Bug 1482798

[Alexander Ploumistos]

Hello,

About ten days ago, Daiki Ueno posted some patched nss-softokn builds
to address an i686-specific bug:
https://bugzilla.redhat.com/show_bug.cgi?id=1482798#c25
Has anyone had any time to test them on non-SSE2 hardware?
I am away for a conference, but if nobody can look into that in the
meantime, I could probably run some tests come Thursday or Friday.

Cheers
Alex

[Jeff Backus]

On Mon, Oct 2, 2017 at 10:31 AM, Alexander Ploumistos
 wrote:
> Hello,
>
> About ten days ago, Daiki Ueno posted some patched nss-softokn builds
> to address an i686-specific bug:
> https://bugzilla.redhat.com/show_bug.cgi?id=1482798#c25
> Has anyone had any time to test them on non-SSE2 hardware?
> I am away for a conference, but if nobody can look into that in the
> meantime, I could probably run some tests come Thursday or Friday.

Thanks for bringing this up.

Anyone have a chance to test Daiki's build?

jeff

[Alexander Ploumistos]

On Sat, Oct 14, 2017 at 6:29 PM, Jeff Backus  wrote:
> Anyone have a chance to test Daiki's build?

By the time I got F27 on my Athlon system and managed to get through a
number of other problems, he had released nss-softokn-3.33.0-1.0, with
which I could not reproduce the reported bug, even though the patch is
not included yet. I haven't figured out a way to trigger the same
error.

[Jeff Backus]

On Sat, Oct 14, 2017 at 12:17 PM, Alexander Ploumistos
 wrote:
> On Sat, Oct 14, 2017 at 6:29 PM, Jeff Backus  wrote:
> > Anyone have a chance to test Daiki's build?
>
> By the time I got F27 on my Athlon system and managed to get through a
> number of other problems, he had released nss-softokn-3.33.0-1.0, with
> which I could not reproduce the reported bug, even though the patch is
> not included yet. I haven't figured out a way to trigger the same
> error.

Hmm... Interesting. I didn't get the impression that the original
reporter was doing anything exotic, so it shouldn't be too hard to
trigger the problem code. We'll wait and see if the original reporter
is still having problems.

Thanks for taking a look at it!

jeff

[Ralf Corsepius]

On 10/14/2017 06:17 PM, Alexander Ploumistos wrote:
> On Sat, Oct 14, 2017 at 6:29 PM, Jeff Backus  wrote:
> > Anyone have a chance to test Daiki's build?
>
> By the time I got F27 on my Athlon system and managed to get through a
> number of other problems, he had released nss-softokn-3.33.0-1.0, with
> which I could not reproduce the reported bug, even though the patch is
> not included yet.

With nss-softokn-3.33.0-1.0, I could not produce the reported bug, either.
> I haven't figured out a way to trigger the same
> error.

For me, launching thunderbird is a deterministic trigger.

It crashes with a SIGILL with nss-softokn-3.33.0,
but doesn't crash with 3.32.1.

Seems to me as if BZ is N/A ATM, but I had reported details on this to 
https://bugzilla.redhat.com/show_bug.cgi?id=1482798

Ralf