aesni(?) corrupts data on 8.2-BETA1

[jhell]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1



On Fri, 17 Dec 2010 09:01, Mike Tancsa wrote:
In Message-Id: <4D0B6D26.7030808 at sentex.net>

> On 12/16/2010 8:14 PM, jhell wrote:
>> On 12/16/2010 20:10, jhell wrote:
>>> Regarding this patch(r216455) and r216162 I have had to back both of
>>> them out of my local tree to avoid panics on a ZFS & UFS2 i386 system.
>
>
> I have zfs & UFS on AMD64 and have not seen any issues.  How do you
> trigger the issue in i386 ?   Or does it just randomly happen ?
>

Looks as if I have it narrowed down to that its only effecting prefetch on 
i386. Seeing as prefetch is not really suggested below 4G of RAM and as 
well on i386 it is not really much concern to me ATM.

Whatever did change "Not in the ZFS code" & "Not in r216162 or r216455" 
has effected prefetch on i386 directly though or at least as far as I can 
see for the moment.

As of right now Ill look into a few other commits starting with r216305 
which was a MFC of r216194.

I would like to see this resolved as I currently only have i386's to run 
from unless that is someone would like to donate some hardware ;)


Regards,

- -- 

  jhell,v

-----BEGIN PGP SIGNATURE-----

iQEcBAEBAgAGBQJNDD0AAAoJEJBXh4mJ2FR+SqIH/jkWf8Z69lKC3wlSYjwwNqTL
gX/rEuAU/2xOQj2bKR4/BEfeNXtSy9n47IbbCZApGuBABf2lkytSUk2n1XgQ1sB1
GRJl8gCm5BtmhZh0oTc9CIwUcJy+h/w3dU34WXp5riegk0MwpPJlbx9JX4huC+6s
j+wG0oJpxJZq1v/0uWM7ks1D6yYzgBXFyp9j+8ZGoeeXIeraqglj57eB0+Ck1Jq2
JjQ9uLTqnkLS2Zzf/Vgbju4O4bV6geAd+j6Td6Fwt+c3yHCWOHSFQcakyrdJzntl
feWJPAl9ZZSGBhvdAH4pumLt01BVyESIe/rsbtD68pAIGWqxrL3NsI6bIDg3cvI=
=apgr
-----END PGP SIGNATURE-----