[uml-devel] Re: 2.6.14-bs1

[uml-devel] Re: 2.6.14-bs1

[Blaisorblade]

On Friday 04 November 2005 22:39, Antoine Martin wrote:
> Hi,

> I thought you might be interested in this regression:

> make ARCH=um SUBARCH=i386 defconfig
> make ARCH=um SUBARCH=i386 vmlinux
> (...)
>   CC      arch/um/kernel/initrd.o
>   CC      arch/um/kernel/skas/clone.o
> arch/um/kernel/skas/clone.c: In function `stub_clone_handler':
> arch/um/include/sysdep/stub.h:32: error: can't find a register in class
> `BREG' while reloading `asm'
> arch/um/include/sysdep/stub.h:53: error: can't find a register in class
> `BREG' while reloading `asm'
> arch/um/include/sysdep/stub.h:42: error: can't find a register in class
> `BREG' while reloading `asm'
> arch/um/include/sysdep/stub.h:75: error: can't find a register in class
> `BREG' while reloading `asm'
> make[2]: *** [arch/um/kernel/skas/clone.o] Error 1
> make[1]: *** [arch/um/kernel/skas] Error 2
> make: *** [arch/um/kernel] Error 2

Ok, today I've had time to look at the compile error, and while they are 
"legitimate" (aka you didn't do anything wrong, and I don't think it can 
be .config-dependant), they're still totally bogus (aka GCC is drunk).

And what's going on is possibly related to the hardening patches both you 
Gentoo GCC (3.4.4) says to have and that the FC4 GCC (4.0) I know to have.

In fact, I compile with 3.4.4 and it works; plus, the source code is IMHO 
legitimate.

However, I just remembered I have an Ubuntu with GCC 4.0... so I could test 
the thing.
> I've had quite a few problems with 2.6.14 not booting on 2.6.13 and
> 2.6.14 hosts. (errors have been reported by other users - unfortunately
> the strace isn't very useful)

I've not looked deeply at the strace, but I don't see anything much clear.

Also, I suppose the problem is with SKAS0 and no regression is seen in other 
modes.

However, given that you're having this compilation problem (caused by 
uml-fix-misassembling-skas0-stub-segv, which you can unapply for testing), 
you could test the 32-bit binaries I've built and uploaded on my homepage.

For at least one people they run better than home-compiled binaries, and 
there's no report of the opposite.

Also, there is at least another GCC-triggered bug in SKAS0 sources, so if 
you're hitting this one then the binaries will fix it.

> Antoine

-- 
Inform me of my mistakes, so I can keep imitating Homer Simpson's "Doh!".
Paolo Giarrusso, aka Blaisorblade (Skype ID "PaoloGiarrusso", ICQ 215621894)
http://www.user-mode-linux.org/~blaisorblade

	

	
		
___________________________________ 
Yahoo! Mail: gratis 1GB per i messaggi e allegati da 10MB 
http://mail.yahoo.it



-------------------------------------------------------
SF.Net email is sponsored by:
Tame your development challenges with Apache's Geronimo App Server. Download
it for free - -and be entered to win a 42" plasma tv or your very own
Sony(tm)PSP.  Click here to play: http://sourceforge.net/geronimo.php
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel

[uml-devel] Re: 2.6.14-bs1

[Antoine Martin]

> However, I just remembered I have an Ubuntu with GCC 4.0... so I could test 
> the thing.
Or you could use a gentoo chroot. (if you leave it to build overnight..)

> > I've had quite a few problems with 2.6.14 not booting on 2.6.13 and
> > 2.6.14 hosts. (errors have been reported by other users - unfortunately
> > the strace isn't very useful)
> 
> I've not looked deeply at the strace, but I don't see anything much clear.
> 
> Also, I suppose the problem is with SKAS0 and no regression is seen in other 
> modes.
> 
> However, given that you're having this compilation problem (caused by 
> uml-fix-misassembling-skas0-stub-segv, which you can unapply for testing), 
I tried removing this patch (and its sibling ^segv^) but patch -R moaned
at me, and after unmerging it by hand I still got the same error so I
gave up (for now anyway)

> you could test the 32-bit binaries I've built and uploaded on my homepage.
I couldn't find the link (I didn't look very long).

> For at least one people they run better than home-compiled binaries, and 
> there's no report of the opposite.
I'll try that.

I'm currently working on those root_fs images, coming along nicely but
it takes *a lot* of time. I should have something ready within a few
days.

Antoine



-------------------------------------------------------
SF.Net email is sponsored by:
Tame your development challenges with Apache's Geronimo App Server. Download
it for free - -and be entered to win a 42" plasma tv or your very own
Sony(tm)PSP.  Click here to play: http://sourceforge.net/geronimo.php
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel

[uml-devel] Re: 2.6.14-bs1

[Blaisorblade]

On Friday 11 November 2005 21:49, Antoine Martin wrote:
> > However, I just remembered I have an Ubuntu with GCC 4.0... so I could
> > test the thing.

> Or you could use a gentoo chroot. (if you leave it to build overnight..)
I have Gentoo installed, no problem... it's that actually they don't support 
4.0 yet and I agree with them.

However, I _did_ test with that GCC and things work ok. So it's definitely the 
hardening patches.

However, I _can_ test them:

gentoo # gcc-config -l
 [1] x86_64-pc-linux-gnu-3.4.4 *
 [2] x86_64-pc-linux-gnu-3.4.4-hardened
 [3] x86_64-pc-linux-gnu-3.4.4-hardenednopie
 [4] x86_64-pc-linux-gnu-3.4.4-hardenednopiessp
 [5] x86_64-pc-linux-gnu-3.4.4-hardenednossp

I will test the whole thing tomorrow.

> > > I've had quite a few problems with 2.6.14 not booting on 2.6.13 and
> > > 2.6.14 hosts. (errors have been reported by other users - unfortunately
> > > the strace isn't very useful)
> >
> > I've not looked deeply at the strace, but I don't see anything much
> > clear.
> >
> > Also, I suppose the problem is with SKAS0 and no regression is seen in
> > other modes.
> >
> > However, given that you're having this compilation problem (caused by
> > uml-fix-misassembling-skas0-stub-segv, which you can unapply for
> > testing),
>
> I tried removing this patch (and its sibling ^segv^) but patch -R moaned
> at me, 

Strange - the proper order is remove "-segv" and then (if needed but it 
shouldn't) remove the other - the proper patch order is documented in the 
"series" file.

Indeed, I missed that you should also remove (first)
patches/uml-fix-64compile-misassembling-skas0-stub-segv

Using quilt makes that fast, however:

*) untar the broken-out tarball

*) mv the directory to "patches

*) $ quilt pop patches/uml-fix-misassembling-skas0-stub-segv

*) (comment the patch out in patches/series)

*) $ quilt push bs-extraversion-bs1 (or quilt push -a, should be the same).

> and after unmerging it by hand I still got the same error so I 
> gave up (for now anyway)

-- 
Inform me of my mistakes, so I can keep imitating Homer Simpson's "Doh!".
Paolo Giarrusso, aka Blaisorblade (Skype ID "PaoloGiarrusso", ICQ 215621894)
http://www.user-mode-linux.org/~blaisorblade

		
___________________________________ 
Yahoo! Messenger: chiamate gratuite in tutto il mondo 
http://it.messenger.yahoo.com



-------------------------------------------------------
SF.Net email is sponsored by:
Tame your development challenges with Apache's Geronimo App Server. Download
it for free - -and be entered to win a 42" plasma tv or your very own
Sony(tm)PSP.  Click here to play: http://sourceforge.net/geronimo.php
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel

[uml-devel] Re: 2.6.14-bs1

[Antoine Martin]

> However, I _did_ test with that GCC and things work ok. So it's definitely the 
> hardening patches.
Probably.

> However, I _can_ test them:
> gentoo # gcc-config -l
>  [1] x86_64-pc-linux-gnu-3.4.4 *
>  [2] x86_64-pc-linux-gnu-3.4.4-hardened
>  [3] x86_64-pc-linux-gnu-3.4.4-hardenednopie
>  [4] x86_64-pc-linux-gnu-3.4.4-hardenednopiessp
>  [5] x86_64-pc-linux-gnu-3.4.4-hardenednossp
I've got similar settings - on a x86 box this time:
# gcc-config -l
 [1] i686-pc-linux-gnu-3.3.6 *
 [2] i686-pc-linux-gnu-3.3.6-hardenednopie
 [3] i686-pc-linux-gnu-3.3.6-hardenednopiessp
 [4] i686-pc-linux-gnu-3.3.6-hardenednossp
 [5] i686-pc-linux-gnu-3.3.6-vanilla

 CC      arch/um/kernel/skas/clone.o
arch/um/kernel/skas/clone.c: In function `stub_clone_handler':
arch/um/include/sysdep/stub.h:32: error: can't find a register in class
`BREG' while reloading `asm'
arch/um/include/sysdep/stub.h:53: error: can't find a register in class
`BREG' while reloading `asm'
arch/um/include/sysdep/stub.h:42: error: can't find a register in class
`BREG' while reloading `asm'
arch/um/include/sysdep/stub.h:75: error: can't find a register in class
`BREG' while reloading `asm'
make[2]: *** [arch/um/kernel/skas/clone.o] Error 1
make[1]: *** [arch/um/kernel/skas] Error 2
make: *** [arch/um/kernel] Error 2

So I built a 2.6.14 instead. Then tried running in skas0:

It ended up spinning at 100% cpu usage (and eventually detecting it):
[42949375.230000] VFS: Mounted root (reiserfs filesystem) readonly.
[42949384.260000] BUG: soft lockup detected on CPU#0!
[42949384.260000]
[42949384.260000] EIP: 0073:[<a02554c1>] CPU: 0 Not tainted ESP: 007b:b7f75f6c EFLAGS: 00000292
[42949384.260000]     Not tainted
[42949384.260000] EAX: 00000000 EBX: 00002136 ECX: 00000013 EDX: a035d364
[42949384.260000] ESI: a0437940 EDI: 00002132 EBP: b7f75f98 DS: 007b ES: 007b
[42949384.260000] a10a6f70:  [<a00501d4>] show_regs+0x214/0x220
[42949384.260000] a10a6fa0:  [<a00785c7>] softlockup_tick+0x57/0x60
[42949384.260000] a10a6fc0:  [<a0062037>] do_timer+0x47/0xd0
[42949384.260000] a10a6fd0:  [<a001bee4>] um_timer+0x14/0x50
[42949384.260000] a10a6fe0:  [<a0078763>] handle_IRQ_event+0x33/0x80
[42949384.260000] a10a7010:  [<a0078805>] __do_IRQ+0x55/0xb0
[42949384.260000] a10a7040:  [<a00152f0>] do_IRQ+0x30/0x40
[42949384.260000] a10a7050:  [<a001be33>] timer_irq+0x113/0x170
[42949384.260000] a10a7080:  [<a001c1f0>] timer_handler+0x70/0x90
[42949384.260000] a10a70a0:  [<a0025254>] sig_handler_common_skas+0xc4/0x150
[42949384.260000] a10a70f0:  [<a00496a4>] alarm_handler+0xa4/0xc0
[42949384.260000] a10a7120:  [<ffffe420>] _etext+0x5fd5795d/0x0
[42949384.260000] a10a7428:  [<a001ca6c>] flush_tlb_kernel_range_common+0xbc/0x160
[42949384.260000] a10a7458:  [<a001cc8e>] flush_tlb_kernel_vm+0x2e/0x30
[42949384.260000] a10a7468:  [<a001d261>] segv+0x281/0x2d0
[42949384.260000] a10a7558:  [<a001d63c>] segv_handler+0xfc/0x150
[42949384.260000] a10a75c8:  [<a0025254>] sig_handler_common_skas+0xc4/0x150
[42949384.260000] a10a7618:  [<a00495f6>] sig_handler+0x76/0x80
[42949384.260000] a10a7648:  [<ffffe420>] _etext+0x5fd5795d/0x0
[42949384.260000] a10a79a0:  [<a00c86b6>] load_elf_binary+0x6c6/0x1040
[42949384.260000] a10a7af0:  [<a00a5fe3>] search_binary_handler+0x63/0xf0
[42949384.260000] a10a7b20:  [<a00a6202>] do_execve+0x192/0x260
[42949384.260000] a10a7b50:  [<a0014908>] execve1+0x38/0x80
[42949384.260000] a10a7b80:  [<a0014972>] um_execve+0x22/0x60
[42949384.260000] a10a7ba0:  [<a00141bc>] run_init_process+0x4c/0x80
[42949384.260000] a10a7bd0:  [<a00142b8>] init+0xc8/0x170
[42949384.260000] a10a7bf0:  [<a00494ff>] run_kernel_thread+0x9f/0xb0
[42949384.260000] a10a7ce0:  [<a00248c3>] new_thread_handler+0xc3/0x120
[42949384.260000] a10a7d20:  [<ffffe420>] _etext+0x5fd5795d/0x0
[42949384.260000]
After that I tried reducing the guest ram, the clue was seeing this message:
Kernel virtual memory size shrunk to 258998272 bytes
It continuously got stuck on:
[42949375.540000] Kernel panic - not syncing: copy_context_skas0 : failed to wait for SIGUSR1/SIGTRAP, pid = 8484, n = 8484, errno = 0, status = 0xb7f

skas3 works absolutely  fine.
The only problem I've got with skas3 is that it requires /proc/mm in
the chroot and the only way that I found to do this is to mount it in
there. (SELinux will need some work to allow me to restrict write
access to just /proc/mm and not the whole /proc...)

Antoine



-------------------------------------------------------
This SF.Net email is sponsored by the JBoss Inc.  Get Certified Today
Register for a JBoss Training Course.  Free Certification Exam
for All Training Attendees Through End of 2005. For more info visit:
http://ads.osdn.com/?ad_id=7628&alloc_id=16845&op=click
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel

[uml-devel] Re: 2.6.14-bs1

[Blaisorblade]

On Wednesday 16 November 2005 00:57, Antoine Martin wrote:

> skas3 works absolutely  fine.
> The only problem I've got with skas3 is that it requires /proc/mm in
> the chroot and the only way that I found to do this is to mount it in
> there. (SELinux will need some work to allow me to restrict write
> access to just /proc/mm and not the whole /proc...)
For the rest I have no definite answer yet, BUT:

mkdir /chroot/proc
touch /chroot/proc/mm
mount --bind /proc/mm /chroot/proc/mm

is reported to work (and I tested it once). On 2.4 kernels too. Surprise 
surprise :-) !

Bye
-- 
Inform me of my mistakes, so I can keep imitating Homer Simpson's "Doh!".
Paolo Giarrusso, aka Blaisorblade (Skype ID "PaoloGiarrusso", ICQ 215621894)
http://www.user-mode-linux.org/~blaisorblade

		
___________________________________ 
Yahoo! Messenger: chiamate gratuite in tutto il mondo 
http://it.messenger.yahoo.com



-------------------------------------------------------
This SF.Net email is sponsored by the JBoss Inc.  Get Certified Today
Register for a JBoss Training Course.  Free Certification Exam
for All Training Attendees Through End of 2005. For more info visit:
http://ads.osdn.com/?ad_id=7628&alloc_id=16845&op=click
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel