Re: 2.5.46-mm2

[William Lee Irwin III <wli@holomorphy.com>]

William Lee Irwin III wrote:
>> Go for it, I'm just trying to get tiobench to actually run (seems to
>> have new/different "die from too many threads" behavior wrt. --threads).
>> Dropping me a fresh kernel shouldn't slow anything down.

On Sun, Nov 10, 2002 at 08:58:28AM -0800, Andrew Morton wrote:
> It could be the procps thing?  `tiobench --threads 256' shows up as a
> single process in top and ps due to the new thread consolidation feature.
> If you run `ps auxm' or hit 'H' in top, all is revealed.  Not my fave
> feature that.

Turns out monitoring things via /proc/ slowed it down by some ridiculous
factor while it was trying to spawn threads. 9 hours became less than 1s
when I stopped looking.


On Sun, Nov 10, 2002 at 08:58:28AM -0800, Andrew Morton wrote:
> BLKELVGET/SET was removed

Okay, looks like there's an fs to use to get at it with.


On Sun, Nov 10, 2002 at 08:58:28AM -0800, Andrew Morton wrote:
>> P.P.S:  kgdb broke wchan reporting... investigating

On Sun, Nov 10, 2002 at 08:58:28AM -0800, Andrew Morton wrote:
> ?

The wchan reporting needs to get taught about kern_schedule()
and user_schedule() so it can trim them off the stack, which
consists of moving them between scheduling_functions_start_here()
and scheduling_functions_end_here().


Bill