Vm won't start with vgt=1 i3-4130 #1
Comments
|
which platform are you using? XenGT only supports HSW formally now. From: slis [mailto:notifications@github.com] hello... i am looking to get xengt running but i cant start it... got this in dmesg [ 1.376813] vGT error:(vgt_set_device_type:432) Unknown chip 0x41e — |
|
what do u mean by HSW? in pdf guide it says only 4th gen intel processor is requirment? i changed source devtable.c or smth to include my chip number... now it almost works... i tried in ubuntu 12.04 and when i enable VGT=1 this happens.... but i cant get screen on output, only brief windows green loading bar / or 0x0000003B BSOD on SDL or VNC,.... nothing on native screen. thx here is dmesg: 33.223610] xenbr0: port 1(eth0) entered forwarding state thx |
|
last part here is first part |
|
yes, it’s Haswell the 4th processor. Looks your box is not included in current device list, so adding your device ID is right way to go, but it also means not validated so some trouble may be caused by hardware difference. Better to find a platform with device ID in existing list. I didn’t see below log in your full log in another mail, so not sure whether it has all the information there. By default a new VM is in the background after it’s created. So it may already boot but you didn’t note it. One way is to add a boot option “hvm_boot_foreground”, or use sysfs control interface to switch foreground/background VM. You can check setup guide for detail. From: slis [mailto:notifications@github.com] what do u mean by HSW? in pdf guide it says only 4th gen intel processor is requirment? i changed source devicemap.c or smth to include my chip number... now it almost works... i tried in ubuntu 12.04 and when i enable VGT=1 this happens.... but i cant get screen on output, only brief windows green loading bar on SDL or VNC,.... thx here is dmesg: 33.223610] xenbr0: port 1(eth0) entered forwarding state thx — |
|
thx for response... i have been struggling for 2 days...yeah... i tried with “hvm_boot_foreground”... or sysfs .... it says display not ready. and in xl list it says that vm is blocked. same on 12.04 and 14.04 here is full dmesg... to me this looks strange: [ 3.230753] vGT warning:(vgt_emulate_read:351) vGT: untracked MMIO read: vm_id(0), offset=0x651b4,len=4, val=0x0!!! so it only depends on processor/gpu? not mbo? still can't belive that my i3 is not on the list lol what processor do u suggest that works? i am looking for cheap desktop one... |
|
i just got 4770k and i got same problem... so it's me... or smth is missing in pdf sorry to bother you |
|
Hi Slis, did you copy vgt_mgr to your /usr/bin, as well as copy vgt.rules to /etc/udev/rules.d/ ? Thanks, On 09/11/2014 06:07 PM, slis wrote:
|
|
Hi, i think i did... i installed it twice (12.04 and 14.04), now i go 3rd time with graphic stack on 12.04. btw there is error in pdf at building kernel last line and program "bc" is missing form kernel build deps |
|
WORKING now :) i think keymap was problem.. now i go do some tests :P thank you one more question... if i want use 2 vm's with 2 monitors, i should use indirect mode? and write monitor.conf for each vm with specific output port? is that even possible? using 2 vms at same time on 2 monitors? |
|
I can confirm it works only with i7 4770k, but NOT with i3 4130 (0x041e id added to devtable.h)... With i3 there is no sdl bios screen, only green "safe mode" windows loading, then black screen with only windows arrow on vnc or bsod 0x00000003B. Here is full dmesg from i3 but if i can't use 2 separate vm's on 2 monitors at same time, my setup is doomed anyway :) |
|
Hi Slis:
Thanks for reporting this! I fixed that doc and update the repo :) |
/* * FIXME: One weird issue observed when switching between dom0 * and win8 VM. The video ring #1 is not used by both dom0 and * win8 (head=tail=0), however sometimes after switching back * to win8 the video ring may enter a weird state that VCS cmd * parser continues to parse the whole ring (fulled with ZERO). * Sometimes it ends for one whole loop when head reaches back * to 0. Sometimes it may parse indefinitely so that there's * no way to wait for the ring empty. * * Add a posted read works around the issue. In the future we * can further optimize by not switching unused ring. */ the ringbuffer looks like: [ 1015.518426] vGT(wait-empty): wait 20 seconds for ring(1) [ 1015.523865] vGT-cur(1): head(0), tail(0), start(7eee000) [ 1015.529339] vGT-dom0(0): head(0), tail(0), start(8023000) [ 1015.534899] debug registers(ring-1, reg maked with <*> may not apply to every ring): [ 1015.542864] ....EIR: 0 [ 1015.545337] ....ESR: 0 [ 1015.547808] ....blit EIR: 0 [ 1015.550724] ....blit ESR: 0 [ 1015.553635] ....IPEHR(last executed inst): 0 [ 1015.558049] ....INSTPS* (parser state): 0 : [ 1015.562375] ....ACTHD(active header): 18800468 [ 1015.566966] ....DMA_FADD_P(current fetch DMA): 7eee400 [ 1015.572264] ....CSCMDOP* (instruction DWORD): 0 [ 1015.576942] ....CSCMDVLD* (command buffer valid): 0 [ 1015.581975] (informative) [ 1015.584712] ....INSTDONE_1(FYI)*: ff8fffff [ 1015.588949] ....INSTDONE_2*: 0 [ 1015.592130] ring buffer(1): head (82c003c4) tail(10000), start(7eee000) [ 1015.598924] p_contents(ffffc900189ee3c4) [ 1015.602986] [ 1015.603025] [00000384]: 0 0 0 0 0 0 0 0 [ 1015.613228] [000003a4]: 0 0 0 0 0 0 0 0 [ 1015.622144] [000003c4]: 0(*) 0 0 0 0 0 0 0 [ 1015.631326] [000003e4]: 0 0 0 0 0 0 0 0 by ktian1
hello... i am looking to get xengt running but i cant start it... got this in dmesg
[ 1.376813] vGT error:(vgt_set_device_type:432) Unknown chip 0x41e
[ 1.376912] i915: xen_start_vgt done
[ 1.400181] Wrong vgt_if magic number
The text was updated successfully, but these errors were encountered: