drm/i915: There is only one fault register from GEN8 onwards
Subsystems (2)
PATCH 1/1 drm/i915: There is only one fault register from GEN8 onwards
Until Haswell/Baytrail, the hardware used to have a per engine fault register (e.g. 0x4094 - render fault register, 0x4194 - media fault register and so on). But since Broadwell, all these registers were combined into a singe one and the engine id stored in bits 14:12.
Not only we should not been reading (and writing to) registers that do not exist, in platforms with VCS2 (SKL), the address that would belong this engine (0x4494, VCS2_HW = 4) is already assigned to other register.
References: IHD-OS-BDW-Vol 2c-11.15, page 75.
References: IHD-OS-SKL-Vol 2c-05.16, page 350.
--- drivers/gpu/drm/i915/i915_gem_gtt.c | 38 ++++++++++++++++++++++++++++++----- drivers/gpu/drm/i915/i915_gpu_error.c | 8 +++++--- drivers/gpu/drm/i915/i915_reg.h | 2 ++ 3 files changed, 40 insertions(+), 8 deletions(-)
drivers/gpu/drm/i915/i915_gem_gtt.c +33 -5
@@ -2256,16 +2256,13 @@ static bool needs_idle_maps(struct drm_i915_private *dev_priv)return IS_GEN5 (dev_priv) &&IS_MOBILE (dev_priv) &&intel_vtd_active (); } -void i915_check_and_clear_faults (struct drm_i915_private *dev_priv) +void __check_and_clear_faults (struct drm_i915_private *dev_priv) {struct intel_engine_cs *engine;enum intel_engine_id id; - -if (INTEL_INFO (dev_priv)->gen <6 ) -return ; +u32 fault_reg;for_each_engine (engine, dev_priv, id) { -u32 fault_reg; fault_reg =I915_READ (RING_FAULT_REG (engine));if (fault_reg &RING_FAULT_VALID ) { DRM_DEBUG_DRIVER("Unexpected fault\n"
@@ -2281,6 +2278,37 @@ void i915_check_and_clear_faults(struct drm_i915_private *dev_priv) fault_reg & ~RING_FAULT_VALID); } } +} + +void __gen8_check_and_clear_faults (struct drm_i915_private *dev_priv) +{ +u32 fault_reg =I915_READ (GEN8_RING_FAULT_REG ); + +if (fault_reg &RING_FAULT_VALID ) { +DRM_DEBUG_DRIVER ("Unexpected fault\n" +"\tAddr: 0x%08lx\n" +"\tEngine ID: %d\n" +"\tSource ID: %d\n" +"\tType: %d\n" , + fault_reg &PAGE_MASK , +GEN8_RING_FAULT_ENGINE_ID (fault_reg), +RING_FAULT_SRCID (fault_reg), +RING_FAULT_FAULT_TYPE (fault_reg)); +I915_WRITE (GEN8_RING_FAULT_REG , + fault_reg & ~RING_FAULT_VALID ); + } +} + +void i915_check_and_clear_faults (struct drm_i915_private *dev_priv) +{ +if (INTEL_GEN (dev_priv) <6 ) +return ; + +/* From GEN8 onwards we only have one 'All Engine Fault Register' */ +if (INTEL_GEN (dev_priv) >=8 ) +__gen8_check_and_clear_faults (dev_priv); +else +__check_and_clear_faults (dev_priv);/* Engine specific init may not have been done till this point. */ if (dev_priv->engine[RCS ])
drivers/gpu/drm/i915/i915_gpu_error.c +5 -3
@@ -1217,11 +1217,13 @@ static void error_record_engine_registers(struct i915_gpu_state *error,if (INTEL_GEN (dev_priv) >=6 ) { ee->rc_psmi =I915_READ (RING_PSMI_CTL (engine->mmio_base)); - ee->fault_reg =I915_READ (RING_FAULT_REG (engine)); -if (INTEL_GEN (dev_priv) >=8 ) +if (INTEL_GEN (dev_priv) >=8 ) {gen8_record_semaphore_state (error, engine, ee); -else + ee->fault_reg =I915_READ (GEN8_RING_FAULT_REG ); + }else {gen6_record_semaphore_state (engine, ee); + ee->fault_reg =I915_READ (RING_FAULT_REG (engine)); + } }if (INTEL_GEN (dev_priv) >=4 ) {
drivers/gpu/drm/i915/i915_reg.h +2 -0
@@ -2329,6 +2329,8 @@ enum i915_power_well_id {#define ARB_MODE_SWIZZLE_BDW (1<<1)#define RENDER_HWS_PGA_GEN7 _MMIO(0x04080)#define RING_FAULT_REG (engine) _MMIO(0x4094 + 0x100*(engine)->hw_id) +#define GEN8_RING_FAULT_REG _MMIO(0x4094) +#define GEN8_RING_FAULT_ENGINE_ID (x) (((x) >> 12) & 0x7)#define RING_FAULT_GTTSEL_MASK (1<<11)#define RING_FAULT_SRCID (x) (((x) >> 3) & 0xff)#define RING_FAULT_FAULT_TYPE (x) (((x) >> 1) & 0x3)
Signature · 6 lines
2.15.0 _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Patchwork
CI✓ Fi.CI.BAT: success for drm/i915: There is only one fault register from Gen8 onwards (rev3)
== Series Details ==
== Series Details ==
Series: drm/i915: There is only one fault register from Gen8 onwards (rev3)
URL : https://patchwork.freedesktop.org/series/26317/
State : success
== Summary ==
Series 26317v3 drm/i915: There is only one fault register from Gen8 onwards
https://patchwork.freedesktop.org/api/1.0/series/26317/revisions/3/mbox/
Test gem_exec_suspend:
Subgroup basic-s3:
dmesg-fail -> PASS (fi-kbl-7560u) fdo#103039
Subgroup basic-s4-devices:
dmesg-fail -> PASS (fi-kbl-7560u) fdo#102846 +1
Test gem_flink_basic:
Subgroup bad-flink:
dmesg-warn -> PASS (fi-kbl-7560u) fdo#103049 +4
Test gem_linear_blits:
Subgroup basic:
incomplete -> PASS (fi-kbl-7560u) fdo#103163
Test kms_cursor_legacy:
Subgroup basic-busy-flip-before-cursor-legacy:
fail -> PASS (fi-gdg-551) fdo#102618fdo#103039 https://bugs.freedesktop.org/show_bug.cgi?id=103039
fdo#102846 https://bugs.freedesktop.org/show_bug.cgi?id=102846
fdo#103049 https://bugs.freedesktop.org/show_bug.cgi?id=103049
fdo#103163 https://bugs.freedesktop.org/show_bug.cgi?id=103163
fdo#102618 https://bugs.freedesktop.org/show_bug.cgi?id=102618
fi-bdw-5557u total:289 pass:268 dwarn:0 dfail:0 fail:0 skip:21 time:443s fi-bdw-gvtdvm total:289 pass:265 dwarn:0 dfail:0 fail:0 skip:24 time:456s fi-blb-e6850 total:289 pass:223 dwarn:1 dfail:0 fail:0 skip:65 time:382s fi-bsw-n3050 total:289 pass:243 dwarn:0 dfail:0 fail:0 skip:46 time:548s fi-bwr-2160 total:289 pass:183 dwarn:0 dfail:0 fail:0 skip:106 time:274s fi-bxt-j4205 total:289 pass:260 dwarn:0 dfail:0 fail:0 skip:29 time:510s fi-byt-j1900 total:289 pass:254 dwarn:0 dfail:0 fail:0 skip:35 time:497s fi-byt-n2820 total:289 pass:250 dwarn:0 dfail:0 fail:0 skip:39 time:489s fi-elk-e7500 total:289 pass:229 dwarn:0 dfail:0 fail:0 skip:60 time:428s fi-gdg-551 total:289 pass:178 dwarn:1 dfail:0 fail:1 skip:109 time:263s fi-glk-1 total:289 pass:261 dwarn:0 dfail:0 fail:0 skip:28 time:544s fi-hsw-4770 total:289 pass:262 dwarn:0 dfail:0 fail:0 skip:27 time:436s fi-hsw-4770r total:289 pass:262 dwarn:0 dfail:0 fail:0 skip:27 time:443s fi-ilk-650 total:289 pass:228 dwarn:0 dfail:0 fail:0 skip:61 time:430s fi-ivb-3520m total:289 pass:260 dwarn:0 dfail:0 fail:0 skip:29 time:474s fi-ivb-3770 total:289 pass:260 dwarn:0 dfail:0 fail:0 skip:29 time:461s fi-kbl-7500u total:289 pass:264 dwarn:1 dfail:0 fail:0 skip:24 time:489s fi-kbl-7560u total:247 pass:230 dwarn:0 dfail:0 fail:0 skip:16 fi-kbl-7567u total:289 pass:269 dwarn:0 dfail:0 fail:0 skip:20 time:476s fi-kbl-r total:289 pass:262 dwarn:0 dfail:0 fail:0 skip:27 time:544s fi-pnv-d510 total:289 pass:222 dwarn:1 dfail:0 fail:0 skip:66 time:568s fi-skl-6260u total:289 pass:269 dwarn:0 dfail:0 fail:0 skip:20 time:455s fi-skl-6600u total:289 pass:262 dwarn:0 dfail:0 fail:0 skip:27 time:549s fi-skl-6700hq total:289 pass:263 dwarn:0 dfail:0 fail:0 skip:26 time:566s fi-skl-6700k total:289 pass:265 dwarn:0 dfail:0 fail:0 skip:24 time:523s fi-skl-6770hq total:289 pass:269 dwarn:0 dfail:0 fail:0 skip:20 time:503s fi-skl-gvtdvm total:289 pass:266 dwarn:0 dfail:0 fail:0 skip:23 time:460s fi-snb-2520m total:289 pass:250 dwarn:0 dfail:0 fail:0 skip:39 time:565s fi-snb-2600 total:289 pass:249 dwarn:0 dfail:0 fail:0 skip:40 time:425s Blacklisted hosts: fi-cfl-s total:289 pass:254 dwarn:3 dfail:0 fail:0 skip:32 time:546s fi-glk-dsi total:62 pass:45 dwarn:0 dfail:0 fail:0 skip:16
fi-bxt-dsi failed to connect after reboot
6f3dca38f7a939db9658bd11d9c37357089cabef drm-tip: 2017y-11m-10d-12h-56m-44s UTC integration manifest bf772f904fe3 drm/i915: There is only one fault register from GEN8 onwards
== Logs ==
For more details see: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_7067/
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Patchwork
CI✓ Fi.CI.IGT: success for drm/i915: There is only one fault register from Gen8 onwards (rev3)
In reply to Michel Thierry== Series Details ==
== Series Details ==
Series: drm/i915: There is only one fault register from Gen8 onwards (rev3)
URL : https://patchwork.freedesktop.org/series/26317/
State : success
== Summary ==
Test kms_flip:
Subgroup modeset-vs-vblank-race-interruptible:
fail -> PASS (shard-hsw) fdo#103060
Subgroup plain-flip-ts-check-interruptible:
pass -> FAIL (shard-hsw) fdo#100368
Test perf:
Subgroup blocking:
fail -> PASS (shard-hsw) fdo#102252
Test kms_frontbuffer_tracking:
Subgroup fbc-farfromfence:
skip -> PASS (shard-hsw)
Test kms_busy:
Subgroup extended-modeset-hang-oldfb-with-reset-render-a:
pass -> DMESG-WARN (shard-hsw) fdo#102249fdo#103060 https://bugs.freedesktop.org/show_bug.cgi?id=103060
fdo#100368 https://bugs.freedesktop.org/show_bug.cgi?id=100368
fdo#102252 https://bugs.freedesktop.org/show_bug.cgi?id=102252
fdo#102249 https://bugs.freedesktop.org/show_bug.cgi?id=102249
shard-hsw total:2584 pass:1471 dwarn:2 dfail:1 fail:11 skip:1099 time:9454s
Blacklisted hosts:
shard-apl total:2584 pass:1620 dwarn:4 dfail:1 fail:23 skip:936 time:13241s shard-kbl total:2556 pass:1688 dwarn:17 dfail:1 fail:23 skip:825 time:9926s shard-snb total:2584 pass:1211 dwarn:2 dfail:1 fail:12 skip:1358 time:7801s
== Logs ==
For more details see: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_7067/shards.html
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Quoting Michel Thierry (2017-11-10 19:01:16)
Quoted 23 lines
Until Haswell/Baytrail, the hardware used to have a per engine fault register (e.g.
0x4094- render fault register,0x4194- media fault register and so on). But since Broadwell, all these registers were combined into a singe one and the engine id stored in bits 14:12.Not only we should not been reading (and writing to) registers that do not exist, in platforms with VCS2 (SKL), the address that would belong this engine (
0x4494,VCS2_HW= 4) is already assigned to other register.References: IHD-OS-BDW-Vol 2c-11.15, page 75.
References: IHD-OS-SKL-Vol 2c-05.16, page 350.Signed-off-by Michel Thierry <[email protected]>--- drivers/gpu/drm/i915/i915_gem_gtt.c | 38 ++++++++++++++++++++++++++++++----- drivers/gpu/drm/i915/i915_gpu_error.c | 8 +++++--- drivers/gpu/drm/i915/i915_reg.h | 2 ++ 3 files changed, 40 insertions(+), 8 deletions(-)drivers/gpu/drm/i915/i915_gem_gtt.c +0
-0
drivers/gpu/drm/i915/i915_gem_gtt.c static bool needs_idle_maps(struct drm_i915_private *dev_priv) · line 2259
Quoted 5 lines
drivers/gpu/drm/i915/i915_gem_gtt.c +1
-1@@ -2256,16 +2256,13 @@ static bool needs_idle_maps(struct drm_i915_private *dev_priv)return IS_GEN5 (dev_priv) &&IS_MOBILE (dev_priv) &&intel_vtd_active (); } -void i915_check_and_clear_faults (struct drm_i915_private *dev_priv) +void __check_and_clear_faults (struct drm_i915_private *dev_priv)
I am amazed that -Wextra doesn't complain. Try with sparse.
There's an old thread where this was raised and how we are not clearing the faults early enough.
-Chris
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Quoted 27 lines · Chris Wilson
Quoting Michel Thierry (2017-11-10 19:01:16)
Quoted 26 lines
Until Haswell/Baytrail, the hardware used to have a per engine fault register (e.g.
0x4094- render fault register,0x4194- media fault register and so on). But since Broadwell, all these registers were combined into a singe one and the engine id stored in bits 14:12.Not only we should not been reading (and writing to) registers that do not exist, in platforms with VCS2 (SKL), the address that would belong this engine (
0x4494,VCS2_HW= 4) is already assigned to other register.References: IHD-OS-BDW-Vol 2c-11.15, page 75.
References: IHD-OS-SKL-Vol 2c-05.16, page 350.Signed-off-by Michel Thierry <[email protected]>--- drivers/gpu/drm/i915/i915_gem_gtt.c | 38 ++++++++++++++++++++++++++++++----- drivers/gpu/drm/i915/i915_gpu_error.c | 8 +++++--- drivers/gpu/drm/i915/i915_reg.h | 2 ++ 3 files changed, 40 insertions(+), 8 deletions(-)drivers/gpu/drm/i915/i915_gem_gtt.c +0
-0@@ -2256,16 +2256,13 @@ static bool needs_idle_maps(struct drm_i915_private *dev_priv)return IS_GEN5 (dev_priv) &&IS_MOBILE (dev_priv) &&intel_vtd_active (); }
Quoted 5 lines
drivers/gpu/drm/i915/i915_gem_gtt.c static bool needs_idle_maps(struct drm_i915_private *dev_priv) · line 2259
Quoted 2 lines
drivers/gpu/drm/i915/i915_gem_gtt.c +1
-1-void i915_check_and_clear_faults (struct drm_i915_private *dev_priv) +void __check_and_clear_faults (struct drm_i915_private *dev_priv)I am amazed that -Wextra doesn't complain. Try with sparse.
Hmm, no they didn't... gen6_check_and_clear_faults and gen8_check_and_clear_faults are less controversial.
Quoted 2 lines
There's an old thread where this was raised and how we are not clearing the faults early enough.
Ah, I found that thread [1], and as you said there, right now the for_each_engine is a nop (with this patch at least gen8+ would do the right thing).
Since gen6/gen7 engines are well known and can't change, do we really need for_each_engine?
[1] https://patchwork.freedesktop.org/patch/101331/
_______________________________________________
Intel-gfx mailing list [email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Quoting Michel Thierry (2017-11-10 23:42:31)
Quoted 28 lines · Chris Wilson
Quoted 27 lines
Quoting Michel Thierry (2017-11-10 19:01:16)
Quoted 26 lines
Until Haswell/Baytrail, the hardware used to have a per engine fault register (e.g.
0x4094- render fault register,0x4194- media fault register and so on). But since Broadwell, all these registers were combined into a singe one and the engine id stored in bits 14:12.Not only we should not been reading (and writing to) registers that do not exist, in platforms with VCS2 (SKL), the address that would belong this engine (
0x4494,VCS2_HW= 4) is already assigned to other register.References: IHD-OS-BDW-Vol 2c-11.15, page 75.
References: IHD-OS-SKL-Vol 2c-05.16, page 350.Signed-off-by Michel Thierry <[email protected]>--- drivers/gpu/drm/i915/i915_gem_gtt.c | 38 ++++++++++++++++++++++++++++++----- drivers/gpu/drm/i915/i915_gpu_error.c | 8 +++++--- drivers/gpu/drm/i915/i915_reg.h | 2 ++ 3 files changed, 40 insertions(+), 8 deletions(-)drivers/gpu/drm/i915/i915_gem_gtt.c +0
-0@@ -2256,16 +2256,13 @@ static bool needs_idle_maps(struct drm_i915_private *dev_priv)return IS_GEN5 (dev_priv) &&IS_MOBILE (dev_priv) &&intel_vtd_active (); }
Quoted 8 lines
Quoted 5 lines
drivers/gpu/drm/i915/i915_gem_gtt.c static bool needs_idle_maps(struct drm_i915_private *dev_priv) · line 2259
Quoted 2 lines
drivers/gpu/drm/i915/i915_gem_gtt.c +1
-1-void i915_check_and_clear_faults (struct drm_i915_private *dev_priv) +void __check_and_clear_faults (struct drm_i915_private *dev_priv)I am amazed that -Wextra doesn't complain. Try with sparse.
Hmm, no they didn't...
gen6_check_and_clear_faultsandgen8_check_and_clear_faultsare less controversial.
The missing static and lack of extern declaration. I know sparse catches it, I guess for K&R the order is just right for it not to complain.
Quoted 9 lines
Quoted 2 lines
There's an old thread where this was raised and how we are not clearing the faults early enough.
Ah, I found that thread [1], and as you said there, right now the
for_each_engineis a nop (with this patch at least gen8+ would do the right thing).Since gen6/gen7 engines are well known and can't change, do we really need
for_each_engine?
For consistency, yeah it would be nice to keep for_each_engine. We should be ok to do the clear around i915_driver_init_mmio (it's mmio, it has to be there ;), maybe worth pulling it into intel_engines_init_mmio() or i915_gem_init_mmio(). I'm learning towards sanitization from within intel_engines_init_mmio()
-Chris
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Quoted 48 lines · Chris Wilson
Quoting Michel Thierry (2017-11-10 23:42:31)
Quoted 28 lines · Chris Wilson
Quoted 27 lines
Quoting Michel Thierry (2017-11-10 19:01:16)
Quoted 26 lines
Until Haswell/Baytrail, the hardware used to have a per engine fault register (e.g.
0x4094- render fault register,0x4194- media fault register and so on). But since Broadwell, all these registers were combined into a singe one and the engine id stored in bits 14:12.Not only we should not been reading (and writing to) registers that do not exist, in platforms with VCS2 (SKL), the address that would belong this engine (
0x4494,VCS2_HW= 4) is already assigned to other register.References: IHD-OS-BDW-Vol 2c-11.15, page 75.
References: IHD-OS-SKL-Vol 2c-05.16, page 350.Signed-off-by Michel Thierry <[email protected]>--- drivers/gpu/drm/i915/i915_gem_gtt.c | 38 ++++++++++++++++++++++++++++++----- drivers/gpu/drm/i915/i915_gpu_error.c | 8 +++++--- drivers/gpu/drm/i915/i915_reg.h | 2 ++ 3 files changed, 40 insertions(+), 8 deletions(-)drivers/gpu/drm/i915/i915_gem_gtt.c +0
-0@@ -2256,16 +2256,13 @@ static bool needs_idle_maps(struct drm_i915_private *dev_priv)return IS_GEN5 (dev_priv) &&IS_MOBILE (dev_priv) &&intel_vtd_active (); }Quoted 8 lines
Quoted 5 lines
drivers/gpu/drm/i915/i915_gem_gtt.c static bool needs_idle_maps(struct drm_i915_private *dev_priv) · line 2259
Quoted 2 lines
drivers/gpu/drm/i915/i915_gem_gtt.c +1
-1-void i915_check_and_clear_faults (struct drm_i915_private *dev_priv) +void __check_and_clear_faults (struct drm_i915_private *dev_priv)I am amazed that -Wextra doesn't complain. Try with sparse.
Hmm, no they didn't...
gen6_check_and_clear_faultsandgen8_check_and_clear_faultsare less controversial.The missing static and lack of extern declaration. I know sparse catches it, I guess for K&R the order is just right for it not to complain.
Quoted 7 lines
Quoted 2 lines
There's an old thread where this was raised and how we are not clearing the faults early enough.
Ah, I found that thread [1], and as you said there, right now the
for_each_engineis a nop (with this patch at least gen8+ would do the right thing).
Quoted 8 lines
Quoted 2 lines
Since gen6/gen7 engines are well known and can't change, do we really need
for_each_engine?For consistency, yeah it would be nice to keep
for_each_engine. We should be ok to do the clear aroundi915_driver_init_mmio(it's mmio, it has to be there ;), maybe worth pulling it intointel_engines_init_mmio()ori915_gem_init_mmio(). I'm learning towards sanitization from withinintel_engines_init_mmio()
Ok, I'll move the check_and_clear_faults from intel_uncore_init to intel_engines_init_mmio (that's the one checking at driver load time).
Thanks,
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx