New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Osx threadfix #23967
Osx threadfix #23967
Conversation
Is there some way of limiting the number of writes instead of doing this? |
sure, but that runs a bit counter to the purpose of the test, they idea (in part) was to not only test the sanity of rw and rcu locks, but to provide a level of performance comparison. What we could do is, instead of wrapping, when we hit MAX_ULLONG, just leave the counter there so that we never tripped over the wrapping case. Would that be acceptable? |
If you are running over the end of the unsigned long long then you need to implement a "carry" into another variable (perhaps even another unsigned long long) each time you reach the wrapping point. That does sound like overkill for this context - but I also find the implementation here more than a little strange. |
@nhorman I do not believe a wrap around 2^64 bit value is happening here. Or, the starting point was not a zero value. If the test run for 1000 seconds (which it is not running) it would mean doing 18 milions iterations in every nanosecond - that's obviously impossible. There must be some other issue in play. |
If we're worried for a wrap around at 2^64, why not clear the first bit? |
Stochastic failures in the RCU test on MACOSX are occuring. Due to beta release, disabling this test on MACOSX until post 3.3 release
Based on OTC discussion, for the purposes of the beta release, since we are not yet using rcu, I've updated this PR to disable the rcu torture test on macosx. Will re-enable prior to introduction of usage post release |
OK with urgent |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Approved. I have one minor comment - but approved either way.
Agree urgent.
@@ -292,7 +293,6 @@ static int writer2_iterations = 0; | |||
static uint64_t *writer_ptr = NULL; | |||
static uint64_t global_ctr = 0; | |||
static int rcu_torture_result = 1; | |||
|
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This blank line removal looks odd. In my mind it is better with it.
Have we raised an issue to remember to come back and fix this? If so what is the number? |
merged to master and 3.3 |
Stochastic failures in the RCU test on MACOSX are occuring. Due to beta release, disabling this test on MACOSX until post 3.3 release Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from #23967)
Stochastic failures in the RCU test on MACOSX are occuring. Due to beta release, disabling this test on MACOSX until post 3.3 release Reviewed-by: Tomas Mraz <tomas@openssl.org> Reviewed-by: Matt Caswell <matt@openssl.org> (Merged from openssl#23967) Fixing Format, Grammar, and Spelling. - Changed many comment blocks to be the correct "/* */" format. - Fixed general typos, mispellings, or incorrect grammar. CLA: trivial
test build for debugging, as local operation on the development m1 doesn't reproduce failure