Date: Wed, 6 Nov 2002 20:43:13 +0100 From: Gerhard Sittig <Gerhard.Sittig@gmx.net> To: freebsd-stable@freebsd.org Subject: ar(4) and sysinstall(8): cannot create _any_ partitions Message-ID: <20021106204313.B4807@shell.gsinet.sittig.org>
next in thread | raw e-mail | index | archive | help
The problem in short: sysinstall's "Disklabel editor" refuses to create partitions on an ar0 slice. This happens with 4.7-RELEASE, 4.7-STABLE, and -CURRENT. The controler is a Promise FastTrak100 (a PDC20267 chip), the RAID volume has two 120GB IBM disks in "mirror" mode (RAID1). And now for the details. The dmesg output was captured by means of a serial console. But since -CURRENT refuses to accept my key strokes (and thus won't let me specify "-h" in the early stage) and -STABLE / -RELEASE does not like my cursor / function keys I'm unable to capture anything after the "do you want to configure the kernel?" menu. So the partition and disklabel editor screens have been transfered manually, the content is verbatim but the layout might miss whitespaces. - mobo "Tyan Tiger200T V2.08 062102(R4)" (the spec can be found at http://www.tyan.com/products/html/tiger200t.html), with an onboard "Promise FastTrak100" (BIOS 2.00.0.24), the PDC20267's PCI ids are 0x105A for the vendor and 0x4D30 for the device parts and revision is 2 - two IBM IC35L120AVVA07-0 disks, each on a channel's master position, forming a "mirror" volume (RAID1), run with "U5" according to the FastBuild tool - Award BIOS with "06/21/2002-694T-686B-6A6LJT5JC-00" in the boot screen bottom line - SMP configuration with two PIII and 1024MB RAM (in case it matters but I doubt this since it's still the install stage) The disk layout was created with SuSE Linux' rescue system (no matter what OS I'm going to install later, this rescue system is by far the best and most handy tool for hardware diagnosis and machine preparation prior to installations I know of ... :) I usually `dd if=/dev/zero of=/dev/hda bs=8388608` any disk several times before partitioning it and `fdisk /mbr` plus DOS install it afterwards. The `fdisk -l /dev/ataraid/d0` output looks like this: | Disk /dev/ataraid/d0: 255 heads, 63 sectors, 15017 cylinders | Units = cylinders of 16065 * 512 bytes | | Device Boot Start End Blocks Id System | /dev/ataraid/d0p1 * 1 33 265041 6 FAT16 | /dev/ataraid/d0p2 34 99 530145 83 Linux | /dev/ataraid/d0p3 100 165 530145 82 Linux swap | ... The "Start" and "End" numbers are cylinders, "Blocks" have 1KB size. Don't think too long about the type fields. :] I'm used to keep a DOS partition for maintenance reasons on every machine I setup. And the "Linux" partitions are leftovers from some experiment. (BTW did SuSE Linux install and run, even in d0p5 which was located after the first three partitions) When I boot 4.7-RELEASE from the kern.flp and mfsroot.flp floppies I turn on debugging in the "Options" screen. sysinstall(8)'s "FDISK Partition Editor" looks like this: | Disk name: ar0 | DISK Geometry: 15017 cyls/255 heads/63 sectors = 241248105 sectors (117796MB) | | Offset Size(KB) End name PType Desc Subtype Flags | 0 31 62 - 6 unused 0 | 63 265041 530144 ar0s1 2 fat 6 | 530145 530145 1590434 ar0s2 3 freebsd 165 | 1590435 530145 2650724 ar0s3 1 linux_swap 130 | ... This is after I changed the ar0s2 type to "FreeBSD" since I intend to install the OS (its root) there. Usually I do this to keep everything necessary for booting in the very first cylinders due to BIOSes and their limits. And now comes the interesting part. The "Disklabel editor" refuses to create any partitions inside the ar0s2 slice. There is always a "Message" window popping up with "Unable to create the root partition. Too big?" (when I try the (A)uto layout) or "Unable to create the partition. Too big?" (when I try to (C)reate a 64M FS partition for the / filesystem). The screen looks like this: | Disk: ar0 Partition name: ar0s2 Free: 1060290 blocks (517MB) | | Part Mount Size Newfs Part Mount Size Newfs | ar0s1 (none) 258MB DOS When I switch to the debug output by means of Alt-F2 I see a lot of repeated "DEBUG: Scanning disk ar0 for root filesystem" and "DEBUG: Scanning disk ar0 for swap filesystem" lines (some ten times these two lines per try to create a partition? I'm not absolutely sure and they're difficult to count ...). Here are some dmesg snippets I believe to be relevant (sorry, the complete log has more than 30KB and I'm reluctant to post it unrequested to the list). The "ar0:" part seems the most interesting one -- the RAID1 volume gets recognized and is in the "READY" state, the installer happily offers to prepare the "ar0" disk, everything looks like things *should* work ... | [ ... ] | FreeBSD 4.7-STABLE-20021106-JPSNAP #0: Tue Nov 5 23:44:20 GMT 2002 | root@tora.jp.freebsd.org:/usr/obj/usr/src/sys/BOOTMFS | [ ... ] | isa0: <ISA bus> on isab0 | atapci0: <VIA 82C686 ATA100 controller> port 0xc400-0xc40f at device 7.1 on pci0 | ata0: iobase=0x01f0 altiobase=0x03f6 bmaddr=0xc400 | ata0: mask=02 ostat0=ff ostat2=50 | ata0-slave: ATAPI 14 eb | ata0: mask=02 stat0=00 stat1=00 | ata0: devices=08 | ata0: at 0x1f0 irq 14 on atapci0 | ata1: iobase=0x0170 altiobase=0x0376 bmaddr=0xc408 | ata1: at 0x170 irq 15 on atapci0 | uhci0: <VIA 83C572 USB controller> port 0xc800-0xc81f irq 11 at device 7.2 on pci0 | [ ... ] | pci0: <unknown card> (vendor=0x1106, dev=0x3057) at 7.4 | atapci1: <Promise ATA100 controller> port 0xe000-0xe03f,0xdc00-0xdc03,0xd800-0xd807,0xd400-0xd403,0xd000-0xd007 mem 0xf6200000-0xf621ffff irq 11 at device 12.0 on pci0 | ata-: ata2 exists, using next available unit number | ata2: iobase=0xd000 altiobase=0xd402 bmaddr=0xe000 | ata2: mask=03 ostat0=50 ostat2=00 | ata2-master: ATAPI 00 00 | ata2-slave: ATAPI 00 00 | ata2: mask=03 stat0=50 stat1=00 | ata2-master: ATA 01 a5 | ata2: devices=01 | ata2: at 0xd000 on atapci1 | ata3: iobase=0xd800 altiobase=0xdc02 bmaddr=0xe008 | ata3: mask=03 ostat0=50 ostat2=00 | ata3-master: ATAPI 00 00 | ata3-slave: ATAPI 00 00 | ata3: mask=03 stat0=50 stat1=00 | ata3-master: ATA 01 a5 | ata3: devices=01 | ata3: at 0xd800 on atapci1 | fxp0: <Intel Pro 10/100B/100+ Ethernet> port 0xe400-0xe43f mem 0xf6000000-0xf60fffff,0xf6222000-0xf6222fff irq 5 at device 13.0 on pci0 | [ ... ] | fd0: <1440-KB 3.5" drive> on fdc0 drive 0 | ata4 failed to probe at port 0x1f0 irq 14 on isa0 | ata5 failed to probe at port 0x170 irq 15 on isa0 | adv0 failed to probe on isa0 | [ ... ] | isa_probe_children: probing PnP devices | BIOS Geometries: | 0:03fffe3f 0..1023=1024 cylinders, 0..254=255 heads, 1..63=63 sectors | 0 accounted for | Device configuration finished. | bpf: sl0 attached | bpf: lo0 attached | ad4: success setting UDMA5 on Promise chip | Creating DISK ad4 | ad6: success setting UDMA5 on Promise chip | Creating DISK ad6 | Creating DISK ar0 | ar0: 117800MB <ATA RAID1 array> [15017/255/63] status: READY subdisks: | 0 READY ad4: <IC35L120AVVA07-0/VA6OA52A> ATA-5 disk at ata2-master | ad4: 117800MB (241254720 sectors), 239340 C, 16 H, 63 S, 512 B | ad4: 16 secs/int, 1 depth queue, UDMA100 | ad4: piomode=4 dmamode=2 udmamode=5 cblid=1 | 1 READY ad6: <IC35L120AVVA07-0/VA6OA52A> ATA-5 disk at ata3-master | ad6: 117800MB (241254720 sectors), 239340 C, 16 H, 63 S, 512 B | ad6: 16 secs/int, 1 depth queue, UDMA100 | ad6: piomode=4 dmamode=2 udmamode=5 cblid=1 | ata0-slave: piomode=4 dmamode=2 udmamode=2 dmaflag=1 | ata0-slave: success setting PIO4 on generic chip | acd0: <CD-224E/1.9A> CDROM drive at ata0 as slave | acd0: read 4134KB/s (34515KB/s), 128KB buffer, PIO4 | acd0: Reads: CD-R, CD-RW, CD-DA stream, packet | acd0: Writes: | acd0: Audio: play, 256 volume levels | acd0: Mechanism: ejectable tray, unlocked | acd0: Medium: CD-ROM 120mm data disc | Mounting root from ufs:/dev/md0c | [ ... ] | ar0s2: type 0x83, start 530145, end = 1590434, size 1060290 : OK | ar0s3: type 0x82, start 1590435, end = 2650724, size 1060290 : OK | ar0s4: type 0x5, start 2650725, end = 241248104, size 238597380 : OK | ar0s5: type 0x83, start 2650788, end = 5799464, size 3148677 : OK | ar0s1: type 0x6, start 63, end = 530144, size 530082 : OK | [ ... many repetitions of these ... ] | ar0s5: type 0x83, start 2650788, end = 5799464, size 3148677 : OK | ar0s1: type 0x6, start 63, end = 530144, size 530082 : OK | ar0s2: type 0x83, start 530145, end = 1590434, size 1060290 : OK | ar0s3: type 0x82, start 1590435, end = 2650724, size 1060290 : OK | ar0s4: type 0x5, start 2650725, end = 241248104, size 238597380 : OK | ar0s5: type 0x83, start 2650788, end = 5799464, size 3148677 : OK | [ and here comes the "sysinstall Main Menu" with its "Welcome to | the FreeBSD installation ..." text and many escape sequences ] This behaviour (the "Unable to create the partition. Too big?" message) is all the same in -RELEASE (from ftp7.de.freebsd.org), today's -STABLE and -CURRENT snapshots (both from current.freebsd.org). I think the error message is bogus and the failure is coming from somewhere else. But what keeps this setup from being useful with FreeBSD??? Searching with the web frontend in the gnats database for the words "promise", "fasttrak" and "pdc202" (separatedly, not in combination, in the one line headers, including closed reports) didn't reveal an open PR for this behaviour. Now I wonder if this is a hardware failure of my machine or a software problem. I'm willing to help diagnose the problem and try whatever tool or patch you ask me to run. But I will have the machine around for no longer than some week before it will go productive and thus slips out of my reach. virtually yours 82D1 9B9C 01DC 4FB4 D7B4 61BE 3F49 4F77 72DE DA76 Gerhard Sittig true | mail -s "get gpg key" Gerhard.Sittig@gmx.net -- If you don't understand or are scared by any of the above ask your parents or an adult to help you. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20021106204313.B4807>