Re: Panic on boot with today's CURRENT, ata related

Radek Kozlowski wrote:
> On Fri, Aug 06, 2004 at 01:33:46PM -0700, Nate Lawson wrote:
> 
>>I took a quick look at this ATA panic.  The exact same one occurs for 
>>Ceri.  A quick dissassemble shows that the testb is the check for the 
>>DMA flag at the very end of ata_generic_transaction().  The bug appears 
>>to be that this may be a PIO request (since the DMA check is outside the 
>>switch() statement).  The fix is to make sure it's a DMA request before 
>>dereferencing an element of the DMA struct.  Try the attached patch.
> 
> 
> I applied your changes to ata-lowlevel.c revision 1.41 manually and they
> seem to solve the issue--my system now boots succesfully, thank you.

You're welcome.  I committed the fix.

> I would also like to point out that since around the beggining of June
> the ad0 here no longer boots in UDMA100 mode but uses PIO4 instead
> (http://lists.freebsd.org/pipermail/freebsd-current/2004-June/029434.html).
> 
> Trying to change the transfer mode to UDMA100 using atacontrol results
> in a panic. I don't know if this backtrace will be in any way useful but
> here it goes:
> 
> Fatal trap 12: page fault while in kernel mode
> fault virtual address   = 0x48
> fault code              = supervisor read, page not present
> instruction pointer     = 0x8:0xc0449391
> stack pointer           = 0x10:0xd302dba0
> frame pointer           = 0x10:0xd302dbd4
> code segment            = base 0x0, limit 0xfffff, type 0x1b
>                         = DPL 0, pres 1, def32 1, gran 1
> processor eflags        = interrupt enabled, resume, IOPL = 0
> current process         = 4 (g_down)
> 
> (kgdb) bt
> #0  doadump () at pcpu.h:159
> #1  0xc04427f5 in db_fncall (dummy1=0, dummy2=0, dummy3=3999, dummy4=0xd302d9ac " `q<C0>\f")
>     at /usr/src/sys/ddb/db_command.c:531
> #2  0xc0442572 in db_command (last_cmdp=0xc0715724, cmd_table=0x0, aux_cmd_tablep=0xc06e2d90,
>     aux_cmd_tablep_end=0xc06e2d94) at /usr/src/sys/ddb/db_command.c:349
> #3  0xc044267a in db_command_loop () at /usr/src/sys/ddb/db_command.c:455
> #4  0xc0444695 in db_trap (type=12, code=0) at /usr/src/sys/ddb/db_main.c:221
> #5  0xc052c057 in kdb_trap (type=0, code=0, tf=0xd302db60) at /usr/src/sys/kern/subr_kdb.c:401
> #6  0xc068e5c8 in trap_fatal (frame=0xd302db60, eva=0) at /usr/src/sys/i386/i386/trap.c:807
> #7  0xc068e2f3 in trap_pfault (frame=0xd302db60, usermode=0, eva=72) at /usr/src/sys/i386/i386/trap.c:730
> #8  0xc068de82 in trap (frame=
>       {tf_fs = 24, tf_es = 16, tf_ds = 16, tf_edi = -1047235072, tf_esi = -1045272284, tf_ebp = -754787372, tf_isp = -754787444, tf_ebx = -1
> 047235072, tf_edx = 0, tf_ecx = 9167, tf_eax = -1047234904, tf_trapno = 12, tf_err = 0, tf_eip = -1069247599, tf_cs = 8, tf_eflags = 66050,
> tf_esp = -1047234904, tf_ss = -805982208}) at /usr/src/sys/i386/i386/trap.c:417
> #9  0xc0449391 in ata_generic_transaction (request=0xc1b26924) at /usr/src/sys/dev/ata/ata-lowlevel.c:144
> #10 0xc0448379 in ata_start (ch=0xc1947600) at /usr/src/sys/dev/ata/ata-queue.c:194
> #11 0xc0447f0c in ata_queue_request (request=0xc1b26924) at /usr/src/sys/dev/ata/ata-queue.c:91
> #12 0xc04595d8 in ad_start (atadev=0xc19476a8) at /usr/src/sys/dev/ata/ata-disk.c:310
> #13 0xc04481c8 in ata_start (ch=0xc1947600) at /usr/src/sys/dev/ata/ata-queue.c:168
> #14 0xc045922e in adstrategy (bp=0xc1e5f738) at /usr/src/sys/dev/ata/ata-disk.c:229
> #15 0xc04cf8f6 in g_disk_start (bp=0xc1b29e70) at /usr/src/sys/geom/geom_disk.c:277
> #16 0xc04d17c0 in g_io_schedule_down (tp=0xc188f000) at /usr/src/sys/geom/geom_io.c:365
> #17 0xc04d1b1e in g_down_procbody () at /usr/src/sys/geom/geom_kern.c:113
> #18 0xc04f4ed0 in fork_exit (callout=0xc04d1b00 <g_down_procbody>, arg=0x0, frame=0x0)
>     at /usr/src/sys/kern/kern_fork.c:819

It looks like the DMA object for the channel is not being initialized. 
However, that is in another path (this backtrace is just a symptom 
triggered by the first DMA access to the drive).  Someone who knows more 
about the control path for going from PIO -> DMA mode will have to look 
into this.

-Nate