This is the mail archive of the systemtap@sourceware.org mailing list for the systemtap project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Results of 20070419 testing


Frank Ch. Eigler wrote:
mmlnx wrote:

I'm working on a fix for the socket test failure. Should be checked
in today. The problem is that do_sock_write() is called in one place
instead of two as of 2.6.19 and, thus, gets optimized away. Same for
do_sock_read().

So I assume function("do_sock_write") succeeds but function().call doesn't?

Yes, but I should have been more specific. The problem is in the return probes.



Is there any way to mark a symbol so that the compiler doesn't
optimize it away?

Or more precisely, does not inline it? Yeah, but we can't require its use.

Understood.


Thanks,
Mike


- FChE


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]