5.2-RC ataraid woes

[Doug White]

On Tue, 23 Dec 2003, Tomi Vainio - Sun Finland wrote:

> This is exactly the same problem I see with my system.  Promise
> PDC20267 is causing these "READ_DMA recovered from missing interrupt"
> -problems but only on secondary channel.  Please try to run
> atacontrol pio4 3 if your problem goes away.  This could finally prove
> that something s*cks on Promise chip or code for it.

Its more likely there is an interrupt routing problem on this hardware.
You shouldn't be getting dropped interrupts, which is what you're seeing.

I don't have any machines with this board to test with right now,
unfortunately.  I do have an S845DP1 that works, however.



>
>   Tomppa
>
> Hendrik Scholz writes:
>  > Hi!
>  >
>  > I'm running an Intel SE7500CW2 board (Dual Xeon 2.4 w/ ht and 4GB RAM).
>  > It has an on-board Promise PDC20267 UDMA Raid controller that
>  > recognizes a mirror set and finds both disks but looses the second one
>  > at boot:
>  >
>  > ...
>  > FreeBSD 5.2-RC #0: Mon Dec 22 10:20:12 EST 2003
>  >     hscholz at cave.wormulon.net:/usr/src/sys/i386/compile/CAVE
>  > ...
>  > atapci0: <Promise PDC20267 UDMA100 controller> port 0x8480-0x84bf,0x84c0-0x84c3,
>  > 0x84c8-0x84cf,0x84c4-0x84c7,0x84d0-0x84d7 mem 0xfc340000-0xfc35ffff irq 22 at de
>  > vice 6.0 on pci4
>  > atapci0: [MPSAFE]
>  > ata2: at 0x84d0 on atapci0
>  > ata2: [MPSAFE]
>  > ata3: at 0x84c8 on atapci0
>  > ata3: [MPSAFE]
>  > isab0: <PCI-ISA bridge> at device 31.0 on pci0
>  > isa0: <ISA bus> on isab0
>  > atapci1: <Intel ICH3 UDMA100 controller> port 0x7040-0x704f,0-0x3,0-0x7,0-0x3,0-
>  > 0x7 at device 31.1 on pci0
>  > ata0: at 0x1f0 irq 14 on atapci1
>  > ata0: [MPSAFE]
>  > ata1: at 0x170 irq 15 on atapci1
>  > ata1: [MPSAFE]
>  > ...
>  > ata0-master: DMA limited to UDMA33, non-ATA66 cable or device
>  > GEOM: create disk ad0 dp=0xcb1a8360
>  > ad0: 156334MB <Maxtor 4G160J8> [317632/16/63] at ata0-master UDMA33
>  > acd0: CDROM <CD-224E> at ata1-master PIO4
>  > GEOM: create disk ad4 dp=0xcb1a7a60
>  > ad4: 156334MB <Maxtor 4G160J8> [317632/16/63] at ata2-master UDMA100
>  > GEOM: create disk ad6 dp=0xcb1a7860
>  > ad6: 156334MB <Maxtor 4G160J8> [317632/16/63] at ata3-master UDMA100
>  > ad6: WARNING - READ_DMA recovered from missing interrupt
>  > ar0: WARNING - mirror lost
>  > GEOM: create disk ar0 dp=0xcae399e0
>  > ar0: 156334MB <ATA RAID1 array> [19929/255/63] status: DEGRADED subdisks:
>  >  disk0 READY on ad4 at ata2-master
>  >  disk1 DOWN no device found for this disk
>  > SMP: AP CPU #1 Launched!
>  > SMP: AP CPU #3 Launched!
>  > SMP: AP CPU #2 Launched!
>  > ad6: WARNING - READ_DMA recovered from missing interrupt
>  >
>  > ad0 is connected to the on-board controller that only supports UDMA33 so
>  > that is no problem.
>  >
>  > The problem is the raid:
>  > [0.04] root at cave:~ > atacontrol status ar0
>  > ar0: ATA RAID1 subdisks: ad4 DOWN status: DEGRADED
>  > [0.03] root at cave:~ > atacontrol list
>  > ATA channel 0:
>  >     Master:  ad0 <Maxtor 4G160J8/GAK81CK0> ATA/ATAPI rev 6
>  >     Slave:       no device present
>  > ATA channel 1:
>  >     Master: acd0 <CD-224E/2.9B> ATA/ATAPI rev 4
>  >     Slave:       no device present
>  > ATA channel 2:
>  >     Master:  ad4 <Maxtor 4G160J8/GAK819K0> ATA/ATAPI rev 6
>  >     Slave:       no device present
>  > ATA channel 3:
>  >     Master:  ad6 <Maxtor 4G160J8/GAK819K0> ATA/ATAPI rev 6
>  >     Slave:       no device present
>  > [0.03] root at cave:~ >
>  >
>  > Using atacontrol detach und attach I can rescan channel 3 but the drive
>  > automatically reports the same error and thus the mirror is not rebuild.
>  >
>  > The same behaviour (but due to missing ATAng with different messages) can
>  > be observed on 5.1-RELEASE.
>  >
>  > Hendrik
>  >
>  > --
>  > Hendrik Scholz - <hscholz at raisdorf.net> - http://raisdorf.net/
>  > cell phone: 404-606-5324 (US) 0160-1570-272 (DE)
>  > _______________________________________________
>  > freebsd-current at freebsd.org mailing list
>  > http://lists.freebsd.org/mailman/listinfo/freebsd-current
>  > To unsubscribe, send any mail to "freebsd-current-unsubscribe at freebsd.org"
> _______________________________________________
> freebsd-current at freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-current
> To unsubscribe, send any mail to "freebsd-current-unsubscribe at freebsd.org"
>

-- 
Doug White                    |  FreeBSD: The Power to Serve
dwhite at gumbysoft.com          |  www.FreeBSD.org