First Last Prev Next    No search results available      Search page      Enter new bug
Bug#: 75254
Alias:
Product:
Component:
Status: RESOLVED
Resolution: FIXED
Assigned To: PPC Porters <ppc@gentoo.org>
Hardware:
OS:
Version:
Priority:
Severity:
Reporter: Marcin Kurek <morgoth6@gmail.com>
Add CC:
CC:
URL:
Summary:
Status Whiteboard:
Keywords:

Filename Description Type Creator Created Size Actions
mol-0.9.70-nopriority.patch Less intrusive, only removes the main thread priority change patch Joe Jezak 2005-01-07 05:34 0000 429 bytes Details | Diff
mol-0.9.70-nopriority2.patch Removes all priority changes patch Joe Jezak 2005-01-07 05:37 0000 1.81 KB Details | Diff
Create a New Attachment (proposed patch, testcase, etc.) View All

Bug 75254 depends on: Show dependency tree
Show dependency graph
Bug 75254 blocks:
Votes: 0    Show votes for this bug    Vote for this bug

Additional Comments: (this is where you put emerge --info)







View Bug Activity   |   Format For Printing   |   XML   |   Clone This Bug


Description:   Opened: 2004-12-21 16:14 0000
Some time ago I give a try to mac on linux emulator and discovered that it
generaly works fine but the mouse pointer jumps badly on my MacOSX
installation. Generaly it stops at all system activities (Opening a window for
example)

I made a small investigation and noticed that mol process has nice value set to
-19 (sigh) and this seems to cause my problem. renice mol to -5 for example
cure my problem and mouse works just fine.


Reproducible: Always
Steps to Reproduce:
1.
2.
3.

------- Comment #1 From Marcin Kurek 2004-12-31 06:40:25 0000 -------
It seems not only I have similar problem. I was able to help a guy at #mol
channel that can't install MacOS9 because the installator takes 30min's to show
up on his system.

Renice mol to 0 helps.

Currently I change the mol sources and remove all setpriority() calls and it
seems to work just fine. I think this can be related to some recent 2.6.x
scheduler changes.

------- Comment #2 From Joe Jezak 2005-01-07 05:34:38 0000 -------
Created an attachment (id=47842) [edit]
Less intrusive, only removes the main thread priority change

------- Comment #3 From Joe Jezak 2005-01-07 05:37:37 0000 -------
Created an attachment (id=47843) [edit]
Removes all priority changes

I've tested both here, I don't see much of a difference between the two, but
someone else might.  Since I didn't see much of a difference, I'd say go for
the less intrusive patch.  I'm running an nptl glibc, so maybe linuxthreads
reacts differently.  Comments appreciated!

------- Comment #4 From Joe Jezak 2005-01-07 05:40:33 0000 -------
Forgot to mention that I do see the behaviour mentioned in the bug report and
either renicing or removing the priority changes do fix the problem.  I suspect
it's related to the scheduler in 2.6.

------- Comment #5 From Marcin Kurek 2005-01-16 15:42:24 0000 -------
Yes this is scheduler for sure

Generaly using O(1) scheduler or staricase scheduler normal apps should not touch nice value at all. It's reserved for system and critical apps.

My system use ntpl and 2.6.9 kernal image too. And remove all setpriorit() calls definitly fix that problem.

------ This note is from starcase scheduler author:

The tuning in the 2.6 kernel scheduler and the staircase scheduler are
specifically designed to not need -ve nice levels for any normal userspace
tasks, but for critical things like root shells to use in case of emergency, or
video and audio capture

------- Comment #6 From Joe Jezak 2005-01-27 10:17:39 0000 -------
Added the patch that removes all priority changes for both versions of mol in
CVS.  Reopen the bug if it's still a problem or if there is a problem with the
fix.  Thanks!

First Last Prev Next    No search results available      Search page      Enter new bug