Date: Thu, 14 Dec 2006 16:15:00 -0800 (PST) From: John Polstra <jdp@polstra.com> To: David Christensen <davidch@broadcom.com> Cc: John Polstra <jdp@FreeBSD.org>, cvs-src@FreeBSD.org, src-committers@FreeBSD.org, cvs-all@FreeBSD.org Subject: RE: cvs commit: src/sys/dev/bge if_bge.c if_bgereg.h Message-ID: <XFMail.20061214161500.jdp@polstra.com> In-Reply-To: <09BFF2FA5EAB4A45B6655E151BBDD903029CEDDE@NT-IRVA-0750.brcm.ad.broadcom.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On 15-Dec-2006 David Christensen wrote: > MSI capabilities registers are present on virtually all "bge" > controllers but many of the "older" controllers don't correctly support > MSI due to a chip errata that prevents MSI interrupts from actually > being disabled during an ISR. Check the Linux code for the test used in > tg3_open() which determines which controllers can actually use MSI > correctly. > > Even with the bug the older controllers will appear to work correctly on > a heavily loaded system for short periods of time (sometimes even days) > but they will eventually fail. Thanks for the heads up, Dave. Scott Long also cautioned me about this. I'll disable the MSI support in the driver until I have time to add the necessary checks. John
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?XFMail.20061214161500.jdp>