Date: Thu, 6 May 2010 21:41:30 -0600 From: Scott Long <scottl@samsco.org> To: ulf@Alameda.net Cc: "freebsd-scsi@freebsd.org" <freebsd-scsi@freebsd.org>, "Erich Jenkins, Fuujin Group Ltd" <erich@fuujingroup.com>, Sean Bruno <seanbru@yahoo-inc.com> Subject: Re: kern/146287: [ciss] ciss(4) cannot see more than one SmartArray controller Message-ID: <A2041F45-911E-4F4C-BA74-0B3C4BAC04DF@samsco.org> In-Reply-To: <20100506213349.GU30353@evil.alameda.net> References: <20100505203116.GP30353@evil.alameda.net> <20100505204652.GQ30353@evil.alameda.net> <1273092737.2487.5.camel@localhost.localdomain> <20100506014748.GR30353@evil.alameda.net> <1273165130.2465.4.camel@localhost.localdomain> <20100506174113.GS30353@evil.alameda.net> <1273169980.2465.6.camel@localhost.localdomain> <4BE318A8.6050108@fuujingroup.com> <1273172053.2465.8.camel@localhost.localdomain> <20100506211713.GT30353@evil.alameda.net> <20100506213349.GU30353@evil.alameda.net>
next in thread | previous in thread | raw e-mail | index | archive | help
On May 6, 2010, at 3:33 PM, Ulf Zimmermann wrote: > On Thu, May 06, 2010 at 02:17:13PM -0700, Ulf Zimmermann wrote: >> On Thu, May 06, 2010 at 11:54:13AM -0700, Sean Bruno wrote: >>>=20 >>>>>=20 >>>>=20 >>>> Not sure if this helps or not, but we've got about 20 of these = machines,=20 >>>> as well as the DL580 version in production. I've seen this very = same=20 >>>> issue on the ML series machines also. If I remember correctly, = there=20 >>>> were two different PCI bridges used in the same production models, = one=20 >>>> was from IBM and the other was a ServerWorks (Broadcom) chipset. = One=20 >>>> didn't show up at all (not even when probed) and the other one was=20= >>>> displayed. Can't remember which was which, but I know it had to do = with=20 >>>> hot-swap PCI subsystem, and I suspect this is what you're = experiencing.=20 >>>> We had a similar experience with MyriCom 10GE cards in an ML570 = with a=20 >>>> hot-swap PCI option. The card wouldn't show up in slot closest to = the=20 >>>> Mem card (can't remember if the DL380's have a riser card for ram = or if=20 >>>> this is onboard), but it worked in the next slot over. >>>>=20 >>>>=20 >>>> Erich M. Jenkins >>>> Fuujin Group Limited >>>>=20 >>>>=20 >>>> "You should never, never doubt what no one is sure about." >>>> -- Gene Wilder >>>>=20 >>>=20 >>> It does help a bit. >>>=20 >>> Basically, we need an AML dump of the ACPI information from one of = the >>> machines that has an IBM based bridge chip. >>>=20 >>> acpidump -dt >>>=20 >>> sean >>>=20 >>>=20 >>=20 >> I wonder what bridge chip is used in the DL380 g3, as far I know the = ones >> we have should be hotswap PCI-X. On the DL380 g4 the standard cage = was not >> hotswap and I think we don't have any of those. >>=20 >> Looking at the product specs of a DL380 g3, it shows different >> bus for slot 1 (bottom at 64-bit/133Mhz PCI-X) against slot 2+3 = (middle >> and top slot at 64-bit/100Mhz PCI-X). >>=20 >> I got one DL380 g3 myself, which isn't running much right now, so I = will >> check out what PCI devices I can see. >>=20 >=20 > 03:01.0 PCI bridge: IBM PCI-X to PCI-X Bridge (rev 02) (prog-if 00 = [Normal decode]) > Flags: bus master, 66Mhz, medium devsel, latency 64 > Bus: primary=3D03, secondary=3D04, subordinate=3D04, = sec-latency=3D64 > I/O behind bridge: 00003000-00003fff > Memory behind bridge: f7e00000-f7efffff > Prefetchable memory behind bridge: = 00000000f1100000-00000000f11fffff > Capabilities: [80] PCI-X bridge device > Capabilities: [90] Power Management version 2 >=20 > Is that the same bridge chip? This comes from a DL380 g3 with hot-swap = cage. >=20 >=20 Yup, that's the bridge chip that we suspected from the original poster. Scott
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?A2041F45-911E-4F4C-BA74-0B3C4BAC04DF>