* [uml-devel] Re: 2.6.9-bb3 creates zombie threads in SKAS
2004-11-22 22:46 ` [uml-devel] Re: 2.6.9-bb3 creates zombie threads in SKAS Blaisorblade
@ 2004-11-22 23:33 ` Christopher S. Aker
2004-11-22 23:56 ` roland
2004-11-23 14:48 ` [uml-devel] " Bodo Stroesser
2004-11-23 17:49 ` Christopher S. Aker
2 siblings, 1 reply; 14+ messages in thread
From: Christopher S. Aker @ 2004-11-22 23:33 UTC (permalink / raw)
To: Blaisorblade; +Cc: user-mode-linux-devel, Bodo Stroesser
> On Monday 22 November 2004 23:23, Christopher S. Aker wrote:
> > Hello,
> Btw, question: what about your I/O token limiter? Actively using it? Could
> start being worked on for merge (there is at least one bug-report on it by
> "roland" on a strange but valid test-case).
Yes, I use it in all my kernels -- it's invaluable.to me for stopping swap-thrashing
UMLs from taking down the entire host. What bug did Roland find? If you want to
merge it, you probably want to add some way to turn it off completely (perhaps by
default). I'm sure there are other improvements to be made...
> > I thought I'd give 2.6.9-bb3 a shot, and I had a bunch of users report the
> > accumulation of zombies.
> What the hell! This should *not* happen. What did we (I mean developers) do?
> > This is on a 2.6.7 host, using your
> > host-skas3-2.6.7-v1 patch.
> ??? There is a ton of bugs in that patch (see the changelogs on my site).
> Actually, it's likely you're not going to be hurt by them if you're using
> NPTL on the host (it is *hard* to get them in that case), but you'd better
> upgrade.
No doubt... I've been getting ready for The Great Reboot (38 hosts), but it takes
careful planning with that many hosts. I've been testing a SKAS-V7 based host kernel
and things look very good. All the weird bugs have gone away (mysql instability,
Gentoo not able to complete emerges, this is all on either V1 or V6 skas patches).
Looking forward to the upgrade to V7, and the latest CFQ changes.
> I have been releasing tons of updates, and even too quickly (most times I was
> not introducing new bugs, just fixing the ones I and Bodo Stroesser was
> finding) but -V1 is "a bit" old. That said, this should not happen.
Changes have been coming so quickly, it's hard to keep up :)
> > I'm not sure if it happened on newer host
> > kernels, since I had to revert it back quickly.
>
> Back to what? To 2.6.9-bb2 or to 2.6.9?
Vanilla 2.6.9-um
> I'll check in these days what's going on... it was hyper-tested on 2.6.9 host
> (not by me, but I trust Bodo Stroesser normally), but not on other hosts, I
> guess.
Thanks,
-Chris
-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://productguide.itmanagersjournal.com/
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel
^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [uml-devel] Re: 2.6.9-bb3 creates zombie threads in SKAS
2004-11-22 23:33 ` Christopher S. Aker
@ 2004-11-22 23:56 ` roland
2004-11-23 7:28 ` [uml-devel] Updated token-limiter patch, was " Christopher S. Aker
0 siblings, 1 reply; 14+ messages in thread
From: roland @ 2004-11-22 23:56 UTC (permalink / raw)
To: Christopher S. Aker, Blaisorblade; +Cc: user-mode-linux-devel, Bodo Stroesser
Hi Christopher,
> > Btw, question: what about your I/O token limiter? Actively using it? Could
> > start being worked on for merge (there is at least one bug-report on it by
> > "roland" on a strange but valid test-case).
see http://sourceforge.net/mailarchive/message.php?msg_id=8560363
that`s probably just a minor one.
roland
----- Original Message -----
From: "Christopher S. Aker" <caker@theshore.net>
To: "Blaisorblade" <blaisorblade_spam@yahoo.it>
Cc: <user-mode-linux-devel@lists.sourceforge.net>; "Bodo Stroesser" <bstroesser@fujitsu-siemens.com>
Sent: Tuesday, November 23, 2004 12:33 AM
Subject: [uml-devel] Re: 2.6.9-bb3 creates zombie threads in SKAS
> > On Monday 22 November 2004 23:23, Christopher S. Aker wrote:
> > > Hello,
> > Btw, question: what about your I/O token limiter? Actively using it? Could
> > start being worked on for merge (there is at least one bug-report on it by
> > "roland" on a strange but valid test-case).
>
> Yes, I use it in all my kernels -- it's invaluable.to me for stopping swap-thrashing
> UMLs from taking down the entire host. What bug did Roland find? If you want to
> merge it, you probably want to add some way to turn it off completely (perhaps by
> default). I'm sure there are other improvements to be made...
>
> > > I thought I'd give 2.6.9-bb3 a shot, and I had a bunch of users report the
> > > accumulation of zombies.
> > What the hell! This should *not* happen. What did we (I mean developers) do?
> > > This is on a 2.6.7 host, using your
> > > host-skas3-2.6.7-v1 patch.
> > ??? There is a ton of bugs in that patch (see the changelogs on my site).
> > Actually, it's likely you're not going to be hurt by them if you're using
> > NPTL on the host (it is *hard* to get them in that case), but you'd better
> > upgrade.
>
> No doubt... I've been getting ready for The Great Reboot (38 hosts), but it takes
> careful planning with that many hosts. I've been testing a SKAS-V7 based host kernel
> and things look very good. All the weird bugs have gone away (mysql instability,
> Gentoo not able to complete emerges, this is all on either V1 or V6 skas patches).
> Looking forward to the upgrade to V7, and the latest CFQ changes.
>
> > I have been releasing tons of updates, and even too quickly (most times I was
> > not introducing new bugs, just fixing the ones I and Bodo Stroesser was
> > finding) but -V1 is "a bit" old. That said, this should not happen.
>
> Changes have been coming so quickly, it's hard to keep up :)
>
> > > I'm not sure if it happened on newer host
> > > kernels, since I had to revert it back quickly.
> >
> > Back to what? To 2.6.9-bb2 or to 2.6.9?
>
> Vanilla 2.6.9-um
>
> > I'll check in these days what's going on... it was hyper-tested on 2.6.9 host
> > (not by me, but I trust Bodo Stroesser normally), but not on other hosts, I
> > guess.
>
> Thanks,
> -Chris
>
>
>
> -------------------------------------------------------
> SF email is sponsored by - The IT Product Guide
> Read honest & candid reviews on hundreds of IT Products from real users.
> Discover which products truly live up to the hype. Start reading now.
> http://productguide.itmanagersjournal.com/
> _______________________________________________
> User-mode-linux-devel mailing list
> User-mode-linux-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel
>
-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://productguide.itmanagersjournal.com/
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel
^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [uml-devel] Updated token-limiter patch, was 2.6.9-bb3 creates zombie threads in SKAS
2004-11-22 23:56 ` roland
@ 2004-11-23 7:28 ` Christopher S. Aker
2004-11-23 20:10 ` Blaisorblade
0 siblings, 1 reply; 14+ messages in thread
From: Christopher S. Aker @ 2004-11-23 7:28 UTC (permalink / raw)
To: roland, Blaisorblade; +Cc: user-mode-linux-devel, Bodo Stroesser
> > > Btw, question: what about your I/O token limiter? Actively using it? Could
> > > start being worked on for merge (there is at least one bug-report on it by
> > > "roland" on a strange but valid test-case).
> see http://sourceforge.net/mailarchive/message.php?msg_id=8560363
>
> that`s probably just a minor one.
>
> roland
From that post:
> ps:
> the token limiter seems to have a little bug:
> set token_refill=1
> do a "dd if=/dev/zero of=test.dat bs=1024k count=whatever"
> then "sync".
> for me, io isn`t limited here - io_tokens just becomes negative - nothing more.
> ok - io_refill=1 is no "real world" value anyway - so not really important to be
fixed - but interesting :)
I see. nanosleep doesn't accept an tv_nsec value over 999999999, so there's a simple
fix in this patch:
http://www.theshore.net/~caker/uml/patches/token-limiter-v2.patch
This patch also includes the proc-io_status patch. It's against 2.6.9, but applies
to 2.4-um as well.
-Chris
-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://productguide.itmanagersjournal.com/
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel
^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [uml-devel] Updated token-limiter patch, was 2.6.9-bb3 creates zombie threads in SKAS
2004-11-23 7:28 ` [uml-devel] Updated token-limiter patch, was " Christopher S. Aker
@ 2004-11-23 20:10 ` Blaisorblade
0 siblings, 0 replies; 14+ messages in thread
From: Blaisorblade @ 2004-11-23 20:10 UTC (permalink / raw)
To: user-mode-linux-devel; +Cc: Christopher S. Aker, roland, Bodo Stroesser
On Tuesday 23 November 2004 08:28, Christopher S. Aker wrote:
> > > > Btw, question: what about your I/O token limiter? Actively using it?
> > > > Could start being worked on for merge (there is at least one
> > > > bug-report on it by "roland" on a strange but valid test-case).
> > see http://sourceforge.net/mailarchive/message.php?msg_id=8560363
> > that`s probably just a minor one.
> > roland
> From that post:
> > ps:
> > the token limiter seems to have a little bug:
> > set token_refill=1
> > do a "dd if=/dev/zero of=test.dat bs=1024k count=whatever"
> > then "sync".
> I see. nanosleep doesn't accept an tv_nsec value over 999 999 999, so
> there's a simple fix in this patch:
> http://www.theshore.net/~caker/uml/patches/token-limiter-v2.patch
> This patch also includes the proc-io_status patch. It's against 2.6.9, but
> applies to 2.4-um as well.
>
> -Chris
Ok, seen that - to be cleaner, I'll change when I remember this to set,
instead, tv_sec to 1 (tv_nsec cannot be one billion because one billion of
nsec is one sec). I see that there is no actual difference and that doing
this is a bit more invasive, so I guess that you fixed it this way on
purpose, but, for merging, such things deserve cleaning. Also the proc status
will have to be moved to a uml-only file.
Just a little thing, anyway - the patch is very nice.
--
Paolo Giarrusso, aka Blaisorblade
Linux registered user n. 292729
http://www.user-mode-linux.org/~blaisorblade
-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://productguide.itmanagersjournal.com/
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel
^ permalink raw reply [flat|nested] 14+ messages in thread
* [uml-devel] Re: 2.6.9-bb3 creates zombie threads in SKAS
2004-11-22 22:46 ` [uml-devel] Re: 2.6.9-bb3 creates zombie threads in SKAS Blaisorblade
2004-11-22 23:33 ` Christopher S. Aker
@ 2004-11-23 14:48 ` Bodo Stroesser
2004-11-23 17:49 ` Christopher S. Aker
2 siblings, 0 replies; 14+ messages in thread
From: Bodo Stroesser @ 2004-11-23 14:48 UTC (permalink / raw)
To: Blaisorblade; +Cc: Christopher S. Aker, user-mode-linux-devel
Blaisorblade wrote:
> On Monday 22 November 2004 23:23, Christopher S. Aker wrote:
>
>>Hello,
>
> Btw, question: what about your I/O token limiter? Actively using it? Could
> start being worked on for merge (there is at least one bug-report on it by
> "roland" on a strange but valid test-case).
>
>>I thought I'd give 2.6.9-bb3 a shot, and I had a bunch of users report the
>>accumiluation of zombies.
>
> What the hell! This should *not* happen. What did we (I mean developers) do?
Don't know, on my 2.6.9-skas3-v7 host it doesn't happen. And on the
2.6.7-skas3-v7 it doesn't also. Not in SKAS and not in TT.
By the way: why should SKAS produce lots of zombies? Maybe in TT we could have
missed something, but in SKAS I can't see any possible reason.
>
>>This is on a 2.6.7 host, using your
>>host-skas3-2.6.7-v1 patch.
>
> ??? There is a ton of bugs in that patch (see the changelogs on my site).
> Actually, it's likely you're not going to be hurt by them if you're using
> NPTL on the host (it is *hard* to get them in that case), but you'd better
> upgrade.
>
> I have been releasing tons of updates, and even too quickly (most times I was
> not introducing new bugs, just fixing the ones I and Bodo Stroesser was
> finding) but -V1 is "a bit" old. That said, this should not happen.
>
>>I'm not sure if it happened on newer host
>>kernels, since I had to revert it back quickly.
>
>
> Back to what? To 2.6.9-bb2 or to 2.6.9?
>
> I'll check in these days what's going on... it was hyper-tested on 2.6.9 host
> (not by me, but I trust Bodo Stroesser normally), but not on other hosts, I
> guess.
I tested on 2.6.9 and 2.6.7, but not on 2.4. For info: my latest patch, as
included in bb3, changes more than necessary. I tested, which parts of the
previous patches from BlaisorBlade and me really are needed. But the tests were
wrong, since I missed the fact, that make doesn't work correctly for USER_OBJS.
Thus I had a inconsistent kernel and saw UML exiting in boot. Since UML accidentally
worked fine after inserting a further part of my changes, I incorrectly assumed
this part to be needed.
Thus, Jeff posted a new, further stripped patch with the subject
"The current fix-kill patch". It has to be applied on top of BlaisorBlade's
"uml-hang-on-2.6.9-host.patch". I have tested this again, with hosts
2.6.7-skas3-v7 and 2.6.9-skas3-v7 in SKAS and TT. For me it works fine.
Bodo
-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://productguide.itmanagersjournal.com/
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel
^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [uml-devel] Re: 2.6.9-bb3 creates zombie threads in SKAS
2004-11-22 22:46 ` [uml-devel] Re: 2.6.9-bb3 creates zombie threads in SKAS Blaisorblade
2004-11-22 23:33 ` Christopher S. Aker
2004-11-23 14:48 ` [uml-devel] " Bodo Stroesser
@ 2004-11-23 17:49 ` Christopher S. Aker
2004-11-23 17:53 ` dopez
` (2 more replies)
2 siblings, 3 replies; 14+ messages in thread
From: Christopher S. Aker @ 2004-11-23 17:49 UTC (permalink / raw)
To: Blaisorblade; +Cc: user-mode-linux-devel, Bodo Stroesser
> On Monday 22 November 2004 23:23, Christopher S. Aker wrote:
> > Hello,
> Btw, question: what about your I/O token limiter? Actively using it? Could
> start being worked on for merge (there is at least one bug-report on it by
> "roland" on a strange but valid test-case).
> > I thought I'd give 2.6.9-bb3 a shot, and I had a bunch of users report the
> > accumiluation of zombies.
> What the hell! This should *not* happen. What did we (I mean developers) do?
Just to be clear, these were zombies *inside* UML, not on the host. I'm happy to
blame it on the older SKAS version.
-Chris
-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://productguide.itmanagersjournal.com/
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel
^ permalink raw reply [flat|nested] 14+ messages in thread* Re: [uml-devel] Re: 2.6.9-bb3 creates zombie threads in SKAS
2004-11-23 17:49 ` Christopher S. Aker
@ 2004-11-23 17:53 ` dopez
2004-11-23 19:30 ` Blaisorblade
2004-11-23 18:22 ` Blaisorblade
2004-11-30 18:55 ` Christopher S. Aker
2 siblings, 1 reply; 14+ messages in thread
From: dopez @ 2004-11-23 17:53 UTC (permalink / raw)
To: user-mode-linux-devel
On Tuesday 23 November 2004 18:49, Christopher S. Aker wrote:
> > What the hell! This should *not* happen. What did we (I mean developers)
> > do?
>
> Just to be clear, these were zombies *inside* UML, not on the host. I'm
> happy to blame it on the older SKAS version.
i am running 2.6.9-skas3-v7 with 2.6.9-bb3 and i too am seeing zombies in skas
mode, inside the UML
i also saw this with 2.6.9-bb2 btw, is there anything usefull i can send to
the list ? (kernel configs, dmesg?)
fwiw, running debian sarge on both the host and uml's
-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://productguide.itmanagersjournal.com/
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel
^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [uml-devel] Re: 2.6.9-bb3 creates zombie threads in SKAS
2004-11-23 17:53 ` dopez
@ 2004-11-23 19:30 ` Blaisorblade
0 siblings, 0 replies; 14+ messages in thread
From: Blaisorblade @ 2004-11-23 19:30 UTC (permalink / raw)
To: user-mode-linux-devel; +Cc: dopez
On Tuesday 23 November 2004 18:53, dopez wrote:
> On Tuesday 23 November 2004 18:49, Christopher S. Aker wrote:
> > > What the hell! This should *not* happen. What did we (I mean
> > > developers) do?
> >
> > Just to be clear, these were zombies *inside* UML, not on the host. I'm
> > happy to blame it on the older SKAS version.
> i am running 2.6.9-skas3-v7 with 2.6.9-bb3 and i too am seeing zombies in
> skas mode, inside the UML
> i also saw this with 2.6.9-bb2 btw, is there anything usefull i can send to
> the list ? (kernel configs, dmesg?)
For now let me test this, if I cannot reproduce it then I'll ask you.
> fwiw, running debian sarge on both the host and uml's
--
Paolo Giarrusso, aka Blaisorblade
Linux registered user n. 292729
http://www.user-mode-linux.org/~blaisorblade
-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://productguide.itmanagersjournal.com/
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel
^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [uml-devel] Re: 2.6.9-bb3 creates zombie threads in SKAS
2004-11-23 17:49 ` Christopher S. Aker
2004-11-23 17:53 ` dopez
@ 2004-11-23 18:22 ` Blaisorblade
2004-11-23 18:29 ` Christopher S. Aker
2004-11-30 18:55 ` Christopher S. Aker
2 siblings, 1 reply; 14+ messages in thread
From: Blaisorblade @ 2004-11-23 18:22 UTC (permalink / raw)
To: Christopher S. Aker; +Cc: user-mode-linux-devel, Bodo Stroesser
On Tuesday 23 November 2004 18:49, Christopher S. Aker wrote:
> > On Monday 22 November 2004 23:23, Christopher S. Aker wrote:
> > > Hello,
> >
> > Btw, question: what about your I/O token limiter? Actively using it?
> > Could start being worked on for merge (there is at least one bug-report
> > on it by "roland" on a strange but valid test-case).
> >
> > > I thought I'd give 2.6.9-bb3 a shot, and I had a bunch of users report
> > > the accumiluation of zombies.
> >
> > What the hell! This should *not* happen. What did we (I mean developers)
> > do?
>
> Just to be clear, these were zombies *inside* UML, not on the host.
*Inside*? Well, this is *VERY* different from what we were understanding. And
nobody checked something like that...
> I'm
> happy to blame it on the older SKAS version.
I'm not, unless your sentence mean that you have a test server on which you
verified this is true... and even in that case, it can mean that this bug
shows on hosts not supporting SYSEMU (like SKAS3/2.6-v1 hosts). Which would
be anyway a bug.
And UML *should* be happy even without SYSEMU (if not for anything else,
because you can pass "nosysemu" to it).
--
Paolo Giarrusso, aka Blaisorblade
Linux registered user n. 292729
http://www.user-mode-linux.org/~blaisorblade
-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://productguide.itmanagersjournal.com/
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel
^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [uml-devel] Re: 2.6.9-bb3 creates zombie threads in SKAS
2004-11-23 18:22 ` Blaisorblade
@ 2004-11-23 18:29 ` Christopher S. Aker
0 siblings, 0 replies; 14+ messages in thread
From: Christopher S. Aker @ 2004-11-23 18:29 UTC (permalink / raw)
To: Blaisorblade; +Cc: user-mode-linux-devel, Bodo Stroesser
> > I'm
> > happy to blame it on the older SKAS version.
> I'm not, unless your sentence mean that you have a test server on which you
> verified this is true...
Nope -- I made an assumption, given that I have what could be seen as similar
problems with this host kernel (mysql segfaults on other UML kernels, but it zombies
under -bb3)... I have not tested -bb3 under the latest skas revision. But, see the
other post from "dopez".
> and even in that case, it can mean that this bug
> shows on hosts not supporting SYSEMU (like SKAS3/2.6-v1 hosts). Which would
> be anyway a bug.
The host kernel in question did have the SYSEMU patches.
-Chris
-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://productguide.itmanagersjournal.com/
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel
^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [uml-devel] Re: 2.6.9-bb3 creates zombie threads in SKAS
2004-11-23 17:49 ` Christopher S. Aker
2004-11-23 17:53 ` dopez
2004-11-23 18:22 ` Blaisorblade
@ 2004-11-30 18:55 ` Christopher S. Aker
2004-11-30 19:23 ` Bodo Stroesser
2 siblings, 1 reply; 14+ messages in thread
From: Christopher S. Aker @ 2004-11-30 18:55 UTC (permalink / raw)
To: user-mode-linux-devel; +Cc: Bodo Stroesser, Blaisorblade, dopez
I've been able to narrow this down to the "signal-delivery" patch. 2.6.10-rc2-mm3
without the patch is fine, whereas adding the signal-delivery patch and you'll get
mysql zombies inside the UML. This happens on a variety of hosts, including
2.6.10-rc2-bk7 + skas-v7.
I just noticed 2.6.10-rc2-mm4 is out, and the patch there is named
"uml-redo-the-signal-delivery-mechanism.patch".
-Chris
-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://productguide.itmanagersjournal.com/
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel
^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [uml-devel] Re: 2.6.9-bb3 creates zombie threads in SKAS
2004-11-30 18:55 ` Christopher S. Aker
@ 2004-11-30 19:23 ` Bodo Stroesser
2004-11-30 20:59 ` Christopher S. Aker
0 siblings, 1 reply; 14+ messages in thread
From: Bodo Stroesser @ 2004-11-30 19:23 UTC (permalink / raw)
To: Christopher S. Aker; +Cc: user-mode-linux-devel, Blaisorblade, dopez
[-- Attachment #1: Type: text/plain, Size: 632 bytes --]
Christopher S. Aker wrote:
> I've been able to narrow this down to the "signal-delivery" patch. 2.6.10-rc2-mm3
> without the patch is fine, whereas adding the signal-delivery patch and you'll get
> mysql zombies inside the UML. This happens on a variety of hosts, including
> 2.6.10-rc2-bk7 + skas-v7.
>
> I just noticed 2.6.10-rc2-mm4 is out, and the patch there is named
> "uml-redo-the-signal-delivery-mechanism.patch".
>
> -Chris
>
Have found a bug in the signal-delivery patch a few days ago.
Could you please try the appended patch? (It's relevant for usage of RTSIGs
only, don't know, whether this could help you)
Bodo
[-- Attachment #2: extramask.patch --]
[-- Type: text/x-diff, Size: 1287 bytes --]
From: Bodo Stroesser <bstroesser@fujitsu-siemens.com>
Restoring of current->blocked in sys_sigreturn is wrong.
The first (long ) of the mask correctly is fetched from sc->oldmask.
The further longs again come from there, but correctly should be
taken from frame->extramask.
Signed-off-by: Bodo Stroesser <bstroesser@fujitsu-siemens.com>
---
diff -puN arch/um/sys-i386/signal.c~extramask arch/um/sys-i386/signal.c
--- linux-2.6.10-rc2/arch/um/sys-i386/signal.c~extramask 2004-11-26 14:03:10.770096974 +0100
+++ linux-2.6.10-rc2-root/arch/um/sys-i386/signal.c 2004-11-26 14:11:03.264403385 +0100
@@ -307,11 +307,12 @@ long sys_sigreturn(struct pt_regs regs)
struct sigframe __user *frame = (struct sigframe *)(sp - 8);
sigset_t set;
struct sigcontext __user *sc = &frame->sc;
- unsigned long __user *mask = &sc->oldmask;
+ unsigned long __user *oldmask = &sc->oldmask;
+ unsigned long __user *extramask = &frame->extramask;
int sig_size = (_NSIG_WORDS - 1) * sizeof(unsigned long);
- if(copy_from_user(&set.sig[0], mask, sizeof(&set.sig[0])) ||
- copy_from_user(&set.sig[1], mask, sig_size))
+ if(copy_from_user(&set.sig[0], oldmask, sizeof(&set.sig[0])) ||
+ copy_from_user(&set.sig[1], extramask, sig_size))
goto segfault;
sigdelsetmask(&set, ~_BLOCKABLE);
_
^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [uml-devel] Re: 2.6.9-bb3 creates zombie threads in SKAS
2004-11-30 19:23 ` Bodo Stroesser
@ 2004-11-30 20:59 ` Christopher S. Aker
0 siblings, 0 replies; 14+ messages in thread
From: Christopher S. Aker @ 2004-11-30 20:59 UTC (permalink / raw)
To: Bodo Stroesser; +Cc: user-mode-linux-devel, Blaisorblade, dopez
> Christopher S. Aker wrote:
> > I've been able to narrow this down to the "signal-delivery" patch.
2.6.10-rc2-mm3
> > without the patch is fine, whereas adding the signal-delivery patch and you'll
get
> > mysql zombies inside the UML. This happens on a variety of hosts, including
> > 2.6.10-rc2-bk7 + skas-v7.
> >
> > I just noticed 2.6.10-rc2-mm4 is out, and the patch there is named
> > "uml-redo-the-signal-delivery-mechanism.patch".
> >
> > -Chris
> >
> Have found a bug in the signal-delivery patch a few days ago.
> Could you please try the appended patch? (It's relevant for usage of RTSIGs
> only, don't know, whether this could help you)
That fixes it. Nice work!
-Chris
-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://productguide.itmanagersjournal.com/
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel
^ permalink raw reply [flat|nested] 14+ messages in thread