Re: kern/59859: NetBSD-10.1 amd64 kernel panic under X

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

The following reply was made to PR kern/59859; it has been noted by GNATS.

From: Ramiro Aceves <ea1abz%gmail.com@localhost>
To: Nat Sloss <nathanialsloss%yahoo.com.au@localhost>
Cc: gnats-bugs%netbsd.org@localhost
Subject: Re: kern/59859: NetBSD-10.1 amd64 kernel panic under X
Date: Mon, 5 Jan 2026 12:33:22 +0100

 El 5/1/26 a las 11:55, Nat Sloss escribió:
 >> I think that the procedure would be:
 > 
 >> -Change "src/sys/dev/usb/usbdevs" to add Vendor and Product of the Fast 
 > 
 >> Track Pro.
 > 
 >> -Execute "make -f Makefile.usbdevs" in "src/sys/dev/usb" to regenerate 
 > 
 >> "src/sys/dev/usb/usbdevs.h" and "src/sys/dev/usb/usbdevs_data.h"
 > 
 >> -Edit usb_quirks.c and add
 > 
 >> UQ_AU_NO_FRAC and/or UQ_AU_INP_ASYNC lines for my card.
 > 
 >> -Recompile de kernel and reboot into it.
 > 
 > 
 > That's correct.
 > 
 > 
 > 
 >> What should I expect from that change?
 > 
 > 
 > I think there's a possibilty that the card stalls in the record 
 > interrupt due to fractional
 > 
 > packets being read from the device so I'd try AU_NO_FRAC as a quirk first.
 > 
 > 
 > Hopefully It will allow for the device to not issue IOERRORS and 
 > continue working.
 > 
 >> Thanks.
 > 
 >> Ramiro.
 > 
 >> PS: or perhaps may I modify usb_quirks.c directly placing the numbers 
 > 
 >> and recompile the kernel? I wait for your advice.
 > 
 > 
 > Yes you can do that, if it works I'll prepare a formal patch to commit.
 > 
 > 
 > Best regards,
 > 
 > 
 > Nat
 > 
 > 
 > 
 
 Hi, I did that:
 
 netbsd-nuc$ diff usb_quirks.c.orig usb_quirks.c
 256a257,258
  >  { 0x0763,0x2012,			0x100,
  > 	{ UQ_AU_NO_FRAC, NULL }},
 netbsd-nuc$
 
 recompiled with:
 
 netbsd-nuc$ ./build.sh -u -U kernel=MYKERNEL
 
 
 netbsd-nuc# cp /usr/src/sys/arch/amd64/compile/obj/MYKERNEL/netbsd 
 /netbsdnofrac
 
 I will boot that netbsdnofrac kernel and see what happens.
 
 Regards.