Date: Wed, 17 Nov 2004 15:56:31 GMT From: John Baldwin <jhb@FreeBSD.org> To: Perforce Change Reviews <perforce@freebsd.org> Subject: PERFORCE change 65336 for review Message-ID: <200411171556.iAHFuVeK072624@repoman.freebsd.org>
next in thread | raw e-mail | index | archive | help
http://perforce.freebsd.org/chv.cgi?CH=65336 Change 65336 by jhb@jhb_slimer on 2004/11/17 15:55:51 Update. Affected files ... .. //depot/projects/smpng/sys/notes#26 edit Differences ... ==== //depot/projects/smpng/sys/notes#26 (text+ko) ==== @@ -49,12 +49,13 @@ buys is being able to swi_sched() from a fast handler, but that doesn't make sense because you can't actually enqueue a task onto its queue from a fast interrupt context -- "Optimize" atomic_store_rel() on i386 (it actually seems to be slower, - go figure) - + It is faster on UP, but slower on SMP, so tweak the patch so that we use - the non-MP safe 386 versions on all UP kernels. - Optimize spin locks on UP such that they don't do atomic operations - Untested - Unbenchmarked +- Try to fix problems with stuck IPIs by raising TPR and enabling interrupts + in spin loops. Need to perhaps generalize this so that critical sections + do this. That really needs to only be done in spinlock_*() (and maybe some + frobbing in intr_execute_handlers()) and needs to happen after the + spinlock_*() changes happen. Space reserved for child branches:
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200411171556.iAHFuVeK072624>