You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Is your feature request related to a problem? Please describe.
We currently have a bunch of ARM64 SyzOS tests in sys/linux/test. I run them locally using syz-manager -mode run-tests, but want to continuously run them somewhere to detect regressions.
Describe the solution you'd like
Because the manager already has all the building blocks to execute those tests, I was thinking about adding a config option that would run the tests at manager startup and report their failures as bugs.
This can be very helpful for verifying the behavior of SyzOS guest code, but can be also used to validate pseudo-syscalls that can silently stop working now.
An alternative approach would be to add a separate Github CI test target that would effectively run syz-manager -mode run-tests.
Additional context
In general, tests in sys/linux/tests do not always validate any functionality. In certain modes, some syscalls may fail leading to test failures. So we'll need to split off the tests meant to be run continuously.
The text was updated successfully, but these errors were encountered:
Is your feature request related to a problem? Please describe.
We currently have a bunch of ARM64 SyzOS tests in
sys/linux/test
. I run them locally usingsyz-manager -mode run-tests
, but want to continuously run them somewhere to detect regressions.Describe the solution you'd like
Because the manager already has all the building blocks to execute those tests, I was thinking about adding a config option that would run the tests at manager startup and report their failures as bugs.
This can be very helpful for verifying the behavior of SyzOS guest code, but can be also used to validate pseudo-syscalls that can silently stop working now.
An alternative approach would be to add a separate Github CI test target that would effectively run
syz-manager -mode run-tests
.Additional context
In general, tests in
sys/linux/tests
do not always validate any functionality. In certain modes, some syscalls may fail leading to test failures. So we'll need to split off the tests meant to be run continuously.The text was updated successfully, but these errors were encountered: