On Sun, 20 Mar 2005 23:20:04 +0100, viking wrote:
-blather-
> at least I *think* they're Linus' patches - no Visual Basic in them yet.
Guess that means Linus hasn't flipped his lid. So we won't see Linux-4.0
for MSVC any time soon.
Heh he - the ultimate virus would be if a Linux kernel could install
itself onto a NTFS partition and take over from the Windows kernel code.
Then MS would have way less support issues, we'd be hailed saviours, and
achieve world domination through the back door.
-meat-
> I've also noted .10 is quicker off the blocks than 2.6.11 seems to be.
Namely seems to happen around the times when I'm doing something like
mounting devfs (takes nearly 30 secs), and when I'm accessing files from
disc (bash$ less some-random-file.txt) - this can take about two seconds
for Linux to actually notice I've done something. I've no idea where the
error is here, either. i.e. is bash waiting around for me? is the
filesystem code waiting for some reason? Is the kernel in a tailspin?
[shrug]
CC if you wish, but I ought to be watching the list again for this one.
--
/| _,.:*^*:., |\ Cheers from the Viking family, including Pippin, our cat
| |_/' viking@ `\_| |
| flying-brick | $FunnyMail : What do you mean, I've lost the plot?
\_.caverock.net.nz_/ 5.40 : I planted them carrots right here!!
Okay - so I see I'm not the only one to see significant slowdowns in 2.6.11.x
compared to 2.6.10 - guess I'll have to wait until the 4-level table thing
sorts itself out...
/me removes foot out of gob.
--
/| _,.:*^*:., |\ Cheers from the Viking family, including Pippin, our cat
| |_/' viking@ `\_| |
| flying-brick | $FunnyMail : What do you mean, I've lost the plot?
\_.caverock.net.nz_/ 5.40 : I planted them carrots right here!!
viking wrote:
> Okay - so I see I'm not the only one to see significant slowdowns in 2.6.11.x
> compared to 2.6.10 - guess I'll have to wait until the 4-level table thing
> sorts itself out...
>
> /me removes foot out of gob.
>
The 4-level page tables slowdowns don't explain the problems you
are seeing. 2.6.12-rc1 is very close to 2.6.10 speed. The slowdown
we are talking about is on the order of microseconds at process
exit time, and unmap.
So please continue to investigate your problems. I think there may
have been some small CPU scheduler changes between 2.6.10 and
2.6.11 from Con. Seems unlikely that they would be causing your
problems, but that's my suggestion for a starting point.
Thanks,
Nick
>-meat-
>> I've also noted .10 is quicker off the blocks than 2.6.11 seems to be.
>
>Namely seems to happen around the times when I'm doing something like
>mounting devfs (takes nearly 30 secs), and when I'm accessing files from
>disc (bash$ less some-random-file.txt) - this can take about two seconds
>for Linux to actually notice I've done something. I've no idea where the
>error is here, either. i.e. is bash waiting around for me? is the
>filesystem code waiting for some reason? Is the kernel in a tailspin?
>[shrug]
Use sysrq+t and see where it "hangs".
Jan Engelhardt
--