Programming

Does anyone know what __read_nocancel does and why it would go into an infinite loop?

What I have gathered in my searches is that it pertains to server code. Yet, I'm not running this application in server mode.

NOTE: There are server functions in the shared object, but the specific code I'm executing for this run is all batch mode.

Here's the backtrace:

Code:

0x000000366c9bc1c2 in __read_nocancel () from /lib64/tls/libc.so.6
(gdb) bt
#0  0x000000366c9bc1c2 in __read_nocancel () from /lib64/tls/libc.so.6
#1  0x0000002a9574382c in SPload () from /local/agility/RegTest/SP3/bin/libagam.so
#2  0x0000002a9574775b in AGAM_SP_Lookup () from /local/agility/RegTest/SP3/bin/libagam.so
#3  0x0000000000449fb8 in ExternalProc ()
#4  0x0000000000403533 in main ()

Any help you can provide would be greatly appreciated.
~Mark

Documentation for what this is seems extremely hard to find, but it seems related to kernel-level non-blocking I/O, so if I would hazard a guess, this non-blocking call is trying to read, returning an error which the program assumes is 'data not yet available' and keeps trying, but the error is in fact something more severe like EOF so it never gets more data no matter how many times it polls.

Can you provide more information as to what this application is about and what it's trying to do...also what platform/OS are you on.

__read_nocancel() is used by libc when a cancellation point (a POSIX threads concept) is not permitted. Otherwise the code path is via __libc_enable_asynccancel().