Hi, Trying to install 4.2-release (and snapshot) on an IBM xServer 345 with ServeRaid/5i controller. I've used the Linux based bootable CD from IBM and created an array with RAID1 Everything seems fine but when I boot up the OpenBSD cd it does not see any disks. dmesg shows: pci5 at pchb7 bus 8 "IBM ServeRAID" rev 0x00 at pci5 dev 2 function 0 not configured Reading the man page for ips(4) I read this: ----- Although the controllers are actual RAID controllers, the driver makes them look just like SCSI controllers. All RAID configuration is done through the controllers' BIOSes. ----- Can anyone shine some light on how to actually make this controller work with OpenBSD? Or am I overlooking something obvious? Thanks! Matt
The "not configured" means the device is not claimed by any kernel driver. In this case it seems that ips(4) is in the GENERIC kernel but not the ramdisk kernels used for installation. Not quite sure if this is a case of the driver not being ready or just an oversight.
ips is not included in the ramdisk kernels. I had the same problem and my solution was to copy the two lines for ips from /usr/src/sys/arch/i386/conf/GENERIC to /usr/src/sys/arch/i386/conf/RAMDISC_CD and then build a release. -- Fredrik
Thanks, that worked. I had never build a release before so I did it as described in the FAQ [1] on a different non-important machine, including a complete build of the userland binaries. Then I build a release from that. Eventually I did end up with the ips(4) enabled cd42.iso in the /usr/rel directory. Reading release(8) I could not find if there was a shortcut in doing this? Could one just build parts of the system needed for cd42.iso ? Thanks again, Matt [1] http://www.openbsd.org/faq/faq5.html
grange@ the author of the driver says there is still a data corruption bug in ips, so rather than enabling it on ramdisks I'll disable it in GENERIC tonight.
Am I understanding correctly grange's info made you decide the ips driver should not be used _at all_ ? I was planning to put this box in production today but the 'data corruption' part got me worried. Is there more info available on what the author said? And is there anything we can / should do to accomodate a fix? Thanks.
"unfixed data corruption bug" is the extent of my knowledge, I don't think grange currently has the time to work on this further right now.
| Linus Torvalds | Re: Slow DOWN, please!!! |
| Andrea Arcangeli | [PATCH 03 of 11] invalidate_page outside PT lock |
| Mark Lord | Re: 2.6.25-rc8: FTP transfer errors |
| Artem Bityutskiy | [RFC PATCH 03/26] UBIFS: add flash scanning |
git: | |
| Linus Torvalds | irc usage.. |
| Junio C Hamano | [ANNOUNCE] GIT 1.5.3-rc4 |
| Dmitry Kakurin | Windows support |
| Mark Levedahl | Re: mingw, windows, crlf/lf, and git |
| Kevin Neff | Patching a SSH 'Weakness' |
| Frank Bax | QEMU crashes |
| Chris Kuethe | Re: Logging failed SSH users and the passwords they typed |
| Edd Barrett | Re: OpenBSD in the webcomic XKCD |
| Suresh Siddha | Re: Kernel oops with 2.6.26, padlock and ipsec: probably problem with fpu state ch... |
| Eric Dumazet | Re: [PATCH] fs: pipe/sockets/anon dentries should not have a parent |
| Matheos Worku | 2.6.24 BUG: soft lockup - CPU#X |
| Steffen Klassert | [RFC PATCH 4/5] crypto: allow allocation of percpu crypto transforms |
