Date: Wed, 21 Oct 2009 23:58:02 +0300 From: Alexander Motin <mav@FreeBSD.org> To: Steve Polyack <korvus@comcast.net> Cc: freebsd-stable <freebsd-stable@FreeBSD.org>, freebsd-hardware@freebsd.org Subject: Re: FreeBSD and SATA Port Multipliers Message-ID: <4ADF75DA.8040707@FreeBSD.org> In-Reply-To: <4ADF44CE.3030605@comcast.net> References: <4ADE060C.5050605@comcast.net> <4ADF44CE.3030605@comcast.net>
next in thread | previous in thread | raw e-mail | index | archive | help
Steve Polyack wrote: > Alexander Motin wrote: >> Have you tried new CAM-based ATA subsystem present in 8.x by siis(4) >> driver? Work is still in progress there, but it should work with port >> multipliers much better then previous implementation. > > Indeed, the siis(4) driver works MUCH better. However, it's still not > always detecting every drive if I fill a port multiplier with 5/5 > drives. Sometimes it boots up with 4, sometimes with 5. > > Snippet from the dmesg in my other reply: > siisch4: Timeout on slot 25 > siisch4: siis_timeout is 00000000 ss 02000000 rs 02000000 es 00000000 > sts 80192000 serr 00000000 > (aprobe0:siisch4:0:1:0): SIGNATURE: 0000 > (aprobe0:siisch4:0:2:0): SIGNATURE: 0000 > (aprobe0:siisch4:0:3:0): SIGNATURE: 0000 > (aprobe0:siisch4:0:4:0): SIGNATURE: 0000 > ada0 at siisch4 bus 0 target 1 lun 0 > ada0: <ST31500341AS CC1H> ATA/ATAPI-8 SATA 2.x device > ada0: 300.000MB/s transfers > ada0: 1430799MB (2930277168 512 byte sectors: 16H 63S/T 16383C) > ada0: Native Command Queueing enabled > > siisch4 bus 0 target 0 lun 0 is missing here, due to some kind of > timeout. The drives are fine and the missing device does not follow a > specific physical drive. Lastly, using camcontrol to rescan port > multipliers and attached drives only leads to failure. Rescanning > everythign simply loses all but one drive attached to each port multiplier. As I have said, infrastructure is not finished yet, especially part of recovery from timeout state. That's why probably rescan does not solve situation. From other side, this timeout may be called by some driver issue, as it looks too repeatable. If your conditions permit, you can try to upgrade to recent HEAD to look how will it go with newer code. I am periodically merging my work there from Perforce. Also I am going to generate new test patch for HEAD, which includes reworked PMP support, tomorrow. I will put it to http://people.freebsd.org/~mav/ as usual, so you can test it. -- Alexander Motin
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4ADF75DA.8040707>