Re: [BISECTED]: Black screen on 7.0-rc1

[David Gow <david@davidgow.net>]

Le 25/02/2026 à 12:57 AM, Linus Torvalds a écrit :
> On Mon, 23 Feb 2026 at 23:59, Harald Arnesen <linux@skogtun.org> wrote:
>>
>>>>> In fact, I would expect it to be related to one (or both) of
>>>>>
>>>>>       f2ae92780ab9 ("firewire: ohci: split page allocation from dma mapping")
>>>>>       993ab48006b3 ("firewire: ohci: stop using page private to store
>>>>> DMA mapping address")
>>>>
>>>> Should I try reverting one or both of these?
>>>
>>> Yes. From a quick check, they revert cleanly (revert 993ab48006b3
>>> first) and the result seems to compile cleanly too.
>>
>> I can confirm that the kernel boots with these two reverted.
>> I have attached the new dmesg output.
> 
> Thanks. I guess it's up to Takashi to figure out what exactly goes
> wrong, but it's good to have this confirmed. I hope you'll be willing
> to test any patches that Takashi comes up with.
> 
> And if we can't figure it out, I guess reverting them is the right
> thing, but let's wait until at least rc2 is imminent before we do
> that.
> 

I am also seeing this, and can confirm that reverting the whole firewire 
pull fixes it here.

Interestingly, I was seeing the same firewire_ohci/vmap errors during 
the bisection (with those firewire patches included), but it was only 
the merge which actually caused the system to not boot (again, with udev 
seeming to hang).

Feel free to send any patches my way to test.

-- David