Date: Tue, 5 Apr 2005 09:31:23 -0400 From: John Baldwin <jhb@FreeBSD.org> To: freebsd-current@FreeBSD.org Cc: Harry Coin <harrycoin@qconline.com> Subject: Re: atapicam hangs boot on 5.3 5.4pre1 5.4pre2 Message-ID: <200504050931.23981.jhb@FreeBSD.org> In-Reply-To: <4.3.2.7.2.20050317141121.01dcc770@mail.qconline.com> References: <4.3.2.7.2.20050317141121.01dcc770@mail.qconline.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thursday 17 March 2005 03:11 pm, Harry Coin wrote: > On my P4 system with 'umass0' solid-state disk (photography chips) on the > USB port, if 'device atapicam' is included Freebsd 5.3 release, 5.4 pre1 > and 5.4 pre 2 will hang just before the SMP startup notice in the dmesg > trace. If atapicam is removed then the system boots normally. > > The message it hangs on in place of the SMP message is an > > interrupt storm on int 18: em0 and uhci2++. > > Scrolling still works, no recovery ever happens. I read somewhere back in > 4.x a developer thought the problem might be failure to restore an > interrupt level in a system fork routine, but I really have no idea what > the fix is. I've tried all the obvious boot changes, no joy. if atapicam > is included, hangs at the same spot every time. If not, then a normal > boot. That's the only variable I can twiddle that made any difference. > > System is a P4 Asus P4C800-E, no AGP video, digital camera disks on the USB > port . > > The lock occurs right after this line in the dmesg > > ata4-master SATA150 > ad10: 35304MB <WDC WD360GD-00FNA0/35.06K35> [71730/16/63] at ata5-master > SATA150 > > The line never seen is > > SMP: AP CPU #1 Launched! > > The error message is that an interrupt storm was detected on int18: em0 and > uhci2++. I can still use the scroll lock to scan up and down, but the > system never moves on from that point -- ever. Well, your ATA controllers and most of your USB controllers are all on IRQ 18 along with em0. Given that, I don't think this is an interrupt routing issue, but rather, atapicam has managed to make an ATA controller very upset to the point that the ATA controller is interrupting in a way that the driver isn't acknowledging somehow. Can you try some snapshots of 6.0 to see if 1) they fare better when the interrupt storm happens, and 2) if the new ATA code added to HEAD last weekend works better? Note that atapicam doesn't work with the new ATA code until last night, so you might have to install yesterday's snapshot and use cvsup to update to the latest to test atapicam. -- John Baldwin <jhb@FreeBSD.org> <>< http://www.FreeBSD.org/~jhb/ "Power Users Use the Power to Serve" = http://www.FreeBSD.org
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200504050931.23981.jhb>