Skip to contentSkip to conversation

2.6.21-rc1 and 2.6.21-rc2 kwin dies silently

DISCUSSION Sid Boyce 8 participants 22 messages Feb 28, 2007 – Mar 30, 2007

lore
Share & tools
lore

Sid Boyce

author
#0 lore raw

openSUSE 10.3 Alpha and KDE-3.5.6, xorg-x11-7.2. KDE is setup not to require a password to unlock, but it asks for password. When the screen unlocks, kwin is gone with no errors logged in /var/log/kdm or /var/log/messages. No problems with 2.6.20.

Same problem on openSUSE 10.2 x86_64, KDE-3.5.5 and 2.6.21-rc2. Regards
Sid.

Signature · 3 lines
Sid Boyce ... Hamradio License G3VBV, Licensed Private Pilot
Emeritus IBM/Amdahl Mainframes and Sun/Fujitsu Servers Tech Support Specialist, Cricket Coach
Microsoft Windows Free Zone - Linux used for all Computing Tasks
#1 lore raw
Quoted 8 lines · Sid Boyce

openSUSE 10.3 Alpha and KDE-3.5.6, xorg-x11-7.2. KDE is setup not to require a password to unlock, but it asks for password. When the screen unlocks, kwin is gone with no errors logged in /var/log/kdm or /var/log/messages. No problems with 2.6.20.

Same problem on openSUSE 10.2 x86_64, KDE-3.5.5 and 2.6.21-rc2. Regards
Sid.

This is the linux kernel mailing list. Perhaps you should post your problem to the opensuse mailing list.

Regards

#2 lore raw

(cc restored. Please always do reply-to-all)

Quoted 7 lines · On Wed, 28 Feb 2007 18:05:13 +0200 [email protected] wrote:
Quoted 5 lines · Sid Boyce

openSUSE 10.3 Alpha and KDE-3.5.6, xorg-x11-7.2. KDE is setup not to require a password to unlock, but it asks for password. When the screen unlocks, kwin is gone with no errors logged in /var/log/kdm or /var/log/messages. No problems with 2.6.20.

Quoted 6 lines
Quoted 3 lines

Same problem on openSUSE 10.2 x86_64, KDE-3.5.5 and 2.6.21-rc2. Regards
Sid.

This is the linux kernel mailing list. Perhaps you should post your problem to the opensuse mailing list.

2.6.20 worked.

2.6.20-rc2 did not.

Working theory: the kernel broke.

Sid, the chances that anyone can work out what caused this are pretty low. It would be great if you could perform a git bisection search sometime in the next few weeks, work out which commit caused this.

Thanks.

author
#3 lore raw
Quoted 20 lines · Andrew Morton

(cc restored. Please always do reply-to-all)

Quoted 8 lines · On Wed, 28 Feb 2007 18:05:13 +0200 [email protected] wrote:
Quoted 5 lines · Sid Boyce

openSUSE 10.3 Alpha and KDE-3.5.6, xorg-x11-7.2. KDE is setup not to require a password to unlock, but it asks for password. When the screen unlocks, kwin is gone with no errors logged in /var/log/kdm or /var/log/messages. No problems with 2.6.20.

Quoted 6 lines
Quoted 4 lines

Same problem on openSUSE 10.2 x86_64, KDE-3.5.5 and 2.6.21-rc2. Regards
Sid.

This is the linux kernel mailing list. Perhaps you should post your problem to the opensuse mailing list.

2.6.20 worked.

Quoted 12 lines

2.6.20-rc2 did not.

Working theory: the kernel broke.

Sid, the chances that anyone can work out what caused this are pretty low. It would be great if you could perform a git bisection search sometime in the next few weeks, work out which commit caused this.

Thanks.

I shall go back to 2.6.20-git3 and work forward. Up to 2.6.20-git2 was OK. Regards
Sid.

Signature · 3 lines
Sid Boyce ... Hamradio License G3VBV, Licensed Private Pilot
Emeritus IBM/Amdahl Mainframes and Sun/Fujitsu Servers Tech Support Specialist, Cricket Coach
Microsoft Windows Free Zone - Linux used for all Computing Tasks
author
#4 lore raw
Quoted 28 lines · Sid Boyce
Quoted 27 lines · Andrew Morton

(cc restored. Please always do reply-to-all)

Quoted 8 lines · On Wed, 28 Feb 2007 18:05:13 +0200 [email protected] wrote:
Quoted 5 lines · Sid Boyce

openSUSE 10.3 Alpha and KDE-3.5.6, xorg-x11-7.2. KDE is setup not to require a password to unlock, but it asks for password. When the screen unlocks, kwin is gone with no errors logged in /var/log/kdm or /var/log/messages. No problems with 2.6.20.

Quoted 6 lines
Quoted 4 lines

Same problem on openSUSE 10.2 x86_64, KDE-3.5.5 and 2.6.21-rc2. Regards
Sid.

This is the linux kernel mailing list. Perhaps you should post your problem to the opensuse mailing list.

2.6.20 worked.

2.6.20-rc2 did not.

Working theory: the kernel broke.

Sid, the chances that anyone can work out what caused this are pretty low. It would be great if you could perform a git bisection search sometime in the next few weeks, work out which commit caused this.

Quoted 9 lines
Quoted 5 lines

Thanks.

I shall go back to 2.6.20-git3 and work forward. Up to 2.6.20-git2 was OK. Regards
Sid.

I tracked the problem down to 2.6.20-git11. Up to 2.6.20-git10 is OK, but from 2.6.20-git11 up to current 2.6.21-rc4-git2 all exhibit the problem. Regards
Sid.

Signature · 3 lines
Sid Boyce ... Hamradio License G3VBV, Licensed Private Pilot
Emeritus IBM/Amdahl Mainframes and Sun/Fujitsu Servers Tech Support Specialist, Cricket Coach
Microsoft Windows Free Zone - Linux used for all Computing Tasks
#5 lore raw
Quoted 37 lines · Sid Boyce
Quoted 36 lines · Sid Boyce
Quoted 35 lines · Andrew Morton

(cc restored. Please always do reply-to-all)

Quoted 8 lines · On Wed, 28 Feb 2007 18:05:13 +0200 [email protected] wrote:
Quoted 5 lines · Sid Boyce

openSUSE 10.3 Alpha and KDE-3.5.6, xorg-x11-7.2. KDE is setup not to require a password to unlock, but it asks for password. When the screen unlocks, kwin is gone with no errors logged in /var/log/kdm or /var/log/messages. No problems with 2.6.20.

Quoted 7 lines
Quoted 4 lines

Same problem on openSUSE 10.2 x86_64, KDE-3.5.5 and 2.6.21-rc2. Regards
Sid.

This is the linux kernel mailing list. Perhaps you should post your problem to the opensuse mailing list.

2.6.20 worked.

2.6.20-rc2 did not.

Working theory: the kernel broke.

Sid, the chances that anyone can work out what caused this are pretty low. It would be great if you could perform a git bisection search sometime in
the next few weeks, work out which commit caused this.

Thanks.

Quoted 7 lines
Quoted 4 lines

I shall go back to 2.6.20-git3 and work forward. Up to 2.6.20-git2 was OK. Regards
Sid.

I tracked the problem down to 2.6.20-git11. Up to 2.6.20-git10 is OK, but from 2.6.20-git11 up to current 2.6.21-rc4-git2 all exhibit the problem.

Thanks for this search.

Looking at the changes between 2.6.20-git10 and 2.6.20-git11, the only suspicious changes are the 60 sysctl patches by Eric.

Eric, can you look at this issue?

Quoted 2 lines

Regards
Sid.

cu
Adrian

Signature · 4 lines
       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed
#6 lore raw
Quoted 46 lines · Adrian Bunk
Quoted 37 lines · Sid Boyce
Quoted 36 lines · Sid Boyce
Quoted 35 lines · Andrew Morton

(cc restored. Please always do reply-to-all)

Quoted 8 lines · On Wed, 28 Feb 2007 18:05:13 +0200 [email protected] wrote:
Quoted 5 lines · Sid Boyce

openSUSE 10.3 Alpha and KDE-3.5.6, xorg-x11-7.2. KDE is setup not to require a password to unlock, but it asks for password. When the screen unlocks, kwin is gone with no errors logged in /var/log/kdm or /var/log/messages. No problems with 2.6.20.

Quoted 7 lines
Quoted 4 lines

Same problem on openSUSE 10.2 x86_64, KDE-3.5.5 and 2.6.21-rc2. Regards
Sid.

This is the linux kernel mailing list. Perhaps you should post your problem to the opensuse mailing list.

2.6.20 worked.

2.6.20-rc2 did not.

Working theory: the kernel broke.

Sid, the chances that anyone can work out what caused this are pretty low. It would be great if you could perform a git bisection search sometime in
the next few weeks, work out which commit caused this.

Thanks.

Quoted 7 lines
Quoted 4 lines

I shall go back to 2.6.20-git3 and work forward. Up to 2.6.20-git2 was OK. Regards
Sid.

I tracked the problem down to 2.6.20-git11. Up to 2.6.20-git10 is OK, but from 2.6.20-git11 up to current 2.6.21-rc4-git2 all exhibit the problem.

Quoted 6 lines

Thanks for this search.

Looking at the changes between 2.6.20-git10 and 2.6.20-git11, the only suspicious changes are the 60 sysctl patches by Eric.

Eric, can you look at this issue?

git bisect between git10 (ac98695d6c1508b724f246f38ce57fb4e3cec356) and git11 (86a71dbd3e81e8870d0f0e56b87875f57e58222b) is likely the most productive thing that can be done right now.

I can't think of anything in my sysctl patches that would kill an application. My sysctl work is right on the border with user space so it is a good candidate but at the same time there should have been no user visible changes. There were a few places where I removed sys_sysctl support (but not /proc/sys support) but I don't think any of those were on x86, and they were is such a messed up state I don't think anyone could have reasonably used them anyway.

So I think either we poke blindly making random guess by hand or we let git-bisect do it.

Sid do you think you can figure out git-bisect? git-bisect start
git-bisect bad 86a71dbd3e81e8870d0f0e56b87875f57e58222b git-bisect good ac98695d6c1508b724f246f38ce57fb4e3cec356

It should narrow the problem down to a single commit in 6-8 tries after which point we should have enough information to start making intelligent guesses.

Eric

author
#7 lore raw
Quoted 79 lines · Eric W. Biederman
Quoted 47 lines · Adrian Bunk
Quoted 45 lines · Sid Boyce
Quoted 43 lines · Sid Boyce
Quoted 23 lines · Andrew Morton

(cc restored. Please always do reply-to-all)

Quoted 9 lines · On Wed, 28 Feb 2007 18:05:13 +0200 [email protected] wrote:
Quoted 5 lines · Sid Boyce

openSUSE 10.3 Alpha and KDE-3.5.6, xorg-x11-7.2. KDE is setup not to require a password to unlock, but it asks for password. When the screen unlocks, kwin is gone with no errors logged in /var/log/kdm or /var/log/messages. No problems with 2.6.20.

Quoted 8 lines
Quoted 5 lines

Same problem on openSUSE 10.2 x86_64, KDE-3.5.5 and 2.6.21-rc2. Regards
Sid.

This is the linux kernel mailing list. Perhaps you should post your problem to the opensuse mailing list.

2.6.20 worked.

Quoted 13 lines

2.6.20-rc2 did not.

Working theory: the kernel broke.

Sid, the chances that anyone can work out what caused this are pretty low. It would be great if you could perform a git bisection search sometime in the next few weeks, work out which commit caused this.

Thanks.

I shall go back to 2.6.20-git3 and work forward. Up to 2.6.20-git2 was OK. Regards
Sid.

Quoted 8 lines
Quoted 2 lines

I tracked the problem down to 2.6.20-git11. Up to 2.6.20-git10 is OK, but from 2.6.20-git11 up to current 2.6.21-rc4-git2 all exhibit the problem.

Thanks for this search.

Looking at the changes between 2.6.20-git10 and 2.6.20-git11, the only suspicious changes are the 60 sysctl patches by Eric.

Eric, can you look at this issue?

git bisect between git10 (ac98695d6c1508b724f246f38ce57fb4e3cec356) and git11 (86a71dbd3e81e8870d0f0e56b87875f57e58222b) is likely the most productive thing that can be done right now.

I can't think of anything in my sysctl patches that would kill an application. My sysctl work is right on the border with user space so it is a good candidate but at the same time there should have been no user visible changes. There were a few places where I removed sys_sysctl support (but not /proc/sys support) but I don't think any of those were on x86, and they were is such a messed up state I don't think anyone could have reasonably used them anyway.

So I think either we poke blindly making random guess by hand or we let git-bisect do it.

Sid do you think you can figure out git-bisect? git-bisect start
git-bisect bad 86a71dbd3e81e8870d0f0e56b87875f57e58222b git-bisect good ac98695d6c1508b724f246f38ce57fb4e3cec356

Quoted 8 lines

It should narrow the problem down to a single commit in 6-8 tries after which point we should have enough information to start making intelligent guesses.

Eric

Reading the manpage doesn't help, so I shall have to delve into the docs or futher help is needed. :/usr/src/linux-2.6.20-git11 # git-bisect good ac98695d6c1508b724f246f38ce57fb4e3cec356 No revs to be shown.

:/usr/src/linux-2.6.20-git11 # ls .git/refs/bisect/
bad good-ac98695d6c1508b724f246f38ce57fb4e3cec356

:/usr/src/linux-2.6.20-git11 # less .git/refs/bisect/good-ac98695d6c1508b724f246f38ce57fb4e3cec356 ac98695d6c1508b724f246f38ce57fb4e3cec356

:/usr/src/linux-2.6.20-git11 # less .git/refs/bisect/bad 86a71dbd3e81e8870d0f0e56b87875f57e58222b Regards
Sid.

Signature · 3 lines
Sid Boyce ... Hamradio License G3VBV, Licensed Private Pilot
Emeritus IBM/Amdahl Mainframes and Sun/Fujitsu Servers Tech Support Specialist, Cricket Coach
Microsoft Windows Free Zone - Linux used for all Computing Tasks
#8 lore raw
Quoted 89 lines · On Thu, 22 Mar 2007 01:32:36 +0000 Sid Boyce wrote:
Quoted 88 lines · Eric W. Biederman
Quoted 53 lines · Adrian Bunk
Quoted 43 lines · Sid Boyce
Quoted 41 lines · Sid Boyce
Quoted 39 lines · Andrew Morton

(cc restored. Please always do reply-to-all)

Quoted 9 lines · On Wed, 28 Feb 2007 18:05:13 +0200 [email protected] wrote:
Quoted 5 lines · Sid Boyce

openSUSE 10.3 Alpha and KDE-3.5.6, xorg-x11-7.2. KDE is setup not to require a password to unlock, but it asks for password. When the screen unlocks, kwin is gone with no errors logged in /var/log/kdm or /var/log/messages. No problems with 2.6.20.

Quoted 9 lines
Quoted 5 lines

Same problem on openSUSE 10.2 x86_64, KDE-3.5.5 and 2.6.21-rc2. Regards
Sid.

This is the linux kernel mailing list. Perhaps you should post your problem to the opensuse mailing list.

2.6.20 worked.

2.6.20-rc2 did not.

Working theory: the kernel broke.

Sid, the chances that anyone can work out what caused this are pretty low. It would be great if you could perform a git bisection search sometime in
the next few weeks, work out which commit caused this.

Thanks.

Quoted 8 lines
Quoted 5 lines

I shall go back to 2.6.20-git3 and work forward. Up to 2.6.20-git2 was OK. Regards
Sid.

I tracked the problem down to 2.6.20-git11. Up to 2.6.20-git10 is OK, but from 2.6.20-git11 up to current 2.6.21-rc4-git2 all exhibit the problem.

Quoted 7 lines

Thanks for this search.

Looking at the changes between 2.6.20-git10 and 2.6.20-git11, the only suspicious changes are the 60 sysctl patches by Eric.

Eric, can you look at this issue?

git bisect between git10 (ac98695d6c1508b724f246f38ce57fb4e3cec356) and git11 (86a71dbd3e81e8870d0f0e56b87875f57e58222b) is likely the most productive thing that can be done right now.

I can't think of anything in my sysctl patches that would kill an application. My sysctl work is right on the border with user space so it is a good candidate but at the same time there should have been no user visible changes. There were a few places where I removed sys_sysctl support (but not /proc/sys support) but I don't think any of those were on x86, and they were is such a messed up state I don't think anyone could have reasonably used them anyway.

So I think either we poke blindly making random guess by hand or we let git-bisect do it.

Sid do you think you can figure out git-bisect? git-bisect start
git-bisect bad 86a71dbd3e81e8870d0f0e56b87875f57e58222b git-bisect good ac98695d6c1508b724f246f38ce57fb4e3cec356

It should narrow the problem down to a single commit in 6-8 tries after which point we should have enough information to start making intelligent guesses.

Quoted 6 lines
Quoted 4 lines

Eric

Reading the manpage doesn't help, so I shall have to delve into the docs or futher help is needed.

There's not a lot of docs out there.

The man-page: http://www.kernel.org/pub/software/scm/git/docs/git-bisect.html

Linus's email doc:
http://www.kernel.org/pub/software/scm/git/docs/howto/isolate-bugs-with-bisect.txt

I worked on something over last weekend, but it doesn't really add much to the references above.

Quoted 3 lines

:/usr/src/linux-2.6.20-git11 # git-bisect good
ac98695d6c1508b724f246f38ce57fb4e3cec356
No revs to be shown.

Did you tell git where to begin with good and bad? I.e., you have to tell it the bracketing of where to do the binary search.

Quoted 10 lines

:/usr/src/linux-2.6.20-git11 # ls .git/refs/bisect/ bad good-ac98695d6c1508b724f246f38ce57fb4e3cec356

:/usr/src/linux-2.6.20-git11 # less .git/refs/bisect/good-ac98695d6c1508b724f246f38ce57fb4e3cec356 ac98695d6c1508b724f246f38ce57fb4e3cec356

:/usr/src/linux-2.6.20-git11 # less .git/refs/bisect/bad 86a71dbd3e81e8870d0f0e56b87875f57e58222b

---
~Randy
*** Remember to use Documentation/SubmitChecklist when testing your code ***

#9 lore raw
Quoted 5 lines · Randy Dunlap
Quoted 1 line · On Thu, 22 Mar 2007 01:32:36 +0000 Sid Boyce wrote:

...

There's not a lot of docs out there.

Quoted 8 lines

The man-page: http://www.kernel.org/pub/software/scm/git/docs/git-bisect.html

Linus's email doc:
http://www.kernel.org/pub/software/scm/git/docs/howto/isolate-bugs-with-bisect.txt

I worked on something over last weekend, but it doesn't really add much to the references above.
...

FWIW:
My standard instructions for users who are asked to bisect are [1]:

<-- snip -->

# install git and cogito on your computer
# clone Linus' tree:
cg-clone \ 
git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git
# start bisecting:
cd linux-2.6
git bisect start
git bisect bad 86a71dbd3e81e8870d0f0e56b87875f57e58222b
git bisect good ac98695d6c1508b724f246f38ce57fb4e3cec356
# start round
cp /path/to/.config .
make oldconfig
make
# install kernel, check whether it's good or bad, then:
git bisect [bad|good]
# start next round

After at about 8 reboots, you'll have found the guilty commit ("... is first bad commit").

More information on git bisecting:
man git-bisect

<-- snip -->

Quoted 1 line

~Randy

cu
Adrian

[1] the start and end commits and the number of reboots are variable, but the rest of the text is generic

Signature · 4 lines
       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed
author
#10 lore raw
Quoted 57 lines · Adrian Bunk
Quoted 7 lines · Randy Dunlap
Quoted 2 lines · On Thu, 22 Mar 2007 01:32:36 +0000 Sid Boyce wrote:

...

There's not a lot of docs out there.

Quoted 9 lines

The man-page: http://www.kernel.org/pub/software/scm/git/docs/git-bisect.html

Linus's email doc:
http://www.kernel.org/pub/software/scm/git/docs/howto/isolate-bugs-with-bisect.txt

I worked on something over last weekend, but it doesn't really add much to the references above.
...

FWIW:
My standard instructions for users who are asked to bisect are [1]:

<-- snip -->

# install git and cogito on your computer
# clone Linus' tree:
cg-clone \ git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git
# start bisecting:
cd linux-2.6
git bisect start
git bisect bad 86a71dbd3e81e8870d0f0e56b87875f57e58222b
git bisect good ac98695d6c1508b724f246f38ce57fb4e3cec356
# start round
cp /path/to/.config .
make oldconfig
make
# install kernel, check whether it's good or bad, then:
git bisect [bad|good]
# start next round

After at about 8 reboots, you'll have found the guilty commit ("... is first bad commit").

More information on git bisecting:
man git-bisect

<-- snip -->

Quoted 10 lines
Quoted 2 lines

~Randy

cu
Adrian

[1] the start and end commits and the number of reboots are variable, but the rest of the text is generic

Kernel built and installed, so I shall have something to report in the next couple of days. Regards
Sid.

Signature · 3 lines
Sid Boyce ... Hamradio License G3VBV, Licensed Private Pilot
Emeritus IBM/Amdahl Mainframes and Sun/Fujitsu Servers Tech Support Specialist, Cricket Coach
Microsoft Windows Free Zone - Linux used for all Computing Tasks
#11 lore raw
Quoted 2 lines · Sid Boyce

Kernel built and installed, so I shall have something to report in the next couple of days.

I see this kwin bug too
on a Thinkpad (x86_64) as well as on a Pmac (powerpc64). It is only triggerd when the screen is locked AND! set to standby mode (display/backlight is turned off - acpi).

boris

author Acked-by Tony Luck
#12 lore raw
In reply to Adrian Bunk
Quoted 57 lines · Adrian Bunk
Quoted 7 lines · Randy Dunlap
Quoted 2 lines · On Thu, 22 Mar 2007 01:32:36 +0000 Sid Boyce wrote:

...

There's not a lot of docs out there.

Quoted 9 lines

The man-page: http://www.kernel.org/pub/software/scm/git/docs/git-bisect.html

Linus's email doc:
http://www.kernel.org/pub/software/scm/git/docs/howto/isolate-bugs-with-bisect.txt

I worked on something over last weekend, but it doesn't really add much to the references above.
...

FWIW:
My standard instructions for users who are asked to bisect are [1]:

<-- snip -->

# install git and cogito on your computer
# clone Linus' tree:
cg-clone \ git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git
# start bisecting:
cd linux-2.6
git bisect start
git bisect bad 86a71dbd3e81e8870d0f0e56b87875f57e58222b
git bisect good ac98695d6c1508b724f246f38ce57fb4e3cec356
# start round
cp /path/to/.config .
make oldconfig
make
# install kernel, check whether it's good or bad, then:
git bisect [bad|good]
# start next round

After at about 8 reboots, you'll have found the guilty commit ("... is first bad commit").

More information on git bisecting:
man git-bisect

<-- snip -->

Quoted 10 lines
Quoted 2 lines

~Randy

cu
Adrian

[1] the start and end commits and the number of reboots are variable, but the rest of the text is generic

This is what I've got so far on the first boot, I shall have to check the manpage for git-bisect again to see if there is anything else to be added, nothing enlightening seen so far - further reboots to be done. barrabas:/usr/src/GIT/linux-2.6 # git bisect bad
Bisecting: 22 revisions left to test after this
[3fc605a2aa38899c12180ca311f1eeb61a6d867e] knfsd: allow the server to provide a gid list when using AUTH_UNIX authentication
rrabas:/usr/src/GIT/linux-2.6 # git bisect bad
Bisecting: 11 revisions left to test after this
[ed8b4d4d7a31923db32f4684535944d69eb43677] qconf: hide empty list items
barrabas:/usr/src/GIT/linux-2.6 #
rrabas:/usr/src/GIT/linux-2.6 # git bisect bad
Bisecting: 6 revisions left to test after this
[66e7c7230fed159e138fc1292ee662b8bbdb74d6] qconf: relocate Search Command
barrabas:/usr/src/GIT/linux-2.6 #

barrabas:/usr/src/GIT/linux-2.6 # git bisect bad
Bisecting: 3 revisions left to test after this [895a39a084e5478121a74752a291165c4502378f] Make mkcompile_h use LANG=C and LC_ALL=C for $CC -v

barrabas:/usr/src/GIT/linux-2.6 # git bisect bad
Bisecting: 2 revisions left to test after this [8c7e4498adfdb4aea5a0d056590ec18d099ba062] search a little harder for mkimage

barrabas:/usr/src/GIT/linux-2.6 # git bisect bad
Bisecting: 1 revisions left to test after this [b7de8e7e3c7b79a72c20c7fd58bd65df3d146b13] ia64: fix noncoherent DMA API so devres builds

# git bisect bad
b7de8e7e3c7b79a72c20c7fd58bd65df3d146b13 is first bad commit
commit b7de8e7e3c7b79a72c20c7fd58bd65df3d146b13
Author: Roland Dreier <[email protected]>
Date:   Wed Feb 14 00:32:53 2007 -0800

[PATCH] ia64: fix noncoherent DMA API so devres builds

On ia64, drivers/base/dma-mapping.c doesn't build because it calls dma_alloc_noncoherent() and dma_free_noncoherent(), which appear to be terminally broken; the calls end up generating errors like

drivers/base/dma-mapping.c: In function 'dmam_noncoherent_release':
drivers/base/dma-mapping.c:32: error: 'struct
ia64_machine_vector' has no member named 'platform_dma_free_coherent'

because the multiple levels of macro expansion in
<asm/dma-mapping.h> and
<asm/machvec.h> end up turning a call to dma_free_noncoherent() into ia64_mv.platform_dma_free_coherent (instead of the intended ia64_mv.dma_free_coherent).

This patch fixes this by converting dma_{alloc,free}_noncoherent() into inline functions that call the corresponding coherent functions,
instead of
trying to do this with macros.

:040000 040000 ac36019577dc1846a5655918cf9d6b2541c398cf ba536f54303cc9495b7839763ed1121d3a22154c M include

barrabas:/usr/src/GIT/linux-2.6 # git bisect bad 86a71dbd3e81e8870d0f0e56b87875f57e58222b
Bisecting: 44 revisions left to test after this [208367eeec8d6f6fcee1a0672af2149aeee7b658] sysctl: cdrom: remove unnecessary insert_at_head flag

barrabas:/usr/src/GIT/linux-2.6 # git bisect good ac98695d6c1508b724f246f38ce57fb4e3cec356
Bisecting: 44 revisions left to test after this [208367eeec8d6f6fcee1a0672af2149aeee7b658] sysctl: cdrom: remove unnecessary insert_at_head flag

Regards
Sid.

Signature · 3 lines
Sid Boyce ... Hamradio License G3VBV, Licensed Private Pilot
Emeritus IBM/Amdahl Mainframes and Sun/Fujitsu Servers Tech Support Specialist, Cricket Coach
Microsoft Windows Free Zone - Linux used for all Computing Tasks
#13 lore raw
Quoted 3 lines · Sid Boyce

This is what I've got so far on the first boot, I shall have to check the manpage for git-bisect again to see if there is anything else to be added, nothing enlightening seen so far - further reboots to be done.

I'm a little confused at what I am seeing below.
Normally what happens is that after a kernel compile and test. You call "git-bisect bad" or "git-bisect good" and then git-bisect picks the next kernel to test. Then you compile and test that one. It looks like you called git-bisect bad several times in a row without testing....

Eric

Quoted 80 lines

barrabas:/usr/src/GIT/linux-2.6 # git bisect bad
Bisecting: 22 revisions left to test after this [3fc605a2aa38899c12180ca311f1eeb61a6d867e] knfsd: allow the server to provide a gid list when using AUTH_UNIX authentication rrabas:/usr/src/GIT/linux-2.6 # git bisect bad
Bisecting: 11 revisions left to test after this [ed8b4d4d7a31923db32f4684535944d69eb43677] qconf: hide empty list items barrabas:/usr/src/GIT/linux-2.6 #
rrabas:/usr/src/GIT/linux-2.6 # git bisect bad
Bisecting: 6 revisions left to test after this [66e7c7230fed159e138fc1292ee662b8bbdb74d6] qconf: relocate Search Command barrabas:/usr/src/GIT/linux-2.6 #

barrabas:/usr/src/GIT/linux-2.6 # git bisect bad
Bisecting: 3 revisions left to test after this [895a39a084e5478121a74752a291165c4502378f] Make mkcompile_h use LANG=C and LC_ALL=C for $CC -v

barrabas:/usr/src/GIT/linux-2.6 # git bisect bad
Bisecting: 2 revisions left to test after this [8c7e4498adfdb4aea5a0d056590ec18d099ba062] search a little harder for mkimage

barrabas:/usr/src/GIT/linux-2.6 # git bisect bad
Bisecting: 1 revisions left to test after this [b7de8e7e3c7b79a72c20c7fd58bd65df3d146b13] ia64: fix noncoherent DMA API so devres builds

# git bisect bad
b7de8e7e3c7b79a72c20c7fd58bd65df3d146b13 is first bad commit
commit b7de8e7e3c7b79a72c20c7fd58bd65df3d146b13
Author: Roland Dreier <[email protected]>
Date:   Wed Feb 14 00:32:53 2007 -0800

[PATCH] ia64: fix noncoherent DMA API so devres builds

On ia64, drivers/base/dma-mapping.c doesn't build because it calls dma_alloc_noncoherent() and dma_free_noncoherent(), which appear to be terminally broken; the calls end up generating errors like

drivers/base/dma-mapping.c: In function 'dmam_noncoherent_release':
drivers/base/dma-mapping.c:32: error: 'struct ia64_machine_vector' has no
member named 'platform_dma_free_coherent'

because the multiple levels of macro expansion in <asm/dma-mapping.h> and <asm/machvec.h> end up turning a call to dma_free_noncoherent() into ia64_mv.platform_dma_free_coherent (instead of the intended ia64_mv.dma_free_coherent).

This patch fixes this by converting dma_{alloc,free}_noncoherent() into inline functions that call the corresponding coherent functions, instead of trying to do this with macros.

:040000 040000 ac36019577dc1846a5655918cf9d6b2541c398cf
ba536f54303cc9495b7839763ed1121d3a22154c M include

barrabas:/usr/src/GIT/linux-2.6 # git bisect bad 86a71dbd3e81e8870d0f0e56b87875f57e58222b
Bisecting: 44 revisions left to test after this [208367eeec8d6f6fcee1a0672af2149aeee7b658] sysctl: cdrom: remove unnecessary insert_at_head flag

barrabas:/usr/src/GIT/linux-2.6 # git bisect good ac98695d6c1508b724f246f38ce57fb4e3cec356
Bisecting: 44 revisions left to test after this [208367eeec8d6f6fcee1a0672af2149aeee7b658] sysctl: cdrom: remove unnecessary insert_at_head flag

Regards
Sid.

Signature · 4 lines
Sid Boyce ... Hamradio License G3VBV, Licensed Private Pilot
Emeritus IBM/Amdahl Mainframes and Sun/Fujitsu Servers Tech Support Specialist,
Cricket Coach
Microsoft Windows Free Zone - Linux used for all Computing Tasks
author
#14 lore raw
Quoted 8 lines · Eric W. Biederman
Quoted 4 lines · Sid Boyce

This is what I've got so far on the first boot, I shall have to check the manpage for git-bisect again to see if there is anything else to be added, nothing enlightening seen so far - further reboots to be done.

Quoted 10 lines

I'm a little confused at what I am seeing below.
Normally what happens is that after a kernel compile and test. You call "git-bisect bad" or "git-bisect good" and then git-bisect picks the next kernel to test. Then you compile and test that one. It looks like you called git-bisect bad several times in a row without testing....

Eric

This is the procedure I followed.
<-- snip -->

# install git and cogito on your computer
# clone Linus' tree:
cg-clone  git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git
# start bisecting:
cd linux-2.6
git bisect start
git bisect bad 86a71dbd3e81e8870d0f0e56b87875f57e58222b
git bisect good ac98695d6c1508b724f246f38ce57fb4e3cec356
# start round
cp /path/to/.config .
make oldconfig
make
# install kernel, check whether it's good or bad, then:
git bisect [bad|good]
========================================================
I built and installed the kernel downloaded using "cg-clone", booted it up  and ran "git bisect bad".  I'm also somewhat confused as the first failing kernel is 2.6.20-git11, the Makefile says the kernel got with cg-clone is actually 2.6.21-rc4 and it built as  2.6.20-g208367ee.
Regards
Sid.
Quoted 84 lines
Quoted 81 lines

barrabas:/usr/src/GIT/linux-2.6 # git bisect bad
Bisecting: 22 revisions left to test after this [3fc605a2aa38899c12180ca311f1eeb61a6d867e] knfsd: allow the server to provide a gid list when using AUTH_UNIX authentication rrabas:/usr/src/GIT/linux-2.6 # git bisect bad
Bisecting: 11 revisions left to test after this [ed8b4d4d7a31923db32f4684535944d69eb43677] qconf: hide empty list items barrabas:/usr/src/GIT/linux-2.6 #
rrabas:/usr/src/GIT/linux-2.6 # git bisect bad
Bisecting: 6 revisions left to test after this [66e7c7230fed159e138fc1292ee662b8bbdb74d6] qconf: relocate Search Command barrabas:/usr/src/GIT/linux-2.6 #

barrabas:/usr/src/GIT/linux-2.6 # git bisect bad
Bisecting: 3 revisions left to test after this [895a39a084e5478121a74752a291165c4502378f] Make mkcompile_h use LANG=C and LC_ALL=C for $CC -v

barrabas:/usr/src/GIT/linux-2.6 # git bisect bad
Bisecting: 2 revisions left to test after this [8c7e4498adfdb4aea5a0d056590ec18d099ba062] search a little harder for mkimage

barrabas:/usr/src/GIT/linux-2.6 # git bisect bad
Bisecting: 1 revisions left to test after this [b7de8e7e3c7b79a72c20c7fd58bd65df3d146b13] ia64: fix noncoherent DMA API so devres builds

# git bisect bad
b7de8e7e3c7b79a72c20c7fd58bd65df3d146b13 is first bad commit
commit b7de8e7e3c7b79a72c20c7fd58bd65df3d146b13
Author: Roland Dreier <[email protected]>
Date:   Wed Feb 14 00:32:53 2007 -0800

[PATCH] ia64: fix noncoherent DMA API so devres builds

On ia64, drivers/base/dma-mapping.c doesn't build because it calls dma_alloc_noncoherent() and dma_free_noncoherent(), which appear to be terminally broken; the calls end up generating errors like

drivers/base/dma-mapping.c: In function 'dmam_noncoherent_release':
drivers/base/dma-mapping.c:32: error: 'struct ia64_machine_vector' has no
member named 'platform_dma_free_coherent'

because the multiple levels of macro expansion in <asm/dma-mapping.h> and <asm/machvec.h> end up turning a call to dma_free_noncoherent() into ia64_mv.platform_dma_free_coherent (instead of the intended ia64_mv.dma_free_coherent).

This patch fixes this by converting dma_{alloc,free}_noncoherent() into inline functions that call the corresponding coherent functions, instead of trying to do this with macros.

:040000 040000 ac36019577dc1846a5655918cf9d6b2541c398cf
ba536f54303cc9495b7839763ed1121d3a22154c M include

barrabas:/usr/src/GIT/linux-2.6 # git bisect bad 86a71dbd3e81e8870d0f0e56b87875f57e58222b
Bisecting: 44 revisions left to test after this [208367eeec8d6f6fcee1a0672af2149aeee7b658] sysctl: cdrom: remove unnecessary insert_at_head flag

barrabas:/usr/src/GIT/linux-2.6 # git bisect good ac98695d6c1508b724f246f38ce57fb4e3cec356
Bisecting: 44 revisions left to test after this [208367eeec8d6f6fcee1a0672af2149aeee7b658] sysctl: cdrom: remove unnecessary insert_at_head flag

Regards
Sid.

Signature · 4 lines
Sid Boyce ... Hamradio License G3VBV, Licensed Private Pilot
Emeritus IBM/Amdahl Mainframes and Sun/Fujitsu Servers Tech Support Specialist,
Cricket Coach
Microsoft Windows Free Zone - Linux used for all Computing Tasks
Signature · 3 lines
Sid Boyce ... Hamradio License G3VBV, Licensed Private Pilot
Emeritus IBM/Amdahl Mainframes and Sun/Fujitsu Servers Tech Support Specialist, Cricket Coach
Microsoft Windows Free Zone - Linux used for all Computing Tasks
#15 lore raw
Quoted 48 lines · Sid Boyce
Quoted 8 lines · Eric W. Biederman
Quoted 4 lines · Sid Boyce

This is what I've got so far on the first boot, I shall have to check the manpage for git-bisect again to see if there is anything else to be added, nothing enlightening seen so far - further reboots to be done.

Quoted 10 lines

I'm a little confused at what I am seeing below.
Normally what happens is that after a kernel compile and test. You call "git-bisect bad" or "git-bisect good" and then git-bisect picks the next kernel to test. Then you compile and test that one. It looks like you called git-bisect bad several times in a row without testing....

Eric

This is the procedure I followed.
<-- snip -->

# install git and cogito on your computer
# clone Linus' tree:
cg-clone  
git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git
# start bisecting:
cd linux-2.6
git bisect start
git bisect bad 86a71dbd3e81e8870d0f0e56b87875f57e58222b
git bisect good ac98695d6c1508b724f246f38ce57fb4e3cec356
# start round
cp /path/to/.config .
make oldconfig
make
# install kernel, check whether it's good or bad, then:
git bisect [bad|good]
========================================================
I built and installed the kernel downloaded using "cg-clone", booted it 
up  and ran "git bisect bad".  I'm also somewhat confused as the first 
failing kernel is 2.6.20-git11, the Makefile says the kernel got with 
cg-clone is actually 2.6.21-rc4 and it built as  2.6.20-g208367ee.
Regards
Sid.
Quoted 7 lines
Quoted 6 lines
Quoted 6 lines

barrabas:/usr/src/GIT/linux-2.6 # git bisect bad
Bisecting: 22 revisions left to test after this [3fc605a2aa38899c12180ca311f1eeb61a6d867e] knfsd: allow the server to provide a
gid list when using AUTH_UNIX authentication rrabas:/usr/src/GIT/linux-2.6 # git bisect bad

...

I think I got your problem:

After rebooting, you do "git bisect [bad|good]" *once*.

Then recompile the kernel from the current tree, reboot, and again *once* "git bisect [bad|good]".

etc.

cu
Adrian

Signature · 4 lines
       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed
#16 lore raw
Quoted 8 lines · Adrian Bunk

I think I got your problem:

After rebooting, you do "git bisect [bad|good]" *once*.

Then recompile the kernel from the current tree, reboot, and again *once* "git bisect [bad|good]".

etc.

Sounds right.

Someone else doing the bisect suggested that 0475ac0845f9295bc5f69af45f58dff2c104c8d1 is the problem commit. Unfortunately I can't see anything wrong there.

Sid when you get that far if you can confirm or deny that 0475ac0845f9295bc5f69af45f58dff2c104c8d1 is the problem commit I would appreciate the extra data point. Having two people report the problem at the exact same place would confirm for me that we actually have a good trace, and that I really am blind.

Eric

#19 lore raw
In reply to Sid Boyce

Sid I think I have found the problem. Could you try the following patch. I believe I accidentally switched the sense of a test....

kernel/exit.c +1 -1
index f132349..b55ed4c 100644
@@ -790,7 +790,7 @@ static void exit_notify(struct task_struct *tsk)
 	
 	pgrp = task_pgrp(tsk);
 	if ((task_pgrp(t) != pgrp) &&
-	    (task_session(t) != task_session(tsk)) &&
+	    (task_session(t) == task_session(tsk)) &&
 	    will_become_orphaned_pgrp(pgrp, tsk) &&
 	    has_stopped_jobs(pgrp)) {
 		__kill_pgrp_info(SIGHUP, SEND_SIG_PRIV, pgrp);
author
#17 lore raw
In reply to Adrian Bunk
Quoted 63 lines · Adrian Bunk
Quoted 48 lines · Sid Boyce
Quoted 9 lines · Eric W. Biederman
Quoted 5 lines · Sid Boyce

This is what I've got so far on the first boot, I shall have to check the manpage for git-bisect again to see if there is anything else to be added, nothing enlightening seen so far - further reboots to be done.

Quoted 11 lines

I'm a little confused at what I am seeing below.
Normally what happens is that after a kernel compile and test. You call "git-bisect bad" or "git-bisect good" and then git-bisect picks the next kernel to test. Then you compile and test that one. It looks like you called git-bisect bad several times in a row without testing....

Eric

This is the procedure I followed.
<-- snip -->

# install git and cogito on your computer
# clone Linus' tree:
cg-clone  git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git
# start bisecting:
cd linux-2.6
git bisect start
git bisect bad 86a71dbd3e81e8870d0f0e56b87875f57e58222b
git bisect good ac98695d6c1508b724f246f38ce57fb4e3cec356
# start round
cp /path/to/.config .
make oldconfig
make
# install kernel, check whether it's good or bad, then:
git bisect [bad|good]
========================================================
I built and installed the kernel downloaded using "cg-clone", booted it up  and ran "git bisect bad".  I'm also somewhat confused as the first failing kernel is 2.6.20-git11, the Makefile says the kernel got with cg-clone is actually 2.6.21-rc4 and it built as  2.6.20-g208367ee.
Regards
Sid.
Quoted 8 lines
Quoted 6 lines
Quoted 6 lines

barrabas:/usr/src/GIT/linux-2.6 # git bisect bad
Bisecting: 22 revisions left to test after this [3fc605a2aa38899c12180ca311f1eeb61a6d867e] knfsd: allow the server to provide a gid list when using AUTH_UNIX authentication rrabas:/usr/src/GIT/linux-2.6 # git bisect bad

...

I think I got your problem:

After rebooting, you do "git bisect [bad|good]" *once*.

Quoted 9 lines

Then recompile the kernel from the current tree, reboot, and again *once* "git bisect [bad|good]".

etc.

cu
Adrian

If I boot up on 2.6.20-git10 and run "git bisect bad", "git bisect good v2.6.20-git10" or "git bisect bad v2.6.20-git11", it says it's not a git repo.

I am still no clearer what's required. If I do "cg-clone" I get 2.6.21-rc5. The previous one got me 2.6.21-rc4. The manpages and the tutorials I've looked at are quite cryptic and the commands do not produce the expected results. As indicated above, I followed the instructions bullet by bullet without any clear understanding of the tools.
1) Ran the commands as above,cg-clone; git start; git bisect bad <checksum>;git bisect good <checksum of 2.6.20-git10 > I presume.
2) 2.6.20-git10 is the last good kernel.
3) 2.6.20-git11 and later are all bad.
4) I do cg-clone which gives a kernel whose Makefile is 2.6.21-rc4 (or whatever is current).
5) I get a kernel named 2.6.20-g208367e after make - rebooted and it fails the same way as 2.6.20-git11.
6) ran "git bisect bad".
Looking through the directories after cg-clone
barrabas:/ftp/mar07/GIT/linux-2.6 # ls .git/refs/tags/

v2.6.11       v2.6.12-rc5  v2.6.13-rc4  v2.6.14-rc2  v2.6.15-rc2  v2.6.16      v2.6.16-rc6  v2.6.17-rc5  v2.6.18-rc4  v2.6.19-rc2  v2.6.20-rc1  v2.6.20-rc7
v2.6.11-tree  v2.6.12-rc6  v2.6.13-rc5  v2.6.14-rc3  v2.6.15-rc3  v2.6.16-rc1  v2.6.17      v2.6.17-rc6  v2.6.18-rc5  v2.6.19-rc3  v2.6.20-rc2  v2.6.21-rc1
v2.6.12       v2.6.13      v2.6.13-rc6  v2.6.14-rc4  v2.6.15-rc4  v2.6.16-rc2  v2.6.17-rc1  v2.6.18      v2.6.18-rc6  v2.6.19-rc4  v2.6.20-rc3  v2.6.21-rc2
v2.6.12-rc2   v2.6.13-rc1  v2.6.13-rc7  v2.6.14-rc5  v2.6.15-rc5  v2.6.16-rc3  v2.6.17-rc2  v2.6.18-rc1  v2.6.18-rc7  v2.6.19-rc5  v2.6.20-rc4  v2.6.21-rc3
v2.6.12-rc3   v2.6.13-rc2  v2.6.14      v2.6.15      v2.6.15-rc6  v2.6.16-rc4  v2.6.17-rc3  v2.6.18-rc2  v2.6.19      v2.6.19-rc6  v2.6.20-rc5  v2.6.21-rc4
v2.6.12-rc4   v2.6.13-rc3  v2.6.14-rc1  v2.6.15-rc1  v2.6.15-rc7  v2.6.16-rc5  v2.6.17-rc4  v2.6.18-rc3  v2.6.19-rc1  v2.6.20      v2.6.20-rc6  v2.6.21-rc5

barrabas:/ftp/mar07/GIT/linux-2.6 # ls .git/refs/bisect/ bad good-ac98695d6c1508b724f246f38ce57fb4e3cec356 barrabas:/ftp/mar07/GIT/linux-2.6 # ls .git/refs/heads/ bisect master origin barrabas:/ftp/mar07/GIT/linux-2.6 # less .git/refs/heads/bisect 208367eeec8d6f6fcee1a0672af2149aeee7b658

barrabas:/ftp/mar07/GIT/linux-2.6 # ls .git/branches/

barrabas:/ftp/mar07/GIT/linux-2.6 # ls

arch   COPYING  crypto         drivers  .git        include  ipc     kernel  .mailmap     Makefile  net     REPORTING-BUGS  security  usr
block  CREDITS  Documentation  fs       .gitignore  init     Kbuild  lib     MAINTAINERS  mm        README  scripts         sound
# less BISECT_LOG
git-bisect start
# bad: [86a71dbd3e81e8870d0f0e56b87875f57e58222b] sysctl: hide the sysctl proc inodes from selinux
git-bisect bad 86a71dbd3e81e8870d0f0e56b87875f57e58222b
# good: [ac98695d6c1508b724f246f38ce57fb4e3cec356] Merge branch 'master' of /home/trondmy/kernel/linux-2.6/
git-bisect good ac98695d6c1508b724f246f38ce57fb4e3cec356

Regards
Sid.

Signature · 3 lines
Sid Boyce ... Hamradio License G3VBV, Licensed Private Pilot
Emeritus IBM/Amdahl Mainframes and Sun/Fujitsu Servers Tech Support Specialist, Cricket Coach
Microsoft Windows Free Zone - Linux used for all Computing Tasks
author
#20 lore raw
In reply to Eric W. Biederman
Quoted 12 lines · Eric W. Biederman

Sid I think I have found the problem. Could you try the following patch. I believe I accidentally switched the sense of a test....

kernel/exit.c +0 -0
index f132349..b55ed4c 100644

kernel/exit.c static void exit_notify(struct task_struct *tsk) · line 796

Quoted 8 lines
kernel/exit.c +1 -1
@@ -790,7 +790,7 @@ static void exit_notify(struct task_struct *tsk)
 
 	pgrp = task_pgrp(tsk);
 	if ((task_pgrp(t) != pgrp) &&
-	    (task_session(t) != task_session(tsk)) &&
+	    (task_session(t) == task_session(tsk)) &&
 	    will_become_orphaned_pgrp(pgrp, tsk) &&
 	    has_stopped_jobs(pgrp)) {
 		__kill_pgrp_info(SIGHUP, SEND_SIG_PRIV, pgrp);

AOK I shall apply to 2.6.21-rc5-git4 and test.
Regards
Sid.

Signature · 3 lines
Sid Boyce ... Hamradio License G3VBV, Licensed Private Pilot
Emeritus IBM/Amdahl Mainframes and Sun/Fujitsu Servers Tech Support Specialist, Cricket Coach
Microsoft Windows Free Zone - Linux used for all Computing Tasks
#21 lore raw
Quoted 9 lines · Sid Boyce
Quoted 8 lines · Eric W. Biederman

Sid I think I have found the problem. Could you try the following patch. I believe I accidentally switched the sense of a test....

kernel/exit.c +0 -0
index f132349..b55ed4c 100644
@@ -790,7 +790,7 @@ static void exit_notify(struct task_struct *tsk)
 
Quoted 16 lines

kernel/exit.c static void exit_notify(struct task_struct *tsk) · line 796

Quoted 12 lines
kernel/exit.c +1 -1
     pgrp = task_pgrp(tsk);
     if ((task_pgrp(t) != pgrp) &&
-        (task_session(t) != task_session(tsk)) &&
+        (task_session(t) == task_session(tsk)) &&
         will_become_orphaned_pgrp(pgrp, tsk) &&
         has_stopped_jobs(pgrp)) {
         __kill_pgrp_info(SIGHUP, SEND_SIG_PRIV, pgrp);

AOK I shall apply to 2.6.21-rc5-git4 and test.
Regards
Sid.

Thanks, the patch works - tested on 2.6.21-rc5-git4. Regards
Sid.

Signature · 3 lines
Sid Boyce ... Hamradio License G3VBV, Licensed Private Pilot
Emeritus IBM/Amdahl Mainframes and Sun/Fujitsu Servers Tech Support Specialist, Cricket Coach
Microsoft Windows Free Zone - Linux used for all Computing Tasks
#18 lore raw
In reply to Sid Boyce
Quoted 8 lines · Sid Boyce

...
1) Ran the commands as above,cg-clone; git start; git bisect bad <checksum>;git bisect good <checksum of 2.6.20-git10 > I presume.
2) 2.6.20-git10 is the last good kernel.
3) 2.6.20-git11 and later are all bad.
4) I do cg-clone which gives a kernel whose Makefile is 2.6.21-rc4 (or whatever is current).
...

No, that's what should be repeated:

# start round
cp /path/to/.config .
make oldconfig
make
# install kernel, check whether it's good or bad, then:
git bisect [bad|good]
# start next round
Quoted 2 lines

Regards
Sid.

cu
Adrian

Signature · 4 lines
       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed