> It results in a leap forward, but still falls short :) > I did need the VGA BIOS (0x277FFF) which results in > X starting up(ish) but displaying a black screen with > a white square near the middle, then stalling. > A ctrl-C in the starting terminal causes X to exit, > with the error: > -------------- copy of end of X log starts -------------- > Error in I810WaitLpRing(), now is 1534883518, start is 1534881517 > pgetbl_ctl: 0x0 pgetbl_err: 0x10 > ipeir: 0 iphdr: 0 > LP ring tail: 40 head: 0 len: f001 start 342000 > eir: 0 esr: 10 emr: 3d > instdone: ff7a instpm: 0 > memmode: 4 instps: c10 > hwstam: 9ac7 ier: 0 imr: 9ac7 iir: 0 > space: 65464 wanted 65528 > > Fatal server error: > lockupThis is pretty weird. I wander if its relying on an interrupt that isn''t getting plumbed through? Is there any mention of an interrupt in the X startup logs? Please could you post these. [In fact, how does the X server get hold an interrupt from user space (hence enabling it to block)? ] Does the i810 driver cause any nasty kernel modules to be loaded such as agpgart or something? The domU would need to be given access rights to the GART, but I can''t immediately think why it wouldn''t work. Obviously, the GART couldn''t be shared between domU''s. One other option might be to try and use the fbdev driver in the kernel (and X''s fbdev module) rather than X''s i810 driver. Thanks, Ian _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Ian Pratt wrote:>> It results in a leap forward, but still falls short :) >> I did need the VGA BIOS (0x277FFF) which results in >> X starting up(ish) but displaying a black screen with >> a white square near the middle, then stalling. >> A ctrl-C in the starting terminal causes X to exit, >> with the error: >> -------------- copy of end of X log starts -------------- >> Error in I810WaitLpRing(), now is 1534883518, start is 1534881517 >> pgetbl_ctl: 0x0 pgetbl_err: 0x10 >> ipeir: 0 iphdr: 0 >> LP ring tail: 40 head: 0 len: f001 start 342000 >> eir: 0 esr: 10 emr: 3d >> instdone: ff7a instpm: 0 >> memmode: 4 instps: c10 >> hwstam: 9ac7 ier: 0 imr: 9ac7 iir: 0 >> space: 65464 wanted 65528 >> >> Fatal server error: >> lockup > > > This is pretty weird. I wander if its relying on an interrupt that isn''t > getting plumbed through? Is there any mention of an interrupt in the X > startup logs? Please could you post these.I''ve attached the Xorg.0.log file> [In fact, how does the X server get hold an interrupt from user space > (hence enabling it to block)? ]Pass.> Does the i810 driver cause any nasty kernel modules to be loaded such as > agpgart or something? The domU would need to be given access rights to > the GART, but I can''t immediately think why it wouldn''t work. Obviously, > the GART couldn''t be shared between domU''s.opac1a:~# lsmod Module Size Used by i810 24768 0 drm 76788 1 i810 ipv6 277568 10 dm_snapshot 19200 0 dm_mirror 24368 0 dm_mod 61272 2 dm_snapshot,dm_mirror intel_agp 24764 1 agpgart 37960 2 drm,intel_agp ohci_hcd 22500 0 uhci_hcd 35440 0 genrtc 11008 0 So yes, agpgart gets loaded. Do I need to setup something for the DomU to get access to it? (I''m somewhat hazy on what, exactly, needs to be done where :) Sharing shouldn''t be a problem as (I understand) agpgart only applies to AGP cards and any/all additional video cards will be PCI.> One other option might be to try and use the fbdev driver in the kernel > (and X''s fbdev module) rather than X''s i810 driver.I''ll see how that goes. -- Julian Davison Note: 1) This may have come from an address @cbhs.school.nz but isn''t necessarily the (or even an) official view of Christchurch Boys'' High School 2) While replying to this address may get into my mailbox it will almost certainly be filtered into a mailing list folder. To actually reach actual me, strip off the bit after the ''-'' in the name. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
> Ian Pratt wrote: > >> It results in a leap forward, but still falls short :) > >> I did need the VGA BIOS (0x277FFF) which results in > >> X starting up(ish) but displaying a black screen with > >> a white square near the middle, then stalling. > >> A ctrl-C in the starting terminal causes X to exit, > >> with the error: > >> -------------- copy of end of X log starts -------------- > >> Error in I810WaitLpRing(), now is 1534883518, start is 1534881517 > >> pgetbl_ctl: 0x0 pgetbl_err: 0x10 > >> ipeir: 0 iphdr: 0 > >> LP ring tail: 40 head: 0 len: f001 start 342000 > >> eir: 0 esr: 10 emr: 3d > >> instdone: ff7a instpm: 0 > >> memmode: 4 instps: c10 > >> hwstam: 9ac7 ier: 0 imr: 9ac7 iir: 0 > >> space: 65464 wanted 65528 > >> > >> Fatal server error: > >> lockup > > > > > > This is pretty weird. I wander if its relying on an interrupt thatisn''t> > getting plumbed through? Is there any mention of an interrupt in theX> > startup logs? Please could you post these. > > I''ve attached the Xorg.0.log fileIt might be helpful to diff the Xorg log file between the driver working successfully in dom0, and failing to work in domU. Thanks, Ian _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel