[Bug runtime/31117] transport failures

["wcohen at redhat dot com" <sourceware-bugzilla@sourceware.org>]

https://sourceware.org/bugzilla/show_bug.cgi?id=31117

--- Comment #3 from William Cohen <wcohen at redhat dot com> ---
I did some investigation on this with the map_hash_SSI.stp transport failures. 
Some fixes of the diagnostic print code for the transport layer were made so it
compiles on fedora 39.  The __stp_print_flush code expects there to be enough
room to write out the _stp_trace header (12 bytes) in the space returned by
_stp_data_write_reserve().  However, there are cases where
_stp_data_write_reserve() returned less space (5 and 9 bytes).  The causes the
_stp_print_flush to fail and increment _stp_transport_failures.  That certainly
needs to be fixed.

I had some additional code in the runtime to print out each time
_stp_print_flush had a failure.  I didn't see that triggering with the sys.stp.
Looks like will need to additional diagnostic prints to figure out what is
going on with the sys.stp output.

-- 
You are receiving this mail because:
You are the assignee for the bug.