2008-06-27 21:54:26

by Vitaly Mayatskih

Subject: [PATCH 0/3] Fix copy_user on x86_64

I've moved functionality of copy_user_64.S and copy_user_nocache_64.S to
copy_user_64.c. Tail handling is now in plain C routine.

copy_user_generic_string, copy_user_generic_unrolled and
__copy_user_nocache were tested by me with a bit modified systemtap's
script which was in attach in my previous letter.
--
wbr, Vitaly

2008-06-28 06:22:43

by Cyrill Gorcunov

Subject: Re: [PATCH 0/3] Fix copy_user on x86_64

[Vitaly Mayatskikh - Fri, Jun 27, 2008 at 11:48:24PM +0200]
| I've moved functionality of copy_user_64.S and copy_user_nocache_64.S to
| copy_user_64.c. Tail handling is now in plain C routine.
|
| copy_user_generic_string, copy_user_generic_unrolled and
| __copy_user_nocache were tested by me with a bit modified systemtap's
| script which was in attach in my previous letter.
| --
| wbr, Vitaly
|

Hi Vitaly, actually I don't see quite benefit from 'wrapping'
assembler code by C. At least it was more readable before. Sorry.
But lets wait for another optinions. Maybe I'm just too hypercritical ;)

A few Cc added.

- Cyrill -