| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
We used to generate a choice table and do its normalization for all
present syscalls, also it was not considered during the /prio page
generation.
Enabled/disabled syscalls were accounted for in the choice table
construction, but there's a chance that the resulting table was still
somewhat skewed.
The situation must have become worse after several thousands of auto
syscalls were added.
|
| |
|
|
|
| |
Fix checking of Logf, it has string in 0-th arg.
Add checking of t.Errorf/Fatalf.
|
| |
|
|
|
|
|
|
|
|
|
|
| |
Two changes:
1) Instead of multiplying static and dynamic prios, add them 1:1.
2) For dynamic priorities, limit the effect of too frequent call
combinations by taking a sqrt() of the value.
On syz-testbed experiments, the updated algorithm triggers 5-10%
more different crash types.
As before, there is no big theory behind the approach :)
|
| |
|
|
|
|
|
|
| |
Add package with RaceEnabled const that can be used in test
to skip long tests in race mode.
Switch existing tests to use the new package.
Update #2886
|
| |
|
|
|
|
|
| |
Switch from float32 to int32.
Float32 is super slow in arm emulation.
Plus flats are generally non-deterministic due to order of operations,
so we needed to do additional sorts to deal with that. Now we don't.
|
| | |
|
| |
|
|
|
|
| |
The test is random and needs some large number of iterations to pass.
It failed for me after an unrelated change in descriptions.
So bump number of iterations.
|
| |
|
|
| |
Make it faster + disable in race mode (still too slow).
|
| |
|
|
|
|
|
| |
Floats bite.
We interated over uses map non-deterministically,
which would be fine overall except that it may
break floats due to rounding.
|
| |
|
|
| |
Reduce size of tests that run for more than 1s in short mode.
|
| |
|
|
|
| |
Update #477
Update #502
|
| |
|
|
|
|
|
|
| |
Linter says:
prog/prio_test.go:68:15: `probablistic` is a misspelling of `probabilistic` (misspell)
// for this probablistic test.
^
|
| |
|
|
|
|
| |
With -short and -race we get only 10 iterations
which is not enough for this probablistic test.
Use at least 100 interations always.
|
| |
|
|
| |
Update #1380
|
| | |
|
|
|
Based on twitter bug report:
https://twitter.com/panicaII/status/1035058001269248000
|
