[Xen-users] Re: Xend crashes repeatedly starting DomU - please help...

[Milan Holzäpfel <listen@xxxxxxxx>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

> Hi, I have booted into a Dom0 kernel on a fairly stripped down machine
> just running the basics needed.  However, I haven't yet managed to
> get a DomU booting.  I constantly get the following error "Error:
> Received invalid response from Xend, twice." on trying to start the
> domain

I seem to have the same problem.  It's gentoo's uClibc-hardened on
x86_32 over here. (system with uClibc as its libc and security
enhancements.) 

Output from 

# strace -s 256 -f -o str /etc/init.d/xend start &

(output is put in the file called "str")

[...]
12893 bind(18, {sa_family=AF_FILE, path="/var/lib/xend/xend-socket"}, 27) = 0
12893 listen(18, 5)                     = 0
12893 accept(18,  <unfinished ...>
[...]
[ loads of old_mmap-ioctl-munmap-... calls from 12893 ]
12893 old_mmap(NULL, 4096, PROT_WRITE, MAP_SHARED, 8, 0) = 0xb75ad000
12893 ioctl(8, SNDCTL_DSP_SPEED or SOUND_PCM_READ_RATE, 0xbe9fbad8) = 0
12893 munmap(0xb75ad000, 4096)          = 0
12893 old_mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_SHARED, 8, 0) = 0xb75ad000
12893 ioctl(8, SNDCTL_DSP_SPEED or SOUND_PCM_READ_RATE, 0xbe9fbb08) = 0
12893 ioctl(8, SNDCTL_DSP_RESET, 0xbe9fbaf8) = 196608
12893 munmap(0xb75ad000, 4096)          = 0
12893 old_mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_SHARED, 8, 0) = 0xb75ad000
12893 ioctl(8, SNDCTL_DSP_SPEED or SOUND_PCM_READ_RATE, 0xbe9fbb08) = 0
12893 munmap(0xb75ad000, 4096)          = 0
12893 mlock(0xbe9fccb8, 136)            = 0
12893 ioctl(8, SNDCTL_DSP_RESET, 0xbe9fbaf8) = 0
12893 munlock(0xbe9fccb8, 136)          = 0
12893 munmap(0xb75ae000, 4251648)       = 0
12893 mlock(0xbe9fd8c8, 136)            = 0
12893 ioctl(8, SNDCTL_DSP_RESET, 0xbe9fbaf8) = 0
12893 munlock(0xbe9fd8c8, 136)          = 0
12893 rt_sigaction(SIGPIPE, {SIG_IGN}, {SIG_IGN}, 8) = 0
12893 write(16, "\10\0\0\0\0\0\0\0\0\0\0\0\v\0\0\0", 16) = 16
12893 write(16, "17\0", 3)              = 3
12893 write(16, "32451\0", 6)           = 6
12893 write(16, "1\0", 2 <unfinished ...>
12890 <... read resumed> "\10\0\0\0\0\0\0\0\0\0\0\0\3\0\0\0", 16) = 16
12893 <... write resumed> )             = 2
12890 read(16,  <unfinished ...>
12893 rt_sigprocmask(SIG_SETMASK, NULL,  <unfinished ...>
12890 <... read resumed> "OK\0", 3)     = 3
12893 <... rt_sigprocmask resumed> [RTMIN], 8) = 0
12890 kill(12893, SIGRTMIN <unfinished ...>
12893 rt_sigsuspend([] <unfinished ...>
12890 <... kill resumed> )              = 0
12893 --- SIGRTMIN (Unknown signal 32) @ 0 (0) ---
12890 read(16,  <unfinished ...>
12893 <... rt_sigsuspend resumed> )     = -1 EINTR (Interrupted system call)
12893 sigreturn()                       = ? (mask now [RTMIN])
12893 rt_sigaction(SIGPIPE, {SIG_IGN}, NULL, 8) = 0
12893 --- SIGSEGV (Segmentation fault) @ 0 (0) ---
12892 <... accept resumed> 0xbebfecc8, [16]) = ? ERESTARTSYS (To be restarted)
12887 <... accept resumed> 0xbf3ff190, [110]) = ? ERESTARTSYS (To be restarted)
12888 <... accept resumed> 0xbf1ff190, [16]) = ? ERESTARTSYS (To be restarted)
12886 <... poll resumed> [{fd=5, events=POLLIN}], 1, 2000) = -1 EINTR 
(Interrupted system call)
12890 <... read resumed> 0x800c6590, 16) = ? ERESTARTSYS (To be restarted)
12889 +++ killed by SIGKILL +++
12885 +++ killed by SIGKILL +++
12892 +++ killed by SIGKILL +++
12887 +++ killed by SIGKILL +++
12888 +++ killed by SIGKILL +++
12886 +++ killed by SIGKILL +++
12890 +++ killed by SIGKILL +++


Because no process sent the kill signal, I thougt it could possibly be
the stack-smashing-handler getting active.  The gentoo's glibc's one
printed a message when killing the process iirc, the gentoo's uClibc's
doesn't seem to do so, as xend doesn't get killed at this specific
location any longer after having recompiled Python with vanilla-gcc
(with no SSP, that is).   (On gentoo uclibc-hardened, you can verify
that SSP isn't used by a specific binary any longer by doing 
"readelf -a /usr/bin/python | grep stack_smash_handler", it doens't
show anything if SSP is off.) 


Now however xm says:

| Error: Device 2049 (vbd) could not be connected. Hotplug scripts not working. 

Yesssssssss...

Regards, 
Milan

(who is terribly happy that he can investigate in yet another error
message from xend.... arg.  I hate xend. haaaaaate.)

(PS: I didn't intend to provide useful information on the hotplug
issue I'm encountering, so don't spend time on it for now... but if you
happen to have some tips ;) ) 





- -- 
Milan Holzaepfel <mail(a)mjh(d)name>             <URL:http://mjh.name/>
pub  4096R/C790FC23  EB8E 5E81 81E3 53A9 9B74  B895 5179 54C0 C790 FC23 

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2.1 (GNU/Linux)
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=l248
-----END PGP SIGNATURE-----


_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users