Skip to contentSkip to conversation

i-g-t/tests: Drop gem_seqno_wrap, gem_pin, gem_hangcheck_forcewake

PATCH In review Abdiel Janulgue 5 participants 6 messages Oct 10, 2017 – Oct 12, 2017 Arm/Amlogic Meson SoC support

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 i-g-t/tests: Drop gem_seqno_wrap, gem_pin, gem_hangcheck_forcewake

This improves the GEM tests section of I-G-T to make it more suitable for CI testing

---
 tests/Makefile.sources          |   3 -
 tests/gem_hangcheck_forcewake.c | 123 ----------
 tests/gem_pin.c                 | 248 ---------------------
 tests/gem_seqno_wrap.c          | 483 ----------------------------------------
 tests/meson.build               |   3 -
 5 files changed, 860 deletions(-)
 delete mode 100644 tests/gem_hangcheck_forcewake.c
 delete mode 100644 tests/gem_pin.c
 delete mode 100644 tests/gem_seqno_wrap.c
tests/Makefile.sources +0 -3
index bb6652e..7f52298 100644
@@ -99,7 +99,6 @@ TESTS_progs = \
 	gem_gtt_cpu_tlb \
 	gem_gtt_hog \
 	gem_gtt_speed \
-	gem_hangcheck_forcewake \
 	gem_largeobject \
 	gem_linear_blits \
 	gem_lut_handle \
@@ -112,7 +111,6 @@ TESTS_progs = \
 	gem_mocs_settings \
 	gem_partial_pwrite_pread \
 	gem_persistent_relocs \
-	gem_pin \
 	gem_pipe_control_store_loop \
 	gem_ppgtt \
 	gem_pread \
@@ -134,7 +132,6 @@ TESTS_progs = \
 	gem_ring_sync_copy \
 	gem_ring_sync_loop \
 	gem_ringfill \
-	gem_seqno_wrap \
 	gem_set_tiling_vs_blt \
 	gem_set_tiling_vs_gtt \
 	gem_set_tiling_vs_pwrite \
tests/gem_hangcheck_forcewake.c deleted +0 -123

123 lines not on this page · raw message

tests/gem_pin.c deleted +0 -248

248 lines not on this page · raw message

tests/gem_seqno_wrap.c deleted +0 -483

483 lines not on this page · raw message

tests/meson.build +0 -3
index 6cb3584..8abc06e 100644
@@ -79,7 +79,6 @@ test_progs = [
 	'gem_gtt_cpu_tlb',
 	'gem_gtt_hog',
 	'gem_gtt_speed',
-	'gem_hangcheck_forcewake',
 	'gem_largeobject',
 	'gem_linear_blits',
 	'gem_lut_handle',
@@ -92,7 +91,6 @@ test_progs = [
 	'gem_mocs_settings',
 	'gem_partial_pwrite_pread',
 	'gem_persistent_relocs',
-	'gem_pin',
 	'gem_pipe_control_store_loop',
 	'gem_ppgtt',
 	'gem_pread',
@@ -114,7 +112,6 @@ test_progs = [
 	'gem_ring_sync_copy',
 	'gem_ring_sync_loop',
 	'gem_ringfill',
-	'gem_seqno_wrap',
 	'gem_set_tiling_vs_blt',
 	'gem_set_tiling_vs_gtt',
 	'gem_set_tiling_vs_pwrite',
Signature · 6 lines
2.7.4

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Reviewed-by on v1 1/1
#1 lore raw

+ Chris and Mika for R-b's as they're the test authors
+ Arek, Tomi for CI blacklist reduction

Quoted 5 lines · Abdiel Janulgue

This improves the GEM tests section of I-G-T to make it more suitable for CI testing

For future reference, lets have the tags in chronological order.

(Suggested-by: if any) S-o-b, Cc, Reviewed-by etc...

This checks with the test blacklisting TODO items.

Arek, Tomi; After this is merged, the the following lines can be removed from CI blacklist.

igt@gem_pin(@.*)?
igt@gem_seqno_wrap(@.*)?
igt@gem_hangcheck_forcewake(@.*)?

Regards, Joonas

Signature · 7 lines
Joonas Lahtinen
Open Source Technology Center
Intel Corporation
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#2 lore raw
In reply to Abdiel Janulgue
Quoted 2 lines · Abdiel Janulgue

This improves the GEM tests section of I-G-T to make it more suitable for CI testing

Can you provide a little more details on what is the rationale behind this choice. Are the tests being removed only because they do not fit the CI guidelines or is it because they do not offer any valuable coverage?

Maybe having three separate patches, one per test would be better in my opinion since the tests look to have not much in common.

-Antonio

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

#3 lore raw

Quoting Antonio Argenziano (2017-10-10 18:12:17)

Quoted 9 lines · Abdiel Janulgue
Quoted 2 lines

This improves the GEM tests section of I-G-T to make it more suitable for CI testing

Can you provide a little more details on what is the rationale behind this choice. Are the tests being removed only because they do not fit the CI guidelines or is it because they do not offer any valuable coverage?

Hear, hear.

gem_seqno_wrap is defunct as the debug API withered away. The handling of wraparound under many different workloads is tested by gem_exec_whisper. The intention is to may seqno wrap handling a kselftest. That hasn't happened yet, gem_exec_whisper is not run wholy by CI, but nevertheless gem_seqno_wrap is still defunct.

gem_pin is a useful ABI I still shed a tear over. But it's time has passed and it's not coming back.

gem_hangcheck_forcewake has been superseded by drv_hangman, and a lot of other very extensive hangchecking that is not run by CI. Something about hang testing taking longer than the sun to go nova, and still will never be completely reliable...
-Chris
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

#4 lore raw
Quoted 19 lines · Chris Wilson

Quoting Antonio Argenziano (2017-10-10 18:12:17)

Quoted 9 lines · Abdiel Janulgue
Quoted 2 lines

This improves the GEM tests section of I-G-T to make it more suitable for CI testing

Can you provide a little more details on what is the rationale behind this choice. Are the tests being removed only because they do not fit the CI guidelines or is it because they do not offer any valuable coverage?

Hear, hear.

gem_seqno_wrap is defunct as the debug API withered away. The handling of wraparound under many different workloads is tested by gem_exec_whisper. The intention is to may seqno wrap handling a kselftest. That hasn't happened yet, gem_exec_whisper is not run wholy by CI, but nevertheless gem_seqno_wrap is still defunct.

Quoted 9 lines

gem_pin is a useful ABI I still shed a tear over. But it's time has passed and it's not coming back.

gem_hangcheck_forcewake has been superseded by drv_hangman, and a lot of other very extensive hangchecking that is not run by CI. Something about hang testing taking longer than the sun to go nova, and still will never be completely reliable...
-Chris

Thanks for the explanation Chris.

This convinces me more that the commit needs to be split into three separate patches. One for each test with an explanation of why it goes away.

-Antonio
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

#5 lore raw
In reply to Abdiel Janulgue
Quoted 6 lines · Abdiel Janulgue

This improves the GEM tests section of I-G-T to make it more suitable for CI testing

---

When resending this patch, please use the correct subjectprefix so CI can pick it up.

Signature · 5 lines
Petri Latvala
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx