On Wed, 09 Sep 2009 13:30:44 +0200 Florian Haas <florian.haas@xxxxxxxxxx> wrote: > On 09/09/2009 01:29 PM, FUJITA Tomonori wrote: > > On Wed, 09 Sep 2009 13:17:10 +0200 > > Florian Haas <florian.haas@xxxxxxxxxx> wrote: > > > >> On 09/09/2009 12:44 PM, FUJITA Tomonori wrote: > >>> We use threads and pipe for event notification (I/O completion); I/O > >>> threads signal the main thread when I/Os are complete. signalfd is > >>> faster than it (signalfd is supported as of 2.6.22). > >> Tomo-san, > >> > >> AFAICS signalfd has not been backported to Red Hat's RHEL 5 kernels. So > >> if I understand correctly, this patch would break building tgt on that > >> platform (and presumably others, like Debian etch and SLES 10). Would > >> you mind employing ifdefs so that when signalfd is not available on the > >> platform tgt is being built on, it reverts to the old behavior? > > > > This patch works in the old way on a box where signalfd is not > > supported. I have not try this patch with systems that doesn't support > > signalfd. Can anyone test this patch on such system? > > I can try and see if it builds. What would I need to look out for when I > actually use it? If it works on such system (an initiator connects to it and does I/Os), then should be ok. Thanks, -- To unsubscribe from this list: send the line "unsubscribe stgt" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
