Is valgrind reporting a false positive inside aesni_cbc_hmac_sha1_cipher.c? #23780
slontis
started this conversation in
General Discussion
Replies: 3 comments
-
Thats odd. I can reproduce your errors running your exact command line, but setting LD_LIBRARY_PATH on the valgrind command is not the way I've read that you should run the command (in the event that the environment doesn't get passed on fork). If I run this command instead:
The errors are not reproduced. Additionally, if I build the library with enable-asan and just run:
I get no errors. As to weather or not its a true positive, I can't being to tell |
Beta Was this translation helpful? Give feedback.
0 replies
-
I feel fairly sure I have seen this before and we treated it as a false positive. But I don't recall any of the details. |
Beta Was this translation helpful? Give feedback.
0 replies
-
Note that asan doesn't report uninitialized values. I think you need msan for that.
|
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
Running the following on master
Results in valgrind results of
'Use of uninitialised value of size 8'
When it does
Where
Has anyone seen this before, and is it expected?
I littered the code with printf statements and the issue propagated further upwards in the file to here
There are all sorts of things going on in this function including the passed in/out parameters being the same, as well as assembler function calls and constant time (inline assembler) calls.
Beta Was this translation helpful? Give feedback.
All reactions