Date: Wed, 22 Apr 2009 16:34:25 -0700 From: "David Christensen" <davidch@broadcom.com> To: "Marius Strobl" <marius@alchemy.franken.de> Cc: Jeff Blank <jb000003@mr-happy.com>, "freebsd-stable@freebsd.org" <freebsd-stable@freebsd.org> Subject: RE: bge problems under 7.2-PRERELEASE Message-ID: <5D267A3F22FD854F8F48B3D2B523819339DA0429DB@IRVEXCHCCR01.corp.ad.broadcom.com> In-Reply-To: <20090422203232.GC50221@alchemy.franken.de> References: <20090420204637.GA1236@mr-happy.com> <20090421184924.GA36542@alchemy.franken.de> <20090421193129.GA4869@mr-happy.com> <20090421195510.GC33994@alchemy.franken.de> <20090422150612.GA1231@mr-happy.com> <20090422192644.GA52151@alchemy.franken.de> <5D267A3F22FD854F8F48B3D2B523819339D9F5A64D@IRVEXCHCCR01.corp.ad.broadcom.com> <20090422203232.GC50221@alchemy.franken.de>
next in thread | previous in thread | raw e-mail | index | archive | help
> > > this is the second report that using MSI with a BCM5754/ > > > chip=3D0x167a14e4 results in no interrupts, is there any=20 > errata about=20 > > > this? > >=20 > > Nope, no errata for MSI on the 5754, nor can I recall any=20 > MSI problems=20 > > on any of the PCIe based devices. Do the systems share a=20 > common root=20 > > complex that might be the source of the issue? > >=20 >=20 > This one useses a: > none0@pci0:0:0:0: class=3D0x050000 card=3D0x02f010de=20 > chip=3D0x02f010de rev=3D0xa2 hdr=3D0x00 > vendor =3D 'Nvidia Corp' > device =3D 'C51 Host Bridge' > class =3D memory > subclass =3D RAM >=20 > At least Linux doesn't seem to avoid using MSI with these. > I didn't get any further regarding the other report so far. Are there any other errors reported by the 5754 such as a PCIe=20 completion timeout in the Advanced Error Uncorrectable Status Register? I recall an issue with another Nvidia chipset and=20 the 5754 where the system BIOS incorrectly configured the Hypertransport unitID field to a value of 0x20 which would prevent our diagnostics from running. PCI configuration and memory-mapped I/O worked correctly but DMA operations such as sending RX'd frames would fail. MSI might fall into this latter category and could be causing this error. The important thing would be to check for a Completion Timeout in the 5754. Dave
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?5D267A3F22FD854F8F48B3D2B523819339DA0429DB>