Currently, each process holding an open TegraDRM channel reserves
for itself one of the limited number of hardware memory contexts.
Attempting to allocate a channel when all contexts are in use
results in failure.
While we cannot have more contexts than the hardware supports in
active use, idle channels don't necessarily need to have a backing
memory context. As such, in this patch, we add another layer
to allow hardware memory contexts to be "stolen away" by channels
that are in active use, from idle processes.
The way this is implemented, is by keeping track of memory
mappings on each abstracted memory context. If we need to
steal that memory context's backing hardware context, we unmap
everything from it and give it away. When that abstracted
memory context is needed again (re-activated), we attempt
to allocate or steal another hardware context and re-map
the previously unmapped buffers.
Unfortunately, this means additional overhead and unpredictability
at submit time. Submit can fail if we cannot re-allocate a
backing memory context. Future work includes a provision for
un-stealable backing hardware memory contexts for processes
requiring more determinism, as well as optimization and cosmetic
improvements.
Bug 4403250
Bug 4399310
Signed-off-by: Mikko Perttunen <mperttunen@nvidia.com>
Change-Id: I3d13e3476f1bff3c4757152254496cddaaafd76a
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3058905
Reviewed-by: Santosh BS <santoshb@nvidia.com>
GVS: Gerrit_Virtual_Submit <buildbot_gerritrpt@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3227749
GVS: buildbot_gerritrpt <buildbot_gerritrpt@nvidia.com>
Tested-by: Dane Liu <danel@nvidia.com>
Reviewed-by: Bibek Basu <bbasu@nvidia.com>
Reduce free memory reported by QueryHeapParams by 0.1% to avoid
OOM issues for Iovmm_heap. For the overall free memory reported
by the query heap params api, page align it to the previous page
to avoid any failures in allocattr due to page align of requested
size, when requested size is equal to reported free size. Update
allocattr api as well so that it checks the available free memory
before allocating.
Bug 4719292
Change-Id: I7ecd69216d58c36ee5d0750107546601400e722d
Signed-off-by: N V S Abhishek <nabhishek@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3165599
(cherry picked from commit 3fdb84d2c8)
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3230712
Reviewed-by: Pritesh Raithatha <praithatha@nvidia.com>
Reviewed-by: Ketan Patil <ketanp@nvidia.com>
GVS: buildbot_gerritrpt <buildbot_gerritrpt@nvidia.com>
Allocate the buffer based on the request instead of a fixed buffer
length. In operations which may require larger buffer size, a fixed
buffer may fail. Similar patch was added for AES algorithms. Fix the
same for HASH algorithms as well.
Bug 4908156
Signed-off-by: Akhil R <akhilrajeev@nvidia.com>
Change-Id: Idd2c1ceae1a85434a5a51154a17dce8c927bb66c
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3234055
GVS: buildbot_gerritrpt <buildbot_gerritrpt@nvidia.com>
Reviewed-by: Laxman Dewangan <ldewangan@nvidia.com>
In Linux v6.8, the 'g_frame_interval' and 's_frame_interval' function
pointers were removed from the 'v4l2_subdev_video_ops' structure and
replaced by 'get_frame_interval' and 'set_frame_interval' that were
added to the 'v4l2_subdev_pad_ops' structure.
This change was previously fixed for the Tegra CSI driver by adding the
necessary conftest changes. A new update to the Tegra Cam V4L2 causes
the build to fail for Linux v6.8+ kernels because of the same issue. Fix
up the Tegra Cam V4L2 driver in the same way as the Tegra CSI driver,
but update the name of the conftest test to indicate that this is
applicable for both the 'get_frame_interval' and 'set_frame_interval'
function pointers.
Bug 4448428
Bug 4807063
Change-Id: Ica47f3aaece0bf81e2adc3c431056ea51b64a893
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3231775
Reviewed-by: svcacv <svcacv@nvidia.com>
Reviewed-by: Brad Griffis <bgriffis@nvidia.com>
GVS: buildbot_gerritrpt <buildbot_gerritrpt@nvidia.com>
The RealTek drivers do not build against all Linux distributions because
some Linux distributions have back-ported upstream changes and so the
KERNEL_VERSION checks in the driver do not work for these distributions.
Add a compilation flag for the rtl8852ce driver so that we can skip
building of this driver for certain Linux distributions.
Bug 4667769
Change-Id: Ie4ec2308b458e4d5e6a3fa29d5e247c5a709172d
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3231107
Reviewed-by: svcacv <svcacv@nvidia.com>
Reviewed-by: Brad Griffis <bgriffis@nvidia.com>
GVS: buildbot_gerritrpt <buildbot_gerritrpt@nvidia.com>
In Linux v6.12, commit 88a2f6468d01 ("struct fd: representation change")
removed the 'struct file' pointer from 'struct fd'. This breaks building
the NVSCIIPC driver that tries to directly access the 'file' pointer
from the 'fd' structure. Fix this by using the helper macros 'fd_empty'
and 'fd_file' as necessary to fix the build.
Bug 4593750
Change-Id: I0f84736b408f533a732476175a8745091bc8542f
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3217429
(cherry picked from commit 0f21c4134eb08dfcafad832d97308f2d98601f1d)
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3226681
Reviewed-by: Brad Griffis <bgriffis@nvidia.com>
GVS: buildbot_gerritrpt <buildbot_gerritrpt@nvidia.com>
In Linux v6.12, commit 8c045ca534d0 ("gpiolib: legacy: Kill GPIOF_DIR_*
definitions") removed the GPIOF_DIR_* definitions and updated drivers to
use the equivalent GPIOF_* definitions instead. The GPIOF_* definitions
were added in Linux v3.0 and so update the appropriate drivers to use
these definitions.
Note that when calling devm_gpio_request_one() with GPIOF_DIR_OUT for
the flags, then because no explicit output level is specified, the GPIO
driver core defaults to low. Hence, in this case we replace
GPIOF_DIR_OUT with GPIOF_OUT_INIT_LOW.
Bug 4593750
Change-Id: I05664fd4e0abf5755c9514dffe64b239266c92fa
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3217397
(cherry picked from commit 0c4fc214d92f2a3c91e9c3b9d1245b4c1ee968f4)
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3226680
GVS: buildbot_gerritrpt <buildbot_gerritrpt@nvidia.com>
Reviewed-by: Brad Griffis <bgriffis@nvidia.com>
In Linux v6.12, commit f9ecc2febf6f ("pwm: Don't export pwm_capture()")
made pwm_capture an internal function and this broke the build for the
Tegra Tachometer driver. The pwm_capture() function simply calls the
drivers '.capture' callback and so fix this by directly calling the
function pwm_tegra_tacho_capture() instead. Note that the rpm_show()
function is also moved to after the declaration of the
pwm_tegra_tacho_capture() function.
Bug 4876974
Change-Id: Idf7fbc16382a9077c651755d9907ded7652610cc
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3217391
(cherry picked from commit 1f752196cab197f708d75bf7a319e23716a685e0)
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3226679
GVS: buildbot_gerritrpt <buildbot_gerritrpt@nvidia.com>
Reviewed-by: Brad Griffis <bgriffis@nvidia.com>
Building the rtl8852ce driver with various different Linux v6.x kernels
fail for various reasons.
For Linux v6.6 the build fails with errors such as ...
drivers/net/wireless/realtek/rtl8852ce/phl/phl_sta.c: In function
'phl_cmd_set_seciv_hdl':
drivers/net/wireless/realtek/rtl8852ce/phl/phl_sta.c:4907:16: error:
implicit conversion from 'enum rtw_hal_status' to
'enum rtw_phl_status' [-Werror=enum-conversion]
4907 | return rtw_hal_set_dctrl_tbl_seciv((void*)phl_info->hal,
| sta, sta->sec_iv);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
For Linux v6.8 the build fails with the above and the following ...
drivers/net/wireless/realtek/rtl8852ce/phl/phl_sta.c:301:5: error:
no previous declaration for '_phl_get_macid'
[-Werror=missing-declarations]
301 | u16 _phl_get_macid(struct phl_info_t *phl_info,
| ^~~~~~~~~~~~~~
For Linux v6.10 the build fails with the above and the following ...
drivers/net/wireless/realtek/rtl8852ce/core/rtw_ap.c:6265:9: error:
suggest braces around empty body in an 'else' statement
[-Werror=empty-body]
6265 | ;
| ^
drivers/net/wireless/realtek/rtl8852ce/core/rtw_sta_mgt.c:742:1: error:
'static' is not at beginning of declaration
[-Werror=old-style-declaration]
742 | u32 static _rtw_free_core_stainfo(_adapter *padapter , struct
| sta_info *psta, u8 aid,
| const u8 *hwaddr)
| ^~~
For Linux v6.12, the driver build is completely broken because of the
following build error ...
drivers/net/wireless/realtek/rtl8852ce/os_dep/linux/wifi_regd.c:1007:17:
error: too few arguments to function 'cfg80211_cac_event'
1007 | cfg80211_cac_event(evt->netdev, &evt->chandef,
| evt->event, GFP_KERNEL);
| ^~~~~~~~~~~~~~~~~~
For Linux v6.6+ mark the driver as broken and do not build this for
Linux v6.6+ kernels until these issues are addressed.
Bug 4667769
Change-Id: Id6b060f6b39ba4ef64d6388e06ef1e038c19206f
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3226276
Reviewed-by: svcacv <svcacv@nvidia.com>
GVS: buildbot_gerritrpt <buildbot_gerritrpt@nvidia.com>
Reviewed-by: Brad Griffis <bgriffis@nvidia.com>
Some distros might use old kernel source but with some latest upstream
kernel patches backported to their kernel source tree. To deal with this
scenario and avoid kernel compilation failure, use conftest to check the
existence of features against the kernel source tree which the OOT
modules are built upon and do the conditional build based on the test
result generated with the conftest tool.
Use NV_DEVFREQ_HAS_FREQ_TABLE to determine whether freq_table field is
there in struct devfreq data strcuture, and choose the correct path for
building the module.
Use tegra_wmark-specific devfreq_get_freq_range always to avoid kernel
version check and conftest check. Since devfreq_get_freq_range exists in
the devfreq-specific governor.h (e.g. drivers/devfreq/governor.h) file
instead of globabl linux kernel include header files
(e.g. include/linux/devfreq.h), conftest cannot be used to the existence
of devfreq_get_freq_range kernel function.
Bug 4884092
Change-Id: I5bde4c712f59f38de74c1d8d95135c9b25d621b1
Signed-off-by: Johnny Liu <johnliu@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3220896
Tested-by: Paritosh Dixit <paritoshd@nvidia.com>
GVS: buildbot_gerritrpt <buildbot_gerritrpt@nvidia.com>
Reviewed-by: Jon Hunter <jonathanh@nvidia.com>
For Linux v6.11, the 'pci_epc_event_ops' structure 'core_init' callback
was renamed to 'epc_init' and a new 'epc_deinit' callback was added.
Update the PCI endpoint drivers for Linux v6.11 by:
1. Adding tests to conftest to detect the presence of 'epc_init' and
'epc_deinit'.
2. Update the drivers as necessary to use the appropriate callback
depending upon what is present in the kernel.
Bug 4749580
Change-Id: Ife7e2d5d18332bf84f22142c9727dc1ece913c49
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3186686
(cherry picked from commit b0ff4f1358)
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3219305
GVS: buildbot_gerritrpt <buildbot_gerritrpt@nvidia.com>
Reviewed-by: Brad Griffis <bgriffis@nvidia.com>
In Linux v6.11, the 'platform_driver' structure 'remove' callback was
updated to return void instead of 'int'. This change impacts all OOT
platform drivers and so instead of editing each manually, add a
coccinelle script to handle updating all platform drivers. This script
is ran by executing the following command ...
$ spatch --in-place \
--sp-file /path/to/scripts/coccinelle/platform_driver_remove.cocci
<driver-source-directory>
Finally, add a test to the conftest script that is used to detect if the
platform driver 'remove' callback returns 'void'.
Bug 4749580
Change-Id: I53d25cc36cc3be97a6eeeb4ccbdf9b7dfda338a9
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3176834
(cherry picked from commit 24826db83f)
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3202689
Reviewed-by: Brad Griffis <bgriffis@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
Reviewed-by: Brad Griffis <bgriffis@nvidia.com>
Reviewed-by: Pritesh Raithatha <praithatha@nvidia.com>
GVS: buildbot_gerritrpt <buildbot_gerritrpt@nvidia.com>
Reviewed-by: Ashish Mhetre <amhetre@nvidia.com>
Release Notes:
1. Add extra queue to handle EAPOL
Mark include/autoconf.h RTW_EAPOL_QUEUE to disable it
2. Use xmit_ext queue to TX eapol packet
3. Flush roam_buf_pkt after roaming is fail
4. Do NOT roam if previous roam does NOT finish
5. Report to WPS after all roam retries are failed
6. Support 11K beacon report fragmentation
7. Fix compile error on Kernel 5.19.2
Bug 4556940
Change-Id: I578ec882584d5d18dfbb8b5a5513ddbe733367b9
Signed-off-by: Shobek Attupurath <sattupurath@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3140393
Reviewed-by: Ashutosh Jha <ajha@nvidia.com>
GVS: buildbot_gerritrpt <buildbot_gerritrpt@nvidia.com>
Reviewed-by: Revanth Kumar Uppala <ruppala@nvidia.com>
In Linux v6.11, commit 1a251f52cfdc ("minmax: make generic MIN() and
MAX() macros available everywhere") causes the PVA driver build to fail
with the following error:
include/uapi/linux/nvpva_ioctl.h:601: error: "MAX" redefined [-Werror]
601 | #define MAX(a, b) ((a) > (b) ? (a) : (b))
Fix this by add guards around the definition of 'MAX' in the
nvpva_ioctl.h file.
Bug 4749580
Change-Id: If7960e0617e659f4e264b1e638e9f8c98a027a73
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3184093
(cherry picked from commit d8bc5883f3)
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3185779
GVS: buildbot_gerritrpt <buildbot_gerritrpt@nvidia.com>
Reviewed-by: Laxman Dewangan <ldewangan@nvidia.com>
The Tegra virtual storage driver fails to build for Linux v6.11 because
of the following issues:
1. The functions blk_queue_logical_block_size(),
blk_queue_physical_block_size() and blk_queue_max_discard_sectors()
have been removed.
2. The function blk_queue_write_cache() has too many arguments for
Linux v6.11.
3. The flag QUEUE_FLAG_NONROT is not defined for Linux v6.11.
Disable this driver for Linux v6.11 until this driver is fixed.
Bug 4311184
Bug 4749580
Change-Id: I56f811d6ee29dd0025a5dc199ae84e3fc0f37b21
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3185049
Reviewed-by: Brad Griffis <bgriffis@nvidia.com>
Reviewed-by: Laxman Dewangan <ldewangan@nvidia.com>
GVS: buildbot_gerritrpt <buildbot_gerritrpt@nvidia.com>
In Linux v6.11, commit 1a251f52cfdc ("minmax: make generic MIN() and
MAX() macros available everywhere") causes the Tegra ASoC Utils driver
build to fail with the following error:
sound/soc/tegra/tegra_asoc_utils.c:18: error: "MAX" redefined [-Werror]
18 | #define MAX(X, Y) ((X > Y) ? (X) : (Y))
Fix this by add guards around the definition of 'MAX' in the
Tegra ASoC Utils file.
Bug 4749580
Change-Id: I63df434627e6ee7ab16d865c296b07b91405fdfd
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3185143
Reviewed-by: Sameer Pujar <spujar@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>
Bottom two lines in IMX390 SDR mode are completely
black. Fix issue in SDR mode table. Also change the
image resolution to 1936x1096.
Last two line are removed by disabling SMPG(Safety Mechanism Pattern Generator) using register 0x2DFC, 0x2E24.
Bug 4505240
Change-Id: I92047ea83d20f1b4ca4e712718051a9c6e69cfcb
Signed-off-by: Ankur Pawar <ankurp@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/3147214
GVS: buildbot_gerritrpt <buildbot_gerritrpt@nvidia.com>
Reviewed-by: Praveen AC <pac@nvidia.com>
Reviewed-by: Laxman Dewangan <ldewangan@nvidia.com>
