Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S262131AbUKDNej (ORCPT ); Thu, 4 Nov 2004 08:34:39 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S262220AbUKDNej (ORCPT ); Thu, 4 Nov 2004 08:34:39 -0500 Received: from sartre.ispvip.biz ([209.118.182.154]:7356 "HELO sartre.ispvip.biz") by vger.kernel.org with SMTP id S262131AbUKDNec (ORCPT ); Thu, 4 Nov 2004 08:34:32 -0500 Subject: Re: [patch] Real-Time Preemption, -RT-2.6.10-rc1-mm2-V0.7.7 From: "Michael J. Cohen" To: Ingo Molnar Cc: "K.R. Foley" , sboyce@blueyonder.co.uk, linux-kernel@vger.kernel.org In-Reply-To: <1099573171.7876.0.camel@optie.uni.325i.org> References: <4189108C.2050804@blueyonder.co.uk> <41892899.6080400@cybsft.com> <41897119.6030607@blueyonder.co.uk> <418988A6.4090902@cybsft.com> <20041104100634.GA29785@elte.hu> <1099563805.30372.2.camel@localhost> <1099567061.7911.4.camel@localhost> <20041104114545.GA3722@elte.hu> <1099573171.7876.0.camel@optie.uni.325i.org> Content-Type: text/plain Date: Thu, 04 Nov 2004 08:34:22 -0500 Message-Id: <1099575262.8110.1.camel@optie.uni.325i.org> Mime-Version: 1.0 X-Mailer: Evolution 2.0.2 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 2116 Lines: 67 On Thu, 2004-11-04 at 07:59 -0500, Michael J. Cohen wrote: > On Thu, 2004-11-04 at 12:45 +0100, Ingo Molnar wrote: > > * Michael J. Cohen wrote: > > > > > Turned off the debugging stuff to fix this one :/ > > > > > > might_sleep issue at swap_on and firefox causes oopsen. > > > > > > dmesg is 120k+ so here: > > > > > > http://325i.org/software/2.6.10-rc1-mm2-RT-V0.7.8.dmesg > > > > does the patch below fix the fbcon problem? (if any new oops happens or > > old one triggers again then please re-post the syslog or serial console > > capture) > > > > Ingo > > > > --- linux/drivers/video/console/fbcon.c.orig > > +++ linux/drivers/video/console/fbcon.c > > @@ -1051,7 +1051,14 @@ static void fbcon_cursor(struct vc_data > > struct display *p = &fb_display[vc->vc_num]; > > int y = real_y(p, vc->vc_y); > > int c = scr_readw((u16 *) vc->vc_pos); > > +#ifdef CONFIG_PREEMPT_REALTIME > > + unsigned long flags; > > +#endif > > > > +#ifdef CONFIG_PREEMPT_REALTIME > > + local_save_flags(flags); > > + local_irq_enable(); > > +#endif > > ops->cursor_flash = 1; > > if (mode & CM_SOFTBACK) { > > mode &= ~CM_SOFTBACK; > > @@ -1069,6 +1076,9 @@ static void fbcon_cursor(struct vc_data > > ops->cursor(vc, info, p, mode, get_color(vc, info, c, 1), > > get_color(vc, info, c, 0)); > > vbl_cursor_cnt = CURSOR_DRAW_DELAY; > > +#ifdef CONFIG_PREEMPT_REALTIME > > + local_irq_restore(flags); > > +#endif > > } > > > > static int scrollback_phys_max = 0; > > Works fine so far. still have to deal with e1000 though... > > ------ I just managed to hardlock it reproducibly with no useful output on serial. light network activity, bonnie++, and using xmms w/ jack. Locks up within a few seconds of starting bonnie++ tried my hardest (sysreq, even kgdb) and I couldn't get it to let me do anything :( ------ Michael Cohen - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/