Re: Xorg server segmentation fault when monitor is not connected.

On Sat, Jan 19, 2019 at 7:54 AM Jan Beich <[email protected]> wrote:

> Chung-Yeol Lee <[email protected]> writes:
>
> > Hi!
> >
> > Since I updated Xorg server from 1.18.4_10,1 -> 1.18.4_11,1,
> > the server does not start up when the monitor is not connected.
> > (I actually updated bunch of other ports at that time, including mesa-dri
> > etc, so it might be something else).
>
> Can you track down the exact cause? Try downgrading mesa-dri to 18.2.8
> first
> then to 18.1.9.
>

I tried to install mesa-dri 18.1.9, but it showed the same error.
Now, I'm not sure when the problem actually started.

From my testing, this does not happen once I connect the monitor even I
disconnect it later.
The problem re-appears only after hard power cycle. Simple reboot did not
cause the issue again.
(I haven't fully turned down the computer for a while, and the monitor was
connected long time ago.)

Also, in case this can be useful, I got the following log when I connect
monitor (and X don't crash anymore).
 kernel: [drm:radeon_dvi_detect] DVI-D-1: probed a monitor but no|invalid
EDID
 kernel: [drm] fb mappable at 0xC03DA000
 kernel: [drm] vram apper at 0xC0000000
 kernel: [drm] size 3145728
 kernel: [drm] fb depth is 24
 kernel: [drm]    pitch is 4096
 kernel: VT: Replacing driver "vga" with new "fb".
 kernel: start FB_INFO:
 kernel: type=11 height=768 width=1024 depth=32
 kernel: cmsize=16 size=3145728
 kernel: pbase=0xc03da000 vbase=0xfffff800c03da000
 kernel: name=drmn0 flags=0x0 stride=4096 bpp=32
 kernel: cmap[0]=0 cmap[1]=7f0000 cmap[2]=7f00 cmap[3]=c4a000
 kernel: end FB_INFO
 kernel: drmn0: fb0: radeondrmfb frame buffer device


> > I'm on 12-STABLE / AMD Radeon 7450 with radeonkms.ko.
> > The server works fine as soon as I connect the monitor.
>
> radeonkms from drm-fbsd12.0-kmod with xf86-video-ati? Can you try
> drm-fbsd11.2-kmod first then xf86-video-ati-legacy? Can you try
> "DRI" "2" or "AccelMethod" "EXA" in xorg.conf per radeon(4x) manpage?
>
> Can you show pciconf output to avoid ambiguity?
> $ pciconf -lvcb | sed -n '/vgapci/,/@pci/p'


 vgapci0@pci0:0:1:0:     class=0x030000 card=0x99931849 chip=0x99931002
rev=0x00 hdr=0x00
    vendor     = 'Advanced Micro Devices, Inc. [AMD/ATI]'
    device     = 'Trinity 2 [Radeon HD 7480D]'
    class      = display
    subclass   = VGA
    bar   [10] = type Prefetchable Memory, range 32, base 0xc0000000, size
268435456, enabled
    bar   [14] = type I/O Port, range 32, base 0xf000, size 256, enabled
    bar   [18] = type Memory, range 32, base 0xfeb00000, size 262144,
enabled
    cap 01[50] = powerspec 3  supports D0 D1 D2 D3  current D0
    cap 10[58] = PCI-Express 2 root endpoint max data 128(128) RO NS
    cap 05[a0] = MSI supports 1 message, 64 bit
    ecap 000b[100] = Vendor 1 ID 1
none0@pci0:0:1:1:       class=0x040300 card=0x99021849 chip=0x99021002
rev=0x00 hdr=0x00

>
> > Is this a known issue?
>
> Not enough details. The crash fingerprint is obscured by lack of symbols,
> so it's not possible to search if anyone reported a similar issue upstream.
>

Sorry for the lack of context. I hoped this is known issue or something
obvious from its symptom.
here's better stack trace with debug symbol:
#0  memcpy () at /usr/src/lib/libc/amd64/string/memmove.S:306
#1  0x0000000802834af8 in r600_create_vertex_fetch_shader (ctx=0x8010fa340,
count=2, elements=0x7fffffffd8c0) at r600_asm.c:2783
#2  0x0000000802690636 in util_blitter_create (pipe=0x8010fa340) at
util/u_blitter.c:302
#3  0x000000080283cd5f in r600_create_context (screen=0x806674100,
priv=0x0, flags=0) at r600_pipe.c:231
#4  0x000000080283c7bf in r600_screen_create (ws=0x801ba5c00,
config=0x7fffffffdbb0) at r600_pipe.c:825
#5  0x0000000802ac4b5c in radeon_drm_winsys_create (fd=9,
config=0x7fffffffdbb0, screen_create=0x80283c360 <r600_screen_create>) at
radeon_drm_winsys.c:941
#6  0x0000000801f7d2e2 in pipe_r600_create_screen (fd=9,
config=0x7fffffffdbb0) at
../../../../src/gallium/auxiliary/target-helpers/drm_helper.h:144
#7  0x00000008027b5ffc in pipe_loader_drm_create_screen (dev=0x80105a800,
config=0x7fffffffdbb0) at pipe_loader_drm.c:362
#8  0x00000008027b4fd1 in pipe_loader_create_screen (dev=0x80105a800) at
pipe_loader.c:137
#9  0x00000008025c4b79 in dri2_init_screen (sPriv=0x80110a200) at
dri2.c:2130
#10 0x00000008025bdf89 in driCreateNewScreen2 (scrn=0, fd=8,
extensions=0x801b32030, driver_extensions=0x802b3ede0
<galliumdrm_driver_extensions>, driver_configs=0x801129478,
data=0x801129280) at dri_util.c:153
#11 0x0000000801b2cf52 in ?? () from /usr/local/lib/libgbm.so.1
#12 0x0000000801b2b975 in ?? () from /usr/local/lib/libgbm.so.1
#13 0x0000000801b2b1be in gbm_create_device () from
/usr/local/lib/libgbm.so.1
#14 0x0000000801b0299d in radeon_glamor_pre_init (scrn=0x801ba5000) at
radeon_glamor.c:135
#15 0x0000000801aee262 in RADEONPreInitAccel_KMS (pScrn=0x801ba5000) at
radeon_kms.c:1358
#16 0x0000000801aecffe in RADEONPreInit_KMS (pScrn=0x801ba5000, flags=0) at
radeon_kms.c:1773
#17 0x00000000002d359f in InitOutput ()
#18 0x000000000028f34b in ?? ()
#19 0x000000000027910b in _start ()
_______________________________________________
[email protected] mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-x11
To unsubscribe, send any mail to "[email protected]"