lkml.org 
[lkml]   [2017]   [Oct]   [11]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
SubjectRE: [PATCH v4 2/2] pid: Remove pidhash
Date
>        DEFINE(IA64_UPID_SHIFT, 5);
>
> Grepping for IA64_UPID_SHIFT leads us to some assembly
> code implementing fsys_getpid (why is that in assembly?!):

The fast system call path has a whole host of serious restrictions on what it can
touch. See Documentation/ia64/fsys.txt. Why is getpid() a fast system call? I think
there was some application (or perhaps benchmark) that used it a lot.

> add r8=IA64_PID_LEVEL_OFFSET,r17
> ;;
> ld4 r8=[r8] // r8 = pid->level
> add r17=IA64_PID_UPID_OFFSET,r17 // r17 = &pid->numbers[0]
> ;;
> shl r8=r8,IA64_UPID_SHIFT
> ;;
> add r17=r17,r8 // r17 = &pid->numbers[pid->level]
> ;;
> ld4 r8=[r17] // r8 = pid->numbers[pid->level].nr
> ;;
> mov r17=0
>
> Luckily it looks like this is only referencing the first members of struct upid,
> and you are removing the last member, so I suspect you will be fine changing the IA64
> to this:
>
>        BUILD_BUG_ON(sizeof(struct upid) != 16);
>        DEFINE(IA64_UPID_SHIFT, 4);
>
> Tony, does that look ok to you?

I think so. Respin and Cc: me on both patches in the series and I'll take it for a spin
on h/w to make sure.

-Tony
\
 
 \ /
  Last update: 2017-10-11 23:22    [W:0.189 / U:0.176 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site