[Dri-users] Instability with RX 470

[Larry Reaves]

I just got a new RX 470 and when playing some games (Team Fortress 2 in
particular) the entire system locks up.   I was using the standard arch
linux packages, but I have updated everything I can think of and I'm still
seeing the instability.  I am using the drm-next branch (git hash g28a3965)
of the kernel, mesa from git (hash 1d466b9) and linux-firmware
(hash c4c07a8).  When the system locks up, I restart, and and I investigate
the logs from the previous boot there is nothing useful at the end.  The
only thing amdgpu related is this, which happens repeatedly when I log in
via gdm:


Oct 01 00:43:55 larry-swarga kernel: ------------[ cut here ]------------
Oct 01 00:43:55 larry-swarga kernel: WARNING: CPU: 1 PID: 2984 at
drivers/gpu/drm/drm_irq.c:1168 drm_vblank_put+0xb2/0xb8 [drm]
Oct 01 00:43:55 larry-swarga kernel: Modules linked in: fuse ipt_MASQUERADE
nf_nat_masquerade_ipv4 xfrm_user xfrm_algo iptable_nat nf_conntrack_ipv4
nf_defrag_ipv4 nf_nat_ipv4 xt_addrtype iptable_filter xt_conntrack nf_nat
nf_conntrack br_netfilter bridge stp llc raid1 amdkfd amd_i
Oct 01 00:43:55 larry-swarga kernel:  fjes soundcore tpm_tis wmi
tpm_tis_core tpm button shpchp sch_fq_codel ip_tables x_tables sd_mod
hid_generic usbhid hid ohci_pci serio_raw atkbd libps2 ahci libahci
xhci_pci ehci_pci ohci_hcd libata xhci_hcd ehci_hcd usbcore scsi_mod
usb_common
Oct 01 00:43:55 larry-swarga kernel: CPU: 1 PID: 2984 Comm: Xorg Tainted: G
       W       4.8.0-rc8-drm-next-02450-g28a3965-dirty #2
Oct 01 00:43:55 larry-swarga kernel: Hardware name: To be filled by O.E.M.
To be filled by O.E.M./M5A97, BIOS 1208 04/18/2012
Oct 01 00:43:55 larry-swarga kernel:  0000000000000286 00000000db090998
ffff8807eda1bc60 ffffffff8140ed00
Oct 01 00:43:55 larry-swarga kernel:  0000000000000000 0000000000000000
ffff8807eda1bca0 ffffffff81079dd3
Oct 01 00:43:55 larry-swarga kernel:  00000490a0261f95 ffff8807eda1bd98
ffff88080f8d6068 00000000ffffffea
Oct 01 00:43:55 larry-swarga kernel: Call Trace:
Oct 01 00:43:55 larry-swarga kernel:  [<ffffffff8140ed00>]
dump_stack+0x63/0x83
Oct 01 00:43:55 larry-swarga kernel:  [<ffffffff81079dd3>] __warn+0xcb/0xe8
Oct 01 00:43:55 larry-swarga kernel:  [<ffffffff81079efd>]
warn_slowpath_null+0x1d/0x20
Oct 01 00:43:55 larry-swarga kernel:  [<ffffffffa024f662>]
drm_vblank_put+0xb2/0xb8 [drm]
Oct 01 00:43:55 larry-swarga kernel:  [<ffffffffa024f67f>]
drm_crtc_vblank_put+0x17/0x20 [drm]
Oct 01 00:43:55 larry-swarga kernel:  [<ffffffffa026a925>]
drm_mode_page_flip_ioctl+0x1b5/0x490 [drm]
Oct 01 00:43:55 larry-swarga kernel:  [<ffffffffa024de0c>]
drm_ioctl+0x21c/0x4c0 [drm]
Oct 01 00:43:55 larry-swarga kernel:  [<ffffffff811f96a4>] ?
__check_object_size+0x54/0x1d5
Oct 01 00:43:55 larry-swarga kernel:  [<ffffffffa026a770>] ?
drm_mode_cursor2_ioctl+0x10/0x10 [drm]
Oct 01 00:43:55 larry-swarga kernel:  [<ffffffffa039c04f>]
amdgpu_drm_ioctl+0x4f/0x88 [amdgpu]
Oct 01 00:43:55 larry-swarga kernel:  [<ffffffff8120f7ab>]
do_vfs_ioctl+0xa3/0x5e8
Oct 01 00:43:55 larry-swarga kernel:  [<ffffffff8121a827>] ?
__fget+0x77/0xa8
Oct 01 00:43:55 larry-swarga kernel:  [<ffffffff8120fd69>]
SyS_ioctl+0x79/0x90
Oct 01 00:43:55 larry-swarga kernel:  [<ffffffff816ecc72>]
entry_SYSCALL_64_fastpath+0x1a/0xa4
Oct 01 00:43:55 larry-swarga kernel: ---[ end trace 270078af45c64169 ]---
Oct 01 00:43:55 larry-swarga /usr/lib/gdm/gdm-x-session[2982]: (WW)
AMDGPU(0): Page flip failed: Invalid argument


Additionally I have noticed that lspci -v lists the card as a RX 480 even
though it is a 470.  On bootup, I see this message:
amdgpu 0000:01:00.0: Invalid PCI ROM header signature: expecting 0xaa55,
got 0xffff


Please let me know if there is anything additional I can do to help debug
this issue.

Thank you,
Larry Reaves
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/dri-users/attachments/20161001/20af427a/attachment.html>