<div><div class="gmail_quote">On Fri, Jun 24, 2011 at 11:46 AM, Leonardo E. Reiter <span dir="ltr"><<a href="mailto:lreiter@vbridges.com">lreiter@vbridges.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<div class="gmail_quote"><div class="im">On Fri, Jun 24, 2011 at 4:33 AM, Gerd Hoffmann <span dir="ltr"><<a href="mailto:kraxel@redhat.com" target="_blank">kraxel@redhat.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Hi,<div><br>
<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
you can see the 2nd QXL device at 00:06.0. I don't know why the first<br>
one says "VGA compatible" and the second one doesn't, but hopefully that<br>
is not an issue.<br>
</blockquote>
<br></div>
The second simply isn't matched by the qxl driver due to the different PCI class. Fixable (see attachent).<br></blockquote></div><div>Thanks... I will try this and keep testing... we are looking into the spicec crash as well, because this was happening for us even with Windows guests. But any more help with the this QXL issue will be much appreciated.</div>
</div></blockquote><div>Definitely confirmed the situation you explained... I had to remove the "Screen" config from the Device sections from the /etc/X11/xorg.conf, at which point I was able to start with 2 SPICE windows but blank. This is after applying your patch (using the 0.13.0 tag in git for the QXL driver). gdb'ing Xorg and getting a backtrace shows:</div>
<div><br></div><div>#0 qxl_ring_push (ring=0xe6c490, new_elt=0x7ffff4578850) at qxl_ring.c:73</div><div>#1 0x00007f5fa413ecd0 in push_surface_cmd (surface=0xed8c60)</div><div> at qxl_surface.c:435</div><div>#2 send_destroy (surface=0xed8c60) at qxl_surface.c:766</div>
<div>#3 0x00007f5fa413ce38 in qxl_destroy_pixmap (pixmap=0x1657670)</div><div> at qxl_driver.c:739</div><div>#4 0x00000000004cc19c in ?? ()</div><div>#5 0x00000000004ab0da in ?? ()</div><div>#6 0x00000000004bb998 in FreePicture ()</div>
<div>#7 0x0000000000455840 in FreeResource ()</div><div>#8 0x00000000004c2f3a in ?? ()</div><div>#9 0x00000000004c014d in ?? ()</div><div>#10 0x00000000004275dc in ?? ()</div><div>#11 0x000000000042208a in _start ()</div>
<div><br></div><div>I'll keep investigating but would appreciate any further help.</div><div><br></div><div>Thanks,</div><div>Leo Reiter</div><div><br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<div class="gmail_quote"><div><div></div><div class="h5">
<div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<br>
That alone doesn't get things going though, guess thats why it isn't supported (yet?) by the driver. I can get X to show the display on one or the other qxl device, but not with xinerama on both.<br>
<br>
Seems to be stuck here:<br>
<br>
0x00007f3f90620992 in qxl_ring_push (ring=0x126e4c0, new_elt=0x7fff74a3afa0) at qxl_ring.c:73<br>
73 while (header->prod - header->cons == header->num_items)<br>
(gdb) bt<br>
#0 0x00007f3f90620992 in qxl_ring_push (ring=0x126e4c0, new_elt=0x7fff74a3afa0) at qxl_ring.c:73<br>
#1 0x00007f3f9061ee5b in push_drawable (drawable=<optimized out>, qxl=<optimized out>)<br>
at qxl_surface.c:516<br>
#2 push_drawable (qxl=<optimized out>, drawable=<optimized out>) at qxl_surface.c:498<br>
#3 0x00007f3f906208d1 in qxl_surface_put_image (dest=0x12f5ce0, x=<optimized out>,<br>
y=<optimized out>, width=24, height=24, src=0x2058468 "", src_pitch=96) at qxl_surface.c:1465<br>
#4 0x00007f3f9062618d in uxa_do_put_image (src_stride=96, bits=0x2058468 "", format=2, h=24,<br>
w=24, y=<optimized out>, x=<optimized out>, depth=<optimized out>, pGC=0x1a2fe00,<br>
pDrawable=0x1a35910) at uxa-accel.c:329<br>
#5 uxa_put_image (pDrawable=0x1a35910, pGC=0x1a2fe00, depth=<optimized out>, x=<optimized out>,<br>
y=0, w=24, h=24, leftPad=0, format=2, bits=0x2058468 "") at uxa-accel.c:367<br>
#6 0x00000000004dc199 in ?? ()<br>
#7 0x000000000042b91d in ?? ()<br>
#8 0x000000000042ec11 in ?? ()<br>
#9 0x0000000000422e1a in ?? ()<br>
#10 0x00000033c1a2139d in __libc_start_main () from /lib64/libc.so.6<br>
#11 0x0000000000423101 in _start ()<br>
<br>
Could be a server side issue too (guest driver spinning, waiting for spice-server to take requests out of the ring). Also note that spicec crashed while the X-Server tried to init the two qxl devices in xinerama mode ...<br>
<br>
cheers,<br><font color="#888888">
Gerd<br></font></blockquote><div><br></div></div></div><div>Thanks,</div><div><br></div><div>- Leo Reiter </div></div><br>
</blockquote></div><br></div>