Hello Felix, Lorenzo et al.
With v6.9 kernel the following card:
01:00.0 Unclassified device [0002]: MEDIATEK Corp. MT7915E 802.11ax PCI Express Wireless Network Adapter [14c3:7915]
doesn't work any more. Upon mt7915e module insertion the following splat happens:
mt7915e 0000:01:00.0: enabling device (0000 -> 0002)
WARNING: CPU: 3 PID: 1648 at net/mac80211/main.c:794 ieee80211_alloc_hw_nm+0x9a3/0x9f0 [mac80211]
Call Trace:
<TASK>
mt76_alloc_device+0x24/0x290 [mt76 46e2c4f11be089903469a4d6045c71cb7842b4cd]
mt7915_mmio_probe+0x33/0x1d0 [mt7915e 9657e1926e619568545a08636674805d5e665c85]
mt7915_pci_probe+0xe1/0x3a0 [mt7915e 9657e1926e619568545a08636674805d5e665c85]
pci_device_probe+0xf0/0x2d0
really_probe+0xde/0x3b0
__driver_probe_device+0x78/0x110
driver_probe_device+0x1f/0x130
__driver_attach+0x91/0x1e0
bus_for_each_dev+0x105/0x160
bus_add_driver+0x137/0x270
driver_register+0x72/0xd0
mt7915_init+0x44/0xff0 [mt7915e 9657e1926e619568545a08636674805d5e665c85]
do_one_initcall+0x5b/0x310
do_init_module+0x60/0x220
init_module_from_file+0x89/0xe0
idempotent_init_module+0x121/0x2b0
__x64_sys_finit_module+0x5e/0xb0
do_syscall_64+0x82/0x160
entry_SYSCALL_64_after_hwframe+0x76/0x7e
mt7915e 0000:01:00.0: probe with driver mt7915e failed with error -12
and the card stays non-functional. Re-inserting the module triggers the same behaviour.
The following condition is triggered:
794 if (WARN_ON(!ops->add_chanctx ||
795 !ops->remove_chanctx ||
796 !ops->change_chanctx ||
797 !ops->assign_vif_chanctx ||
798 !ops->unassign_vif_chanctx))
799 return NULL;
This worked with v6.8 and earlier just fine.
Please help.
Thank you.
--
Oleksandr Natalenko (post-factum)
Also /cc Johannes because of this commit:
6092077ad09ce wifi: mac80211: introduce 'channel request'
On středa 15. května 2024 0:43:40, SELČ Oleksandr Natalenko wrote:
> Hello Felix, Lorenzo et al.
>
> With v6.9 kernel the following card:
>
> 01:00.0 Unclassified device [0002]: MEDIATEK Corp. MT7915E 802.11ax PCI Express Wireless Network Adapter [14c3:7915]
>
> doesn't work any more. Upon mt7915e module insertion the following splat happens:
>
> mt7915e 0000:01:00.0: enabling device (0000 -> 0002)
> WARNING: CPU: 3 PID: 1648 at net/mac80211/main.c:794 ieee80211_alloc_hw_nm+0x9a3/0x9f0 [mac80211]
> Call Trace:
> <TASK>
> mt76_alloc_device+0x24/0x290 [mt76 46e2c4f11be089903469a4d6045c71cb7842b4cd]
> mt7915_mmio_probe+0x33/0x1d0 [mt7915e 9657e1926e619568545a08636674805d5e665c85]
> mt7915_pci_probe+0xe1/0x3a0 [mt7915e 9657e1926e619568545a08636674805d5e665c85]
> pci_device_probe+0xf0/0x2d0
> really_probe+0xde/0x3b0
> __driver_probe_device+0x78/0x110
> driver_probe_device+0x1f/0x130
> __driver_attach+0x91/0x1e0
> bus_for_each_dev+0x105/0x160
> bus_add_driver+0x137/0x270
> driver_register+0x72/0xd0
> mt7915_init+0x44/0xff0 [mt7915e 9657e1926e619568545a08636674805d5e665c85]
> do_one_initcall+0x5b/0x310
> do_init_module+0x60/0x220
> init_module_from_file+0x89/0xe0
> idempotent_init_module+0x121/0x2b0
> __x64_sys_finit_module+0x5e/0xb0
> do_syscall_64+0x82/0x160
> entry_SYSCALL_64_after_hwframe+0x76/0x7e
> mt7915e 0000:01:00.0: probe with driver mt7915e failed with error -12
>
> and the card stays non-functional. Re-inserting the module triggers the same behaviour.
>
> The following condition is triggered:
>
> 794 if (WARN_ON(!ops->add_chanctx ||
> 795 !ops->remove_chanctx ||
> 796 !ops->change_chanctx ||
> 797 !ops->assign_vif_chanctx ||
> 798 !ops->unassign_vif_chanctx))
> 799 return NULL;
>
> This worked with v6.8 and earlier just fine.
>
> Please help.
>
> Thank you.
>
>
--
Oleksandr Natalenko (post-factum)
Oleksandr Natalenko <[email protected]> writes:
> Hello Felix, Lorenzo et al.
>
> With v6.9 kernel the following card:
>
> 01:00.0 Unclassified device [0002]: MEDIATEK Corp. MT7915E 802.11ax
> PCI Express Wireless Network Adapter [14c3:7915]
>
> doesn't work any more. Upon mt7915e module insertion the following splat happens:
>
> mt7915e 0000:01:00.0: enabling device (0000 -> 0002)
> WARNING: CPU: 3 PID: 1648 at net/mac80211/main.c:794
> ieee80211_alloc_hw_nm+0x9a3/0x9f0 [mac80211]
> Call Trace:
> <TASK>
> mt76_alloc_device+0x24/0x290 [mt76 46e2c4f11be089903469a4d6045c71cb7842b4cd]
> mt7915_mmio_probe+0x33/0x1d0 [mt7915e 9657e1926e619568545a08636674805d5e665c85]
> mt7915_pci_probe+0xe1/0x3a0 [mt7915e 9657e1926e619568545a08636674805d5e665c85]
> pci_device_probe+0xf0/0x2d0
> really_probe+0xde/0x3b0
> __driver_probe_device+0x78/0x110
> driver_probe_device+0x1f/0x130
> __driver_attach+0x91/0x1e0
> bus_for_each_dev+0x105/0x160
> bus_add_driver+0x137/0x270
> driver_register+0x72/0xd0
> mt7915_init+0x44/0xff0 [mt7915e 9657e1926e619568545a08636674805d5e665c85]
> do_one_initcall+0x5b/0x310
> do_init_module+0x60/0x220
> init_module_from_file+0x89/0xe0
> idempotent_init_module+0x121/0x2b0
> __x64_sys_finit_module+0x5e/0xb0
> do_syscall_64+0x82/0x160
> entry_SYSCALL_64_after_hwframe+0x76/0x7e
> mt7915e 0000:01:00.0: probe with driver mt7915e failed with error -12
>
> and the card stays non-functional. Re-inserting the module triggers
> the same behaviour.
>
> The following condition is triggered:
>
> 794 if (WARN_ON(!ops->add_chanctx ||
> 795 !ops->remove_chanctx ||
> 796 !ops->change_chanctx ||
> 797 !ops->assign_vif_chanctx ||
> 798 !ops->unassign_vif_chanctx))
> 799 return NULL;
>
> This worked with v6.8 and earlier just fine.
Thanks for the report. I'll add this to regzbot for easier tracking:
#regzbot introduced: v6.8..v6.9 ^
--
https://patchwork.kernel.org/project/linux-wireless/list/
https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches
(please don't top post)
Oleksandr Natalenko <[email protected]> writes:
> Also /cc Johannes because of this commit:
>
> 6092077ad09ce wifi: mac80211: introduce 'channel request'
Have you tried reverting that commit?
--
https://patchwork.kernel.org/project/linux-wireless/list/
https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches
On Wed, 2024-05-15 at 00:51 +0200, Oleksandr Natalenko wrote:
> Also /cc Johannes because of this commit:
>
> 6092077ad09ce wifi: mac80211: introduce 'channel request'
>
> On středa 15. května 2024 0:43:40, SELČ Oleksandr Natalenko wrote:
> > Hello Felix, Lorenzo et al.
> >
> > With v6.9 kernel the following card:
> >
> > 01:00.0 Unclassified device [0002]: MEDIATEK Corp. MT7915E 802.11ax PCI Express Wireless Network Adapter [14c3:7915]
> >
> > doesn't work any more. Upon mt7915e module insertion the following splat happens:
> >
6.9 didn't get commit 2f7cf3b61d85 ("wifi: mt76: mt7915: add missing
chanctx ops")? Huh?
johannes
Hello Johannes.
On středa 15. května 2024 8:23:35, SELČ Johannes Berg wrote:
> On Wed, 2024-05-15 at 00:51 +0200, Oleksandr Natalenko wrote:
> > Also /cc Johannes because of this commit:
> >
> > 6092077ad09ce wifi: mac80211: introduce 'channel request'
> >
> > On středa 15. května 2024 0:43:40, SELČ Oleksandr Natalenko wrote:
> > > Hello Felix, Lorenzo et al.
> > >
> > > With v6.9 kernel the following card:
> > >
> > > 01:00.0 Unclassified device [0002]: MEDIATEK Corp. MT7915E 802.11ax PCI Express Wireless Network Adapter [14c3:7915]
> > >
> > > doesn't work any more. Upon mt7915e module insertion the following splat happens:
> > >
>
> 6.9 didn't get commit 2f7cf3b61d85 ("wifi: mt76: mt7915: add missing
> chanctx ops")? Huh?
Yes, you are right, this commit is not present in v6.9, and I can find it in the "next" branch only.
I can also confirm this commit fixes the regression for me.
Thank you.
> johannes
>
--
Oleksandr Natalenko (post-factum)
Johannes Berg <[email protected]> writes:
> On Wed, 2024-05-15 at 00:51 +0200, Oleksandr Natalenko wrote:
>> Also /cc Johannes because of this commit:
>>
>> 6092077ad09ce wifi: mac80211: introduce 'channel request'
>>
>> On středa 15. května 2024 0:43:40, SELČ Oleksandr Natalenko wrote:
>> > Hello Felix, Lorenzo et al.
>> >
>> > With v6.9 kernel the following card:
>> >
>> > 01:00.0 Unclassified device [0002]: MEDIATEK Corp. MT7915E 802.11ax PCI Express Wireless Network Adapter [14c3:7915]
>> >
>> > doesn't work any more. Upon mt7915e module insertion the following splat happens:
>> >
>
> 6.9 didn't get commit 2f7cf3b61d85 ("wifi: mt76: mt7915: add missing
> chanctx ops")? Huh?
Ah, it will be only in v6.10-rc1. Oleksandr, can you try this commit:
https://git.kernel.org/linus/2f7cf3b61d85
--
https://patchwork.kernel.org/project/linux-wireless/list/
https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches
Oleksandr Natalenko <[email protected]> writes:
> Hello Johannes.
>
> On středa 15. května 2024 8:23:35, SELČ Johannes Berg wrote:
>> On Wed, 2024-05-15 at 00:51 +0200, Oleksandr Natalenko wrote:
>> > Also /cc Johannes because of this commit:
>> >
>> > 6092077ad09ce wifi: mac80211: introduce 'channel request'
>> >
>> > On středa 15. května 2024 0:43:40, SELČ Oleksandr Natalenko wrote:
>> > > Hello Felix, Lorenzo et al.
>> > >
>> > > With v6.9 kernel the following card:
>> > >
>> > > 01:00.0 Unclassified device [0002]: MEDIATEK Corp. MT7915E 802.11ax PCI Express Wireless Network Adapter [14c3:7915]
>> > >
>> > > doesn't work any more. Upon mt7915e module insertion the following splat happens:
>> > >
>>
>> 6.9 didn't get commit 2f7cf3b61d85 ("wifi: mt76: mt7915: add missing
>> chanctx ops")? Huh?
>
> Yes, you are right, this commit is not present in v6.9, and I can find it in the "next" branch only.
>
> I can also confirm this commit fixes the regression for me.
Thanks for testing.
#regzbot fix: 2f7cf3b61d85
--
https://patchwork.kernel.org/project/linux-wireless/list/
https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches
Johannes Berg <[email protected]> writes:
> On Wed, 2024-05-15 at 00:51 +0200, Oleksandr Natalenko wrote:
>> Also /cc Johannes because of this commit:
>>
>> 6092077ad09ce wifi: mac80211: introduce 'channel request'
>>
>> On středa 15. května 2024 0:43:40, SELČ Oleksandr Natalenko wrote:
>> > Hello Felix, Lorenzo et al.
>> >
>> > With v6.9 kernel the following card:
>> >
>> > 01:00.0 Unclassified device [0002]: MEDIATEK Corp. MT7915E
>> > 802.11ax PCI Express Wireless Network Adapter [14c3:7915]
>> >
>> > doesn't work any more. Upon mt7915e module insertion the following splat happens:
>> >
>
> 6.9 didn't get commit 2f7cf3b61d85 ("wifi: mt76: mt7915: add missing
> chanctx ops")? Huh?
Yeah, commit 2f7cf3b61d85 should have been taken to wireless tree but it
fell through the cracks. I didn't realise that it's an important bugfix
and the commit message didn't give any indication of that either.
But commit 2f7cf3b61d85 is now in Linus' tree so it would be good if
someone could submit it to stable releases.
--
https://patchwork.kernel.org/project/linux-wireless/list/
https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches
Kalle Valo <[email protected]> writes:
> Johannes Berg <[email protected]> writes:
>
>> On Wed, 2024-05-15 at 00:51 +0200, Oleksandr Natalenko wrote:
>>> Also /cc Johannes because of this commit:
>>>
>>> 6092077ad09ce wifi: mac80211: introduce 'channel request'
>>>
>>> On středa 15. května 2024 0:43:40, SELČ Oleksandr Natalenko wrote:
>>> > Hello Felix, Lorenzo et al.
>>> >
>>> > With v6.9 kernel the following card:
>>> >
>>> > 01:00.0 Unclassified device [0002]: MEDIATEK Corp. MT7915E
>>> > 802.11ax PCI Express Wireless Network Adapter [14c3:7915]
>>> >
>>> > doesn't work any more. Upon mt7915e module insertion the following splat happens:
>>> >
>>
>> 6.9 didn't get commit 2f7cf3b61d85 ("wifi: mt76: mt7915: add missing
>> chanctx ops")? Huh?
>
> Yeah, commit 2f7cf3b61d85 should have been taken to wireless tree but it
> fell through the cracks. I didn't realise that it's an important bugfix
> and the commit message didn't give any indication of that either.
>
> But commit 2f7cf3b61d85 is now in Linus' tree so it would be good if
> someone could submit it to stable releases.
Johannes pointed out to me that the commit is already in stable:
https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=linux-6.9.y&id=ddb02739a2be54ed922bce3cb57d5f901590ee70
So all should be good now.
--
https://patchwork.kernel.org/project/linux-wireless/list/
https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches
On Sat, May 25, 2024 at 09:26:01AM +0300, Kalle Valo wrote:
> Johannes Berg <[email protected]> writes:
>
> > On Wed, 2024-05-15 at 00:51 +0200, Oleksandr Natalenko wrote:
> >> Also /cc Johannes because of this commit:
> >>
> >> 6092077ad09ce wifi: mac80211: introduce 'channel request'
> >>
> >> On středa 15. května 2024 0:43:40, SELČ Oleksandr Natalenko wrote:
> >> > Hello Felix, Lorenzo et al.
> >> >
> >> > With v6.9 kernel the following card:
> >> >
> >> > 01:00.0 Unclassified device [0002]: MEDIATEK Corp. MT7915E
> >> > 802.11ax PCI Express Wireless Network Adapter [14c3:7915]
> >> >
> >> > doesn't work any more. Upon mt7915e module insertion the following splat happens:
> >> >
> >
> > 6.9 didn't get commit 2f7cf3b61d85 ("wifi: mt76: mt7915: add missing
> > chanctx ops")? Huh?
>
> Yeah, commit 2f7cf3b61d85 should have been taken to wireless tree but it
> fell through the cracks. I didn't realise that it's an important bugfix
> and the commit message didn't give any indication of that either.
>
> But commit 2f7cf3b61d85 is now in Linus' tree so it would be good if
> someone could submit it to stable releases.
It's already in 6.9.1, does it need to go anywhere else?
thanks,
greg k-h
On Mon, 2024-05-27 at 20:29 +0200, Greg KH wrote:
> >
> > Yeah, commit 2f7cf3b61d85 should have been taken to wireless tree but it
> > fell through the cracks. I didn't realise that it's an important bugfix
> > and the commit message didn't give any indication of that either.
> >
> > But commit 2f7cf3b61d85 is now in Linus' tree so it would be good if
> > someone could submit it to stable releases.
>
> It's already in 6.9.1, does it need to go anywhere else?
>
No, but apparently we need a similar commit for mt7915e or something? I
have no idea why I missed those. I'll take a look tomorrow.
johannes