* [PATCH] wifi: mt76: fix oops on non-dbdc mt7986
@ 2024-07-13 13:00 Bjørn Mork
2024-07-31 9:27 ` Kalle Valo
0 siblings, 1 reply; 4+ messages in thread
From: Bjørn Mork @ 2024-07-13 13:00 UTC (permalink / raw)
To: linux-wireless; +Cc: linux-mediatek, Felix Fietkau, Bjørn Mork
mt7915_band_config() sets band_idx = 1 on the main phy for mt7986
with MT7975_ONE_ADIE or MT7976_ONE_ADIE.
Commit 0335c034e726 ("wifi: mt76: fix race condition related to
checking tx queue fill status") introduced a dereference of the
phys array indirectly indexed by band_idx via wcid->phy_idx in
mt76_wcid_cleanup(). This caused the following Oops on affected
mt7986 devices:
Unable to handle kernel read from unreadable memory at virtual address 0000000000000024
Mem abort info:
ESR = 0x0000000096000005
EC = 0x25: DABT (current EL), IL = 32 bits
SET = 0, FnV = 0
EA = 0, S1PTW = 0
FSC = 0x05: level 1 translation fault
Data abort info:
ISV = 0, ISS = 0x00000005
CM = 0, WnR = 0
user pgtable: 4k pages, 39-bit VAs, pgdp=0000000042545000
[0000000000000024] pgd=0000000000000000, p4d=0000000000000000, pud=0000000000000000
Internal error: Oops: 0000000096000005 [#1] SMP
Modules linked in: ... mt7915e mt76_connac_lib mt76 mac80211 cfg80211 ...
CPU: 2 PID: 1631 Comm: hostapd Not tainted 5.15.150 #0
Hardware name: ZyXEL EX5700 (Telenor) (DT)
pstate: 80400005 (Nzcv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--)
pc : mt76_wcid_cleanup+0x84/0x22c [mt76]
lr : mt76_wcid_cleanup+0x64/0x22c [mt76]
sp : ffffffc00a803700
x29: ffffffc00a803700 x28: ffffff80008f7300 x27: ffffff80003f3c00
x26: ffffff80000a7880 x25: ffffffc008c26e00 x24: 0000000000000001
x23: ffffffc000a68114 x22: 0000000000000000 x21: ffffff8004172cc8
x20: ffffffc00a803748 x19: ffffff8004152020 x18: 0000000000000000
x17: 00000000000017c0 x16: ffffffc008ef5000 x15: 0000000000000be0
x14: ffffff8004172e28 x13: ffffff8004172e28 x12: 0000000000000000
x11: 0000000000000000 x10: ffffff8004172e30 x9 : ffffff8004172e28
x8 : 0000000000000000 x7 : ffffff8004156020 x6 : 0000000000000000
x5 : 0000000000000031 x4 : 0000000000000000 x3 : 0000000000000001
x2 : 0000000000000000 x1 : ffffff80008f7300 x0 : 0000000000000024
Call trace:
mt76_wcid_cleanup+0x84/0x22c [mt76]
__mt76_sta_remove+0x70/0xbc [mt76]
mt76_sta_state+0x8c/0x1a4 [mt76]
mt7915_eeprom_get_power_delta+0x11e4/0x23a0 [mt7915e]
drv_sta_state+0x144/0x274 [mac80211]
sta_info_move_state+0x1cc/0x2a4 [mac80211]
sta_set_sinfo+0xaf8/0xc24 [mac80211]
sta_info_destroy_addr_bss+0x4c/0x6c [mac80211]
ieee80211_color_change_finish+0x1c08/0x1e70 [mac80211]
cfg80211_check_station_change+0x1360/0x4710 [cfg80211]
genl_family_rcv_msg_doit+0xb4/0x110
genl_rcv_msg+0xd0/0x1bc
netlink_rcv_skb+0x58/0x120
genl_rcv+0x34/0x50
netlink_unicast+0x1f0/0x2ec
netlink_sendmsg+0x198/0x3d0
____sys_sendmsg+0x1b0/0x210
___sys_sendmsg+0x80/0xf0
__sys_sendmsg+0x44/0xa0
__arm64_sys_sendmsg+0x20/0x30
invoke_syscall.constprop.0+0x4c/0xe0
do_el0_svc+0x40/0xd0
el0_svc+0x14/0x4c
el0t_64_sync_handler+0x100/0x110
el0t_64_sync+0x15c/0x160
Code: d2800002 910092c0 52800023 f9800011 (885f7c01)
---[ end trace 7e42dd9a39ed2281 ]---
Fix by using mt76_dev_phy() which will map band_idx to the correct phy
for all hardware combinations.
Fixes: 0335c034e726 ("wifi: mt76: fix race condition related to checking tx queue fill status")
Link: https://github.com/openwrt/openwrt/issues/14548
Signed-off-by: Bjørn Mork <bjorn@mork.no>
---
drivers/net/wireless/mediatek/mt76/mac80211.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/wireless/mediatek/mt76/mac80211.c b/drivers/net/wireless/mediatek/mt76/mac80211.c
index e8ba2e4e8484..b5dbcf925f92 100644
--- a/drivers/net/wireless/mediatek/mt76/mac80211.c
+++ b/drivers/net/wireless/mediatek/mt76/mac80211.c
@@ -1524,7 +1524,7 @@ EXPORT_SYMBOL_GPL(mt76_wcid_init);
void mt76_wcid_cleanup(struct mt76_dev *dev, struct mt76_wcid *wcid)
{
- struct mt76_phy *phy = dev->phys[wcid->phy_idx];
+ struct mt76_phy *phy = mt76_dev_phy(dev, wcid->phy_idx);
struct ieee80211_hw *hw;
struct sk_buff_head list;
struct sk_buff *skb;
--
2.39.2
^ permalink raw reply related [flat|nested] 4+ messages in thread* Re: [PATCH] wifi: mt76: fix oops on non-dbdc mt7986
2024-07-13 13:00 [PATCH] wifi: mt76: fix oops on non-dbdc mt7986 Bjørn Mork
@ 2024-07-31 9:27 ` Kalle Valo
2024-07-31 10:26 ` Bjørn Mork
0 siblings, 1 reply; 4+ messages in thread
From: Kalle Valo @ 2024-07-31 9:27 UTC (permalink / raw)
To: Bjørn Mork; +Cc: linux-wireless, linux-mediatek, Felix Fietkau
Bjørn Mork <bjorn@mork.no> writes:
> mt7915_band_config() sets band_idx = 1 on the main phy for mt7986
> with MT7975_ONE_ADIE or MT7976_ONE_ADIE.
>
> Commit 0335c034e726 ("wifi: mt76: fix race condition related to
> checking tx queue fill status") introduced a dereference of the
> phys array indirectly indexed by band_idx via wcid->phy_idx in
> mt76_wcid_cleanup(). This caused the following Oops on affected
> mt7986 devices:
>
> Unable to handle kernel read from unreadable memory at virtual address 0000000000000024
> Mem abort info:
> ESR = 0x0000000096000005
> EC = 0x25: DABT (current EL), IL = 32 bits
> SET = 0, FnV = 0
> EA = 0, S1PTW = 0
> FSC = 0x05: level 1 translation fault
> Data abort info:
> ISV = 0, ISS = 0x00000005
> CM = 0, WnR = 0
> user pgtable: 4k pages, 39-bit VAs, pgdp=0000000042545000
> [0000000000000024] pgd=0000000000000000, p4d=0000000000000000, pud=0000000000000000
> Internal error: Oops: 0000000096000005 [#1] SMP
> Modules linked in: ... mt7915e mt76_connac_lib mt76 mac80211 cfg80211 ...
> CPU: 2 PID: 1631 Comm: hostapd Not tainted 5.15.150 #0
> Hardware name: ZyXEL EX5700 (Telenor) (DT)
> pstate: 80400005 (Nzcv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--)
> pc : mt76_wcid_cleanup+0x84/0x22c [mt76]
> lr : mt76_wcid_cleanup+0x64/0x22c [mt76]
> sp : ffffffc00a803700
> x29: ffffffc00a803700 x28: ffffff80008f7300 x27: ffffff80003f3c00
> x26: ffffff80000a7880 x25: ffffffc008c26e00 x24: 0000000000000001
> x23: ffffffc000a68114 x22: 0000000000000000 x21: ffffff8004172cc8
> x20: ffffffc00a803748 x19: ffffff8004152020 x18: 0000000000000000
> x17: 00000000000017c0 x16: ffffffc008ef5000 x15: 0000000000000be0
> x14: ffffff8004172e28 x13: ffffff8004172e28 x12: 0000000000000000
> x11: 0000000000000000 x10: ffffff8004172e30 x9 : ffffff8004172e28
> x8 : 0000000000000000 x7 : ffffff8004156020 x6 : 0000000000000000
> x5 : 0000000000000031 x4 : 0000000000000000 x3 : 0000000000000001
> x2 : 0000000000000000 x1 : ffffff80008f7300 x0 : 0000000000000024
> Call trace:
> mt76_wcid_cleanup+0x84/0x22c [mt76]
> __mt76_sta_remove+0x70/0xbc [mt76]
> mt76_sta_state+0x8c/0x1a4 [mt76]
> mt7915_eeprom_get_power_delta+0x11e4/0x23a0 [mt7915e]
> drv_sta_state+0x144/0x274 [mac80211]
> sta_info_move_state+0x1cc/0x2a4 [mac80211]
> sta_set_sinfo+0xaf8/0xc24 [mac80211]
> sta_info_destroy_addr_bss+0x4c/0x6c [mac80211]
>
> ieee80211_color_change_finish+0x1c08/0x1e70 [mac80211]
> cfg80211_check_station_change+0x1360/0x4710 [cfg80211]
> genl_family_rcv_msg_doit+0xb4/0x110
> genl_rcv_msg+0xd0/0x1bc
> netlink_rcv_skb+0x58/0x120
> genl_rcv+0x34/0x50
> netlink_unicast+0x1f0/0x2ec
> netlink_sendmsg+0x198/0x3d0
> ____sys_sendmsg+0x1b0/0x210
> ___sys_sendmsg+0x80/0xf0
> __sys_sendmsg+0x44/0xa0
> __arm64_sys_sendmsg+0x20/0x30
> invoke_syscall.constprop.0+0x4c/0xe0
> do_el0_svc+0x40/0xd0
> el0_svc+0x14/0x4c
> el0t_64_sync_handler+0x100/0x110
> el0t_64_sync+0x15c/0x160
> Code: d2800002 910092c0 52800023 f9800011 (885f7c01)
> ---[ end trace 7e42dd9a39ed2281 ]---
>
> Fix by using mt76_dev_phy() which will map band_idx to the correct phy
> for all hardware combinations.
>
> Fixes: 0335c034e726 ("wifi: mt76: fix race condition related to checking tx queue fill status")
> Link: https://github.com/openwrt/openwrt/issues/14548
> Signed-off-by: Bjørn Mork <bjorn@mork.no>
Should this go to wireless tree?
--
https://patchwork.kernel.org/project/linux-wireless/list/
https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches
^ permalink raw reply [flat|nested] 4+ messages in thread* Re: [PATCH] wifi: mt76: fix oops on non-dbdc mt7986
2024-07-31 9:27 ` Kalle Valo
@ 2024-07-31 10:26 ` Bjørn Mork
2024-07-31 11:27 ` Kalle Valo
0 siblings, 1 reply; 4+ messages in thread
From: Bjørn Mork @ 2024-07-31 10:26 UTC (permalink / raw)
To: Kalle Valo; +Cc: linux-wireless, linux-mediatek, Felix Fietkau
Kalle Valo <kvalo@kernel.org> writes:
> Bjørn Mork <bjorn@mork.no> writes:
>
>> mt7915_band_config() sets band_idx = 1 on the main phy for mt7986
>> with MT7975_ONE_ADIE or MT7976_ONE_ADIE.
>>
>> Commit 0335c034e726 ("wifi: mt76: fix race condition related to
>> checking tx queue fill status") introduced a dereference of the
>> phys array indirectly indexed by band_idx via wcid->phy_idx in
>> mt76_wcid_cleanup(). This caused the following Oops on affected
>> mt7986 devices:
>>
>> Unable to handle kernel read from unreadable memory at virtual address 0000000000000024
>> Mem abort info:
>> ESR = 0x0000000096000005
>> EC = 0x25: DABT (current EL), IL = 32 bits
>> SET = 0, FnV = 0
>> EA = 0, S1PTW = 0
>> FSC = 0x05: level 1 translation fault
>> Data abort info:
>> ISV = 0, ISS = 0x00000005
>> CM = 0, WnR = 0
>> user pgtable: 4k pages, 39-bit VAs, pgdp=0000000042545000
>> [0000000000000024] pgd=0000000000000000, p4d=0000000000000000, pud=0000000000000000
>> Internal error: Oops: 0000000096000005 [#1] SMP
>> Modules linked in: ... mt7915e mt76_connac_lib mt76 mac80211 cfg80211 ...
>> CPU: 2 PID: 1631 Comm: hostapd Not tainted 5.15.150 #0
>> Hardware name: ZyXEL EX5700 (Telenor) (DT)
>> pstate: 80400005 (Nzcv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--)
>> pc : mt76_wcid_cleanup+0x84/0x22c [mt76]
>> lr : mt76_wcid_cleanup+0x64/0x22c [mt76]
>> sp : ffffffc00a803700
>> x29: ffffffc00a803700 x28: ffffff80008f7300 x27: ffffff80003f3c00
>> x26: ffffff80000a7880 x25: ffffffc008c26e00 x24: 0000000000000001
>> x23: ffffffc000a68114 x22: 0000000000000000 x21: ffffff8004172cc8
>> x20: ffffffc00a803748 x19: ffffff8004152020 x18: 0000000000000000
>> x17: 00000000000017c0 x16: ffffffc008ef5000 x15: 0000000000000be0
>> x14: ffffff8004172e28 x13: ffffff8004172e28 x12: 0000000000000000
>> x11: 0000000000000000 x10: ffffff8004172e30 x9 : ffffff8004172e28
>> x8 : 0000000000000000 x7 : ffffff8004156020 x6 : 0000000000000000
>> x5 : 0000000000000031 x4 : 0000000000000000 x3 : 0000000000000001
>> x2 : 0000000000000000 x1 : ffffff80008f7300 x0 : 0000000000000024
>> Call trace:
>> mt76_wcid_cleanup+0x84/0x22c [mt76]
>> __mt76_sta_remove+0x70/0xbc [mt76]
>> mt76_sta_state+0x8c/0x1a4 [mt76]
>> mt7915_eeprom_get_power_delta+0x11e4/0x23a0 [mt7915e]
>> drv_sta_state+0x144/0x274 [mac80211]
>> sta_info_move_state+0x1cc/0x2a4 [mac80211]
>> sta_set_sinfo+0xaf8/0xc24 [mac80211]
>> sta_info_destroy_addr_bss+0x4c/0x6c [mac80211]
>>
>> ieee80211_color_change_finish+0x1c08/0x1e70 [mac80211]
>> cfg80211_check_station_change+0x1360/0x4710 [cfg80211]
>> genl_family_rcv_msg_doit+0xb4/0x110
>> genl_rcv_msg+0xd0/0x1bc
>> netlink_rcv_skb+0x58/0x120
>> genl_rcv+0x34/0x50
>> netlink_unicast+0x1f0/0x2ec
>> netlink_sendmsg+0x198/0x3d0
>> ____sys_sendmsg+0x1b0/0x210
>> ___sys_sendmsg+0x80/0xf0
>> __sys_sendmsg+0x44/0xa0
>> __arm64_sys_sendmsg+0x20/0x30
>> invoke_syscall.constprop.0+0x4c/0xe0
>> do_el0_svc+0x40/0xd0
>> el0_svc+0x14/0x4c
>> el0t_64_sync_handler+0x100/0x110
>> el0t_64_sync+0x15c/0x160
>> Code: d2800002 910092c0 52800023 f9800011 (885f7c01)
>> ---[ end trace 7e42dd9a39ed2281 ]---
>>
>> Fix by using mt76_dev_phy() which will map band_idx to the correct phy
>> for all hardware combinations.
>>
>> Fixes: 0335c034e726 ("wifi: mt76: fix race condition related to checking tx queue fill status")
>> Link: https://github.com/openwrt/openwrt/issues/14548
>> Signed-off-by: Bjørn Mork <bjorn@mork.no>
>
> Should this go to wireless tree?
I believe it should. If fixes a regression on the affected hardware,
introduced by commit 0335c034e726.
It should also go into any still maintained v6.7, v6.8, v6.9, v6.10
stable trees. But I assume they will pick it up automatically based on
the Fixes tag.
Bjørn
^ permalink raw reply [flat|nested] 4+ messages in thread* Re: [PATCH] wifi: mt76: fix oops on non-dbdc mt7986
2024-07-31 10:26 ` Bjørn Mork
@ 2024-07-31 11:27 ` Kalle Valo
0 siblings, 0 replies; 4+ messages in thread
From: Kalle Valo @ 2024-07-31 11:27 UTC (permalink / raw)
To: Bjørn Mork; +Cc: linux-wireless, linux-mediatek, Felix Fietkau
Bjørn Mork <bjorn@mork.no> writes:
> Kalle Valo <kvalo@kernel.org> writes:
>
>> Bjørn Mork <bjorn@mork.no> writes:
>>
>>> mt7915_band_config() sets band_idx = 1 on the main phy for mt7986
>>> with MT7975_ONE_ADIE or MT7976_ONE_ADIE.
>>>
>>> Commit 0335c034e726 ("wifi: mt76: fix race condition related to
>>> checking tx queue fill status") introduced a dereference of the
>>> phys array indirectly indexed by band_idx via wcid->phy_idx in
>>> mt76_wcid_cleanup(). This caused the following Oops on affected
>>> mt7986 devices:
>>>
>>> Unable to handle kernel read from unreadable memory at virtual address 0000000000000024
>>> Mem abort info:
>>> ESR = 0x0000000096000005
>>> EC = 0x25: DABT (current EL), IL = 32 bits
>>> SET = 0, FnV = 0
>>> EA = 0, S1PTW = 0
>>> FSC = 0x05: level 1 translation fault
>>> Data abort info:
>>> ISV = 0, ISS = 0x00000005
>>> CM = 0, WnR = 0
>>> user pgtable: 4k pages, 39-bit VAs, pgdp=0000000042545000
>>> [0000000000000024] pgd=0000000000000000, p4d=0000000000000000, pud=0000000000000000
>>> Internal error: Oops: 0000000096000005 [#1] SMP
>>> Modules linked in: ... mt7915e mt76_connac_lib mt76 mac80211 cfg80211 ...
>>> CPU: 2 PID: 1631 Comm: hostapd Not tainted 5.15.150 #0
>>> Hardware name: ZyXEL EX5700 (Telenor) (DT)
>>> pstate: 80400005 (Nzcv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--)
>>> pc : mt76_wcid_cleanup+0x84/0x22c [mt76]
>>> lr : mt76_wcid_cleanup+0x64/0x22c [mt76]
>>> sp : ffffffc00a803700
>>> x29: ffffffc00a803700 x28: ffffff80008f7300 x27: ffffff80003f3c00
>>> x26: ffffff80000a7880 x25: ffffffc008c26e00 x24: 0000000000000001
>>> x23: ffffffc000a68114 x22: 0000000000000000 x21: ffffff8004172cc8
>>> x20: ffffffc00a803748 x19: ffffff8004152020 x18: 0000000000000000
>>> x17: 00000000000017c0 x16: ffffffc008ef5000 x15: 0000000000000be0
>>> x14: ffffff8004172e28 x13: ffffff8004172e28 x12: 0000000000000000
>>> x11: 0000000000000000 x10: ffffff8004172e30 x9 : ffffff8004172e28
>>> x8 : 0000000000000000 x7 : ffffff8004156020 x6 : 0000000000000000
>>> x5 : 0000000000000031 x4 : 0000000000000000 x3 : 0000000000000001
>>> x2 : 0000000000000000 x1 : ffffff80008f7300 x0 : 0000000000000024
>>> Call trace:
>>> mt76_wcid_cleanup+0x84/0x22c [mt76]
>>> __mt76_sta_remove+0x70/0xbc [mt76]
>>> mt76_sta_state+0x8c/0x1a4 [mt76]
>>> mt7915_eeprom_get_power_delta+0x11e4/0x23a0 [mt7915e]
>>> drv_sta_state+0x144/0x274 [mac80211]
>>> sta_info_move_state+0x1cc/0x2a4 [mac80211]
>>> sta_set_sinfo+0xaf8/0xc24 [mac80211]
>>> sta_info_destroy_addr_bss+0x4c/0x6c [mac80211]
>>>
>>> ieee80211_color_change_finish+0x1c08/0x1e70 [mac80211]
>>> cfg80211_check_station_change+0x1360/0x4710 [cfg80211]
>>> genl_family_rcv_msg_doit+0xb4/0x110
>>> genl_rcv_msg+0xd0/0x1bc
>>> netlink_rcv_skb+0x58/0x120
>>> genl_rcv+0x34/0x50
>>> netlink_unicast+0x1f0/0x2ec
>>> netlink_sendmsg+0x198/0x3d0
>>> ____sys_sendmsg+0x1b0/0x210
>>> ___sys_sendmsg+0x80/0xf0
>>> __sys_sendmsg+0x44/0xa0
>>> __arm64_sys_sendmsg+0x20/0x30
>>> invoke_syscall.constprop.0+0x4c/0xe0
>>> do_el0_svc+0x40/0xd0
>>> el0_svc+0x14/0x4c
>>> el0t_64_sync_handler+0x100/0x110
>>> el0t_64_sync+0x15c/0x160
>>> Code: d2800002 910092c0 52800023 f9800011 (885f7c01)
>>> ---[ end trace 7e42dd9a39ed2281 ]---
>>>
>>> Fix by using mt76_dev_phy() which will map band_idx to the correct phy
>>> for all hardware combinations.
>>>
>>> Fixes: 0335c034e726 ("wifi: mt76: fix race condition related to checking tx queue fill status")
>>> Link: https://github.com/openwrt/openwrt/issues/14548
>>> Signed-off-by: Bjørn Mork <bjorn@mork.no>
>>
>> Should this go to wireless tree?
>
> I believe it should. If fixes a regression on the affected hardware,
> introduced by commit 0335c034e726.
Ok, so the regression is introduced in v6.7-rc1. But I noticed that
Felix had applied already to his tree so this will go to v6.12:
https://github.com/nbd168/wireless/commit/cc9370fc0d7a83dab7159a3a91d363e6903d8eb2
> It should also go into any still maintained v6.7, v6.8, v6.9, v6.10
> stable trees. But I assume they will pick it up automatically based on
> the Fixes tag.
The fixes tag is just a hint, it does not guarantee that stable
maintainers will pick the commit.
--
https://patchwork.kernel.org/project/linux-wireless/list/
https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2024-07-31 11:27 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2024-07-13 13:00 [PATCH] wifi: mt76: fix oops on non-dbdc mt7986 Bjørn Mork
2024-07-31 9:27 ` Kalle Valo
2024-07-31 10:26 ` Bjørn Mork
2024-07-31 11:27 ` Kalle Valo
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).