Issue: MAC registers are not getting reset on ether_close(),
which lead to old values in MAC registers at the time of
osi_hw_core_init().
Fix: Assert MAC-RST gpio and disable clocks in ether_close().
Enable clock, reset MAC_RST gpio and poll for swr, which will
reset all MAC registers.
Note: To read MAC registers, you must have clocks enable and MAC
out of reset
Bug 200512422
Change-Id: If253eff0ae456702d3cdcbe1f177dd91a5aae20d
Signed-off-by: Rakesh Goyal <rgoyal@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2138031
Reviewed-by: Narayan Reddy <narayanr@nvidia.com>
GVS: Gerrit_Virtual_Submit
Reviewed-by: Sachin Nikam <snikam@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Issue: Tegra-eqos thermal cooling device is registered
during driver probe(). Once this registration is done,
callbacks via cooling_device_ops can be invoked at any time.
This implies even if driver is just probed, callbacks can try
to trigger a pad calibration due to temperature change and
result in failure trying to access MAC registers when the
MAC netdev interface is not up.
Fix: Move cooling device register/unregister to ether_open/
ether_close() routines respectively, so that callbacks can
be invoked only when interface is actually up.
Bug 1679250
Change-Id: Iaf181ceb3af4b9def188171606d9a9c141d06ccc
Signed-off-by: Srinivas Ramachandran <srinivasra@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2138382
Reviewed-by: Narayan Reddy <narayanr@nvidia.com>
GVS: Gerrit_Virtual_Submit
Reviewed-by: Bitan Biswas <bbiswas@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Updated IOCTL path to not get impacted with new API
change. OSI call should be done with correct parameters
1) dma_routing_enable
2) dma_chan and
3) addr_mask (for L2 filtering) to support new IOCTL
4) src_dest (for L2 filtering) to support new IOCTL
Included logic for User priority based RxQ selection. We can have DT
entry "nvidia,rx_queue_prio" to configure this setting
i.e. for valid following Queue mapping
Priorities 6,3 to Queue 3
Priorities 4,5 to Queue 2
Priority 0 to Queue 1 and
Priority 1 to Queue 0.
setting will be nvidia,rx-queue-prio = <0x2 0x1 0x30 0x48>;
If static channel selection, i.e. RXQtoDMA set with value 0x03020100,
Priorities 6,3 to Queue 3 to chan 3
Priorities 4,5 to Queue 2 to chan 2
Priority 0 to Queue 1 to chan 1 and
Priority 1 to Queue 0 to chan 0
else if DCS(dma channel selection) enabled
channel will be selected based on filter rules.
Included logic to take input from User On DCS enable or disable. User can
update dt entry "nvidia,dcs-enable", for dcs enable disable for all queue.
i.e nvidia,dcs-enable = <0x1> will enable DCS for all queue.
Bug 200525721
Change-Id: I4ba820f178b03424d01bb4ddd1f1d6eadde572f7
Signed-off-by: Rakesh Goyal <rgoyal@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2133196
GVS: Gerrit_Virtual_Submit
Reviewed-by: Narayan Reddy <narayanr@nvidia.com>
Reviewed-by: Srinivas Ramachandran <srinivasra@nvidia.com>
Reviewed-by: Sachin Nikam <snikam@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
These stats are read from MAC HW RMON counters as well as SW controlled
control from Core and DMA path.
ethtool -S <interface> is used to get statistics.
There are 3 stats
1) ether_mmc_counters: EQOS HW counters
2) ether_xtra_stat_counters: SW counters from osi/core
3) ether_xtra_dma_stat_counters: SW counters from osi/dma
Bug 200519211
Change-Id: I5b1b6592541c650f01e845667eca2f5c59bcfd08
Signed-off-by: Rakesh Goyal <rgoyal@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2113962
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Ethernet pad calibration needs to be triggered for temperature
changes in steps of 35C from -40C to 110C. In order to get
indication of current operating temperature, register a ethernet
cooling device with thermal zones defined as per above requirement.
When the temperature trips these zones, callback function is invoked
from kernel thermal framework, and ethernet driver can trigger pad
calibration.
Bug 1679250
Change-Id: Iaeccca650e371843fa571f7b368bb5464e106314
Signed-off-by: Srinivas Ramachandran <srinivasra@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2120431
GVS: Gerrit_Virtual_Submit
Reviewed-by: Ajay Gupta <ajayg@nvidia.com>
Reviewed-by: Ashutosh Jha <ajha@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
1. ARP offload can be configured via private ioctl.
Application requesting the arp offload configuration
should provide the IP address to be configured for
ARP offload via ioctl data struct. Refer to
struct ether_ifr_data and
struct arp_offload_param for details.
2. Tx/Rx Checksum offload can be configured via ethtool
3. TCP Segmentation offload can be configured via ethtool
Bug 2571001
Change-Id: If639ab9049db97f200911af456ddb8cb8433fa12
Signed-off-by: Srinivas Ramachandran <srinivasra@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2109676
Reviewed-by: Rakesh Goyal <rgoyal@nvidia.com>
Reviewed-by: Narayan Reddy <narayanr@nvidia.com>
GVS: Gerrit_Virtual_Submit
Reviewed-by: Ashutosh Jha <ajha@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Add dma-mask parsing in device tree for eqos
driver if IOMMU is enabled. This will allow
memory allocations from the specified address
and downwards, while device can reach every part
of physical memory through the IOMMU.
dma-mask will be remain default 32-bit mask if no
IOMMU is present since now DMA mask represents a
fundamental limit of the device.
Bug 200458098
Change-Id: I6eef70f97dba5807d6483cbd76e474d321debe75
Signed-off-by: Narayan Reddy <narayanr@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2096023
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Adds support for dynamic MAC address change.
Gets MAC address from userspace and update in
OSI core structure. Automatically MAC address
will be picked up from OSI core structure and
program the same in HW registers.
Bug 200512251
Change-Id: Ie49fd56c3d219637fc4bf33c818c380a8f1c7bf3
Signed-off-by: Bhadram Varka <vbhadram@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/2080718
Reviewed-by: Rakesh Goyal <rgoyal@nvidia.com>
Reviewed-by: Narayan Reddy <narayanr@nvidia.com>
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Srinivas Ramachandran <srinivasra@nvidia.com>
GVS: Gerrit_Virtual_Submit
Reviewed-by: Ashutosh Jha <ajha@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
1) support for driver probe based on nveqos
2) support for EQOS hw init
3) data transmission support for Tx and Rx
Bug 200507585
Change-Id: I5c8bc88f97b2e61a2b346a23aadcb19a7b8c53a5
Signed-off-by: Bhadram Varka <vbhadram@nvidia.com>
Using this patch we are adding support for
hv-vse driver in oot kernel.
List of commit ids still which we are using file from
nvidia repo.
4eb5988cde vse-safety: align gpcdma buffer to 64 bytes
4852794a4a crypto: fix array out-of-bound access defect
56e33f53c8 crypto: extend features in multi-IVC development
0b7c238968 (multi_ivc) crypto: check crypto ctx before handling
7e35a5739b crypto: check crypto request before handling
446c88152f Revert "crypto: vse: add multi-ivc support-10"
c33753cbd6 crypto: disable Power management for GPCDMA cfg
JIRA ESLC-6885
Signed-off-by: Manish Bhardwaj <mbhardwaj@nvidia.com>
Change-Id: I77c5abcff045e444ff320fbf2709b07f0ab383aa
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/2794878
Reviewed-by: Leo Chiu <lchiu@nvidia.com>
Reviewed-by: Sandeep Trasi <strasi@nvidia.com>
GVS: Gerrit_Virtual_Submit <buildbot_gerritrpt@nvidia.com>
Using this patch we are converting hypervisor driver
as platform driver since insertion of this module
is getting failed on native kernel due to
is_tegra_hypervisor_mode API failure since
hypervisor environment is missing, leading to
nvethernet module insertion causing the error
"Unknown Symbol tegra_hv_ivc_reserve".
Bug 3793131
Change-Id: I626c62aa0ada15c41283fb1875c351cada16ac51
Signed-off-by: Manish Bhardwaj <mbhardwaj@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/2798141
Reviewed-by: Suresh Venkatachalam <skathirampat@nvidia.com>
Reviewed-by: Sandeep Trasi <strasi@nvidia.com>
GVS: Gerrit_Virtual_Submit <buildbot_gerritrpt@nvidia.com>
1. Add tsec device entry in the DT overlay file
2. Copy over the existing tsec driver source code
and compile it when building with oot kernel
3. Add a dummy tsec folder with an empty Makefile
which conveys that common driver source is used
from kernel/nvidia/drivers/video/tegra/tsec
Bug 3817626
Change-Id: I7b3064fc08bf7507b7660985564bc6bc8eba86eb
Signed-off-by: Nikesh Oswal <noswal@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/2793265
Reviewed-by: Bharat Nihalani <bnihalani@nvidia.com>
Reviewed-by: Mikko Perttunen <mperttunen@nvidia.com>
GVS: Gerrit_Virtual_Submit <buildbot_gerritrpt@nvidia.com>
When compiling the MTTCAN driver for Linux v5.19 the build fails because
the can_led_xxx() functions are not found. The can_led_xxx() functions
were removed for Linux v5.19 and not Linux 6.0 and so update the driver
accordingly to fix this.
Bug 3820317
Change-Id: I340cc9cbb44b5cb892c50cc0b5d06444ef894fe9
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/2792549
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Building the Tegra VBLK driver fails on various Linux v5.x versions due
to mismatches in arguments. Reviewing the Linux history the arguments to
the following functions changed in the following Linux versions ...
1. Linux v5.15
- The function __alloc_disk_node() is updated to pass an additional
argument of type 'struct request_queue'.
2. Linux v5.16
- The functions blk_get_request() and blk_put_request() are replaced by
blk_mq_alloc_request() and blk_mq_free_request(), respectively.
3. Linux v5.17
- Removes the argument of type 'struct gendisk' from the function
blk_execute_rq().
4. Linux v5.18
- Removes the header 'linux/genhd.h'.
5. Linux v6.0
- Replaces the function __alloc_disk_node() with
blk_mq_alloc_disk_for_queue().
Update the Tegra VBLK driver to ensure that the correct functions and
header files are used per the exact Linux version.
Bug 3820317
Change-Id: Ia04ea78d1edcf7674030379159599949f13f3120
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/2791669
Reviewed-by: Manish Bhardwaj <mbhardwaj@nvidia.com>
Reviewed-by: Laxman Dewangan <ldewangan@nvidia.com>
GVS: Gerrit_Virtual_Submit <buildbot_gerritrpt@nvidia.com>
The Aurix SPI driver fails to build with Linux v5.18 because the return
type of the remove function is not correct. This was previously fixed
for the Linux v5.19 build, but the change in the upstream kernel
actually occurred in Linux v5.18.
Bug 3820317
Change-Id: I2588fe33af6877c1c5c61a2a1d9195248066bbd1
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/2792537
Reviewed-by: Laxman Dewangan <ldewangan@nvidia.com>
GVS: Gerrit_Virtual_Submit <buildbot_gerritrpt@nvidia.com>
Build the Tegra PCI EPF driver fails with Linux v5.14 because the
pci_epf_alloc/free_space() have the incorrect number of arguments.
The arguments to the pci_epf_alloc/free_space() functions were updated
in Linux v5.12 but the current Tegra PCI EPF driver assumes that they
were updating after Linux v5.14. Update the Tegra PCI EPF driver to
use the correct arguments for v5.12 onwards.
Bug 3820317
Change-Id: Ieba44bfc22e0f978a5ff227b6d07a6a56ad23f3b
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/2791670
Reviewed-by: Nagarjuna Kristam <nkristam@nvidia.com>
Reviewed-by: Laxman Dewangan <ldewangan@nvidia.com>
GVS: Gerrit_Virtual_Submit <buildbot_gerritrpt@nvidia.com>
Upstream Linux commit f59ac19b7f44 ("thermal/of: Remove old OF code")
removes the structure "thermal_zone_of_device_ops" and function
devm_thermal_zone_of_sensor_register() which breaks building the pex9749
driver with Linux v6.1. Update the pex9749 driver as necessary to fix
building with Linux v6.1.
Bug 3831575
Change-Id: I88ed29142a7abab3f00e57f55d0a8b180a55779c
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/2793518
GVS: Gerrit_Virtual_Submit <buildbot_gerritrpt@nvidia.com>
Reviewed-by: Laxman Dewangan <ldewangan@nvidia.com>
Upstream Linux kernel commit f2042ed21da7 ("iommu/dma: Make header
private") made the header "linux/dma-iommu.h" a private header file and
so compiling the Tegra PCI EDMA driver for Linux v6.1-rc1 now fails
because this header is not found. On further review, instead of
including "linux/dma-iommu.h" it is only necessary to include
"linux/dma-mapping.h" instead.
Bug 3831575
Change-Id: I3552eca64071bba98bb322377d7c211e3c22e7b6
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/2793515
GVS: Gerrit_Virtual_Submit <buildbot_gerritrpt@nvidia.com>
Reviewed-by: Laxman Dewangan <ldewangan@nvidia.com>
Upstream Linux kernel commit b48b89f9c189 ("net: drop the weight
argument from netif_napi_add") removes the weight argument from the
netif_napi_add() function and this is breaking the build of a few
drivers that use this function. Fix this by updating these drivers to
use the netif_napi_add_weight() function instead for Linux v6.1.
Bug 3831575
Change-Id: I5ab3c2380db75313f9226c68777e6245e9e9dad5
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/2793511
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Upstream Linux kernel commit b88c48bfdd85 ("pwm: core: Get rid of unused
devm_of_pwm_get()") has removed the function devm_of_pwm_get() which
breaks building the Tegra PWM tachometer driver with Linux v6.1-rc1.
Fix this by updating the driver to use the function
devm_fwnode_pwm_get() instead.
Bug 3831575
Change-Id: I89def599312baa4d3b6b25b1275835233e942af0
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/2793510
Reviewed-by: Laxman Dewangan <ldewangan@nvidia.com>
GVS: Gerrit_Virtual_Submit <buildbot_gerritrpt@nvidia.com>
When building NVGPU and Host1x-Nvhost driver as OOT modules it is not
necessary to copy the host1x-next.h and nvmap_exports.h header files
because these headers files are already present in the source
repository. Fix this by updating the include paths used for building
these drivers and remove the copy of these headers.
Bug 3817518
Change-Id: I0084055a27d3547642c034546a4b7192e8da52e6
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/2785906
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Make mc-hwpm module oot. Below are the reasons why
making a fresh copy in nvdia-oot folder -
- hwpm is introduced from T234 onwards
- mc-hwpm is dependent on mc driver. nvidia-oot
uses upstream mc driver. Hence mc-hwpm should also
use upstream mc driver.
- Thats why porting changes needed in mc-hwpm driver
to be incompatible with upstream mc driver
Bug 3728991
Change-Id: I2d7347d968895e080fa79d100b4c50c094ced1e3
Signed-off-by: Puneet Saxena <puneets@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/2781118
Reviewed-by: Ketan Patil <ketanp@nvidia.com>
Reviewed-by: Bitan Biswas <bbiswas@nvidia.com>
Reviewed-by: Sachin Nikam <snikam@nvidia.com>
GVS: Gerrit_Virtual_Submit <buildbot_gerritrpt@nvidia.com>
Port the downstream nvpmodel driver to the OOT project.
The following items summarize the change:
- Original bwmgr APIs are removed
- Clock cap for other clocks except for EMC is removed since userspace
nvpmodel service will handle clock cap for other clocks
Bug 3779573
Signed-off-by: Johnny Liu <johnliu@nvidia.com>
Change-Id: Ib297dfed808dad03d1450860c24156da1a84b4ca
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nv-oot/+/2773538
Reviewed-by: Laxman Dewangan <ldewangan@nvidia.com>
GVS: Gerrit_Virtual_Submit <buildbot_gerritrpt@nvidia.com>
