Date: Sat, 21 Jul 2007 13:28:40 -0600 From: Scott Long <scottl@samsco.org> To: Tom Judge <tom@tomjudge.com> Cc: freebsd-stable@freebsd.org, Michael Worobcuk <mail@michaelworobcuk.de> Subject: Re: Dell PERC5/i SAS5/5IR - RAID monitoring Message-ID: <46A25E68.2080507@samsco.org> In-Reply-To: <46A14216.5010507@tomjudge.com> References: <4CC93356-890C-43DF-A905-0B9ABC990223@michaelworobcuk.de> <469D5061.8060306@tomjudge.com> <2F1F8117-86EF-49B4-AB4E-8A656727570A@michaelworobcuk.de> <469DE961.9060608@tomjudge.com> <932C632C-A89B-427F-92EC-71B351D66D17@michaelworobcuk.de> <46A134AB.2050803@tomjudge.com> <4837D0B1-93C2-4476-8D91-7D2586722458@michaelworobcuk.de> <46A14216.5010507@tomjudge.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Tom Judge wrote: > Michael Worobcuk wrote: > >> >> Am 21.07.2007 um 00:18 schrieb Tom Judge: >> >>> Michael Worobcuk wrote: >>> >>>>> Tom Judge wrote: >>>>> >>>>> As for performance issues with the SAS5/i, there is a problem in >>>>> the controller. A work arround was created by Scott Long which >>>>> created a sysctl that could be set to cause the controller to turn >>>>> on the on drive write cache's. These changes where commited to >>>>> RELENG_6 on 2007-06-05 21:32:57 UTC. >>>>> >>>>> >>>>> ... >>>>> >>>>>> >>>>>> Here is the original commit log: >>>>>> >>>>>> scottl 2007-06-03 23:13:05 UTC >>>>>> >>>>>> FreeBSD src repository >>>>>> >>>>>> Modified files: >>>>>> sys/dev/mpt mpt.c mpt.h mpt_cam.c >>>>>> Log: >>>>>> mpt.c: >>>>>> mpt.h: >>>>>> Add support for reading extended configuration pages. >>>>>> mpt_cam.c: >>>>>> Do a top level topology scan on the SAS controller. If >>>>>> any SATA device are discovered in this scan, send a passthrough >>>>>> FIS to set the write cache. This is controllable through the >>>>>> following tunable at boot: >>>>>> >>>>>> hw.mpt.enable_sata_wc: >>>>>> -1 = Do not configure, use the controller default >>>>>> 0 = Disable the write cache >>>>>> 1 = Enable the write cache >>>>>> >>>>>> The default is -1. This tunable is just a hack and may be >>>>>> deprecated in the future. >>>>> >>>>> >>>> I set mpt.enable_sata_wc to 1, as hw.mpt.enable_sata_wc is, AFAIK >>>> not tunable in mpt_cam.c. This did not take any effect to the >>>> performance. Is there anything else to change ? >>> >>> >>> Not that I know of, do you have SAS or SATA disks attached to the >>> controller? >> >> >> yes, SAS. >> >> > > In that case the above sysctl is not going to work for you as as the > name suggests it is only for sata devices. > > For someone with a bit more SCSI experience than me: Could this be > solved by setting the WCE (Write cache enable) bit in the modepage (8) > on sas devices if it is not already set? The driver could make this > change on SAS devices during the topology scan in similar way to the way > it does for SATA devices? > There are absolutely no plans for the MPT driver to automatically change WCE settings for SAS, FC, or SCSI drives. The hack that I put in for SATA is exactly what it is, a hack. It gets around the fact that there is no SATA layer within CAM yet and thus no way to send SATA commands from userland like there is with SCSI/FC/SAS. That will be addressed in the future. So yes, the proper way to handle this is via camcontrol. With that you should be able to change the WCE bit both temporarily (i.e. until the next bus reset or power reset) or permanently (i.e. in NVRAM of the drive). Worst case scenario is the the drive doesn't support permanent settings, so you write a simple RC script to run camcontrol on each boot. Scott
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?46A25E68.2080507>