Ping-Ke Shih [Tue, 7 Jan 2025 11:42:54 +0000 (19:42 +0800)]
wifi: rtw89: fix race between cancel_hw_scan and hw_scan completion
The rtwdev->scanning flag isn't protected by mutex originally, so
cancel_hw_scan can pass the condition, but suddenly hw_scan completion
unset the flag and calls ieee80211_scan_completed() that will free
local->hw_scan_req. Then, cancel_hw_scan raises null-ptr-deref and
use-after-free. Fix it by moving the check condition to where
protected by mutex.
Zong-Zhe Yang [Fri, 3 Jan 2025 07:44:12 +0000 (15:44 +0800)]
wifi: rtw89: mcc: consider time limits not divisible by 1024
For each MCC role, time limits, including max_tob_us, max_toa_us, and
mac_dur_us, are calculated if there are NoA attributes. The relation
between these time limits is "max_dur_us = max_tob_us + max_toa_us".
Then, the unit is converted from us to TU. However, originally, each
time limit was divided by 1024 independently. It missed to consider
the cases that max_tob_us or max_toa_us is not divisible by 1024. It
causes the result breaks "max_dur (TU) = max_tob (TU) + max_toa (TU)".
Finally, when MCC calculates pattern parameters based on these kinds
of time limits, it might not perform well.
Ping-Ke Shih [Fri, 3 Jan 2025 02:51:26 +0000 (10:51 +0800)]
wifi: rtw89: pci: treat first receiving part as first segment for 8922AE
For early chips, the RX BD info contains FS/LS bits for first/last segments
of a receiving packet. For 8922AE, the FS bit should be ignored, or it
may throw warning:
rtw89_8922ae 0000:1a:00.0: skb should not be ready before first segment start
To have compatible logic, FS determined by what pending skb is existing
(!new) or not.
Since we expect every single receiving packet in single one segment,
enlarge 4 bytes for RX BD info, and add a debug message to note if
RX buffer is possibly smaller than expected size.
Chih-Kang Chang [Fri, 3 Jan 2025 02:45:00 +0000 (10:45 +0800)]
wifi: rtw89: avoid to init mgnt_entry list twice when WoWLAN failed
If WoWLAN failed in resume flow, the rtw89_ops_add_interface() triggered
without removing the interface first. Then the mgnt_entry list init again,
causing the list_empty() check in rtw89_chanctx_ops_assign_vif()
useless, and list_add_tail() again. Therefore, we have added a check to
prevent double adding of the list.
rtw89_8852ce 0000:01:00.0: failed to check wow status disabled
rtw89_8852ce 0000:01:00.0: wow: failed to check disable fw ready
rtw89_8852ce 0000:01:00.0: wow: failed to swap to normal fw
rtw89_8852ce 0000:01:00.0: failed to disable wow
rtw89_8852ce 0000:01:00.0: failed to resume for wow -110
rtw89_8852ce 0000:01:00.0: MAC has already powered on
i2c_hid_acpi i2c-ILTK0001:00: PM: acpi_subsys_resume+0x0/0x60 returned 0 after 284705 usecs
list_add corruption. prev->next should be next (ffff9d9719d82228), but was ffff9d9719f96030. (prev=ffff9d9719f96030).
------------[ cut here ]------------
kernel BUG at lib/list_debug.c:34!
invalid opcode: 0000 [#1] PREEMPT SMP NOPTI
CPU: 2 PID: 6918 Comm: kworker/u8:19 Tainted: G U O
Hardware name: Google Anraggar/Anraggar, BIOS Google_Anraggar.15217.514.0 03/25/2024
Workqueue: events_unbound async_run_entry_fn
RIP: 0010:__list_add_valid_or_report+0x9f/0xb0
Code: e8 56 89 ff ff 0f 0b 48 c7 c7 3e fc e0 96 48 89 c6 e8 45 89 ff ...
RSP: 0018:ffffa51b42bbbaf0 EFLAGS: 00010246
RAX: 0000000000000075 RBX: ffff9d9719d82ab0 RCX: 13acb86e047a4400
RDX: 3fffffffffffffff RSI: 0000000000000000 RDI: 00000000ffffdfff
RBP: ffffa51b42bbbb28 R08: ffffffff9768e250 R09: 0000000000001fff
R10: ffffffff9765e250 R11: 0000000000005ffd R12: ffff9d9719f95c40
R13: ffff9d9719f95be8 R14: ffff9d97081bfd78 R15: ffff9d9719d82060
FS: 0000000000000000(0000) GS:ffff9d9a6fb00000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 00007e7d029a4060 CR3: 0000000345e38000 CR4: 0000000000750ee0
PKRU: 55555554
Call Trace:
<TASK>
? __die_body+0x68/0xb0
? die+0xaa/0xd0
? do_trap+0x9f/0x170
? __list_add_valid_or_report+0x9f/0xb0
? __list_add_valid_or_report+0x9f/0xb0
? handle_invalid_op+0x69/0x90
? __list_add_valid_or_report+0x9f/0xb0
? exc_invalid_op+0x3c/0x50
? asm_exc_invalid_op+0x16/0x20
? __list_add_valid_or_report+0x9f/0xb0
rtw89_chanctx_ops_assign_vif+0x1f9/0x210 [rtw89_core cbb375c44bf28564ce479002bff66617a25d9ac1]
? __mutex_unlock_slowpath+0xa0/0xf0
rtw89_ops_assign_vif_chanctx+0x4b/0x90 [rtw89_core cbb375c44bf28564ce479002bff66617a25d9ac1]
drv_assign_vif_chanctx+0xa7/0x1f0 [mac80211 6efaad16237edaaea0868b132d4f93ecf918a8b6]
ieee80211_reconfig+0x9cb/0x17b0 [mac80211 6efaad16237edaaea0868b132d4f93ecf918a8b6]
? __pfx_wiphy_resume+0x10/0x10 [cfg80211 572d03acaaa933fe38251be7fce3b3675284b8ed]
? dev_printk_emit+0x51/0x70
? _dev_info+0x6e/0x90
wiphy_resume+0x89/0x180 [cfg80211 572d03acaaa933fe38251be7fce3b3675284b8ed]
? __pfx_wiphy_resume+0x10/0x10 [cfg80211 572d03acaaa933fe38251be7fce3b3675284b8ed]
dpm_run_callback+0x37/0x1e0
device_resume+0x26d/0x4b0
? __pfx_dpm_watchdog_handler+0x10/0x10
async_resume+0x1d/0x30
async_run_entry_fn+0x29/0xd0
worker_thread+0x397/0x970
kthread+0xed/0x110
? __pfx_worker_thread+0x10/0x10
? __pfx_kthread+0x10/0x10
ret_from_fork+0x38/0x50
? __pfx_kthread+0x10/0x10
ret_from_fork_asm+0x1b/0x30
</TASK>
Zong-Zhe Yang [Tue, 31 Dec 2024 00:48:09 +0000 (08:48 +0800)]
wifi: rtw89: chan: fix soft lockup in rtw89_entity_recalc_mgnt_roles()
During rtw89_entity_recalc_mgnt_roles(), there is a normalizing process
which will re-order the list if an entry with target pattern is found.
And once one is found, should have aborted the list_for_each_entry. But,
`break` just aborted the inner for-loop. The outer list_for_each_entry
still continues. Normally, only the first entry will match the target
pattern, and the re-ordering will change nothing, so there won't be
soft lockup. However, in some special cases, soft lockup would happen.
Fix it by `goto fill` to break from the list_for_each_entry.
The following is a sample of kernel log for this problem.
Zong-Zhe Yang [Tue, 31 Dec 2024 00:48:08 +0000 (08:48 +0800)]
wifi: rtw89: fix proceeding MCC with wrong scanning state after sequence changes
When starting/proceeding MCC, it will abort an ongoing hw scan process.
In the proceeding cases, it unexpectedly tries to abort a non-exist hw
scan process. Then, a trace shown at the bottom will happen. This problem
is caused by a previous commit which changed some call sequence inside
rtw89_hw_scan_complete() to fix some coex problems. These changes lead
to our scanning flag was not cleared when proceeding MCC. To keep the
fixes on coex, and resolve the problem here, re-consider the related
call sequence.
The known sequence requirements are listed below.
* the old sequence:
A. notify coex
B. clear scanning flag
C. proceed chanctx
C-1. set channel
C-2. proceed MCC
(the problem: A needs to be after C-1)
* the current sequence:
C. proceed chanctx
C-1. set channel
C-2. proceed MCC
A. notify coex
B. clear scanning flag
(the problem: C-2 needs to be after B)
So, now let hw scan caller pass a callback to proceed chanctx if needed.
Then, the new sequence will be like the below.
C-1. set channel
A. notify coex
B. clear scanning flag
C-2. proceed MCC
The following is the kernel log for the problem in current sequence.
Chin-Yen Lee [Wed, 25 Dec 2024 12:28:04 +0000 (20:28 +0800)]
wifi: rtw89: pci: disable PCI completion timeout control
Realtek's chips follow suggestion of PCIe spec to design the max timeout of
PCI completion, but some PCIe host reply too slow to meet it and lead PCI
AER. Disable PCI completion timeout function via PCI configuration to
avoid the AER.
A user reports that the firmware upload consistently fails when it's
uploaded in chunks of 128 bytes, but it works when uploaded in chunks
of 196 bytes. The official driver uses 196 bytes also.
Chih-Kang Chang [Thu, 19 Dec 2024 01:45:11 +0000 (09:45 +0800)]
wifi: rtw89: adjust thermal protection step and more RTL8852BE-VT ID
To have better tuning of TX duty for fields usage, adjust thermal
protection step to 5, then TX duty reduce to 75% at most. Add RTL8852BE-VT
ID to enable this function by the way.
Bitterblue Smith [Wed, 18 Dec 2024 22:35:49 +0000 (00:35 +0200)]
wifi: rtw88: usb: Preallocate and reuse the RX skbs
The USB driver uses four USB Request Blocks for RX. Before submitting
one, it allocates a 32768 byte skb for the RX data. This allocation can
fail, maybe due to temporary memory fragmentation. When the allocation
fails, the corresponding URB is never submitted again. After four such
allocation failures, all RX stops because the driver is not requesting
data from the device anymore.
Don't allocate a 32768 byte skb when submitting a USB Request Block
(which happens very often). Instead preallocate 8 such skbs, and reuse
them over and over. If all 8 are busy, allocate a new one. This is
pretty rare. If the allocation fails, use a work to try again later.
When there are enough free skbs again, free the excess skbs.
Also, use WQ_BH for the RX workqueue. With a normal or high priority
workqueue the skbs are processed too slowly when the system is even a
little busy, like when opening a new page in a browser, and the driver
runs out of free skbs and allocates a lot of new ones.
This is more or less what the out-of-tree Realtek drivers do, except
they use a tasklet instead of a BH workqueue.
Tested with RTL8723DU, RTL8821AU, RTL8812AU, RTL8812BU, RTL8822CU,
RTL8811CU.
Bitterblue Smith [Wed, 18 Dec 2024 22:34:42 +0000 (00:34 +0200)]
wifi: rtw88: Handle C2H_ADAPTIVITY in rtw_fw_c2h_cmd_handle()
The firmware message C2H_ADAPTIVITY is currently handled in
rtw_fw_c2h_cmd_rx_irqsafe(), which runs in the RX workqueue, but it's
not "irqsafe" with USB because it sleeps (reads hardware registers).
This becomes a problem after the next patch, which will create the RX
workqueue with the flag WQ_BH.
To avoid sleeping when it's not allowed, handle C2H_ADAPTIVITY in
rtw_fw_c2h_cmd_handle(), which runs in the c2h workqueue.
Bitterblue Smith [Wed, 18 Dec 2024 22:33:20 +0000 (00:33 +0200)]
wifi: rtw88: usb: Copy instead of cloning the RX skb
"iperf3 -c 192.168.0.1 -R --udp -b 0" shows about 40% of datagrams
are lost. Many torrents don't download faster than 3 MiB/s, probably
because the Bittorrent protocol uses UDP. This is somehow related to
the use of skb_clone() in the RX path.
Don't use skb_clone(). Instead allocate a new skb for each 802.11 frame
received and copy the data from the big (32768 byte) skb.
With this patch, "iperf3 -c 192.168.0.1 -R --udp -b 0" shows only 1-2%
of datagrams are lost, and torrents can reach download speeds of 36
MiB/s.
Bitterblue Smith [Tue, 17 Dec 2024 22:53:11 +0000 (00:53 +0200)]
wifi: rtlwifi: rtl8821ae: Fix media status report
RTL8821AE is stuck transmitting at the lowest rate allowed by the rate
mask. This is because the firmware doesn't know the device is connected
to a network.
Fix the macros SET_H2CCMD_MSRRPT_PARM_OPMODE and
SET_H2CCMD_MSRRPT_PARM_MACID_IND to work on the first byte of __cmd,
not the second. Now the firmware is correctly notified when the device
is connected to a network and it activates the rate control.
Ping-Ke Shih [Thu, 12 Dec 2024 05:42:03 +0000 (13:42 +0800)]
wifi: rtw88: add __packed attribute to efuse layout struct
The layout struct of efuse should not do address alignment by compiler.
Otherwise it leads unexpected layout and size for certain arch suc as arm.
In x86-64, the results are identical before and after this patch.
Also adjust bit-field to prevent over adjacent byte to avoid warning:
rtw88/rtw8822b.h:66:1: note: offset of packed bit-field `res2` has changed in GCC 4.4
66 | } __packed;
| ^
Dmitry Antipov [Tue, 10 Dec 2024 07:04:41 +0000 (10:04 +0300)]
wifi: brcmsmac: add gain range check to wlc_phy_iqcal_gainparams_nphy()
In 'wlc_phy_iqcal_gainparams_nphy()', add gain range check to WARN()
instead of possible out-of-bounds 'tbl_iqcal_gainparams_nphy' access.
Compile tested only.
Found by Linux Verification Center (linuxtesting.org) with SVACE.
Ping-Ke Shih [Mon, 9 Dec 2024 04:21:27 +0000 (12:21 +0800)]
wifi: rtw89: phy: add dummy C2H event handler for report of TAS power
The newer firmware, lik RTL8852C version 0.27.111.0, will notify driver
report of TAS (Time Averaged SAR) power by new C2H events. This is to
assist in higher accurate calculation of TAS.
For now, driver doesn't use the report yet, so add a dummy handler to
avoid it throws info like:
rtw89_8852ce 0000:03:00.0: c2h class 9 func 6 not support
Also add "MAC" and "PHY" to the message to disambiguate the source of
C2H event.
Ping-Ke Shih [Mon, 9 Dec 2024 04:20:19 +0000 (12:20 +0800)]
wifi: rtw89: 8852c: rfk: refine target channel calculation in _rx_dck_channel_calc()
The channel is not possibly 0, so original code is fine. Still want to
avoid Coverity warning, so ensure -32 offset for the channel number which
is larger than 125 only. Actually, don't change logic at all.
wifi: rtlwifi: pci: wait for firmware loading before releasing memory
At probe error path, the firmware loading work may have already been
queued. In such a case, it will try to access memory allocated by the probe
function, which is about to be released. In such paths, wait for the
firmware worker to finish before releasing memory.
wifi: rtlwifi: fix memory leaks and invalid access at probe error path
Deinitialize at reverse order when probe fails.
When init_sw_vars fails, rtl_deinit_core should not be called, specially
now that it destroys the rtl_wq workqueue.
And call rtl_pci_deinit and deinit_sw_vars, otherwise, memory will be
leaked.
Remove pci_set_drvdata call as it will already be cleaned up by the core
driver code and could lead to memory leaks too. cf. commit 8d450935ae7f
("wireless: rtlwifi: remove unnecessary pci_set_drvdata()") and
commit 3d86b93064c7 ("rtlwifi: Fix PCI probe error path orphaned memory").
wifi: rtlwifi: destroy workqueue at rtl_deinit_core
rtl_wq is allocated at rtl_init_core, so it makes more sense to destroy it
at rtl_deinit_core. In the case of USB, where _rtl_usb_init does not
require anything to be undone, that is fine. But for PCI, rtl_pci_init,
which is called after rtl_init_core, needs to deallocate data, but only if
it has been called.
That means that destroying the workqueue needs to be done whether
rtl_pci_init has been called or not. And since rtl_pci_deinit was doing it,
it has to be moved out of there.
It makes more sense to move it to rtl_deinit_core and have it done in both
cases, USB and PCI.
Since this is a requirement for a followup memory leak fix, mark this as
fixing such memory leak.
Commit 2461c7d60f9f ("rtlwifi: Update header file") introduced a global
list of private data structures.
Later on, commit 26634c4b1868 ("rtlwifi Modify existing bits to match
vendor version 2013.02.07") started adding the private data to that list at
probe time and added a hook, check_buddy_priv to find the private data from
a similar device.
However, that function was never used.
Besides, though there is a lock for that list, it is never used. And when
the probe fails, the private data is never removed from the list. This
would cause a second probe to access freed memory.
Remove the unused hook, structures and members, which will prevent the
potential race condition on the list and its corruption during a second
probe when probe fails.
Fixes: 26634c4b1868 ("rtlwifi Modify existing bits to match vendor version 2013.02.07") Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo@igalia.com> Signed-off-by: Ping-Ke Shih <pkshih@realtek.com> Link: https://patch.msgid.link/20241206173713.3222187-2-cascardo@igalia.com
Zong-Zhe Yang [Fri, 6 Dec 2024 05:57:15 +0000 (13:57 +0800)]
wifi: rtw89: regd: update regulatory map to R68-R51
Sync Realtek Channel Plan R68 and Realtek Regulatory R51.
Configure 6 GHz field of Realtek regd for the following countries.
BO, DO, EG, LS, MZ, NG, OM, ZW, PK, PH, TH, KM, CG, CD,
GE, GI, GU, LR, MH, FM, MP, PW, MF, SX, SZ, TZ, VI
Besides, add entries for the following countries.
CU, SY, SD
Kuan-Chung Chen [Fri, 6 Dec 2024 05:57:14 +0000 (13:57 +0800)]
wifi: rtw89: 8852c: disable ER SU when 4x HE-LTF and 0.8 GI capability differ
Since hardware only has single one register for HE-LTF setting,
to prevent interoperability issues, 8852CE disables ER SU when
the AP can handle SU/MU with 4x HE-LTF and 0.8 GI, but does
not support ER SU with the same settings.
Kuan-Chung Chen [Fri, 6 Dec 2024 05:57:13 +0000 (13:57 +0800)]
wifi: rtw89: disable firmware training HE GI and LTF
Given the performance trade-off associated with firmware training
HE GI/LTF, especially in high attenuation environments, we have
decided to utilize a constant value instead.
Eric Huang [Fri, 6 Dec 2024 05:57:12 +0000 (13:57 +0800)]
wifi: rtw89: ps: update data for firmware and settings for hardware before/after PS
For MLO supported IC, send H2C command to firmware before PS with link
information for each PHY for MLO to work properly. And re-init hardware
settings regarding to RX descriptor information after PS.
Ping-Ke Shih [Fri, 6 Dec 2024 05:57:11 +0000 (13:57 +0800)]
wifi: rtw89: ps: refactor channel info to firmware before entering PS
In PS mode, firmware needs hardware parameters related to channel info
to configure hardware itself. Before entering PS, driver prepares these
info to firmware via firmware H2C command.
Since firmware only consider PS for single one vif, change the argument
of entry function to rtwvif, and only consider first link for this old
H2C command that only support legacy.
Ping-Ke Shih [Fri, 6 Dec 2024 05:57:10 +0000 (13:57 +0800)]
wifi: rtw89: ps: refactor PS flow to support MLO
Firmware can only support PS on single one VIF operating in station mode,
so argument of PS entry rtw89_enter_lps() should be rtwvif insetad of
rtwvif_link.
To enter PS under MLO, for each rtwvif, driver sends H2C command to tell
firmware which mac_id will enter PS one by one, and afterward asks
firmware to enter deep PS.
Kalle Valo [Mon, 9 Dec 2024 18:41:20 +0000 (20:41 +0200)]
Merge tag 'ath-next-20241209' of git://git.kernel.org/pub/scm/linux/kernel/git/ath/ath
ath.git patches for v6.14
This development cycle featured multiple patchsets to ath12k to
support the new 802.11be MLO feature, although the feature is still
incomplete. Also in ath12k, there were other feature patches. In
ath11k, support was added for QCA6698AQ. And there was the usual set
of bug fixes and cleanups across most drivers, notable being the
addition of "noinline_for_stack" to some functions to avoid "stack
frame size" warnings when compiling with clang.
Pin-yen Lin [Wed, 27 Nov 2024 10:55:43 +0000 (18:55 +0800)]
wifi: mwifiex: decrease timeout waiting for host sleep from 10s to 5s
In commit 52250cbee7f6 ("mwifiex: use timeout variant for
wait_event_interruptible") it was noted that sometimes we seemed
to miss the signal that our host sleep settings took effect. A
10 second timeout was added to the code to make sure we didn't
hang forever waiting. It appears that this problem still exists
and we hit the timeout sometimes for Chromebooks in the field.
Recently on ChromeOS we've started setting the DPM watchdog
to trip if full system suspend takes over 10 seconds. Given
the timeout in the original patch, obviously we're hitting
the DPM watchdog before mwifiex gets a chance to timeout.
While we could increase the DPM watchdog in ChromeOS to avoid
this problem, it's probably better to simply decrease the
timeout. Any time we're waiting several seconds for the
firmware to respond it's likely that the firmware won't ever
respond. With that in mind, decrease the timeout in mwifiex
from 10 seconds to 5 seconds.
Suggested-by: Doug Anderson <dianders@chromium.org> Signed-off-by: Pin-yen Lin <treapking@chromium.org> Reviewed-by: Douglas Anderson <dianders@chromium.org> Acked-by: Brian Norris <briannorris@chromium.org> Signed-off-by: Kalle Valo <kvalo@kernel.org> Link: https://patch.msgid.link/20241127105709.4014302-1-treapking@chromium.org
wifi: brcmfmac: fix brcmf_vif_clear_mgmt_ies when stopping AP
This removes the following error log when stopping AP:
ieee80211 phy0: brcmf_vif_set_mgmt_ie: vndr ie set error : -52
It happened if:
1) previously wlan interface was in station mode (wpa_supplicant) and
connected to a hotspot
2) then started AP mode (hostapd)
3) and then stopped AP mode.
The error happened when it tried to clear BRCMF_VNDR_IE_PRBREQ_FLAG.
This flag is not set in `brcmf_config_ap_mgmt_ie`, but
BRCMF_VNDR_IE_ASSOCRSP_FLAG is set instead.
'struct nla_policy' is not modified in this driver.
Constifying this structure moves some data to a read-only section, so
increase overall security, especially when the structure holds some
function pointers.
On a x86_64, with allmodconfig:
Before:
======
text data bss dec hex filename
5062 528 0 5590 15d6 drivers/net/wireless/ti/wlcore/testmode.o
After:
=====
text data bss dec hex filename
5178 404 0 5582 15ce drivers/net/wireless/ti/wlcore/testmode.o
wifi: brcmfmac: fix scatter-gather handling by detecting end of sg list
The scatter-gather handling uses a pre-allocated list (with nents entries).
If the driver runs out of sg entries it will result in an oops. Let's detect
this instead and make the SDIO block transfer fail.
wifi: ath12k: introduce mlo_capable flag for device group
Currently, during mac80211 allocate and register single device is
considered for the registration. But, in future, during multi device
group abstraction is introduced, all the devices has to be combined
together as a single abstraction and then hardware should be allocated.
All the devices in the group should be combined together only if it
supports inter device mlo capability. The decision of whether to combine
the devices or not can be based on the mlo capability flag in
ath12k_hw_group.
By default, mlo_capable flag in the group would be set as false. During
QMI PHY capability exchange, only when we have more than one chip in the
group or if one chip, then that chip supports inter MLO, then mlo_capable
flag in the group will be enabled.
Add changes to introduce mlo_capable flag for device group and refactor
ath12k_mac_hw_allocate() api based on device group (ag) rather than
device (ab).
Signed-off-by: Karthikeyan Periyasamy <quic_periyasa@quicinc.com> Signed-off-by: Harshitha Prem <quic_hprem@quicinc.com> Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com> Link: https://patch.msgid.link/20241204163216.433795-8-kvalo@kernel.org Signed-off-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com>
wifi: ath12k: send QMI host capability after device group is ready
QMI host capability has the information regarding MLO parameters such as
device_id, MLO capability, group id and the information of each devices
involved in the group and sent immediately on QMI server arrive event.
Currently, only one device is involved in the group and hence, single
device information is sent as part of MLO capability of host.
But, in future when multi device group abstraction is introduced, host
should send all the device information involved in the group as part
of QMI MLO host capability rather than single device. Hence, sending
QMI host capability immediately on server arrive of a device might not
be ideal for multi device group abstraction as the details of other
devices in the group would not be available.
Hence, once QMI server arrive event is received, request for QMI PHY
capabilities of device, and defer the host capability send for that device.
After QMI PHY capability is received for all the devices in the group
trigger the host capability event for the deferred devices in the group.
Hence, add changes to defer the QMI host capability event until the device
group is ready and then resume the QMI exchange for all the device with
host capabilities.
Signed-off-by: Karthikeyan Periyasamy <quic_periyasa@quicinc.com> Signed-off-by: Harshitha Prem <quic_hprem@quicinc.com> Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com> Link: https://patch.msgid.link/20241204163216.433795-7-kvalo@kernel.org Signed-off-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com>
wifi: ath12k: move struct ath12k_hw from per device to group
Currently, hardware abstractions (ah) of different radio bands are tightly
coupled to a single device (ab). But, with hardware device group abstraction
(ag), multiple radios across different devices in a group can form different
combinations of hardware abstractions (ah) within the group. Hence, the mapping
between ah to ab can be removed and instead it can be mapped with struct
ath12k_hw_group (ag).
Current mapping between struct ath12k_hw (ah), struct ath12k_base (ab) and struct
ath12k_hw_group (ag):
After hardware device group abstraction moving ah array out of ab to ag:
+----------------------------------------------+
| +---------------+ +---------------+ |
| |ath12k_hw (ah) | |ath12k_hw (ah) | |
| +---------------+ +---------------+ |
| +-------------------------------------+ |
| | +-----------+ +-----------+ | |
| | | ar (2GHz) | | ar (5GHz) | | |
| | +-----------+ +-----------+ | |
| | Dual band device-1 (ab) | |
| +-------------------------------------+ |
| ath12k_hw_group (ag) based on group id |
+----------------------------------------------+
This decoupling of struct ath12k_hw (ah) from struct ath12k_base (ab) and
mapping it to struct ath12k_hw_group (ag) will help in forming different
combinations of multi-link devices.
Signed-off-by: Karthikeyan Periyasamy <quic_periyasa@quicinc.com> Signed-off-by: Harshitha Prem <quic_hprem@quicinc.com> Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com> Link: https://patch.msgid.link/20241204163216.433795-6-kvalo@kernel.org Signed-off-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com>
wifi: ath12k: refactor core start based on hardware group
Currently, mac allocate/register and core_pdev_create are initiated immediately
when QMI firmware ready event is received for a particular device. With
hardware device group abstraction, QMI firmware ready event can be received
simultaneously for different devices in the group and so, it should not be
registered immediately rather it has to be deferred until all devices in the
group has received QMI firmware ready.
To handle this, refactor the code of core start to have registering within
ath12k_core_hw_group_start() and unregistering in ath12k_core_hw_group_stop().
Currently, single device is probed, and once firmware is ready, the device
is registered to mac80211. For Multi-Link Operation, different bands of
different devices or same device are part of a single wiphy and for
this, hardware device group abstraction is needed.
Hardware device group abstraction - when there are multiple devices (with
single radio or dual radio) that are connected by any means of interface
for communicating between them, then these devices can be combined
together as a single group using a group id to form a group abstraction
and register to mac80211.
The grouping information of multiple devices would be based on device tree
during device probe (will be implemented in future patches). If no such
information is available, then a single device will be part of group abstraction
and registered to mac80211, else multiple devices advertised in device tree are
combined and then registered to mac80211.
For device group abstraction, a base structure ath12k_hw_group (ag) and the
helpers are implemented. These helpers are used during device probe and mapping
the group to the devices involved.
An illustration of how multiple devices might be combined together in
future based on group id:
+------------------------------------------------------------------------+
| +-------------------------------------+ +-------------------+ |
| | +-----------+ | | +-----------+ | | +-----------+ | |
| | | ar (2GHz) | | | | ar (5GHz) | | | | ar (6GHz) | | |
| | +-----------+ | | +-----------+ | | +-----------+ | |
| | ath12k_base (ab) | | ath12k_base (ab) | |
| | (Dual band device) | | | |
| +-------------------------------------+ +-------------------+ |
| ath12k_hw_group (ag) based on group id |
+------------------------------------------------------------------------+
In the above representation, two devices are combined into single group
based on group id.
Add base code changes where single device would be part of a group with an
invalid group id forming an group abstraction. Multi device grouping will
be introduced in future.
wifi: ath12k: rename mlo_capable_flags to single_chip_mlo_supp
At present, the mlo_capable_flags in ath12k_base is used to indicate whether
the chip supports inter (QCN9274) or intra (WCN7850) chip MLO. However, it’s
possible that the chip supports neither, especially with older firmware
versions. Additionally, if intra chip MLO is not supported, inter chip MLO will
also be non-functional. Therefore, having two separate flags for this is
unnecessary.
Therefore, rename this flag to single_chip_mlo_supp. At the same time convert
it into a bool data type. Also, get rid of the enums defined earlier.
For the QCN9274 family of chipsets, this will be set only when firmware
advertises the support during the QMI exchange.
For the WCN7850 family of chipsets, since the event is not supported,
assumption is made that single chip MLO is supported.
wifi: ath12k: move ATH12K_FLAG_REGISTERED handling to ath12k_mac_register()
When hardware device group abstraction is introduced, a group abstraction is
registered to mac80211 rather than a particular single device. So we cannot
set the device registered when the QMI firmware ready event is received, only
after all the devices in group have received the event. To do that set and
unset ATH12K_FLAG_REGISTERED flag inside ath12k_mac_register() and
ath12k_mac_unregister() respectively.
Signed-off-by: Karthikeyan Periyasamy <quic_periyasa@quicinc.com> Signed-off-by: Harshitha Prem <quic_hprem@quicinc.com> Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com> Link: https://patch.msgid.link/20241204163216.433795-2-kvalo@kernel.org Signed-off-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com>
Chih-Kang Chang [Thu, 28 Nov 2024 05:54:33 +0000 (13:54 +0800)]
wifi: rtw89: add crystal_cap check to avoid setting as overflow value
In the original flow, the crystal_cap might be calculated as a negative
value and set as an overflow value. Therefore, we added a check to limit
the calculated crystal_cap value. Additionally, we shrank the crystal_cap
adjustment according to specific CFO.
Zong-Zhe Yang [Thu, 28 Nov 2024 05:54:32 +0000 (13:54 +0800)]
wifi: rtw89: refine link handling for link_sta_rc_update
The original handling will iterate all active links under the given sta
and apply the changes to each. Now, stack tweaks ops from sta_rc_update
to link_sta_rc_update, which means targeting a given link. Then, our link
iteration looks redundant. So, refine it to apply the changes to the link
directly.
Chih-Kang Chang [Thu, 28 Nov 2024 05:54:31 +0000 (13:54 +0800)]
wifi: rtw89: 8922a: use RSSI from PHY report in RX descriptor
The PPDU status of probe response will fail to parse the IE due to being
filtered by the to_self check. Therefore, we parse RSSI from PHY report in
RX descriptor.
Po-Hao Huang [Thu, 28 Nov 2024 05:54:30 +0000 (13:54 +0800)]
wifi: rtw89: 8852bt: add beacon filter and CQM support
Declare beacon filter and connection monitor for 8852BT.
This offloads connection monitor mechanism to firmware, and this
is required for the MCC feature.
Po-Hao Huang [Thu, 28 Nov 2024 05:54:29 +0000 (13:54 +0800)]
wifi: rtw89: 8852b: add beacon filter and CQM support
Declare beacon filter and connection monitor for 8852B.
This offloads connection monitor mechanism to firmware, and this
is required for the MCC feature.
Currently, ieee80211_ie_build_he_oper() lacks support for 320 MHz handling
(already noted as a TODO). This is because 320 MHz is not included in
IEEE 802.11-ax. However, IEEE 802.11-be introduces 320 MHz support and if
the chandef indicates a 320 MHz bandwidth and is used directly as it is, it
will result in an incorrect HE Operation Information Element.
In order to support EHT 320 MHz, HE Operation Element should indicate
bandwidth as 160 MHz only. In EHT Operation IE, the correct bandwidth will
be present. Devices capable of EHT can parse EHT Information Element and
connect in 320 MHz and other HE capable devices can parse HE and can
connect in 160 MHz.
Add support to downgrade the bandwidth in ieee80211_ie_build_he_oper()
during 320 MHz operation and advertise it.
When running ethtool on a monitor interface, the channel
wasn't reporting properly. This adds logic to properly report
the channel for monitor interfaces in ethtool.
ML interfaces can have multiple affiliated links to it and
hence there is a need to report tx power of specified link
rather deflink.
Add changes to report tx power of requested link from mac80211,
also pass link id as an argument in get_tx_power op so that supported
drivers can use it to report link's tx power.
wifi: cfg80211: send MLO links tx power info in GET_INTERFACE
Currently, TX power is reported on interface/wdev level as
part of NL80211_CMD_GET_INTERFACE. With MLO, Multiple links
can be part of an interface/wdev and hence its necessary to
report the TX power of each link.
Add support to send tx power for all valid links of an MLD as
part of NL80211_CMD_GET_INTERFACE request.
As far as userspace is concerned, there is no behavioral change
for Non-ML Interfaces. For ML interfaces, userspace should fetch
TX power that is nested inside NL80211_ATTR_MLO_LINKS, similar to
how channel info(NL80211_ATTR_WIPHY_FREQ) is fetched.
Aloka Dixit [Wed, 27 Nov 2024 18:02:55 +0000 (10:02 -0800)]
wifi: mac80211: fix variable used in for_each_sdata_link()
Macro for_each_sdata_link() accepts input '_local' but uses 'local'
in its processing. This currently works because all the functions
calling this macro have declared 'local' as a variable themselves.
But this results in compilation error when a new caller uses
'sdata->local' instead of declaring 'local' variable.
Use '_local' instead of 'local' in for_each_sdata_link().
Linus Torvalds [Sun, 1 Dec 2024 21:38:24 +0000 (13:38 -0800)]
Merge tag 'i2c-for-6.13-rc1-part3' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux
Pull i2c component probing support from Wolfram Sang:
"Add OF component probing.
Some devices are designed and manufactured with some components having
multiple drop-in replacement options. These components are often
connected to the mainboard via ribbon cables, having the same signals
and pin assignments across all options. These may include the display
panel and touchscreen on laptops and tablets, and the trackpad on
laptops. Sometimes which component option is used in a particular
device can be detected by some firmware provided identifier, other
times that information is not available, and the kernel has to try to
probe each device.
Instead of a delicate dance between drivers and device tree quirks,
this change introduces a simple I2C component probe function. For a
given class of devices on the same I2C bus, it will go through all of
them, doing a simple I2C read transfer and see which one of them
responds. It will then enable the device that responds"
* tag 'i2c-for-6.13-rc1-part3' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux:
MAINTAINERS: fix typo in I2C OF COMPONENT PROBER
of: base: Document prefix argument for of_get_next_child_with_prefix()
i2c: Fix whitespace style issue
arm64: dts: mediatek: mt8173-elm-hana: Mark touchscreens and trackpads as fail
platform/chrome: Introduce device tree hardware prober
i2c: of-prober: Add GPIO support to simple helpers
i2c: of-prober: Add simple helpers for regulator support
i2c: Introduce OF component probe function
of: base: Add for_each_child_of_node_with_prefix()
of: dynamic: Add of_changeset_update_prop_string
Linus Torvalds [Sun, 1 Dec 2024 21:10:51 +0000 (13:10 -0800)]
Merge tag 'trace-printf-v6.13' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace
Pull bprintf() removal from Steven Rostedt:
- Remove unused bprintf() function, that was added with the rest of the
"bin-printf" functions.
These are functions that are used by trace_printk() that allows to
quickly save the format and arguments into the ring buffer without
the expensive processing of converting numbers to ASCII. Then on
output, at a much later time, the ring buffer is read and the string
processing occurs then. The bprintf() was added for consistency but
was never used. It can be safely removed.
* tag 'trace-printf-v6.13' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:
printf: Remove unused 'bprintf'
Linus Torvalds [Sun, 1 Dec 2024 20:41:21 +0000 (12:41 -0800)]
Merge tag 'timers_urgent_for_v6.13_rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull timer fixes from Borislav Petkov:
- Fix a case where posix timers with a thread-group-wide target would
miss signals if some of the group's threads are exiting
- Fix a hang caused by ndelay() calling the wrong delay function
__udelay()
- Fix a wrong offset calculation in adjtimex(2) when using ADJ_MICRO
(microsecond resolution) and a negative offset
* tag 'timers_urgent_for_v6.13_rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
posix-timers: Target group sigqueue to current task only if not exiting
delay: Fix ndelay() spuriously treated as udelay()
ntp: Remove invalid cast in time offset math
Linus Torvalds [Sun, 1 Dec 2024 20:37:58 +0000 (12:37 -0800)]
Merge tag 'irq_urgent_for_v6.13_rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull irq fixes from Borislav Petkov:
- Move the ->select callback to the correct ops structure in
irq-mvebu-sei to fix some Marvell Armada platforms
- Add a workaround for Hisilicon ITS erratum 162100801 which can cause
some virtual interrupts to get lost
- More platform_driver::remove() conversion
* tag 'irq_urgent_for_v6.13_rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
irqchip: Switch back to struct platform_driver::remove()
irqchip/gicv3-its: Add workaround for hip09 ITS erratum 162100801
irqchip/irq-mvebu-sei: Move misplaced select() callback to SEI CP domain
Linus Torvalds [Sun, 1 Dec 2024 20:35:37 +0000 (12:35 -0800)]
Merge tag 'x86_urgent_for_v6.13_rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull x86 fixes from Borislav Petkov:
- Add a terminating zero end-element to the array describing AMD CPUs
affected by erratum 1386 so that the matching loop actually
terminates instead of going off into the weeds
- Update the boot protocol documentation to mention the fact that the
preferred address to load the kernel to is considered in the
relocatable kernel case too
- Flush the memory buffer containing the microcode patch after applying
microcode on AMD Zen1 and Zen2, to avoid unnecessary slowdowns
- Make sure the PPIN CPU feature flag is cleared on all CPUs if PPIN
has been disabled
* tag 'x86_urgent_for_v6.13_rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/CPU/AMD: Terminate the erratum_1386_microcode array
x86/Documentation: Update algo in init_size description of boot protocol
x86/microcode/AMD: Flush patch buffer mapping after application
x86/mm: Carve out INVLPG inline asm for use by others
x86/cpu: Fix PPIN initialization
Linus Torvalds [Sun, 1 Dec 2024 17:23:33 +0000 (09:23 -0800)]
strscpy: write destination buffer only once
The point behind strscpy() was to once and for all avoid all the
problems with 'strncpy()' and later broken "fixed" versions like
strlcpy() that just made things worse.
So strscpy not only guarantees NUL-termination (unlike strncpy), it also
doesn't do unnecessary padding at the destination. But at the same time
also avoids byte-at-a-time reads and writes by _allowing_ some extra NUL
writes - within the size, of course - so that the whole copy can be done
with word operations.
It is also stable in the face of a mutable source string: it explicitly
does not read the source buffer multiple times (so an implementation
using "strnlen()+memcpy()" would be wrong), and does not read the source
buffer past the size (like the mis-design that is strlcpy does).
Finally, the return value is designed to be simple and unambiguous: if
the string cannot be copied fully, it returns an actual negative error,
making error handling clearer and simpler (and the caller already knows
the size of the buffer). Otherwise it returns the string length of the
result.
However, there was one final stability issue that can be important to
callers: the stability of the destination buffer.
In particular, the same way we shouldn't read the source buffer more
than once, we should avoid doing multiple writes to the destination
buffer: first writing a potentially non-terminated string, and then
terminating it with NUL at the end does not result in a stable result
buffer.
Yes, it gives the right result in the end, but if the rule for the
destination buffer was that it is _always_ NUL-terminated even when
accessed concurrently with updates, the final byte of the buffer needs
to always _stay_ as a NUL byte.
[ Note that "final byte is NUL" here is literally about the final byte
in the destination array, not the terminating NUL at the end of the
string itself. There is no attempt to try to make concurrent reads and
writes give any kind of consistent string length or contents, but we
do want to guarantee that there is always at least that final
terminating NUL character at the end of the destination array if it
existed before ]
This is relevant in the kernel for the tsk->comm[] array, for example.
Even without locking (for either readers or writers), we want to know
that while the buffer contents may be garbled, it is always a valid C
string and always has a NUL character at 'comm[TASK_COMM_LEN-1]' (and
never has any "out of thin air" data).
So avoid any "copy possibly non-terminated string, and terminate later"
behavior, and write the destination buffer only once.
bprintf() is unused. Remove it. It was added in the commit 4370aa4aa753
("vsprintf: add binary printf") but as far as I can see was never used,
unlike the other two functions in that patch.
Link: https://lore.kernel.org/20241002173147.210107-1-linux@treblig.org Reviewed-by: Andy Shevchenko <andy@kernel.org> Acked-by: Petr Mladek <pmladek@suse.com> Signed-off-by: Dr. David Alan Gilbert <linux@treblig.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Linus Torvalds [Sun, 1 Dec 2024 02:23:05 +0000 (18:23 -0800)]
Merge tag 'pci-v6.13-fixes-1' of git://git.kernel.org/pub/scm/linux/kernel/git/pci/pci
Pull PCI fix from Bjorn Helgaas:
- When removing a PCI device, only look up and remove a platform device
if there is an associated device node for which there could be a
platform device, to fix a merge window regression (Brian Norris)
* tag 'pci-v6.13-fixes-1' of git://git.kernel.org/pub/scm/linux/kernel/git/pci/pci:
PCI/pwrctrl: Unregister platform device only if one actually exists
Linus Torvalds [Sun, 1 Dec 2024 02:14:56 +0000 (18:14 -0800)]
Merge tag 'lsm-pr-20241129' of git://git.kernel.org/pub/scm/linux/kernel/git/pcmoore/lsm
Pull ima fix from Paul Moore:
"One small patch to fix a function parameter / local variable naming
snafu that went up to you in the current merge window"
* tag 'lsm-pr-20241129' of git://git.kernel.org/pub/scm/linux/kernel/git/pcmoore/lsm:
ima: uncover hidden variable in ima_match_rules()
Linus Torvalds [Sat, 30 Nov 2024 23:47:29 +0000 (15:47 -0800)]
Merge tag 'block-6.13-20242901' of git://git.kernel.dk/linux
Pull more block updates from Jens Axboe:
- NVMe pull request via Keith:
- Use correct srcu list traversal (Breno)
- Scatter-gather support for metadata (Keith)
- Fabrics shutdown race condition fix (Nilay)
- Persistent reservations updates (Guixin)
- Add the required bits for MD atomic write support for raid0/1/10
- Correct return value for unknown opcode in ublk
- Fix deadlock with zone revalidation
- Fix for the io priority request vs bio cleanups
- Use the correct unsigned int type for various limit helpers
- Fix for a race in loop
- Cleanup blk_rq_prep_clone() to prevent uninit-value warning and make
it easier for actual humans to read
- Fix potential UAF when iterating tags
- A few fixes for bfq-iosched UAF issues
- Fix for brd discard not decrementing the allocated page count
- Various little fixes and cleanups
* tag 'block-6.13-20242901' of git://git.kernel.dk/linux: (36 commits)
brd: decrease the number of allocated pages which discarded
block, bfq: fix bfqq uaf in bfq_limit_depth()
block: Don't allow an atomic write be truncated in blkdev_write_iter()
mq-deadline: don't call req_get_ioprio from the I/O completion handler
block: Prevent potential deadlock in blk_revalidate_disk_zones()
block: Remove extra part pointer NULLify in blk_rq_init()
nvme: tuning pr code by using defined structs and macros
nvme: introduce change ptpl and iekey definition
block: return bool from get_disk_ro and bdev_read_only
block: remove a duplicate definition for bdev_read_only
block: return bool from blk_rq_aligned
block: return unsigned int from blk_lim_dma_alignment_and_pad
block: return unsigned int from queue_dma_alignment
block: return unsigned int from bdev_io_opt
block: req->bio is always set in the merge code
block: don't bother checking the data direction for merges
block: blk-mq: fix uninit-value in blk_rq_prep_clone and refactor
Revert "block, bfq: merge bfq_release_process_ref() into bfq_put_cooperator()"
md/raid10: Atomic write support
md/raid1: Atomic write support
...
Linus Torvalds [Sat, 30 Nov 2024 23:43:02 +0000 (15:43 -0800)]
Merge tag 'io_uring-6.13-20242901' of git://git.kernel.dk/linux
Pull more io_uring updates from Jens Axboe:
- Remove a leftover struct from when the cqwait registered waiting was
transitioned to regions.
- Fix for an issue introduced in this merge window, where nop->fd might
be used uninitialized. Ensure it's always set.
- Add capping of the task_work run in local task_work mode, to prevent
bursty and long chains from adding too much latency.
- Work around xa_store() leaving ->head non-NULL if it encounters an
allocation error during storing. Just a debug trigger, and can go
away once xa_store() behaves in a more expected way for this
condition. Not a major thing as it basically requires fault injection
to trigger it.
- Fix a few mapping corner cases
- Fix KCSAN complaint on reading the table size post unlock. Again not
a "real" issue, but it's easy to silence by just keeping the reading
inside the lock that protects it.
* tag 'io_uring-6.13-20242901' of git://git.kernel.dk/linux:
io_uring/tctx: work around xa_store() allocation error issue
io_uring: fix corner case forgetting to vunmap
io_uring: fix task_work cap overshooting
io_uring: check for overflows in io_pin_pages
io_uring/nop: ensure nop->fd is always initialized
io_uring: limit local tw done
io_uring: add io_local_work_pending()
io_uring/region: return negative -E2BIG in io_create_region()
io_uring: protect register tracing
io_uring: remove io_uring_cqwait_reg_arg
Linus Torvalds [Sat, 30 Nov 2024 23:36:17 +0000 (15:36 -0800)]
Merge tag 'dma-mapping-6.13-2024-11-30' of git://git.infradead.org/users/hch/dma-mapping
Pull dma-mapping fix from Christoph Hellwig:
- fix physical address calculation for struct dma_debug_entry (Fedor
Pchelkin)
* tag 'dma-mapping-6.13-2024-11-30' of git://git.infradead.org/users/hch/dma-mapping:
dma-debug: fix physical address calculation for struct dma_debug_entry
Linus Torvalds [Sat, 30 Nov 2024 22:51:08 +0000 (14:51 -0800)]
Merge tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm
Pull more kvm updates from Paolo Bonzini:
- ARM fixes
- RISC-V Svade and Svadu (accessed and dirty bit) extension support for
host and guest
* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm:
KVM: riscv: selftests: Add Svade and Svadu Extension to get-reg-list test
RISC-V: KVM: Add Svade and Svadu Extensions Support for Guest/VM
dt-bindings: riscv: Add Svade and Svadu Entries
RISC-V: Add Svade and Svadu Extensions Support
KVM: arm64: Use MDCR_EL2.HPME to evaluate overflow of hyp counters
KVM: arm64: Ignore PMCNTENSET_EL0 while checking for overflow status
KVM: arm64: Mark set_sysreg_masks() as inline to avoid build failure
KVM: arm64: vgic-its: Add stronger type-checking to the ITS entry sizes
KVM: arm64: vgic: Kill VGIC_MAX_PRIVATE definition
KVM: arm64: vgic: Make vgic_get_irq() more robust
KVM: arm64: vgic-v3: Sanitise guest writes to GICR_INVLPIR
Linus Torvalds [Sat, 30 Nov 2024 22:45:29 +0000 (14:45 -0800)]
Merge tag 'sh-for-v6.13-tag1' of git://git.kernel.org/pub/scm/linux/kernel/git/glaubitz/sh-linux
Pull sh updates from John Paul Adrian Glaubitz:
"Two small fixes.
The first one by Huacai Chen addresses a runtime warning when
CONFIG_CPUMASK_OFFSTACK and CONFIG_DEBUG_PER_CPU_MAPS are selected
which occurs because the cpuinfo code on sh incorrectly uses NR_CPUS
when iterating CPUs instead of the runtime limit nr_cpu_ids.
A second fix by Dan Carpenter fixes a use-after-free bug in
register_intc_controller() which occurred as a result of improper
error handling in the interrupt controller driver code when
registering an interrupt controller during plat_irq_setup() on sh"
* tag 'sh-for-v6.13-tag1' of git://git.kernel.org/pub/scm/linux/kernel/git/glaubitz/sh-linux:
sh: intc: Fix use-after-free bug in register_intc_controller()
sh: cpuinfo: Fix a warning for CONFIG_CPUMASK_OFFSTACK
Linus Torvalds [Sat, 30 Nov 2024 22:33:44 +0000 (14:33 -0800)]
Merge tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux
Pull arm64 fixes from Catalin Marinas:
- Deselect ARCH_CORRECT_STACKTRACE_ON_KRETPROBE so that tests depending
on it don't run (and fail) on arm64
- Fix lockdep assert in the Arm SMMUv3 PMU driver
- Fix the port and device ID bits setting in the Arm CMN perf driver
* tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux:
perf/arm-cmn: Ensure port and device id bits are set properly
perf/arm-smmuv3: Fix lockdep assert in ->event_init()
arm64: disable ARCH_CORRECT_STACKTRACE_ON_KRETPROBE tests
Patryk Wlazlyn [Wed, 2 Oct 2024 13:05:15 +0000 (15:05 +0200)]
tools/power turbostat: Add RAPL psys as a built-in counter
Introduce the counter as a part of global, platform counters structure.
We open the counter for only one cpu, but otherwise treat it as an
ordinary RAPL counter, allowing for grouped perf read.
The counter is disabled by default, because it's interpretation may
require additional, platform specific information, making it unsuitable
for general use.
Signed-off-by: Patryk Wlazlyn <patryk.wlazlyn@linux.intel.com> Signed-off-by: Len Brown <len.brown@intel.com>
Add '+' to optstring when early scanning for --no-msr and --no-perf.
It causes option processing to stop as soon as a nonoption argument is
encountered, effectively skipping child's arguments.
Fixes: 3e4048466c39 ("tools/power turbostat: Add --no-msr option") Signed-off-by: Patryk Wlazlyn <patryk.wlazlyn@linux.intel.com> Signed-off-by: Len Brown <len.brown@intel.com>
Patryk Wlazlyn [Thu, 24 Oct 2024 13:17:45 +0000 (15:17 +0200)]
tools/power turbostat: Force --no-perf in --dump mode
Force the --no-perf early to prevent using it as a source. User asks for
raw values, but perf returns them relative to the opening of the file
descriptor.
Signed-off-by: Patryk Wlazlyn <patryk.wlazlyn@linux.intel.com> Signed-off-by: Len Brown <len.brown@intel.com>
Zhang Rui [Thu, 14 Nov 2024 07:59:46 +0000 (15:59 +0800)]
tools/power turbostat: Add support for /sys/class/drm/card1
On some machines, the graphics device is enumerated as
/sys/class/drm/card1 instead of /sys/class/drm/card0. The current
implementation does not handle this scenario, resulting in the loss of
graphics C6 residency and frequency information.
Add support for /sys/class/drm/card1, ensuring that turbostat can
retrieve and display the graphics columns for these platforms.
Signed-off-by: Zhang Rui <rui.zhang@intel.com> Signed-off-by: Len Brown <len.brown@intel.com>
Zhang Rui [Thu, 14 Nov 2024 07:59:45 +0000 (15:59 +0800)]
tools/power turbostat: Cache graphics sysfs file descriptors during probe
Snapshots of the graphics sysfs knobs are taken based on file
descriptors. To optimize this process, open the files and cache the file
descriptors during the graphics probe phase. As a result, the previously
cached pathnames become redundant and are removed.
This change aims to streamline the code without altering its functionality.
No functional change intended.
Signed-off-by: Zhang Rui <rui.zhang@intel.com> Signed-off-by: Len Brown <len.brown@intel.com>
Currently, there is an inconsistency in how graphics sysfs knobs are
accessed: graphics residency sysfs knobs are opened and closed for each
read, while graphics frequency sysfs knobs are opened once and remain
open until turbostat exits. This inconsistency is confusing and adds
unnecessary code complexity.
Consolidate the access method by opening the sysfs files once and
reusing the file pointers for subsequent accesses. This approach
simplifies the code and ensures a consistent method for accessing
graphics sysfs knobs.
Signed-off-by: Zhang Rui <rui.zhang@intel.com> Signed-off-by: Len Brown <len.brown@intel.com>
In various generations, platforms often share a majority of features,
diverging only in a few specific aspects. The current approach of using
hardcoded values in 'platform_features' structure fails to effectively
represent these divergences.
To improve the description of platform divergence:
1. Each newly introduced 'platform_features' structure must have a base,
typically derived from the previous generation.
2. Platform feature values should be inherited from the base structure
rather than being hardcoded.
This approach ensures a more accurate and maintainable representation of
platform-specific features across different generations.
Converts `adl_features` and `lnl_features` to follow this new scheme.
No functional change.
Signed-off-by: Zhang Rui <rui.zhang@intel.com> Signed-off-by: Len Brown <len.brown@intel.com>
Zhang Rui [Tue, 27 Aug 2024 05:07:51 +0000 (13:07 +0800)]
tools/power turbostat: Fix trailing '\n' parsing
parse_cpu_string() parses the string input either from command line or
from /sys/fs/cgroup/cpuset.cpus.effective to get a list of CPUs that
turbostat can run with.
The cpu string returned by /sys/fs/cgroup/cpuset.cpus.effective contains
a trailing '\n', but strtoul() fails to treat this as an error.
That says, for the code below
val = ("\n", NULL, 10);
val returns 0, and errno is also not set.
As a result, CPU0 is erroneously considered as allowed CPU and this
causes failures when turbostat tries to run on CPU0.
get_counters: Could not migrate to CPU 0
...
turbostat: re-initialized with num_cpus 8, allowed_cpus 5
get_counters: Could not migrate to CPU 0
Add a check to return immediately if '\n' or '\0' is detected.
Fixes: 8c3dd2c9e542 ("tools/power/turbostat: Abstrct function for parsing cpu string") Signed-off-by: Zhang Rui <rui.zhang@intel.com> Signed-off-by: Len Brown <len.brown@intel.com>
Patryk Wlazlyn [Tue, 20 Aug 2024 16:47:59 +0000 (18:47 +0200)]
tools/power turbostat: Allow using cpu device in perf counters on hybrid platforms
Intel hybrid platforms expose different perf devices for P and E cores.
Instead of one, "/sys/bus/event_source/devices/cpu" device, there are
"/sys/bus/event_source/devices/{cpu_core,cpu_atom}".
This, however makes it more complicated for the user,
because most of the counters are available on both and had to be
handled manually.
This patch allows users to use "virtual" cpu device that is seemingly
translated to cpu_core and cpu_atom perf devices, depending on the type
of a CPU we are opening the counter for.
Signed-off-by: Patryk Wlazlyn <patryk.wlazlyn@linux.intel.com> Signed-off-by: Len Brown <len.brown@intel.com>
Patryk Wlazlyn [Wed, 7 Aug 2024 11:43:39 +0000 (13:43 +0200)]
tools/power turbostat: Fix column printing for PMT xtal_time counters
If the very first printed column was for a PMT counter of type xtal_time
we would misalign the column header, because we were always printing the
delimiter.
Signed-off-by: Patryk Wlazlyn <patryk.wlazlyn@linux.intel.com> Signed-off-by: Len Brown <len.brown@intel.com>
projects / thirdparty / kernel / linux.git / log
