Skip to contentSkip to conversation

travis.yml: Run tcg tests with tci

PATCH Superseded Thomas Huth 3 participants 7 messages Nov 27, 2019 – Nov 28, 2019

v1 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 travis.yml: Run tcg tests with tci

So far we only have compile coverage for tci. But since commit 2f160e0f9797c7522bfd0d09218d0c9340a5137c ("tci: Add implementation for INDEX_op_ld16u_i64") has been included, we can also run the x86 TCG tests with tci, so let's enable them in Travis now.

---
 .travis.yml | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
.travis.yml +1 -1
index c09b6a0014..b0b634d484 100644
@@ -218,7 +218,7 @@ matrix:
     # We manually include builds which we disable "make check" for
     - env:
         - CONFIG="--enable-debug --enable-tcg-interpreter"
-        - TEST_CMD=""
+        - TEST_CMD="make run-tcg-tests-x86_64-softmmu V=1"
 
 
     # We don't need to exercise every backend with every front-end
Signature · 1 line
2.18.1
Reviewed-by Tested-by on v1 1/1
#1 lore raw
Quoted 6 lines · Thomas Huth

So far we only have compile coverage for tci. But since commit 2f160e0f9797c7522bfd0d09218d0c9340a5137c ("tci: Add implementation for INDEX_op_ld16u_i64") has been included, we can also run the x86 TCG tests with tci, so let's enable them in Travis now.

Ran for 26 min 28 sec:
https://travis-ci.org/philmd/qemu/jobs/617823541

Quoted 18 lines
---
  .travis.yml | 2 +-
  1 file changed, 1 insertion(+), 1 deletion(-)
.travis.yml +1 -1
index c09b6a0014..b0b634d484 100644
@@ -218,7 +218,7 @@ matrix:
      # We manually include builds which we disable "make check" for
      - env:
          - CONFIG="--enable-debug --enable-tcg-interpreter"
-        - TEST_CMD=""
+        - TEST_CMD="make run-tcg-tests-x86_64-softmmu V=1"
 
 
      # We don't need to exercise every backend with every front-end
#2 lore raw
In reply to Thomas Huth
Quoted 15 lines · Thomas Huth

So far we only have compile coverage for tci. But since commit 2f160e0f9797c7522bfd0d09218d0c9340a5137c ("tci: Add implementation for INDEX_op_ld16u_i64") has been included, we can also run the x86 TCG tests with tci, so let's enable them in Travis now.

---
 .travis.yml | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
.travis.yml +0 -0
index c09b6a0014..b0b634d484 100644

.travis.yml matrix: · line 218

Quoted 1 line
.travis.yml +0 -0
@@ -218,7 +218,7 @@ matrix:
     # We manually include builds which we disable "make check" for

this comment is out of date now (or rather has been for a while)

.travis.yml matrix: · line 220

Quoted 2 lines
.travis.yml +0 -0
     - env:
         - CONFIG="--enable-debug --enable-tcg-interpreter"

Perhaps as linux-user doesn't add much to testing the TCG we could limit by MAIN_SOFTMMU_TARGETS here?

.travis.yml matrix: · line 221

Quoted 2 lines
.travis.yml +1 -1
-        - TEST_CMD=""
+        - TEST_CMD="make run-tcg-tests-x86_64-softmmu V=1"

How about "make check-qtest check-tcg"

Which will exercise the moderate boot code tests of various architectures as well as x86_64 (if we include the --disable-docker configure stanza)

Quoted 3 lines
.travis.yml +0 -0
     # We don't need to exercise every backend with every front-end
Signature · 1 line
Alex Bennée
#3 lore raw
Quoted 21 lines · Alex Bennée
Quoted 15 lines · Thomas Huth

So far we only have compile coverage for tci. But since commit 2f160e0f9797c7522bfd0d09218d0c9340a5137c ("tci: Add implementation for INDEX_op_ld16u_i64") has been included, we can also run the x86 TCG tests with tci, so let's enable them in Travis now.

---
  .travis.yml | 2 +-
  1 file changed, 1 insertion(+), 1 deletion(-)
.travis.yml +0 -0
index c09b6a0014..b0b634d484 100644

.travis.yml matrix: · line 218

Quoted 1 line
.travis.yml +0 -0
@@ -218,7 +218,7 @@ matrix:
      # We manually include builds which we disable "make check" for

this comment is out of date now (or rather has been for a while)

Quoted 5 lines

.travis.yml matrix: · line 220

Quoted 2 lines
.travis.yml +0 -0
      - env:
          - CONFIG="--enable-debug --enable-tcg-interpreter"

Perhaps as linux-user doesn't add much to testing the TCG we could limit by MAIN_SOFTMMU_TARGETS here?

Good idea.

Quoted 8 lines

.travis.yml matrix: · line 221

Quoted 2 lines
.travis.yml +1 -1
-        - TEST_CMD=""
+        - TEST_CMD="make run-tcg-tests-x86_64-softmmu V=1"

How about "make check-qtest check-tcg"

Which will exercise the moderate boot code tests of various architectures as well as x86_64 (if we include the --disable-docker configure stanza)

Another good idea!

Quoted 3 lines
Quoted 3 lines
.travis.yml +0 -0
      # We don't need to exercise every backend with every front-end
#4 lore raw
In reply to Alex Bennée
Quoted 22 lines · Alex Bennée
Quoted 15 lines · Thomas Huth

So far we only have compile coverage for tci. But since commit 2f160e0f9797c7522bfd0d09218d0c9340a5137c ("tci: Add implementation for INDEX_op_ld16u_i64") has been included, we can also run the x86 TCG tests with tci, so let's enable them in Travis now.

---
  .travis.yml | 2 +-
  1 file changed, 1 insertion(+), 1 deletion(-)
.travis.yml +0 -0
index c09b6a0014..b0b634d484 100644

.travis.yml matrix: · line 218

Quoted 1 line
.travis.yml +0 -0
@@ -218,7 +218,7 @@ matrix:
      # We manually include builds which we disable "make check" for

this comment is out of date now (or rather has been for a while)

Quoted 5 lines

.travis.yml matrix: · line 220

Quoted 2 lines
.travis.yml +0 -0
      - env:
          - CONFIG="--enable-debug --enable-tcg-interpreter"

Perhaps as linux-user doesn't add much to testing the TCG we could limit by MAIN_SOFTMMU_TARGETS here?

Limiting is a good idea, but I think we should use the architectures that are supported by tests/boot-serial-test.c instead.

Quoted 7 lines

.travis.yml matrix: · line 221

Quoted 2 lines
.travis.yml +1 -1
-        - TEST_CMD=""
+        - TEST_CMD="make run-tcg-tests-x86_64-softmmu V=1"

How about "make check-qtest check-tcg"

Quoted 1 line

Which will exercise the moderate boot code tests of various

architectures as well as x86_64 (if we include the --disable-docker configure stanza)

I tried "check-tcg" before switching to "run-tcg-tests-x86_64-softmmu", but "check-tcg" then resulted in a docker error:

https://travis-ci.com/huth/qemu/jobs/260949876#L11974

... so looks like we indeed have to use --disable-docker in that case.

I'll have a try and send a v2 if that works out fine.

Thomas

#5 lore raw
Quoted 21 lines · Thomas Huth
Quoted 20 lines · Alex Bennée
Quoted 15 lines · Thomas Huth

So far we only have compile coverage for tci. But since commit 2f160e0f9797c7522bfd0d09218d0c9340a5137c ("tci: Add implementation for INDEX_op_ld16u_i64") has been included, we can also run the x86 TCG tests with tci, so let's enable them in Travis now.

---
  .travis.yml | 2 +-
  1 file changed, 1 insertion(+), 1 deletion(-)
.travis.yml +0 -0
index c09b6a0014..b0b634d484 100644

.travis.yml matrix: · line 218

Quoted 1 line
.travis.yml +0 -0
@@ -218,7 +218,7 @@ matrix:
      # We manually include builds which we disable "make check" for

this comment is out of date now (or rather has been for a while)

Quoted 8 lines
Quoted 5 lines

.travis.yml matrix: · line 220

Quoted 2 lines
.travis.yml +0 -0
      - env:
          - CONFIG="--enable-debug --enable-tcg-interpreter"

Perhaps as linux-user doesn't add much to testing the TCG we could limit
by MAIN_SOFTMMU_TARGETS here?

Limiting is a good idea, but I think we should use the architectures that are supported by tests/boot-serial-test.c instead.

Maybe a new env - TESTBOOT_SOFTMMU_TARGETS?

Quoted 19 lines
Quoted 7 lines

.travis.yml matrix: · line 221

Quoted 3 lines
.travis.yml +1 -1
-        - TEST_CMD=""
+        - TEST_CMD="make run-tcg-tests-x86_64-softmmu V=1"

How about "make check-qtest check-tcg"

Quoted 1 line

Which will exercise the moderate boot code tests of various

architectures as well as x86_64 (if we include the --disable-docker configure stanza)

I tried "check-tcg" before switching to "run-tcg-tests-x86_64-softmmu", but "check-tcg" then resulted in a docker error:

https://travis-ci.com/huth/qemu/jobs/260949876#L11974

... so looks like we indeed have to use --disable-docker in that case.

I'll have a try and send a v2 if that works out fine.

Thomas

Signature · 1 line
Alex Bennée
#6 lore raw
Quoted 26 lines · Alex Bennée
Quoted 23 lines · Thomas Huth
Quoted 22 lines · Alex Bennée
Quoted 15 lines · Thomas Huth

So far we only have compile coverage for tci. But since commit 2f160e0f9797c7522bfd0d09218d0c9340a5137c ("tci: Add implementation for INDEX_op_ld16u_i64") has been included, we can also run the x86 TCG tests with tci, so let's enable them in Travis now.

---
   .travis.yml | 2 +-
   1 file changed, 1 insertion(+), 1 deletion(-)
.travis.yml +0 -0
index c09b6a0014..b0b634d484 100644

.travis.yml matrix: · line 218

Quoted 1 line
.travis.yml +0 -0
@@ -218,7 +218,7 @@ matrix:
       # We manually include builds which we disable "make check" for

this comment is out of date now (or rather has been for a while)

Quoted 2 lines
.travis.yml +0 -0
       - env:
           - CONFIG="--enable-debug --enable-tcg-interpreter"
Quoted 8 lines
Quoted 6 lines
Quoted 3 lines

Perhaps as linux-user doesn't add much to testing the TCG we could limit
by MAIN_SOFTMMU_TARGETS here?

Limiting is a good idea, but I think we should use the architectures that are supported by tests/boot-serial-test.c instead.

Maybe a new env - TESTBOOT_SOFTMMU_TARGETS?

I tried something like that - but with all possible targets that are supported by boot-serial-tests, I hit the 50 minutes timeout from Travis. Looks like I've got to skip some, TCI is really slow. So I'll rather specify the target subset in the matrix entry directly.

Thomas