Re: i915 crash

On Wed, Jan 07, 2015 at 01:01:23AM -0500, Allan Jude wrote:
> I grabbed the latest i915.8.patch from kib_at_'s website and compiled it
> against r276774 (today)
> 
> Machine is a Lenovo T530, booted UEFI, with the nvidia GPU disabled.
> 
> CPU: Intel(R) Core(TM) i5-3320M CPU _at_ 2.60GHz (2594.16-MHz K8-class CPU)
> 
> It is an Ivy-bridge CPU/GPU
> 
> I installed xorg and kde, and when I try to start KDE, it loads, and
> gets so far as showing the FreeBSD wallpaper, then panics:
> 
> 
> text dump: http://www.allanjude.com/bsd/i915_core.3.txt
> Full dump: (26mb compressed, 740mb original)
This is useless for anybody except you.

> http://www.allanjude.com/bsd/i915_vmcore.3.xz
> 
> Unread portion of the kernel message buffer:
> panic: In GPU write domain
> cpuid = 3
> KDB: stack backtrace:
> db_trace_self_wrapper() at db_trace_self_wrapper+0x2b/frame
> 0xfffffe045e52e600
> vpanic() at vpanic+0x189/frame 0xfffffe045e52e680
> kassert_panic() at kassert_panic+0x132/frame 0xfffffe045e52e6f0
> i915_gem_pread_ioctl() at i915_gem_pread_ioctl+0x678/frame
> 0xfffffe045e52e790
> drm_ioctl() at drm_ioctl+0x318/frame 0xfffffe045e52e800
> devfs_ioctl_f() at devfs_ioctl_f+0x122/frame 0xfffffe045e52e860
> kern_ioctl() at kern_ioctl+0x2c0/frame 0xfffffe045e52e8c0
> sys_ioctl() at sys_ioctl+0x153/frame 0xfffffe045e52e9a0
> amd64_syscall() at amd64_syscall+0x25a/frame 0xfffffe045e52eab0
> Xfast_syscall() at Xfast_syscall+0xfb/frame 0xfffffe045e52eab0
> --- syscall (54, FreeBSD ELF64, sys_ioctl), rip = 0x8022a56fa, rsp =
> 0x7fffffffe718, rbp = 0x7fffffffe740 ---
> KDB: enter: panic
> Uptime: 9m19s
> Dumping 739 out of 16176
> MB:..3%..11%..22%..31%..42%..52%..61%..72%..81%..91%
> 
> Reading symbols from /boot/kernel/i915kms.ko.symbols...done.
> Loaded symbols for /boot/kernel/i915kms.ko.symbols
> Reading symbols from /boot/kernel/drm2.ko.symbols...done.
> Loaded symbols for /boot/kernel/drm2.ko.symbols
> Reading symbols from /boot/kernel/iicbus.ko.symbols...done.
> Loaded symbols for /boot/kernel/iicbus.ko.symbols
> Reading symbols from /boot/kernel/iic.ko.symbols...done.
> Loaded symbols for /boot/kernel/iic.ko.symbols
> Reading symbols from /boot/kernel/iicbb.ko.symbols...done.
> Loaded symbols for /boot/kernel/iicbb.ko.symbols
> #0  doadump (textdump=Unhandled dwarf expression opcode 0x93
> ) at pcpu.h:219
> 219	pcpu.h: No such file or directory.
> 	in pcpu.h
> (kgdb) #0  doadump (textdump=Unhandled dwarf expression opcode 0x93
> ) at pcpu.h:219
> #1  0xffffffff80965d27 in kern_reboot (howto=Unhandled dwarf expression
> opcode 0x93
> )
>     at /usr/src/sys/kern/kern_shutdown.c:448
> #2  0xffffffff80966318 in vpanic (fmt=<value optimized out>,
>     ap=<value optimized out>) at /usr/src/sys/kern/kern_shutdown.c:747
> #3  0xffffffff80966142 in kassert_panic (fmt=<value optimized out>)
>     at /usr/src/sys/kern/kern_shutdown.c:635
> #4  0xffffffff81e1de88 in i915_gem_pread_ioctl (dev=0xfffff8002ffd1000,
>     data=0xfffffe045e52e8f0, file=<value optimized out>)
>     at
> /usr/src/sys/modules/drm2/i915kms/../../../dev/drm2/i915/i915_gem.c:3010

> #5  0xffffffff81e9a398 in drm_ioctl (kdev=<value optimized out>,
>     cmd=2149606492, data=0xfffffe045e52e8f0 "y", flags=Unhandled dwarf
> expression opcode 0x93
> )
>     at /usr/src/sys/modules/drm2/drm2/../../../dev/drm2/drm_drv.c:942
> #6  0xffffffff80849942 in devfs_ioctl_f (fp=0xfffff80009c15690,
>     com=2149606492, data=0xfffffe045e52e8f0, cred=0xfffffe045e52e8f0,
>     td=0xfffff80009c74000) at /usr/src/sys/fs/devfs/devfs_vnops.c:775
> #7  0xffffffff809c3ad0 in kern_ioctl (td=0xfffff80009c74000,
>     fd=<value optimized out>, com=0, data=<value optimized out>) at
> file.h:318
> #8  0xffffffff809c3763 in sys_ioctl (td=0xfffff80009c74000,
>     uap=0xfffffe045e52ea40) at /usr/src/sys/kern/sys_generic.c:718
> #9  0xffffffff80d8590a in amd64_syscall (td=0xfffff80009c74000, traced=0)
>     at subr_syscall.c:133
> #10 0xffffffff80d632ab in Xfast_syscall ()
>     at /usr/src/sys/amd64/amd64/exception.S:395
> #11 0x00000008022a56fa in ?? ()
> Previous frame inner to this frame (corrupt stack?)
> Current language:  auto; currently minimal
> 

Is this reproducable ?

Try the following patch on top of i915.8.

commit 9af6c652745f551e2b6ce5218e350a5e47999feb
Author: Konstantin Belousov <kib_at_freebsd.org>
Date:   Wed Jan 7 20:21:46 2015 +0200

    Properly move object into gtt domain when needed.

diff --git a/sys/dev/drm2/i915/i915_gem.c b/sys/dev/drm2/i915/i915_gem.c
index 0f72d08..58cbb59 100644
--- a/sys/dev/drm2/i915/i915_gem.c
+++ b/sys/dev/drm2/i915/i915_gem.c
_at__at_ -1251,7 +1251,7 _at__at_ i915_gem_shmem_pread(struct drm_device *dev,
 		 * optimizes for the case when the gpu will dirty the data
 		 * anyway again before the next pread happens. */
 		needs_clflush = !cpu_cache_is_coherent(dev, obj->cache_level);
-		ret = i915_gem_object_wait_rendering(obj);
+		ret = i915_gem_object_set_to_gtt_domain(obj, false);
 		if (ret)
 			return ret;
 	}
_at__at_ -1579,7 +1579,7 _at__at_ i915_gem_shmem_pwrite(struct drm_device *dev,
 		 * optimizes for the case when the gpu will use the data
 		 * right away and we therefore have to clflush anyway. */
 		needs_clflush_after = cpu_write_needs_clflush(obj);
-		ret = i915_gem_object_wait_rendering(obj);
+		ret = i915_gem_object_set_to_gtt_domain(obj, true);
 		if (ret)
 			return ret;
 	}