cvs commit: src/sys/dev/em if_em.c
pyunyh at gmail.com
Tue Aug 1 00:39:41 UTC 2006
On Mon, Jul 31, 2006 at 11:43:11PM +0400, Gleb Smirnoff wrote:
> On Mon, Jul 31, 2006 at 02:49:05PM +0100, Robert Watson wrote:
> R> On Mon, 31 Jul 2006, Gleb Smirnoff wrote:
> R> >On Thu, Jul 20, 2006 at 03:57:58AM +0000, Pyun YongHyeon wrote:
> R> >P> yongari 2006-07-20 03:57:58 UTC
> R> >P>
> R> >P> FreeBSD src repository
> R> >P>
> R> >P> Modified files:
> R> >P> sys/dev/em if_em.c
> R> >P> Log:
> R> >P> Honor IFF_DRV_OACTIVE in em_start_locked().
> R> >
> R> >Isn't it better to shift this change to the em_start(), to avoid double
> R> >check of IFF_DRV_RUNNING? Or may be just merge em_start_locked() into the
> R> >em_start()?
> R> Isn't em_start_locked() called from quite a few different places, including
> R> the interrupt handler, etc, where the mutex is already held?
> It is, so the latter idea won't work. However, in all these places
> we can know the status of IFF_DRV_OACTIVE before calling the start
You're right. Feel free to fix it.
I have to solve re(4) issues on sparc64 at the moment...
More information about the cvs-src