Account NvMap allocated memory into both RSS and CG tracking to make
efficient OOM kill decisions during memory pressure.
NvMap allocates memory via kernel APIs like alloc_pages, the kernel
memory is not accounted on behalf of process who requests the
allocation. Hence in case OOM, the OOM killer never kills the process
who has allocated memory via NvMap even though this process might be
holding most of the memory.
Solve this issue using following approach:
- Use __GFP_ACCOUNT and __GFP_NORETRY flag
- __GFP_NORETRY will not let the current allocation flow to go into OOM
path, so that it will never trigger OOM.
- __GFP_ACCOUNT causes the allocation to be accounted to kmemcg. So any
allocation done by NvMap will be definitely accounted to kmemcg and
cgroups can be used to define memory limits.
- Add RSS counting for the process which allocates by NvMap, so that OOM
score for that process will get updated and OOM killer can pick this
process based upon the OOM score.
- Every process that has a reference to NvMap Handle would have the
memory size accounted into its RSS. On releasing the reference to
handle, the RSS would be reduced.
Bug 5222690
Change-Id: I3fa9b76ec9fc8d7f805111cb96e11e2ab1db42ce
Signed-off-by: Ketan Patil <ketanp@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3427871
(cherry picked from commit 2ca91098aa)
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3453101
Tested-by: Amulya Yarlagadda <ayarlagadda@nvidia.com>
Reviewed-by: Amulya Yarlagadda <ayarlagadda@nvidia.com>
GVS: buildbot_gerritrpt <buildbot_gerritrpt@nvidia.com>
nvmap uses pid of group_leader task to indicate a client process. During
create_client operation, whenever any client with the same group_leader
pid already exists in clients list of nvmap_device, then nvmap
increments the count field of nvmap_client struct. Otherwise, create a
new nvmap_client. Both of the operations i.e. checking the list for
client and incrementing the counter happen inside lock. On the other
hand, during nvmap_release, first the counter is decremented and checked
if it's zero or not. If it's zero then the lock is taken and client is
removed from client list of nvmap_device. As both the operations i.e.
decrementing the counter value and removing client from list (if the
counter becomes 0) are not happening inside a lock, it's resulting into
the following data race scenario.
1) nvmap_release on existing client process 1
- decrement client's counter
- counter value has become zero
- client is yet to be removed from the dev->clients list
- context switch happen to __nvmap_create_client as another
namespace/thread with same with same group_leader pid is created.
2) __nvmap_create_client
- as the client with same pid exists in dev->client list, it
increments counter value to 1, instead of creating a new client struct.
- context switch happen to nvmap_release from step 1
3) nvmap_release
- It calls destroy_client and remove the client from dev->client
list.
- Completes rest of the operations in destroy_client and returns.
- Context switch to remaining operations from step 2
4) nvmap_release
- Now, when the nvmap_release will be called for the thread/namespace
which was created in step 2, then list_del operation would fail as the
client struct was already removed from dev->client list.
Fix the above issue by doing both operations i.e. decrementing the
counter value and removing the client struct from dev->client list in a
single lock.
Bug 4829958
Change-Id: I87ebbcb45b18114d0ec75520443bee010f88d59e
Signed-off-by: Ketan Patil <ketanp@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3209794
(cherry picked from commit cc74d1fe1b)
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3210845
Reviewed-by: Brad Griffis <bgriffis@nvidia.com>
Reviewed-by: Amulya Yarlagadda <ayarlagadda@nvidia.com>
Tested-by: Brad Griffis <bgriffis@nvidia.com>
GVS: buildbot_gerritrpt <buildbot_gerritrpt@nvidia.com>
Upstream commit cdd9a571b7d8 ("fs/proc: move page_mapcount() to
fs/proc/internal.h") made page_mapcount() an internal function for Linux
v6.11. This function is used by the NVMAP driver and so implement this
function in the NVMAP driver directly instead. Note that newer kernels
implement folio_entire_mapcount() for compound page counts where as
older kernels use compound_mapcount(). Use conftest to detect which of
these functions is present in the kernel.
Bug 4749580
Change-Id: I6504448727d6b3e9f2caa8581a66aad464ae2426
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3180097
(cherry picked from commit f327656ea3)
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3183926
GVS: buildbot_gerritrpt <buildbot_gerritrpt@nvidia.com>
Reviewed-by: Ketan Patil <ketanp@nvidia.com>
Upstream commit 690da22dbfa8 ("asm-generic/io.h: kill vmalloc.h
dependency") removed the vmalloc.h header file from io.h and this breaks
building various drivers with the latest -next kernels. Fix this
by ensuring vmalloc.h is included and in most cases slab is not actually
needed and so remove this where possible. Note that it is fine to make
this change for all current supported kernels.
Bug 4593750
Change-Id: I003d1302bda226d356467e6ede99949b2716940a
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3141984
(cherry picked from commit cb8b9f5a53)
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3123218
Reviewed-by: Brad Griffis <bgriffis@nvidia.com>
GVS: buildbot_gerritrpt <buildbot_gerritrpt@nvidia.com>
When the tsec driver is built with the compiler flag
-Werror=old-style-declaration the following error is observed ...
error: 'static' is not at beginning of declaration
[-Werror=old-style-declaration]
393 | const static struct dev_pm_ops tsec_module_pm_ops = {
| ^~~~~
This is observed with the latest linux-next kernels, but could be
observed with any kernel where this flag is set. Fix this by ensuring
that the 'static' keyword is declared first.
Bug 4593750
Change-Id: Ia635188569d827d361bd36f7be943fbf9f1f0a60
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3126898
GVS: Gerrit_Virtual_Submit <buildbot_gerritrpt@nvidia.com>
Reviewed-by: Brad Griffis <bgriffis@nvidia.com>
Kernel freeze is occurring due to incorrect cleanup
in vi5.c.
Fix for kernel freeze:
1 Don't call vi_channel_drv_unregister() and
tegra_vi_media_controller_cleanup() in vi5_remove.
2 capture_vi_probe() is calling vi_channel_drv_register() and
tegra_capture_vi_media_controller_init(), so capture_vi_remove()
should call vi_channel_drv_unregister() and
tegra_vi_media_controller_cleanup()
3 Use correct dev pointer in vi_channel_drv_unregister().
Code cleanup:
1 Remove use_legacy_path variable that was used for T186.
2 Removed unused t194_vi5_file_private struct.
Bug 4415340
Change-Id: I6153f65d62a9f7f4fc5d04c2ace948a29396e404
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3077660
Reviewed-by: Shubham Chandra <shubhamc@nvidia.com>
Reviewed-by: Ankur Pawar <ankurp@nvidia.com>
Reviewed-by: Laxman Dewangan <ldewangan@nvidia.com>
GVS: Gerrit_Virtual_Submit <buildbot_gerritrpt@nvidia.com>
Tested-by: Ankur Pawar <ankurp@nvidia.com>
The function nvhost_get_private_data() is a wrapper around the Linux
platform_get_drvdata() function. To remove all the legacy nvhost code
that is no longer supported or needed, update the camera drivers to use
platform_get_drvdata() directly. Note that there are places in these
drivers that already use platform_get_drvdata() and so this change makes
these drivers more consistent.
Bug 4475969
Change-Id: Ie474c46c438f69c2141f41d0702bf57525597ab5
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3065974
(cherry picked from commit 73a4985dbd)
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3066024
Reviewed-by: Laxman Dewangan <ldewangan@nvidia.com>
GVS: Gerrit_Virtual_Submit <buildbot_gerritrpt@nvidia.com>
When dla resumes back from system sleep, it should not send any
bandwidth request through ICC framework since it is in idle state.
In the job submit path for dla, nvhost will try to wakeup the device
through nvhost_module_busy, and the associated runtime resume callback
is called in the dla driver. This is the time dla driver should send
bandwidth request to BPMP for bandwidth allocation through ICC
framework.
Bug 4478515
Signed-off-by: Johnny Liu <johnliu@nvidia.com>
Change-Id: Ie4b5cfad6593bd058860ee3be467d6a2384f1b5d
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3067369
Reviewed-by: Rajkumar Kasirajan <rkasirajan@nvidia.com>
GVS: Gerrit_Virtual_Submit <buildbot_gerritrpt@nvidia.com>
The compiler option -Wmissing-prototypes is being enabled globally in
the upstream Linux kernel and this causes build failures for various
drivers. The build failures occur because either the driver is missing
an include file which has the prototype or because the function is not
declared statically when it should be (ie. there are no external users).
Fix the various build failures and enable -Wmissing-prototypes to
prevent any new instances from occurring. Note that the only driver that
is not fixed at the moment is rtl8822ce due to large number of failures
and so build this with '-Wno-missing-prototypes' for now, which is not
different to how it was being compiled prior to this change.
Bug 4404965
Change-Id: Ie5572d23659e0346fa035d645d9043b0a6da5fdc
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3027488
(cherry picked from commit d5391a25ab)
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3034644
Reviewed-by: Laxman Dewangan <ldewangan@nvidia.com>
GVS: Gerrit_Virtual_Submit <buildbot_gerritrpt@nvidia.com>
The Nvhost-related APIs have been deprecated and are no longer
supported since our migration from K510 to K515. Therefore, the camera
platform driver should refrain from using "nvhost_*" functions to set
the clock rate.
Integrate the common operation callback interface into the
tegra_camera_dev_info structure. This modification allows individual
drivers to achieve clock set rate operations by registering the
callback and completing the implementation.
Bug 4374126
Signed-off-by: Johnny Liu <johnliu@nvidia.com>
Change-Id: Iebed7cc974f5dbdabddded3c84d0925de3a6c4f9
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3020535
(cherry picked from commit 57a82628aa)
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3023111
Reviewed-by: svcacv <svcacv@nvidia.com>
Reviewed-by: Bibek Basu <bbasu@nvidia.com>
GVS: Gerrit_Virtual_Submit <buildbot_gerritrpt@nvidia.com>
Consider the following execution scenario, it can lead to a data race
Thread 1 Thread 2
-------------------------------------------------------
NvRmMemHandleAllocAttr
NvRmMemGetFd
NvRmMemHandleFromFd close(fd), NvRmMemHandleFree
When thread 1 is executing NvRmMemHandleFromFd while thread 2 is closing
the fd and freeing the handle, then following sequence will lead to
accessing a freed dma-buf and may lead to dereference issue.
- TH1: Get handle from fd and increment handle's refcount.
- TH2: Close the fd and start execution of NvRmMemHandleFree.
- TH2: Decrement ref's dup count, it will become 0, hence ref would be
freed and dma-buf as well, but as handle's refcount is incremented in
step 1, handle won't be freed.
- TH1: Resume HandleFromFd part, call to nvmap_duplicate_handle. Ref is
already freed, so generate new ref and increment dma-buf's count but as
dma-buf is freed already, accessing dma-buf will lead to dereferene
issue. Hence, we need to add a null check here and return error value in
such scenario.
Also, add check for return value of nvmap_handle_get, at missing places.
Bug 4214453
Change-Id: Ib6ef66b4a7126bef2ed1dbb48643445a4ded1bab
Signed-off-by: Ketan Patil <ketanp@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/2994962
(cherry picked from commit 7ba6a3abfe65f7cffeeb4004cf0868468121fc32)
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/2994440
Reviewed-by: Sachin Nikam <snikam@nvidia.com>
GVS: Gerrit_Virtual_Submit <buildbot_gerritrpt@nvidia.com>
Use devm_of_icc_get() in rtcpu base and debug driver.
CONFIG_TEGRA_T23X_GRHOST, CONFIG_TEGRA_ISOMGR,
CONFIG_TEGRA_BWMGR are not enabled in K5.15 so cleanup
tegra_camera_platform driver.
Don't read memory-bw from rtcpu dt, set it to max
during probe.
Bug 3997304
Bug 4311411
Change-Id: Ib76b3338749de5d33e13cd518375a6b55dd17f5b
Signed-off-by: Ankur Pawar <ankurp@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/2899021
Reviewed-by: Jonathan Hunter <jonathanh@nvidia.com>
Reviewed-by: Johnny Liu <johnliu@nvidia.com>
Reviewed-by: Laxman Dewangan <ldewangan@nvidia.com>
GVS: Gerrit_Virtual_Submit <buildbot_gerritrpt@nvidia.com>
remove dependency on 0xFF flag in block height field as
a performance optimization.
fold processing of descriptors into channel processing loop
in order to be able to traverse the descriptor list.
second call to nvpva_task_dma_desc_mapping is to process
descriptors not attached to any channels.
Bug 4185299
Change-Id: Ic5d3c60bca7de13698b83365fe3b0cedbe42c60b
Signed-off-by: omar <onemri@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/2979303
GVS: Gerrit_Virtual_Submit <buildbot_gerritrpt@nvidia.com>
Reviewed-by: Vivek Bangera <vbangera@nvidia.com>
