https://bz.apache.org/bugzilla/show_bug.cgi?id=64096

--- Comment #5 from Jacob Tolar <[email protected]> ---
Ah, I see, yes, use after release in `e = APR_BUCKET_NEXT(e)`; I didn't read
that carefully enough earlier. If the pool decided to free() the bucket
immediately bad things would happen.

Your patch to directly backport the change in trunk works fine with the test
described in my previous comment: no excess memory is retained after the end of
the request.

> See the description of MaxMemFree for what may happen here where large 
> request pool allocations might be held onto for subsequent requests, 
> appearing allocated to the OS.

Makes sense, although I've had it set to the default (looks like 2 MB) in all
my tests, nothing close to the amount of memory that is actually retained after
the end of the request. I didn't trace through very carefully in gdb, though, I
probably just missed something.

-- 
You are receiving this mail because:
You are the assignee for the bug.
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to