Re: CFT/HEADS-UP: linux 2.6.16 emulation

On Mon, 29 Jan 2007 10:10:06 +0100 Divacky Roman wrote:
> On Mon, Jan 29, 2007 at 02:22:53AM +0300, Boris Samorodov wrote:
> > On Mon, 22 Jan 2007 17:49:26 +0100 Divacky Roman wrote:
> > > On Sat, Jan 20, 2007 at 05:07:23PM +0100, Alexander Leidinger wrote:
> > 
> > > > today I committed the last fixes for the showstopper problems (panics)
> > > > in the linux 2.6.16 emulation. I intend to switch the default version
> > > > to 2.6.16 on i386 "soon" (see below), so please help testing it.
> > 
> > > to be more precise.. we want testing on -current on i386...
> > > any other report is useles. except for reports from p4 linuxulator
> > > branch on 2.6/amd64.. 
> > 
> > Here are some test results (for 2.6.16 linux emulation) at:
> > $ uname -a
> > FreeBSD bsam.ru 7.0-CURRENT FreeBSD 7.0-CURRENT #0: Wed Jan 24 01:58:12 MSK 2007     bsam_at_bsam.ru:/usr/obj/usr/src/sys/GENERIC  i386
> > 
> > ...and linux_base-fc6 (a port from my repository), other port
> > dependencies are from FC4:
> > 
> > 1. Just work (sometimes messages about an unimplemented syscall(s) do
> > appear):
> >    - print/acroread7
> >    - www/linux-opera
> >    - www/linux-firefox
> >    - www/linux-flashplugin7
> >    - mail/linux-thunderbird
> >    - multimedia/linux-realplayer

> please post those messages about unimplemented syscalls, thnx

So far I have this one:
-----
Jan 29 17:49:18 bsam kernel: linux: pid 892 (pwd): syscall openat not implemented
-----

> > 2. Works but coredumps when exitting:
> >    - net/skype

> cant reproduce that. I have proceses left when I "click on exit in skype" but I guess
> its normal as skype wants to be able to answer a phone call.

There is no coredump with linux_base-fc4_9 -- I got it with
linux_base-fc6.

> > 3. Is not executed (some new... or old libraries are needed):
> >    - www/linux-mozilla.

> but this doesnt relate to 2.6 emulation, right?

Yes. And this port is OK with linux_base-fc4_9.

> > Sometimes (not stable result) I get the following LOR. This one I got
> > while testing net/skype:
[LOR deletted]
> I'll investigate that. thnx

Just one LOR more (but the same bug?):
-----
Jan 29 17:38:53 bsam kernel: lock order reversal:
Jan 29 17:38:53 bsam kernel: 1st 0xc0bd58a0 emuldata->shared lock (emuldata->shared lock) _at_ /usr/src/sys/modules/linux/../../compat/linux/linux_emul.c:117
Jan 29 17:38:53 bsam kernel: 2nd 0xc0bd58e0 emuldata lock (emuldata lock) _at_ /usr/src/sys/modules/linux/../../compat/linux/linux_emul.c:67
Jan 29 17:38:53 bsam kernel: KDB: stack backtrace:
Jan 29 17:38:53 bsam kernel: db_trace_self_wrapper(c09442e5) at db_trace_self_wrapper+0x25
Jan 29 17:38:53 bsam kernel: kdb_backtrace(0,ffffffff,c0a525d0,c0a525f8,c09ed16c,...) at kdb_backtrace+0x29
Jan 29 17:38:53 bsam kernel: witness_checkorder(c0bd58e0,9,c0bd2066,43) at witness_checkorder+0x586
Jan 29 17:38:53 bsam kernel: _sx_xlock(c0bd58e0,c0bd2066,43,c4597240,312,...) at _sx_xlock+0x50
Jan 29 17:38:53 bsam kernel: em_find(c472c900,1,c0bd58a0,c0bd2066,75,...) at em_find+0x51
Jan 29 17:38:53 bsam kernel: linux_proc_init(c45e8510,312,3d0f00,0,16,...) at linux_proc_init+0xbd
Jan 29 17:38:53 bsam kernel: linux_clone(c45e8510,e6946d00) at linux_clone+0x8b
Jan 29 17:38:53 bsam kernel: syscall(e6946d38) at syscall+0x256
Jan 29 17:38:53 bsam kernel: Xint0x80_syscall() at Xint0x80_syscall+0x20
Jan 29 17:38:53 bsam kernel: --- syscall (758682804), eip = 0x2, esp = 0xe6946d74, ebp = 0xbfbfd304 ---
-----


WBR
-- 
Boris Samorodov (bsam)
Research Engineer, http://www.ipt.ru Telephone & Internet SP
FreeBSD committer, http://www.FreeBSD.org The Power To Serve