MacbookPro4,1, atapci/ata stall during boot

Rick van der Zwet info at
Tue Mar 8 17:22:41 UTC 2011

After 8.1-RELEASE was running fine, upgrading to 8.2-RELEASE caused to
stalls completely without crash during boot. I now went for a fresh
-CURRENT which still stalls/freezes completely without crashing with
last messages (captured via firewire/dcons):
    atapci0: <Intel ICH8M UDMA100 controller> port
irq 21 at device 31.1 on pci0
    ata2: <ATA channel 0> on atapci0
    ata2: reset tp1 mask=03 ostat0=50 ostat1=00
    ata2: stat0=0x00 err=0x01 lsb=0x14 msb=0xeb
    ata2: stat1=0x00 err=0x00 lsb=0x00 msb=0x00
    ata2: reset tp2 stat0=00 stat1=00 devices=0x10000
    atapci1: <Intel ICH8M SATA300 controller> port
irq 18 at device 31.2 on pci0
    ata3: <ATA channel 0> on atapci1

My MacbookPro4,1 is using (S)ATA controller chipset :
        atapci0 at pci0:0:31:1:	class=0x01018f card=0x00a3106b
chip=0x28508086 rev=0x04 hdr=0x00
        vendor     = 'Intel Corporation'
        device     = '82801H (ICH8 Family) Ultra ATA Storage Controllers'
        class      = mass storage
        subclass   = ATA
    atapci1 at pci0:0:31:2:	class=0x01018f card=0x00a3106b
chip=0x28288086 rev=0x04 hdr=0x00
        vendor     = 'Intel Corporation'
        device     = '82801 (ICH8M) 3 port SATA Controller'
        class      = mass storage
        subclass   = ATA

The issue seem to be introduced in r214016 which is a fairly large
commit about 'Set of legacy mode SATA enchancements'.

If  using the tree sys/dev/ata at 214015 it boots fine:
   # svn up -r214015 /usr/src-head/sys/dev/ata
   # make -C /usr/src-head buildkernel installkernel
And the next commit  (which affects my chipset as well) it stops working:
   # svn up -r214016 /usr/src-head/sys/dev/ata
   # make -C /usr/src-head buildkernel installkernel

Does anybody has the same SATA controller chipset (ATA_I82801HBM_S1),
and is able to boot fine?

Any hints on how to debug this issue much appreciated as I having
quite some trouble understanding what the patch changed and how-to
debug this, as it completely freezes the machine, making it impossible
to get into the debugger.

Br. /Rick

More information about the freebsd-current mailing list