1:08 a.m.
I tested 3.19.0-0.rc3.git0.1.fc22 on three machines and the boot hung
pretty early on one of them. The video mode didn't change when it normally
would have shortly before it hung. However the output had scrolled past
where I might have seen an error message. This makes me suspect that this
is an issue with Nouveau and nv28s. I had another machine that has an nv50
boot a mainline kernel from around rc1 which would have the Nouveau changes,
so if it is a Nouveau probelm it only affects a subset of cards.
I'll look at doing some testing on the machine to see if the problem is
really in the changes to Nouveau support in the kernel, but not tonight.
1:18 a.m.
On Wed, Jan 07, 2015 at 01:08:09 -0600,
Bruno Wolff III <bruno(a)wolff.to> wrote:
I'll look at doing some testing on the machine to see if the problem
is really in the changes to Nouveau support in the kernel, but not
tonight.
My plan is to see if Airlie's DRM merge breaks things. I might have a
negative answer tomorrow morning, but I won't get both sides tested
until tomorrow night.
7:55 a.m.
On Wed, Jan 07, 2015 at 01:18:25 -0600,
Bruno Wolff III <bruno(a)wolff.to> wrote:
On Wed, Jan 07, 2015 at 01:08:09 -0600,
Bruno Wolff III <bruno(a)wolff.to> wrote:
>
>I'll look at doing some testing on the machine to see if the problem
>is really in the changes to Nouveau support in the kernel, but not
>tonight.
My plan is to see if Airlie's DRM merge breaks things. I might have a
negative answer tomorrow morning, but I won't get both sides tested
until tomorrow night.
The commit before Ailie's DRM merge also crashed. Not exactly the same,
but similar. So I may or may not be crashing on the same bug there as I am
with rc3. For now I assume so and bisect with that commit is bad.
11:45 a.m.
The bisect for this seems to be working OK, and I should have the commit
late in the weekend and I will file a bug report if it is still broken
at that time.
7:34 a.m.
On 07.01.2015 08:08, Bruno Wolff III wrote:
I tested 3.19.0-0.rc3.git0.1.fc22 on three machines and the boot hung
pretty early on one of them. The video mode didn't change when it normally
would have shortly before it hung. However the output had scrolled past
where I might have seen an error message. This makes me suspect that this
is an issue with Nouveau and nv28s. I had another machine that has an nv50
boot a mainline kernel from around rc1 which would have the Nouveau changes,
so if it is a Nouveau probelm it only affects a subset of cards.
3.19.0-0.rc3.git0.1.fc21.x86_64 on NV50,
so far so good.
10:56 p.m.
On Wed, Jan 07, 2015 at 01:08:09 -0600,
Bruno Wolff III <bruno(a)wolff.to> wrote:
I tested 3.19.0-0.rc3.git0.1.fc22 on three machines and the boot hung
pretty early on one of them. The video mode didn't change when it
normally would have shortly before it hung. However the output had
scrolled past where I might have seen an error message. This makes me
suspect that this is an issue with Nouveau and nv28s. I had another
machine that has an nv50 boot a mainline kernel from around rc1 which
would have the Nouveau changes, so if it is a Nouveau probelm it only
affects a subset of cards.
The bad commit is e8972421623cd10a12cc7343da029b58777fdd2f. I'm going to
make sure it is still a problem in HEAD and then test a build with
this commit reverted. Assuming things go as expected I'll file bugs with
both Nouveau and the Kernel.
3392
days inactive
3395
days old
kernel@lists.fedoraproject.org
5 comments
2 participants
participants (2)
-
Bruno Wolff III -
poma