Re: port-hpcarm/60283

["requiem. via gnats" <gnats-admin%NetBSD.org@localhost>]

The following reply was made to PR port-hpcarm/60283; it has been noted by GNATS.

From: "requiem." <rqm%0xde501a7a.net@localhost>
To: gnats-bugs%netbsd.org@localhost
Cc: 
Subject: Re: port-hpcarm/60283
Date: Tue, 9 Jun 2026 12:55:01 +0100

 Hello uwe, 
 
 Tried booting the NetBSD8.gz kernel; and I am getting stuck with the
 same issue on boot:
 
 wdc0:0:0: lost interrupt
 	type: ata tc_bcount: 512 tc_skip: 0
 wd0d: device timeout reading fsbn 0 (wd0 bn 0; cn 0 tn 0 sn 0), retrying
 wdc0:0:0: lost interrupt
 	type: ata tc_bcount: 512 tc_skip: 0
 wd0d: device timeout reading fsbn 0 (wd0 bn 0; cn 0 tn 0 sn 0), retrying
 wdc0:0:0: lost interrupt
 	type: ata tc_bcount: 512 tc_skip: 0
 wd0d: device timeout reading fsbn 0 (wd0 bn 0; cn 0 tn 0 sn 0), retrying
 wdc0:0:0: lost interrupt
 	type: ata tc_bcount: 512 tc_skip: 0
 wd0d: device timeout reading fsbn 0 (wd0 bn 0; cn 0 tn 0 sn 0), retrying
 wdc0:0:0: recal error (5)
 wd0d: error reading fsbn 0 (wd0 bn 0; cn 0 tn 0 sn 0)
 wd0: (address mark not found, aborted command)
 wdc0:0:0: lost interrupt
 	type: ata tc_bcount: 512 tc_skip: 0
 wd0d: device timeout reading fsbn 0 (wd0 bn 0; cn 0 tn 0 sn 0), retrying
 wdc0:0:0: recal error (60)
 wd0d: error reading fsbn 0 (wd0 bn 0; cn 0 tn 0 sn 0), retrying
 wd0: (media changed, uncorrectable data error)
 wdc0:0:0: recal drive fault
 wd0d: device fault reading fsbn 0 (wd0 bn 0; cn 0 tn 0 sn 0), retrying
 
 
 The CF Card works fine in the machine otherwise; it's an old 1G
 Industrial grade CF card (used to hold a Windows XP Embedded
 installation).
 
 However I just discovered not long ago that what I purchased listed as
 a "8600" is in fact a *8630*. Might this be the cause; is the
 underlying hardware just different enough to cause these problems?
 
 On Mon, 25 May 2026 08:20:02 +0000 (UTC)
 "Valery Ushakov via gnats" <gnats-admin%NetBSD.org@localhost> wrote:
 
 > The following reply was made to PR port-hpcarm/60283; it has been
 > noted by GNATS.
 > 
 > From: Valery Ushakov <uwe%stderr.spb.ru@localhost>
 > To: gnats-bugs%netbsd.org@localhost
 > Cc: 
 > Subject: Re: port-hpcarm/60283
 > Date: Mon, 25 May 2026 11:17:22 +0300
 > 
 >  On Sun, May 24, 2026 at 21:41:06 +0100, requiem. wrote:
 >  [...]
 >  > [   1.0542408] wd0 at atabus0 drive 0
 >  > [   1.1042437] syscall 295 is busy
 >  > wdc0:0:0: lost interrupt
 >  > [   4.0642161] 	type: ata tc_bcount: 512 tc_skip: 0
 >  > [   4.1010932] wd0: <SMART CF>
 >  > [   4.1177233] wd0: drive supports 1-sector PIO transfers, LBA
 >  > addressing [   4.1479389] wd0: 971 MB, 1974 cyl, 16 head, 63 sec,
 >  > 512 bytes/sect x 1989792 sectors  
 >  
 >  This:
 >  
 >    wdc0:0:0: lost interrupt
 >    type: ata tc_bcount: 512 tc_skip: 0
 >  
 >  probably indicates that something is not right when netbsd tries to
 >  talk to the card.
 >  
 >  I'm not familiar with this hardware or wdc(4) internals.  One thing
 >  that you can try _easily_ is using older NetBSD releases.  You don't
 >  have to bother with the installation, just drop an old kernel to the
 >  CF card under a different name and try to boot it.  Some wd/wdc
 >  changes might have broken the driver for hpcarm b/c it, obviously,
 >  doesn't get a lot of testing there...
 >  
 >  The next thing to try (if none of the old kernels work) would be
 > using an older, simpler CF card.  I don't know _anything_ about the
 >  hardware, but HPC is _old_ and who knows what corners they cut.
 >  E.g. you can't use large cards (a 8GB CF card doesn't work in my
 >  Jornada 690) b/c some address lines are not there or something.  I
 > did run into cards that just didn't work in my hpcsh, though it's been
 >  decades ago and I don't remember any details.
 >  
 >  
 >  -uwe
 >