https://bugs.kde.org/show_bug.cgi?id=516289
Bug ID: 516289
Summary: illumos lsframe2 regtest fails
Classification: Developer tools
Product: valgrind
Version First 3.27 GIT
Reported In:
Platform: Compiled Sources
OS: Other
Status: REPORTED
Severity: normal
Priority: NOR
Component: general
Assignee: [email protected]
Reporter: [email protected]
Target Milestone: ---
There is a lot going on in this failure, not sure how it is all related.
The testcase is fairly simple. There is a recursive function that takes a depth
argument that gets decremented from 64 to 0 and then the recursion is undone.
The function has a 1Mbyte local array of ints. Before recursing it gets
initialised to "element_index * depth". After recursing the sum of all elements
in the array is calculated and that is the return value.
Nothing too special there, other than the client stack size exceeding the
default and requiring --main-stacksize to make it bigger. The testcase runs OK
on both FreeBSD and Linux.
On illumos I see (not sure what is related)
1. Debug traces are corrupted with a control-B character appearing in the same
row on longer lines.
2. gdbserver watch point messages are getting printed
3. On the 8th call to the recursive function there are memcheck invalid read
errors.
The result is something like 25 unique errors, 700k total errors.
The lsframe1 test passes OK (like lsframe2 but just 64Mbyte array in main, no
recursion).
These are the only 2 tests that use --main-stacksize (on Linux and Solaris).
--
You are receiving this mail because:
You are watching all bug changes.
