Messages in this thread | | | From | David Howells <> | Subject | Re: [RFC][PATCH 0/4] SunRPC/NFS: Use no_printk() in | Date | Thu, 26 Sep 2013 16:27:39 +0100 |
| |
J. Bruce Fields <bfields@fieldses.org> wrote:
> > Here's a series of patches to make SunRPC/NFS use no_printk() to implement > > its null dfprintk() macro (ie. when RPC_DEBUG is disabled). This prevents > > 'unused variable' errors from occurring when a variable is set only for > > use in debugging statements and renders RPC/NFS_IFDEBUG unnecessary. > > Does this patch series fix any actual warnings? Or does it just change > the way that we prevent the warnings?
It fixes some unused variable warnings introduced by NFS FS-Cache patches that I have (a variable is set up and only passed to dfprintk() a couple of times).
I could change those patches to do something different, but I think changing dfprintk() is actually the right solution as it will catch errors introduced into dfprintk() calls that are currently reduced to do{}while(0) by the preprocessor rather than letting the compiler chew on them and then reducing them to nothing with the optimiser.
David
| |