24.10 nss - #64
Open
johnpaultads wants to merge 281 commits into
Open
Conversation
This is mostly a cosmetic change for users running NSS firmware 11.4 with mesh setups. Rather than fill up kernel logs with warnings, move log warnings to debug. The stats are still captured in '/sys/kernel/debug/qca-nss-drv/stats/*' Signed-off-by: Sean Khan <datapronix@protonmail.com>
Should be 'rx' Signed-off-by: Sean Khan <datapronix@protonmail.com>
While freeing skb in tx completion path for status reinject, inspect or vdevid mismatch, dev_kfree_skb_any() is used but when a function is using functions from mac80211 to free an skb then it should do it consistently and not switch to the generic dev_kfree_skb_any. Otherwise, mac80211 will not be aware of the freed skb and thus not clean up related information in its internal data structures. Hence fix the issue by properly using ieee80211_free_txskb(). Signed-off-by: Sarika Sharma <quic_sarishar@quicinc.com> Signed-off-by: Sean Khan <datapronix@protonmail.com>
Lock is missed while accessing idr_pool of tx_ring which causes Use after free crash in dp_free path when unloading the module. Fix this by adding tx_idr_lock when accessing idr_pool of tx_ring. Signed-off-by: Aishwarya R <quic_aisr@quicinc.com> Signed-off-by: Sean Khan <datapronix@protonmail.com>
…rrays Fixes compile issues on GCC 13.3+ Signed-off-by: Sean Khan <datapronix@protonmail.com>
Signed-off-by: Sean Khan <datapronix@protonmail.com>
1. Fix function to check for bridge master status while checking for Bridge VLAN filter feature is enabled on bridge slave ports. 2. Disable default PVID for bridges during device registration in the system. Change-Id: Ibea6559c1b0700a2300b60e20d57b7818e23a8a8 Signed-off-by: Vishnu Vardhan Bantanahal <quic_vishvard@quicinc.com> bridge: Fix Bridge VLAN stats update This patch fixes Bridge VLAN stats update for both bridge master and bridge slave. Change-Id: Ia26f4c71e83e27dd83336815cda5c05c8c3f24ff Signed-off-by: Vishnu Vardhan Bantanahal <quic_vishvard@quicinc.com> bridge: Add bridge VLAN filter APIs for offload for 6.1 Kernel Change-Id: I54e44c26664f86ae024f54605a032713a9a3eee5 Signed-off-by: Vishnu Vardhan Bantanahal <quic_vishvard@quicinc.com> Signed-off-by: Sean Khan <datapronix@protonmail.com>
This patch adds support for allocating 1MB of reserved memory for ramoops on IPQ807x and IPQ6018 platforms. This is required to get PSTORE option working. Tested with the following options: ``` CONFIG_PACKAGE_kmod-pstore=y CONFIG_PACKAGE_kmod-ramoops=y CONFIG_PACKAGE_kmod-reed-solomon=y ``` dmesg output ``` ➤ dmesg | grep -Ei "psto|ramo" [Sun Jun 9 23:22:01 2024] OF: reserved mem: 0x0000000051200000..0x00000000512fffff (1024 KiB) map non-reusable ramoops@51200000 [Sun Jun 9 23:22:05 2024] pstore: Using crash dump compression: deflate [Sun Jun 9 23:22:05 2024] pstore: Registered ramoops as persistent store backend [Sun Jun 9 23:22:05 2024] ramoops: using 0x100000@0x51200000, ecc: 0 ``` Triggering kernel panic ``` sh -c 'echo 10 > /proc/sys/kernel/panic; echo c > /proc/sysrq-trigger' ``` After rebooting ``` mkdir -p /tmp/ramoops mount -t pstore pstore /tmp/ramoops ``` ``` ➤ l /tmp/ramoops drwxr-x--- 2 root root 0 Jun 9 23:02 . drwxr-xr-x 8 root root 4.0K Jun 9 23:23 .. -r--r--r-- 1 root root 6.6K Jun 9 23:21 dmesg-ramoops-0 ``` Truncated output ``` Panic#1 Part1 <6>[ 21.656119] remoteproc remoteproc0: remote processor cd00000.q6v5_wcss is now up <14>[ 21.657910] kmodloader: done loading kernel modules from /etc/modules.d/* <6>[ 21.710545] ath11k c000000.wifi: chip_id 0x0 chip_family 0x0 board_id 0xff soc_id 0xffffffff <6>[ 21.710594] ath11k c000000.wifi: fw_version 0x270204a5 fw_build_timestamp 2022-08-04 13:05 fw_build_id WLAN.HK.2.7.0.1-01744-QCAHKSWPL_SILICONZ-1 <6>[ 91.416867] br-lan: topology change detected, sending tcn bpdu <6>[ 1256.023874] sysrq: Trigger a crash <0>[ 1256.023911] Kernel panic - not syncing: sysrq triggered crash <4>[ 1256.026176] CPU: 2 PID: 2842 Comm: sh Tainted: G O 6.6.32 #0 <4>[ 1256.031993] Hardware name: Linksys MX5300 (DT) <4>[ 1256.039106] Call trace: <4>[ 1256.043354] dump_backtrace+0xb0/0x10c <4>[ 1256.045701] show_stack+0x18/0x24 <4>[ 1256.049519] dump_stack_lvl+0x48/0x60 <4>[ 1256.052906] dump_stack+0x18/0x24 <4>[ 1256.056550] panic+0x2d0/0x31c <4>[ 1256.059848] sysrq_handle_moom+0x0/0x30 <4>[ 1256.062800] __handle_sysrq+0xe4/0x1e0 <4>[ 1256.066534] write_sysrq_trigger+0xbc/0xfc <4>[ 1256.070354] proc_reg_write+0xb0/0xf8 <4>[ 1256.074434] vfs_write+0xa4/0x298 <4>[ 1256.078165] ksys_write+0x5c/0xe0 <4>[ 1256.081464] __arm64_sys_write+0x1c/0x28 <4>[ 1256.084763] invoke_syscall.constprop.0+0x5c/0x100 <4>[ 1256.088758] do_el0_svc+0x40/0xc8 <4>[ 1256.093355] el0_svc+0x30/0xb8 <4>[ 1256.096740] el0t_64_sync_handler+0x120/0x12c <4>[ 1256.099694] el0t_64_sync+0x178/0x17c <2>[ 1256.104122] SMP: stopping secondary CPUs <0>[ 1256.107773] Kernel Offset: disabled <0>[ 1256.111759] CPU features: 0x0,00000000,00000000,0000400b <0>[ 1256.114975] Memory Limit: none ``` Signed-off-by: Sean Khan <datapronix@protonmail.com>
The following options can now be specified in `pbuf` uci config.
1.) auto_scale (bool) - enable or disable NSS CPU auto-scaling.
default: 'off'.
NOTE: Recommended to leave off, as it locks
NSS core to it's full 1.7Ghz speed. In the past
auto_scaling has been unstable, often causing
kernel panics or connection drops. Qualcomm
themselves disables scaling in QSDK for
ath11k/ath12k/qca-wifi.
It is provided to allow users wanting more
control over power usage. Setting to 'on' also
fixes NSS CPU stats updating properly.
2.) n2h_queue_limit_core0 (int) - set queue limits for CORE 0 and 1.
default: '2048'
For users on low memory systems (<=512M),
that are experiencing issues can try lowering
to the driver default of '256'.
NSS patches should always get applied last as to avoid fuzzing issues with upstream patches. Also cleanup patches that introduced spaces as indents vs. tabs Signed-off-by: Sean Khan <datapronix@protonmail.com> ath11k_nss: drop upstreamed patch A variation of this patch 'wifi: mac80211: Use flexible array in struct ieee80211_tim_ie' was upstreamed in commit 'e6e4d90b69027d74ba7e9a0a586b95a5714e10c0'. Remove it to avoid compilation issues. Signed-off-by: Sean Khan <datapronix@protonmail.com> ath11k_nss: Fix patch conflict Signed-off-by: Sean Khan <datapronix@protonmail.com>
Signed-off-by: Sean Khan <datapronix@protonmail.com> qualcommax: NSS: Refresh patches against 6.6.41 Signed-off-by: Sean Khan <datapronix@protonmail.com> qualcommax: NSS: refresh patches for 6.6.51 Signed-off-by: Sean Khan <datapronix@protonmail.com> qualcommax: add console-size + increase record-size Signed-off-by: Sean Khan <datapronix@protonmail.com> (cherry picked from commit 4ea718f) qualcommax: NSS: refresh patches for 6.6.56 Signed-off-by: Sean Khan <datapronix@protonmail.com> qualcommax: NSS: refresh patches for 6.6.60 Signed-off-by: Sean Khan <datapronix@protonmail.com> (cherry picked from commit 8efd9fedbff95990ffe192f504491fe63e9a7d55) qualcommax: Refresh NSS patches against rebase 863d1a6 Signed-off-by: Sean Khan <datapronix@protonmail.com> qualcommax: NSS: Refresh patches for 6.6.67 Signed-off-by: Sean Khan <datapronix@protonmail.com>
Signed-off-by: Sean Khan <datapronix@protonmail.com>
Revert to using the default kernel timer (CONFIG_HZ_100). Setting the timer to 1000hz is unecessary and just adds CPU overhead. It also causes atomic scheduling issues in various threaded tasks, especially when using threaded NAPI. When the kernel timer frequency is set to 1000 Hz, it generates interrupts 1000 times per second. This increased frequency can lead to more frequent context switches and interrupt handling, which can interfere with other kernel activities. Since network packet processing is offloaded to the NSS cores, the main CPU does not need high-resolution timer interrupts to manage network traffic efficiently. The NSS cores handle this independently. This is the setting used and recommended by Qualcomm for all their IPQ platforms. Signed-off-by: Sean Khan <datapronix@protonmail.com>
Cleaned up erronous settings for disabled features. The syntax should be "# CONFIG_SOME_OPTION is not set", rather than simply commenting it out. Removed left over config options from 5.15, and enabled features recommended by QSDK: ``` CONFIG_IRQ_TIME_ACCOUNTING=y CONFIG_MULTI_IRQ_HANDLER=y CONFIG_HANDLE_DOMAIN_IRQ=y CONFIG_HZ_FIXED=0 CONFIG_NO_HZ=y ``` Signed-off-by: Sean Khan <datapronix@protonmail.com>
Signed-off-by: Sean Khan <datapronix@protonmail.com>
Signed-off-by: Sean Khan <datapronix@protonmail.com>
Signed-off-by: Sean Khan <datapronix@protonmail.com>
These packages are required for basic NSS offloading. So ensure they're selected for all qualcommax targets. * kmod-qca-nss-ecm (Enhanced Connection Manager) * kmod-qca-nss-drv (NSS Driver) * kmod-qca-nss-drv-bridge-mgr (NSS Manager for Bridges) Signed-off-by: Sean Khan <datapronix@protonmail.com>
Signed-off-by: Sean Khan <datapronix@protonmail.com>
Signed-off-by: Sean Khan <datapronix@protonmail.com>
This reverts commit 5cf32cfe3edca047b7347236496c275bc23ae7fb.
Signed-off-by: Sean Khan <datapronix@protonmail.com>
commit 072327212c033baf96365a5045673c5981eb0bdf (HEAD -> next, origin/next)
Author: John Crispin <john@phrozen.org>
AuthorDate: Fri Jul 5 15:42:24 2024 +0200
Commit: John Crispin <john@phrozen.org>
CommitDate: Fri Jul 5 15:42:40 2024 +0200
ath11k: fix monitor mode
Enabling CPTCFG_ATH11K_MEM_PROFILE_512M will change the layout of
struct hal_rx_desc_ipq8074. This broke monitor mode. Fix it by not modifying
the struct even if the 512MB profile is enabled.
Fixes: WIFI-12102
Signed-off-by: John Crispin <john@phrozen.org>
Signed-off-by: Sean Khan <datapronix@protonmail.com>
This patch handles the updating of MU-EDCA parameters indicated by
driver. Driver requests to update the parameters and Parameter Set
Update Count in beacon frames. This patch addresses only AP mode.
Three actions are taken to successfully update MU-EDCA parameters
in beacon:
(1) Update MU-EDCA parameters stored in hostapd
The following Access Categories (AC) are updated:
VO: Voice
VI: Video
BE: Best Effort
BK: Background
(2) Increment Parameter Set Update Count
(3) Update beacon with new parameters
There aren't any options to configure, hostapd simply updates MU-EDCA
published by ath11k driver.
To verify params are getting updated, set log_level to '1' and check
logger
i.e. `uci set wireless.radio0.log_level=1`
```
Sat Jul 6 15:36:10 2024 daemon.debug hostapd: phy0-ap0: IEEE 802.11 MU-EDCA: Updated MU-EDCA parameters for AC 0: BE: 3, BK: 7, VI: 2, VO: 2
Sat Jul 6 15:36:10 2024 daemon.debug hostapd: phy0-ap0: IEEE 802.11 MU-EDCA: Updated MU-EDCA parameters for AC 1: BE: 164, BK: 164, VI: 67, VO: 50
Sat Jul 6 15:36:10 2024 daemon.debug hostapd: phy0-ap0: IEEE 802.11 MU-EDCA: Updated MU-EDCA parameters for AC 2: BE: 255, BK: 2, VI: 255, VO: 255
```
Requires the following patches applied to 'ath11k/subsys'
commit ce7f1ad354869d85abdf8e8ed8d36599e057ba5d
Author: Sean Khan <datapronix@protonmail.com>
AuthorDate: Tue Feb 27 00:45:32 2024 -0500
Commit: Sean Khan <datapronix@protonmail.com>
CommitDate: Sat Jul 6 15:19:26 2024 -0400
ath11k_nss: FW Initiated Dynamic MU-EDCA
package/kernel/mac80211/patches/nss/ath11k/203-mac80211-ath11k-fw-dynamic-muedca.patch
package/kernel/mac80211/patches/nss/subsys/203-mac80211-ath11k-fw-dynamic-muedca.patch
Signed-off-by: Sean Khan <datapronix@protonmail.com>
Signed-off-by: Sean Khan <datapronix@protonmail.com> ath11k_nss: fix missing semicolon ';' Signed-off-by: Sean Khan <datapronix@protonmail.com>
Signed-off-by: Sean Khan <datapronix@protonmail.com>
Signed-off-by: Sean Khan <datapronix@protonmail.com>
Some features are not available in NSS firmware 11.4.0.5 * `NSS_WIFI_VDEV_EXT_DATA_PKT_TYPE_4ADDR` This was introduced in `NHSS.QSDK.11.5`. To avoid unexpected issues, disable the case check for `NSS_WIFI_VDEV_EXT_DATA_PKT_TYPE_4ADDR` if `CONFIG_NSS_FIRMWARE_VERSION_11_4` is selected. Signed-off-by: Sean Khan <datapronix@protonmail.com> ath11k_nss: rename '*.patch.patch' to '.patch' Signed-off-by: Sean Khan <datapronix@protonmail.com>
Signed-off-by: Sean Khan <datapronix@protonmail.com>
New:
* Add UCI option to /etc/config/pbuf
1. `stats_disable` - Allows disabling extended stats
collection. (Default: '1')
2. `scaling_governor` - Allow specifying CPU governor.
(Default: disabled )
Choices depend on what your system is compiled with. Can be found
using:
```sh
cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_available_governors
ondemand performance schedutil
```
Changes:
* Rather than hardcode every device model's memory profile. Simply read the
memory available from `/proc/meminfo`.
* Scaling governor is no longer forced as `performance`. It is now user
configurable.
* Hash bitmap is now calculated from number of CPUs rather than hardcoded to `15`
* Limit `n2h_queue_limit_core` to `256`. This was set to `2048` which is
too high when NSS wifi is also enabled.
Signed-off-by: Sean Khan <datapronix@protonmail.com>
This commit includes the following fixes and improvements: Missing flags: * IFF_EXT_TUN_TAP, IFF_EXT_GRE_V4_TAP and IFF_EXT_GRE_V6_TAP were missing causing issues with proprly offloading GRE/TUN/IPIP tunnels. * Brought in several missing or fixes for ECM/Bridge/VLAN manager. Reworked the skb_recycler-support patch with the latest 6.6 QSDK changes: * Added new flags for tracking recycled SKBs * Implemented fast SKB recycling path for improved performance * Added new helper functions for SKB allocation/recycling * Optimized memory usage with different SKB cache sizes Consolidated L2TP offloading code: * Merged functionality from deleted patch (0603-6) into 0603-2 * Added proper support for L2TP device identification in network stack * Fixed various NSS-ECM networking components: Properly set skb_iif for tunneled traffic * Cleaned up bridge VLAN filtering support * Improved device type identification for offloading decisions * Code organization improvements: Moved device type flags to core patch * Eliminated code duplication between patches * Fixed stats update logic for VLAN devices Updated nf_conntrack_dscpremark_ext.h to match the latest QSDK changes: Quilt refreshed all patches Signed-off-by: Sean Khan <datapronix@protonmail.com>
Adds in missing callbacks for `rx_desc_get_ip_valid` in IPQ8074, IPQ6018 and QCN9074. This was preventing running in non-NSS mode. Signed-off-by: Sean Khan <datapronix@protonmail.com>
When running with NSS mode disabled (nss_offload=0) the driver would crash soon after a client connected with: ``` dog_hb.c:210 DOG_HB detects starvation of task "WLAN RT0", triage with its owner(d.dump 0x4af49e00) ``` Revert the budgeting of the task to the original code to fix the crash. Signed-off-by: Sean Khan <datapronix@protonmail.com>
…WIOTLB issues" This commit resolves applying issues due to the upstream commit (fa50e53: "Fix DMA buffer allocation to resolve SWIOTLB issues".) Signed-off-by: Sean Khan <datapronix@protonmail.com>
no longer needed with "51300e4 - nss-cfi: fix issue with kernel reporting duplicate driver" Signed-off-by: Sean Khan <datapronix@protonmail.com>
Signed-off-by: Sean Khan <datapronix@protonmail.com> ath11k_nss: Restructure patch to avoid dependency on 207-ath11k-Enable-256_512MB-profiles.patch In an effort to potentially reduce the number of patches required, this commit restructures the patch to avoid dependency on `207-ath11k-Enable-256_512MB-profiles.patch`. Signed-off-by: Sean Khan <datapronix@protonmail.com> ath11k_nss: Sync 780-mac80211-Advertise-HW-checksum-offload-only-for-ethm with upstream Apply QSDK 12.5 patch Signed-off-by: Sean Khan <datapronix@protonmail.com>
When using 802.11r Fast Transition the driver will sometimes reject a redundant PTK install even though the handshake completes successfully. Upstream hostapd logs this as "kernel reports: key addition failed" on every roam, flooding syslogs with harmless errors. This patch removes the `GENL_SET_ERR_MSG` for that code path in `nl80211_new_key()`, so we still return the proper error code to userspace but skip the misleading diagnostic string. Now users won't be distracted by known-benign messages during FT. Signed-off-by: Sean Khan <datapronix@protonmail.com>
Acidentaly reverted the following commit when rebasing skb recyler from QSDK ipvs: Always clear ipvs_property flag in skb_scrub_packet() [ Upstream commit de2c211868b9424f9aa9b3432c4430825bafb41b ] Refreshed patches to apply cleanly Signed-off-by: Sean Khan <datapronix@protonmail.com>
Signed-off-by: Sean Khan <datapronix@protonmail.com>
…N filtering. Signed-off-by: qosmio <datapronix@protonmail.com>
Fix race in peer kickout handling by ensuring `base_lock` covers peer checks. Avoids spurious kick events when firmware triggers removal before peer is in mac80211. Also fix a possible deadlock in peer_sta_ps_state_chg by ensuring spin_unlock on all exit paths. Patch-by: Sebastian Gottschall <s.gottschall@dd-wrt.com> Signed-off-by: Sean Khan <datapronix@protonmail.com> (cherry picked from commit 2d43e93)
GRO fraglist causes issues when NSS offload is enabled, paticularly with UDP packets (e.g. DNS queries). Rather than rely on hotplug script disable by default. Signed-off-by: Sean Khan <datapronix@protonmail.com>
* Restructure the README.md to provide concise UCI configs. * Add section for WWAN configuration. Signed-off-by: Sean Khan <datapronix@protonmail.com>
Signed-off-by: Sean Khan <datapronix@protonmail.com>
Signed-off-by: Sean Khan <datapronix@protonmail.com>
qosmio
force-pushed
the
24.10-nss
branch
2 times, most recently
from
December 30, 2025 00:30
617b0b3 to
04665e8
Compare
qosmio
force-pushed
the
24.10-nss
branch
2 times, most recently
from
January 5, 2026 16:58
f1aa967 to
8f05fb6
Compare
qosmio
force-pushed
the
24.10-nss
branch
2 times, most recently
from
February 27, 2026 08:25
fa4c3f9 to
7cbd6aa
Compare
qosmio
force-pushed
the
24.10-nss
branch
2 times, most recently
from
June 4, 2026 18:42
b9c8095 to
2a9a392
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
6 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Thanks for your contribution to OpenWrt!
To help keep the codebase consistent and readable,
and to help people review your contribution,
we ask you to follow the rules you find in the wiki at this link
https://openwrt.org/submitting-patches
Please remove this message before posting the pull request.