* [igt-dev] [PATCH igt] lib: Fix MI_BATCH_BUFFER_START for hang injection
@ 2018-03-02 16:13 Chris Wilson
2018-03-02 17:09 ` [Intel-gfx] " Ville Syrjälä
` (2 more replies)
0 siblings, 3 replies; 5+ messages in thread
From: Chris Wilson @ 2018-03-02 16:13 UTC (permalink / raw)
To: intel-gfx; +Cc: igt-dev
A couple of bugs inside the hang injector, the worst being that the
presumed_offset of the reloc didn't match the batch; so if the reloc was
skipped (as the presumed_offset matched the reloc offset), the batch
wasn't updated and so we may not have generated a hanging batch at all!
Secondly, the MI_BATCH_BUFFER_START was not correct for all gen.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
---
lib/igt_gt.c | 28 +++++++++++++++++++++-------
1 file changed, 21 insertions(+), 7 deletions(-)
diff --git a/lib/igt_gt.c b/lib/igt_gt.c
index e630550b..799ca1ae 100644
--- a/lib/igt_gt.c
+++ b/lib/igt_gt.c
@@ -276,6 +276,7 @@ igt_hang_t igt_hang_ctx(int fd,
uint32_t b[16];
unsigned ban;
unsigned len;
+ int gen;
igt_require_hang_ring(fd, ring);
@@ -310,12 +311,26 @@ igt_hang_t igt_hang_ctx(int fd,
memset(b, 0xc5, sizeof(b));
- len = 2;
- if (intel_gen(intel_get_drm_devid(fd)) >= 8)
+ len = 0;
+ gen = intel_gen(intel_get_drm_devid(fd));
+ if (gen >= 8) {
+ b[len++] = MI_BATCH_BUFFER_START | 1 << 8 | 1;
+ b[len++] = 0;
+ b[len++] = 0;
+ } else if (gen >= 6) {
+ b[len++] = MI_BATCH_BUFFER_START | 1 << 8;
+ b[len++] = 0;
+ } else {
+ b[len++] = MI_BATCH_BUFFER_START | 2 << 6;
+ b[len] = 0;
+ if (gen < 4) {
+ b[len] |= 1;
+ reloc.delta = 1;
+ }
len++;
- b[0] = MI_BATCH_BUFFER_START | (len - 2);
- b[len] = MI_BATCH_BUFFER_END;
- b[len+1] = MI_NOOP;
+ }
+ b[len++] = MI_BATCH_BUFFER_END;
+ b[len] = MI_NOOP;
gem_write(fd, exec.handle, 0, b, sizeof(b));
reloc.offset = sizeof(uint32_t);
@@ -364,8 +379,7 @@ void igt_post_hang_ring(int fd, igt_hang_t arg)
if (arg.handle == 0)
return;
- gem_set_domain(fd, arg.handle,
- I915_GEM_DOMAIN_GTT, I915_GEM_DOMAIN_GTT);
+ gem_sync(fd, arg.handle);
gem_close(fd, arg.handle);
context_set_ban(fd, arg.ctx, arg.ban);
--
2.16.2
_______________________________________________
igt-dev mailing list
igt-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/igt-dev
^ permalink raw reply related [flat|nested] 5+ messages in thread* Re: [Intel-gfx] [PATCH igt] lib: Fix MI_BATCH_BUFFER_START for hang injection
2018-03-02 16:13 [igt-dev] [PATCH igt] lib: Fix MI_BATCH_BUFFER_START for hang injection Chris Wilson
@ 2018-03-02 17:09 ` Ville Syrjälä
2018-03-02 19:06 ` [igt-dev] " Chris Wilson
2018-03-02 20:35 ` [igt-dev] ✓ Fi.CI.BAT: success for " Patchwork
2018-03-03 0:08 ` [igt-dev] ✓ Fi.CI.IGT: " Patchwork
2 siblings, 1 reply; 5+ messages in thread
From: Ville Syrjälä @ 2018-03-02 17:09 UTC (permalink / raw)
To: Chris Wilson; +Cc: igt-dev, intel-gfx
On Fri, Mar 02, 2018 at 04:13:46PM +0000, Chris Wilson wrote:
> A couple of bugs inside the hang injector, the worst being that the
> presumed_offset of the reloc didn't match the batch; so if the reloc was
> skipped (as the presumed_offset matched the reloc offset), the batch
> wasn't updated and so we may not have generated a hanging batch at all!
> Secondly, the MI_BATCH_BUFFER_START was not correct for all gen.
>
> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
> ---
> lib/igt_gt.c | 28 +++++++++++++++++++++-------
> 1 file changed, 21 insertions(+), 7 deletions(-)
>
> diff --git a/lib/igt_gt.c b/lib/igt_gt.c
> index e630550b..799ca1ae 100644
> --- a/lib/igt_gt.c
> +++ b/lib/igt_gt.c
> @@ -276,6 +276,7 @@ igt_hang_t igt_hang_ctx(int fd,
> uint32_t b[16];
> unsigned ban;
> unsigned len;
> + int gen;
>
> igt_require_hang_ring(fd, ring);
>
> @@ -310,12 +311,26 @@ igt_hang_t igt_hang_ctx(int fd,
>
> memset(b, 0xc5, sizeof(b));
>
> - len = 2;
> - if (intel_gen(intel_get_drm_devid(fd)) >= 8)
> + len = 0;
> + gen = intel_gen(intel_get_drm_devid(fd));
> + if (gen >= 8) {
> + b[len++] = MI_BATCH_BUFFER_START | 1 << 8 | 1;
> + b[len++] = 0;
> + b[len++] = 0;
> + } else if (gen >= 6) {
> + b[len++] = MI_BATCH_BUFFER_START | 1 << 8;
> + b[len++] = 0;
ppgtt for gen6+
> + } else {
> + b[len++] = MI_BATCH_BUFFER_START | 2 << 6;
> + b[len] = 0;
> + if (gen < 4) {
> + b[len] |= 1;
> + reloc.delta = 1;
> + }
> len++;
gtt secure for gen4/5
gtt non-secure for gen2/3
How does the security thing work on ctg/ilk for chained batches? The spec
is a wee bit unclear. It says the security bit for chained batches is
ignored, but then it also says non-secure batches can't access the gtt.
That was the case for MI_STORE_DWORD if I recall your earlier patch
correctly. So if we don't execute the first batch as secure the chained
MI_BB_START gets nopped out maybe?
Hmm. Now I wonder how the earlier MI_STORE_DWORD thing worked on pre-ctg
with a non-secure batch? Wasn't the hardware supposed to nop those out
entirely? /me confused
Anyways the new code looks at least more correct than the old one so
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
> - b[0] = MI_BATCH_BUFFER_START | (len - 2);
> - b[len] = MI_BATCH_BUFFER_END;
> - b[len+1] = MI_NOOP;
> + }
> + b[len++] = MI_BATCH_BUFFER_END;
> + b[len] = MI_NOOP;
> gem_write(fd, exec.handle, 0, b, sizeof(b));
>
> reloc.offset = sizeof(uint32_t);
> @@ -364,8 +379,7 @@ void igt_post_hang_ring(int fd, igt_hang_t arg)
> if (arg.handle == 0)
> return;
>
> - gem_set_domain(fd, arg.handle,
> - I915_GEM_DOMAIN_GTT, I915_GEM_DOMAIN_GTT);
> + gem_sync(fd, arg.handle);
> gem_close(fd, arg.handle);
>
> context_set_ban(fd, arg.ctx, arg.ban);
> --
> 2.16.2
>
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/intel-gfx
--
Ville Syrjälä
Intel OTC
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
^ permalink raw reply [flat|nested] 5+ messages in thread* Re: [igt-dev] [Intel-gfx] [PATCH igt] lib: Fix MI_BATCH_BUFFER_START for hang injection
2018-03-02 17:09 ` [Intel-gfx] " Ville Syrjälä
@ 2018-03-02 19:06 ` Chris Wilson
0 siblings, 0 replies; 5+ messages in thread
From: Chris Wilson @ 2018-03-02 19:06 UTC (permalink / raw)
To: Ville Syrjälä; +Cc: igt-dev, intel-gfx
Quoting Ville Syrjälä (2018-03-02 17:09:29)
> On Fri, Mar 02, 2018 at 04:13:46PM +0000, Chris Wilson wrote:
> > A couple of bugs inside the hang injector, the worst being that the
> > presumed_offset of the reloc didn't match the batch; so if the reloc was
> > skipped (as the presumed_offset matched the reloc offset), the batch
> > wasn't updated and so we may not have generated a hanging batch at all!
> > Secondly, the MI_BATCH_BUFFER_START was not correct for all gen.
> >
> > Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
> > ---
> > lib/igt_gt.c | 28 +++++++++++++++++++++-------
> > 1 file changed, 21 insertions(+), 7 deletions(-)
> >
> > diff --git a/lib/igt_gt.c b/lib/igt_gt.c
> > index e630550b..799ca1ae 100644
> > --- a/lib/igt_gt.c
> > +++ b/lib/igt_gt.c
> > @@ -276,6 +276,7 @@ igt_hang_t igt_hang_ctx(int fd,
> > uint32_t b[16];
> > unsigned ban;
> > unsigned len;
> > + int gen;
> >
> > igt_require_hang_ring(fd, ring);
> >
> > @@ -310,12 +311,26 @@ igt_hang_t igt_hang_ctx(int fd,
> >
> > memset(b, 0xc5, sizeof(b));
> >
> > - len = 2;
> > - if (intel_gen(intel_get_drm_devid(fd)) >= 8)
> > + len = 0;
> > + gen = intel_gen(intel_get_drm_devid(fd));
> > + if (gen >= 8) {
> > + b[len++] = MI_BATCH_BUFFER_START | 1 << 8 | 1;
> > + b[len++] = 0;
> > + b[len++] = 0;
> > + } else if (gen >= 6) {
> > + b[len++] = MI_BATCH_BUFFER_START | 1 << 8;
> > + b[len++] = 0;
>
> ppgtt for gen6+
>
> > + } else {
> > + b[len++] = MI_BATCH_BUFFER_START | 2 << 6;
> > + b[len] = 0;
> > + if (gen < 4) {
> > + b[len] |= 1;
> > + reloc.delta = 1;
> > + }
> > len++;
>
> gtt secure for gen4/5
> gtt non-secure for gen2/3
>
> How does the security thing work on ctg/ilk for chained batches? The spec
> is a wee bit unclear. It says the security bit for chained batches is
> ignored, but then it also says non-secure batches can't access the gtt.
> That was the case for MI_STORE_DWORD if I recall your earlier patch
> correctly. So if we don't execute the first batch as secure the chained
> MI_BB_START gets nopped out maybe?
Yes, as soon as we lose the privilege bit (usually when the kernel does
the first MI_BB to us), we can regain privilege in this batch chain. How
this works with the ppgtt is a mystery, but it also functions
differently if it's not enabled.
> Hmm. Now I wonder how the earlier MI_STORE_DWORD thing worked on pre-ctg
> with a non-secure batch? Wasn't the hardware supposed to nop those out
> entirely? /me confused
Aiui, before ctg/ilk they remained unprivileged ops.
-Chris
_______________________________________________
igt-dev mailing list
igt-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/igt-dev
^ permalink raw reply [flat|nested] 5+ messages in thread
* [igt-dev] ✓ Fi.CI.BAT: success for lib: Fix MI_BATCH_BUFFER_START for hang injection
2018-03-02 16:13 [igt-dev] [PATCH igt] lib: Fix MI_BATCH_BUFFER_START for hang injection Chris Wilson
2018-03-02 17:09 ` [Intel-gfx] " Ville Syrjälä
@ 2018-03-02 20:35 ` Patchwork
2018-03-03 0:08 ` [igt-dev] ✓ Fi.CI.IGT: " Patchwork
2 siblings, 0 replies; 5+ messages in thread
From: Patchwork @ 2018-03-02 20:35 UTC (permalink / raw)
To: Chris Wilson; +Cc: igt-dev
== Series Details ==
Series: lib: Fix MI_BATCH_BUFFER_START for hang injection
URL : https://patchwork.freedesktop.org/series/39292/
State : success
== Summary ==
IGT patchset tested on top of latest successful build
bddfb8dd3c1767f13d2af578d5c3d897fddf0dcd igt/gem_ctx_switch: Exercise all engines at once
with latest DRM-Tip kernel build CI_DRM_3867
4f4e4dd52a30 drm-tip: 2018y-03m-02d-16h-28m-21s UTC integration manifest
No testlist changes.
fi-bdw-5557u total:288 pass:267 dwarn:0 dfail:0 fail:0 skip:21 time:418s
fi-bdw-gvtdvm total:288 pass:264 dwarn:0 dfail:0 fail:0 skip:24 time:430s
fi-blb-e6850 total:288 pass:223 dwarn:1 dfail:0 fail:0 skip:64 time:373s
fi-bsw-n3050 total:288 pass:242 dwarn:0 dfail:0 fail:0 skip:46 time:485s
fi-bwr-2160 total:288 pass:183 dwarn:0 dfail:0 fail:0 skip:105 time:279s
fi-bxt-dsi total:288 pass:258 dwarn:0 dfail:0 fail:0 skip:30 time:479s
fi-bxt-j4205 total:288 pass:259 dwarn:0 dfail:0 fail:0 skip:29 time:487s
fi-byt-j1900 total:288 pass:253 dwarn:0 dfail:0 fail:0 skip:35 time:468s
fi-byt-n2820 total:288 pass:249 dwarn:0 dfail:0 fail:0 skip:39 time:460s
fi-cfl-8700k total:288 pass:260 dwarn:0 dfail:0 fail:0 skip:28 time:391s
fi-cfl-s2 total:288 pass:262 dwarn:0 dfail:0 fail:0 skip:26 time:559s
fi-cfl-u total:288 pass:262 dwarn:0 dfail:0 fail:0 skip:26 time:503s
fi-elk-e7500 total:288 pass:229 dwarn:0 dfail:0 fail:0 skip:59 time:409s
fi-gdg-551 total:288 pass:179 dwarn:0 dfail:0 fail:1 skip:108 time:288s
fi-glk-1 total:288 pass:260 dwarn:0 dfail:0 fail:0 skip:28 time:508s
fi-hsw-4770 total:288 pass:261 dwarn:0 dfail:0 fail:0 skip:27 time:394s
fi-ilk-650 total:288 pass:228 dwarn:0 dfail:0 fail:0 skip:60 time:416s
fi-ivb-3520m total:288 pass:259 dwarn:0 dfail:0 fail:0 skip:29 time:447s
fi-ivb-3770 total:288 pass:255 dwarn:0 dfail:0 fail:0 skip:33 time:410s
fi-kbl-7500u total:288 pass:263 dwarn:1 dfail:0 fail:0 skip:24 time:450s
fi-kbl-7560u total:288 pass:269 dwarn:0 dfail:0 fail:0 skip:19 time:492s
fi-kbl-7567u total:288 pass:268 dwarn:0 dfail:0 fail:0 skip:20 time:449s
fi-kbl-r total:288 pass:261 dwarn:0 dfail:0 fail:0 skip:27 time:494s
fi-pnv-d510 total:288 pass:222 dwarn:1 dfail:0 fail:0 skip:65 time:587s
fi-skl-6260u total:288 pass:268 dwarn:0 dfail:0 fail:0 skip:20 time:425s
fi-skl-6600u total:288 pass:261 dwarn:0 dfail:0 fail:0 skip:27 time:502s
fi-skl-6700hq total:288 pass:262 dwarn:0 dfail:0 fail:0 skip:26 time:517s
fi-skl-6700k2 total:288 pass:264 dwarn:0 dfail:0 fail:0 skip:24 time:488s
fi-skl-6770hq total:288 pass:268 dwarn:0 dfail:0 fail:0 skip:20 time:473s
fi-skl-guc total:288 pass:260 dwarn:0 dfail:0 fail:0 skip:28 time:411s
fi-skl-gvtdvm total:288 pass:265 dwarn:0 dfail:0 fail:0 skip:23 time:428s
fi-snb-2520m total:288 pass:248 dwarn:0 dfail:0 fail:0 skip:40 time:526s
fi-snb-2600 total:288 pass:248 dwarn:0 dfail:0 fail:0 skip:40 time:400s
== Logs ==
For more details see: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_1048/issues.html
_______________________________________________
igt-dev mailing list
igt-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/igt-dev
^ permalink raw reply [flat|nested] 5+ messages in thread
* [igt-dev] ✓ Fi.CI.IGT: success for lib: Fix MI_BATCH_BUFFER_START for hang injection
2018-03-02 16:13 [igt-dev] [PATCH igt] lib: Fix MI_BATCH_BUFFER_START for hang injection Chris Wilson
2018-03-02 17:09 ` [Intel-gfx] " Ville Syrjälä
2018-03-02 20:35 ` [igt-dev] ✓ Fi.CI.BAT: success for " Patchwork
@ 2018-03-03 0:08 ` Patchwork
2 siblings, 0 replies; 5+ messages in thread
From: Patchwork @ 2018-03-03 0:08 UTC (permalink / raw)
To: Chris Wilson; +Cc: igt-dev
== Series Details ==
Series: lib: Fix MI_BATCH_BUFFER_START for hang injection
URL : https://patchwork.freedesktop.org/series/39292/
State : success
== Summary ==
---- Possible new issues:
Test pm_rc6_residency:
Subgroup rc6-accuracy:
skip -> PASS (shard-snb)
---- Known issues:
Test kms_chv_cursor_fail:
Subgroup pipe-b-256x256-bottom-edge:
dmesg-warn -> PASS (shard-snb) fdo#105185 +1
Test kms_fbcon_fbt:
Subgroup fbc-suspend:
incomplete -> PASS (shard-hsw) fdo#105087
Test kms_flip:
Subgroup 2x-flip-vs-expired-vblank-interruptible:
pass -> FAIL (shard-hsw) fdo#102887
Test kms_frontbuffer_tracking:
Subgroup fbc-suspend:
pass -> INCOMPLETE (shard-hsw) fdo#103540
Test kms_pipe_crc_basic:
Subgroup suspend-read-crc-pipe-a:
pass -> INCOMPLETE (shard-snb) fdo#103375
Test kms_rotation_crc:
Subgroup primary-rotation-180:
pass -> FAIL (shard-snb) fdo#103925 +1
fdo#105185 https://bugs.freedesktop.org/show_bug.cgi?id=105185
fdo#105087 https://bugs.freedesktop.org/show_bug.cgi?id=105087
fdo#102887 https://bugs.freedesktop.org/show_bug.cgi?id=102887
fdo#103540 https://bugs.freedesktop.org/show_bug.cgi?id=103540
fdo#103375 https://bugs.freedesktop.org/show_bug.cgi?id=103375
fdo#103925 https://bugs.freedesktop.org/show_bug.cgi?id=103925
shard-apl total:3463 pass:1821 dwarn:1 dfail:0 fail:7 skip:1632 time:11902s
shard-hsw total:3407 pass:1742 dwarn:1 dfail:0 fail:2 skip:1660 time:11514s
shard-snb total:3372 pass:1322 dwarn:2 dfail:0 fail:3 skip:2044 time:6805s
Blacklisted hosts:
shard-kbl total:3349 pass:1878 dwarn:1 dfail:0 fail:7 skip:1461 time:8856s
== Logs ==
For more details see: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_1048/shards.html
_______________________________________________
igt-dev mailing list
igt-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/igt-dev
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2018-03-03 0:08 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2018-03-02 16:13 [igt-dev] [PATCH igt] lib: Fix MI_BATCH_BUFFER_START for hang injection Chris Wilson
2018-03-02 17:09 ` [Intel-gfx] " Ville Syrjälä
2018-03-02 19:06 ` [igt-dev] " Chris Wilson
2018-03-02 20:35 ` [igt-dev] ✓ Fi.CI.BAT: success for " Patchwork
2018-03-03 0:08 ` [igt-dev] ✓ Fi.CI.IGT: " Patchwork
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox