resources
it needs are available. Ideally the IRQ handler should be registered
later, however it would mean a quite large change. Instead we just add
some check to avoid running the state machine as long as the
framebuffer hasn't been registered.
While at it also prevent scheduling the state machine if it hasn't been
initialized yet.
Bug 1618089
Change-Id: I8f9dc07d2a4bf7e33e869206ad83ed80af93f566
Signed-off-by: Alban Bedel <alban.bedel@avionic-design.de>
Signed-off-by: Bibek Basu <bbasu@nvidia.com>
Reviewed-on: http://git-master/r/741616
GVS: Gerrit_Virtual_Submit
Reviewed-by: Venkat Moganty <vmoganty@nvidia.com>
|
|
Clean L2 by set/ways doesn't ensure that L1 is
cleaned on A15 and A57.
Clean L1 on all cores before cleaning L2.
Bug 20007733
Change-Id: I8a389317571060ba4716663c03ab75896aa13612
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/742255
GVS: Gerrit_Virtual_Submit
Reviewed-by: Bibek Basu <bbasu@nvidia.com>
|
|
Clean cache during page allocations into page pool to
avoid cache clean overhead at the time of allocation.
Increase page pool refill size to 1MB from 512KB.
Bug 1539190
Change-Id: I6c45782e54879541f7b518bbbb016383b24e376b
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/453197
Reviewed-by: Sri Krishna Chowdary <schowdary@nvidia.com>
GVS: Gerrit_Virtual_Submit
Reviewed-by: Alex Waterman <alexw@nvidia.com>
Reviewed-by: Michael I Gold <gold@nvidia.com>
[ccross: moved on top of background zeroing patches, replaced
atomic with bool since it has to be protected by a lock anyways]
Signed-off-by: Colin Cross <ccross@android.com>
Reviewed-on: http://git-master/r/664676
Reviewed-on: http://git-master/r/736430
Tested-by: Alex Waterman <alexw@nvidia.com>
|
|
lockdep complains about the alloc_page call under the pool lock
in nvmap_page_pool_init because the pool lock is also taken during
reclaim in nvmap_page_pool_shrink. Rewrite nvmap_page_pool_init
to perform the allocation in PENDING_PAGES_SIZE chunks outside the
lock. Also move the global pending_pages to a static copy in
nvmap_pp_do_background_zero_pages and a kcalloc'd array during
init to avoid conflicts when two threads try to use the same array.
[ 57.734407] =================================
[ 57.738786] [ INFO: inconsistent lock state ]
[ 57.743198] 3.10.40-ge3b2801-dirty #145 Tainted: G W
[ 57.749135] ---------------------------------
[ 57.753521] inconsistent {RECLAIM_FS-ON-W} -> {IN-RECLAIM_FS-W} usage.
[ 57.760082] kswapd0/41 [HC0[0]:SC0[0]:HE1:SE1] takes:
[ 57.765159] (&pool->lock#2){+.+.?.}, at: [<ffffffc000404d28>] nvmap_page_pool_shrink+0x48/0x9c
[ 57.773953] {RECLAIM_FS-ON-W} state was registered at:
[ 57.779118] [<ffffffc000106964>] mark_lock+0x1d0/0x7a8
[ 57.784938] [<ffffffc000107004>] mark_held_locks+0xc8/0x184
[ 57.791543] [<ffffffc000107bd0>] lockdep_trace_alloc+0xb4/0xf8
[ 57.797594] [<ffffffc00015ce24>] __alloc_pages_nodemask+0x98/0x7f8
[ 57.803987] [<ffffffc0004051d8>] nvmap_page_pool_init+0x19c/0x220
[ 57.810287] [<ffffffc0003fbf20>] nvmap_probe+0x9c/0x78c
[ 57.815713] [<ffffffc00049c588>] platform_drv_probe+0x18/0x24
[ 57.821662] [<ffffffc00049aadc>] driver_probe_device+0xa8/0x3a8
[ 57.827783] [<ffffffc00049aed0>] __driver_attach+0xa0/0xa8
[ 57.833466] [<ffffffc000498d64>] bus_for_each_dev+0x58/0x9c
[ 57.839238] [<ffffffc00049a9d0>] driver_attach+0x1c/0x28
[ 57.844746] [<ffffffc000499828>] bus_add_driver+0x1d8/0x294
[ 57.850517] [<ffffffc00049ba10>] driver_register+0x68/0x174
[ 57.856289] [<ffffffc00049cfe0>] platform_driver_register+0x54/0x60
[ 57.862759] [<ffffffc000d61354>] nvmap_init_driver+0x24/0x40
[ 57.868617] [<ffffffc000d328d4>] do_one_initcall+0xac/0x164
[ 57.874387] [<ffffffc000d32af8>] kernel_init_freeable+0x158/0x1f8
[ 57.880678] [<ffffffc0009e7d40>] kernel_init+0x10/0x158
[ 57.886103] [<ffffffc000084cfc>] ret_from_fork+0xc/0x1c
[ 57.891526] irq event stamp: 803
[ 57.894763] hardirqs last enabled at (803): [<ffffffc0009fc208>] _raw_spin_unlock_irqrestore+0x64/0x94
[ 57.904211] hardirqs last disabled at (802): [<ffffffc0009fbfa8>] _raw_spin_lock_irqsave+0x24/0x9c
[ 57.913202] softirqs last enabled at (0): [<ffffffc0000a508c>] copy_process.part.55+0x448/0x1150
[ 57.922108] softirqs last disabled at (0): [< (null)>] (null)
[ 57.929530]
[ 57.929530] other info that might help us debug this:
[ 57.936074] Possible unsafe locking scenario:
[ 57.936074]
[ 57.942008] CPU0
[ 57.944456] ----
[ 57.946905] lock(&pool->lock#2);
[ 57.950338] <Interrupt>
[ 57.952960] lock(&pool->lock#2);
[ 57.956565]
[ 57.956565] *** DEADLOCK ***
[ 57.956565]
[ 57.962502] 1 lock held by kswapd0/41:
[ 57.966257] #0: (shrinker_rwsem){++++..}, at: [<ffffffc0001651cc>] shrink_slab+0x54/0x3b0
[ 57.974677]
[ 57.974677] stack backtrace:
[ 57.979063] CPU: 0 PID: 41 Comm: kswapd0 Tainted: G W 3.10.40-ge3b2801-dirty #145
[ 57.987434] Call trace:
[ 57.989888] [<ffffffc0000887e4>] dump_backtrace+0x0/0x188
[ 57.995300] [<ffffffc00008897c>] show_stack+0x10/0x1c
[ 58.000363] [<ffffffc0009f3490>] dump_stack+0x1c/0x28
[ 58.005428] [<ffffffc0009f0d6c>] print_usage_bug.part.37+0x28c/0x2a8
[ 58.011799] [<ffffffc0001068f0>] mark_lock+0x15c/0x7a8
[ 58.016951] [<ffffffc000109f50>] __lock_acquire+0x7c0/0xd24
[ 58.022536] [<ffffffc00010ada8>] lock_acquire+0xa8/0x148
[ 58.027862] [<ffffffc0009f870c>] mutex_lock_nested+0x80/0x3e0
[ 58.033623] [<ffffffc000404d24>] nvmap_page_pool_shrink+0x44/0x9c
[ 58.039731] [<ffffffc000165364>] shrink_slab+0x1ec/0x3b0
[ 58.045056] [<ffffffc000168554>] balance_pgdat+0x4d0/0x5d4
[ 58.050555] [<ffffffc0001687d0>] kswapd+0x178/0x45c
[ 58.055448] [<ffffffc0000d0fa0>] kthread+0xd0/0xdc
[ 69.502055] dhd_set_suspend: Remove extra suspend setting
[ 69.787786] dhd_set_suspend: force extra Suspend setting
Change-Id: I41e908da9c51f353300f32c50354b4f48b2424c5
Signed-off-by: Colin Cross <ccross@android.com>
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/664675
GVS: Gerrit_Virtual_Submit
Reviewed-on: http://git-master/r/736429
Reviewed-by: Alex Waterman <alexw@nvidia.com>
Tested-by: Alex Waterman <alexw@nvidia.com>
|
|
The background allocator rapidly becomes useless once the system
has filled memory with cached pages. It refuses to allocate when
free memory < 128MB (which it always is, the kernel is aiming to
keep very little free), and freed pages are not returned to the
page pool when zero_memory=1.
Remove the background allocator completely, and instead return
freed memory to the page pool in a separate list to be zeroed
in the background. This results in a self-balancing pool of
memory available to graphics, and reduces presure on the kernel's
page allocator. If the pool grows too big it will get reduced
by the shrinker. If it gets too small, the next allocation will
fall back to the page allocator, and then later return those pages
to the pool.
Before this change incremental page pool hit rate reported by
/d/nvmap/pagepool/page_pool_hits vs. /d/nvmap/pagepool/page_pool_misses
goes to 0% after boot. After this change it is near 100% for small
app launches and 75% for larger app launches.
Change-Id: I4bc914498d7d0369eef9e621bda110d9b8be90b2
Signed-off-by: Colin Cross <ccross@android.com>
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/664674
GVS: Gerrit_Virtual_Submit
Reviewed-on: http://git-master/r/736428
Reviewed-by: Alex Waterman <alexw@nvidia.com>
Tested-by: Alex Waterman <alexw@nvidia.com>
|
|
Fix a race condition in the background allocator where
wake_up_process could be called just before set_current_state
changed the state to TASK_INTERRUPTIBLE, causing the thread
not to wake. Use a waitqueue instead.
Also make the background allocator nicer by marking it freezable
so it doesn't compete with suspend, and setting it SCHED_IDLE so
it only runs when no other threads want to run.
Change-Id: If95da005bb1fc4c9b5e802d40730803a57057fe1
Signed-off-by: Colin Cross <ccross@android.com>
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/664673
GVS: Gerrit_Virtual_Submit
Reviewed-on: http://git-master/r/736427
Reviewed-by: Alex Waterman <alexw@nvidia.com>
Tested-by: Alex Waterman <alexw@nvidia.com>
|
|
Combine CONFIG_NVMAP_FORCE_ZEROED_USER_PAGES and the zero_memory
modparam into a single option by forcing zero_memory=1 when
CONFIG_NVMAP_FORCE_ZEROED_USER_PAGES is set, and always using
zero_memory to decided whether to zero or not.
Change-Id: I9ce0106cfaea950bd9494b697916fbc2a03329ea
Signed-off-by: Colin Cross <ccross@android.com>
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/664672
GVS: Gerrit_Virtual_Submit
Reviewed-on: http://git-master/r/736426
Reviewed-by: Alex Waterman <alexw@nvidia.com>
Tested-by: Alex Waterman <alexw@nvidia.com>
|
|
The internal pool lock is exported so that nvmap_handle can lock
it, call a *_locked function, and then unlock it. Provide a
version of the *_locked functions that takes the lock, remove
the lock and unlock helpers, and make the lock private to
the pools again.
Change-Id: I5a99753058e43161d50a0c61f3a984655cd7cd35
Signed-off-by: Colin Cross <ccross@android.com>
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/664671
GVS: Gerrit_Virtual_Submit
Reviewed-on: http://git-master/r/736425
Reviewed-by: Alex Waterman <alexw@nvidia.com>
Tested-by: Alex Waterman <alexw@nvidia.com>
|
|
struct page already has a list node that is available to use by
whoever got the page with alloc_page. Use it to keep the free
pages in the pool in a list instead of a circular buffer in an
array.
Change-Id: I0377633be7d620b59daf34799bd4ebc5fd9443fb
Signed-off-by: Colin Cross <ccross@android.com>
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/664670
GVS: Gerrit_Virtual_Submit
Reviewed-on: http://git-master/r/736424
Reviewed-by: Alex Waterman <alexw@nvidia.com>
Tested-by: Alex Waterman <alexw@nvidia.com>
|
|
This reverts commit b1d8c6c9415df111e4af1425a3d84b25c00a9c06.
Change-Id: Ide7e78780722bdd30426089f38155c7cabf28934
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Reviewed-on: http://git-master/r/664669
GVS: Gerrit_Virtual_Submit
Reviewed-on: http://git-master/r/736423
Reviewed-by: Alex Waterman <alexw@nvidia.com>
Tested-by: Alex Waterman <alexw@nvidia.com>
|
|
Fix the parent clock search for the display
controller. It addresses the corner case when
the divider difference is just smaller than the
chosen precision and wrongly assumed zero.
An explicit ceiling function is applied for
corretness.
Bug 1580265
Bug 1552739
Change-Id: I47312116c19813260941abcae42e4067a57f3d2c
Signed-off-by: Ivan Raul Guadarrama <iguadarrama@nvidia.com>
Reviewed-on: http://git-master/r/714972
GVS: Gerrit_Virtual_Submit
Reviewed-by: Pankaj Dabade <pdabade@nvidia.com>
Reviewed-by: Venkat Moganty <vmoganty@nvidia.com>
|
|
Zeroing out mipical register space during DSI
calibration may lead to timeouts with camera(CSI) interface.
Instead, This fix deselects the required CSI pads
for DSI calibration to succeed
Bug 1581879
Change-Id: I535debc233589924a9548a2a1839b617e8d9da0c
Reviewed-on: http://git-master/r/665124
(cherry picked from commit f01329808d2cb0f01d6add88ed6904dfff665d74)
Signed-off-by: Vineel Kumar Reddy Kovvuri <vineelkumarr@nvidia.com>
Signed-off-by: Pankaj Dabade <pdabade@nvidia.com>
Reviewed-on: http://git-master/r/711996
GVS: Gerrit_Virtual_Submit
Reviewed-by: Matthew Pedro <mapedro@nvidia.com>
Reviewed-by: Venkat Moganty <vmoganty@nvidia.com>
|
|
Framebuffer console init requires atleast one valid
mode to exist for it to succeed.
Virtual terminals relies on the success of fbcon_init
without really checking the return value of this function.
Bug 200036424
Change-Id: Ia34da6ab3814f667a485d23ef88e1a52f8b5519b
Signed-off-by: Pankaj Dabade <pdabade@nvidia.com>
Reviewed-on: http://git-master/r/707309
Reviewed-by: Automatic_Commit_Validation_User
GVS: Gerrit_Virtual_Submit
Reviewed-by: Bibek Basu <bbasu@nvidia.com>
Reviewed-by: Venkat Moganty <vmoganty@nvidia.com>
Reviewed-by: Matthew Pedro <mapedro@nvidia.com>
|
|
Check max clock supported by DC before programming it with that mode.
If the requested pixel clock is greater than the maximum supported,
fall back to default mode.
Bug 200031813
Change-Id: I9c5d4373ff0ee8de039af42f46323909b0bec272
Signed-off-by: Pankaj Dabade <pdabade@nvidia.com>
Reviewed-on: http://git-master/r/676941
Reviewed-by: Automatic_Commit_Validation_User
GVS: Gerrit_Virtual_Submit
Reviewed-by: Matthew Pedro <mapedro@nvidia.com>
Reviewed-by: Bibek Basu <bbasu@nvidia.com>
Reviewed-by: Winnie Hsu <whsu@nvidia.com>
|