Messages in this thread | | | Subject | Re: [PATCH 5/4] futex: silence uninitialized warnings | From | Davidlohr Bueso <> | Date | Mon, 06 Jan 2014 19:02:23 -0800 |
| |
On Tue, 2014-01-07 at 10:55 +0800, Linus Torvalds wrote: > On Mon, Jan 6, 2014 at 9:38 AM, Davidlohr Bueso <davidlohr@hp.com> wrote: > > { > > int lock_taken, ret, force_take = 0; > > - u32 uval, newval, curval, vpid = task_pid_vnr(task); > > + u32 uval, newval, uninitialized_var(curval), vpid = task_pid_vnr(task); > > Do you have some broken compiler?
I only notice this when testing this patchset on our servers with RHEL 6.4 (gcc 4.4.7 20120313 (Red Hat 4.4.7-3)).
> I really tend to hate this kind of > workarounds, because as mentioned, they can actually hide valid > warnings, and it seems to be due to just stupid compilers. Are we > perhaps better off trying to get people off the broken compiler > versions instead?
As Darren points out, this path is unlikely to change, but I have no particular preference otherwise.
| |