Re: Regression in 028abd92 for Sun UltraSPARC T1
John Paul Adrian Glaubitz <[email protected]>
| Newsgroups | gmane.linux.ports.sparc,gmane.linux.debian.ports.sparc |
|---|---|
| Message-ID | <e705404a9f66e0cffc0c474b565b2755d8d560c2.camel@physik.fu-berlin.de> |
Hi Christoph,
On Thu, 2021-03-25 at 08:50 +0100, Christoph Hellwig wrote:
> I have to admit I'm completely lost at this point. This new trace looks
> totally strange to me, and I'm pretty sure whatever symptoms you see are
> due to different alignments / code sections etc just triggered by the
> removal, we need help from the real sparc experts.
This has finally been tracked down to a regression in the copy_{to,from}_user
code on SPARC as it's found in arch/sparc/lib. Patches to fix these issues
have already been posted in [1].
I'm waiting for Frank to test the patches on his SPARC T1 which is the only
SPARC sub-arch we don't have test results for yet.
Adrian
> [1] https://lore.kernel.org/all/[email protected]/
--
.''`. John Paul Adrian Glaubitz
: :' : Debian Developer
`. `' Physicist
`- GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913