Re: Latest i386 DRMKMS results

[Patrick Welche <prlw1%cam.ac.uk@localhost>]

On Sat, Apr 26, 2014 at 11:35:32PM +0900, NONAKA Kimihiro wrote:
> 2014-04-26 22:51 GMT+09:00 Patrick Welche <prlw1%cam.ac.uk@localhost>:
> 
> > Attempting intel driver gets a black screen. At least the system is
> > responsive though!
> 
> ThinkPad E430 (Ivy bridge) work X with intel driver.
> However, accel method uses the UXA instead of SNA.

Ah thanks.

I'm trying the Ivy bridge now:
- vesa is fine but with a high vertical offset
- intel with NoAccel and with AccelMethod UXA just give me a blank screen

intel with UXA:
[   327.517] (II) intel(0): [DRI2] Setup complete
[   327.517] (II) intel(0): [DRI2]   DRI driver: i965
[   327.517] (II) intel(0): Allocated new frame buffer 1280x1024 stride 5120, ti
led 
[   327.520] (II) UXA(0): Driver registered support for the following operations
...
[   327.682] (EE) intel(0): Failed to submit batch buffer, expect rendering corr
uption: No such file or directory.


> Maybe you need to apply patch.
> patch: http://ftp.netbsd.org/pub/NetBSD/misc/nonaka/drmkms/i915drmkms-src.diff

With the last 2 hunks of your patch, I did see an xdm login prompt,
with extra coloured patches appearing.

While logging in:

fatal protection fault in supervisor mode
trap type 4 code 0 rip ffffffff807a866c cs 8 rflags 10202 cr2 7f7ff7e9a000 
ilevel 8 rsp fffffe810ec80b60
curlwp 0xfffffe810ec452a0 pid 0.79 lowest kstack 0xfffffe810ec7d2c0
panic: trap
cpu2: Begin traceback...
vpanic() at netbsd:vpanic+0x158
snprintf() at netbsd:snprintf
startlwp() at netbsd:startlwp
alltraps() at netbsd:alltraps+0x9e
cv_init() at netbsd:cv_init+0x1d
buf_init() at netbsd:buf_init+0x18
getiobuf() at netbsd:getiobuf+0x33
dkstart() at netbsd:dkstart+0xab
dkstrategy() at netbsd:dkstrategy+0xd5
bdev_strategy() at netbsd:bdev_strategy+0x5b
rf_DispatchKernelIO() at netbsd:rf_DispatchKernelIO+0x278
rf_DiskIOEnqueue() at netbsd:rf_DiskIOEnqueue+0x10e
rf_DiskWriteFuncForThreads() at netbsd:rf_DiskWriteFuncForThreads+0xc1
FireNodeList() at netbsd:FireNodeList+0x8d
rf_FinishNode() at netbsd:rf_FinishNode+0x14
rf_DispatchDAG() at netbsd:rf_DispatchDAG+0x194
rf_State_ExecuteDAG() at netbsd:rf_State_ExecuteDAG+0xd0
rf_ContinueRaidAccess() at netbsd:rf_ContinueRaidAccess+0x4b
rf_DoAccess() at netbsd:rf_DoAccess+0x115
raidstart() at netbsd:raidstart+0x239
rf_RaidIOThread() at netbsd:rf_RaidIOThread+0xea

Added the first hunk as well, this time on boot

panic: kernel diagnostic assertion "ptp == NULL || ptp->wire_count > 1" failed: 
file "../../../../arch/x86/x86/pmap.c", line 4022 
cpu1: Begin traceback...
vpanic() at netbsd:vpanic+0x158
kern_assert() at netbsd:kern_assert+0x4f
pmap_enter_ma() at netbsd:pmap_enter_ma+0x71a
pmap_enter_default() at netbsd:pmap_enter_default+0x1d
uvm_fault_internal() at netbsd:uvm_fault_internal+0xa2a
trap() at netbsd:trap+0x253
--- trap (number 6) ---
7f7ff649feca:
cpu1: End traceback...
M
udteumx pienrrg orto:  mdeutv e1x_6v8,ec7 t(oor_fefsnteter=:17 a4s03se83rt1i, 
ons izfae=i2le01d:88 2MU5)TE:X
_OduWmNEp R(mtx->mtx_owner) == curthread

lock address : 0xfffffe82139bae00 type     :     sleep/adaptive
initialized  : 0xffffffff80587657
shared holds :                  0 exclusive:                  0
shares wanted:                  0 exclusive:                  0
current cpu  :                  2 last held:                  2
current lwp  : 0xfffffe821698aa20 last held: 000000000000000000
last locked  : 0xffffffff8059f2c7 unlocked*: 0xffffffff8059f3f2
owner field  : 000000000000000000 wait/spin:                0/0

Turnstile chain at 0xffffffff80f3ce80.
=> No active turnstile for this lock.


381 380 379 378 377 376 375 374 373 372 371 370 369 368 367 366 365 364 363 362 
...
6 5 4 3 2 1 succeeded

yet:

# savecore -z .
savecore: no core dump

The above doesn't really look drm related, but an ordinary kernel from
the same source is fine.

Cheers,

Patrick