[meta-intel] Cedartrail PVR Driver "HWRecoveryResetSGX"

[Chris Tapp]

Does anyone know what causes "PVR_K: HWRecoveryResetSGX: SGX Hardware Recovery triggered" entries in the system log?

I've got a GLES application that generally runs fine, but the system log sometimes shows entries similar to:

   PVR_K: HWRecoveryResetSGX: SGX Hardware Recovery triggered
   PVR_K: SGX debug (pc_i686_linux_XOrg_family sgxddk 17 1.7 at 862890)
   PVR_K: (P0) EUR_CR_CORE_ID:          01150000
   PVR_K: (P0) EUR_CR_CORE_REVISION:    0001000D
   PVR_K: (P0) EUR_CR_EVENT_STATUS:     20000001
   PVR_K: 	(master_irq: no, timer_irq: yes, ta_dpm_fault: no, zls_oom: no, ta_mem_free: no, isp_end_tile: no, dpm_initend: no, zls_csw_finish: no, pbe_end_render: no, isp_vis_fail: no, isp_break: no, sw_event no, ta_finish: no, ta
   PVR_K: (P0) EUR_CR_EVENT_STATUS2:    00000C80
   PVR_K: 	(mte_flush: no, vdm_load: no, vdm_kicked: no, otpm_mem_clear: no,otpm_flush: yes, dcu_invalid: yes, gsg_flush: no, gsg_load: no, lockup_ta: yes, lockup_3d: no, partial_deadlock: no, dpm_dhost_free: nodpm_host_free: no, 
   PVR_K: (P0) EUR_CR_BIF_CTRL:         00000000
   PVR_K: (P0) EUR_CR_BIF_BANK0:        0000202F
   PVR_K: (P0) EUR_CR_BIF_INT_STAT:     00080000
   PVR_K: (P0) EUR_CR_BIF_FAULT:        00000000
   PVR_K: (P0) EUR_CR_BIF_MEM_REQ_STAT: 00000001
   PVR_K: (P0) EUR_CR_CLKGATECTL:       0000A6AA
   PVR_K: Flip Command Complete Data 0 for display device 2:
   PVR_K: SGX Host control:
   PVR_K: 	(HC-0) 0x00000001 0x00000000 0x00000000 0x00000008
   PVR_K: 	(HC-10) 0x00000063 0x0000000A 0x00061A80 0x00000000
   PVR_K: 	(HC-20) 0x00000000 0x00000001 0x00000000 0x00000000
   PVR_K: 	(HC-30) 0x06387B1E 0x21969C60 0x00000000 0x00000000
   PVR_K: 	(HC-40) 0x00000000 0x00000000 0x00000000 0x00000000
   PVR_K: SGX TA/3D control:
   PVR_K: 	(T3C-0) 0xF4003000 0xF4003100 0xF4002000 0x00000000
   PVR_K: 	(T3C-10) 0x00000000 0x00000000 0x00000000 0x00000000
   PVR_K: 	(T3C-20) 0x00000000 0x00000000 0x00000000 0x00000000
   PVR_K: 	(T3C-30) 0x00000000 0x00000000 0x00000000 0x00000000
   PVR_K: 	(T3C-40) 0x00000000 0x00000000 0x00000000 0x00000000
   PVR_K: 	(T3C-50) 0xF4096D98 0xF40969D0 0xF4000000 0xCC788000
   PVR_K: 	(T3C-60) 0xF406F4C0 0xF4096D98 0xF409A200 0xF400AE60
   PVR_K: 	(T3C-70) 0x00000000 0xF4098000 0x00000000 0x00020000
   PVR_K: 	(T3C-80) 0xF4096880 0x00000000 0x00000000 0x00000000
   PVR_K: 	(T3C-90) 0x00000000 0x00000000 0x00000000 0x00000000
   PVR_K: 	(T3C-A0) 0x00000000 0x001158D6 0x001158D5 0xF4004000
   PVR_K: 	(T3C-B0) 0xF400A420 0xF400B000 0xF4030060 0x00000000
   PVR_K: 	(T3C-C0) 0x00000000 0x00000002 0x0000050E 0x0000050E
   PVR_K: 	(T3C-D0) 0x00000001 0x00000000 0x00000001 0x00000003
   PVR_K: 	(T3C-E0) 0x00000000 0x00000001 0x0538C639 0x00000000
   PVR_K: SGX Kernel CCB WO:0xAA RO:0xA6

These don't generally cause any noticeable issues, but a couple of times I've had a burst of these errors which corresponded with the screen appearing to jitter between two frames. The screen resumed normal operation about 10 seconds later with animated content in the position it should have been in by that point (i.e. the application had continued to run and generate frames, but these were either lost or did not make it to the framebuffer).

--

Chris Tapp
opensource at keylevel.com
www.keylevel.com

----
You can tell you're getting older when your car insurance gets real cheap!