Re: port-amd64/53113: NetBSD_8.0_BETA/amd64 fails to boot.

[Utkarsh Anand <utkarsh009%yandex.com@localhost>]

21.03.2018, 15:10, "Utkarsh Anand" <utkarsh009%yandex.com@localhost>:
> The following reply was made to PR port-amd64/53113; it has been noted by GNATS.
>
> From: Utkarsh Anand <utkarsh009%yandex.com@localhost>
> To: Martin Husemann <martin%duskware.de@localhost>
> Cc: "gnats-bugs%netbsd.org@localhost" <gnats-bugs%netbsd.org@localhost>
> Subject: Re: port-amd64/53113: NetBSD_8.0_BETA/amd64 fails to boot.
> Date: Wed, 21 Mar 2018 15:06:57 +0530
>
>  21.03.2018, 14:43, "Utkarsh Anand" <utkarsh009%yandex.com@localhost>:
>  > 21.03.2018, 14:34, "Martin Husemann" <martin%duskware.de@localhost>:
>  >>  On Wed, Mar 21, 2018 at 02:03:06PM +0530, Utkarsh Anand wrote:
>  >>
>  >>>   BTW, looking at this:
>  >>>   https://nxr.netbsd.org/xref/src/distrib/amd64/liveimage/usbimage/Makefile#5
>  >>>   and confirming from the size of the image, it seems that
>  >>>   ${IMAGESECTORS} must be 3145728 (1536*1024*1024/512 taken from
>  >>>   https://nxr.netbsd.org/xref/src/distrib/common/bootimage/Makefile.bootimage#163
>  >>>   ), and that should be the partition d: in
>  >>>   https://nxr.netbsd.org/xref/src/distrib/common/bootimage/diskproto.noswap.mbr.in#24
>  >>>   ). According to freebsd, it is 2867200, which should be wrong. So, any
>  >>>   ideas on how this happened? Can anybody build an install-image, write
>  >>>   it to a flash drive and double check the disklabel? I did not change
>  >>>   any values, everything was set to default.
>  >>
>  >>  I am not sure what image exactly you did use.
>  >
>  > I build 8.0_BETA myself using build.sh with `release install-image` options.
>  >>  I downloaded NetBSD-8.99.14.amd64-install.img.gz from the latest
>  >>  nycdn/daily/HEAD build and extracted it onto a USB stick.
>  >>
>  >>  The disklabel looks like this:
>  >>
>  >>  # /dev/rsd0d:
>  >>  type: ESDI
>  >>  disk: image
>  >>  label:
>  >>  flags:
>  >>  bytes/sector: 512
>  >>  sectors/track: 32
>  >>  tracks/cylinder: 64
>  >>  sectors/cylinder: 2048
>  >>  cylinders: 1450
>  >>  total sectors: 2969600
>  >>  rpm: 3600
>  >>  interleave: 1
>  >>  trackskew: 0
>  >>  cylinderskew: 0
>  >>  headswitch: 0 # microseconds
>  >>  track-to-track seek: 0 # microseconds
>  >>  drivedata: 0
>  >>
>  >>  8 partitions:
>  >>  # size offset fstype [fsize bsize cpg/sgs]
>  >>   a: 2967552 2048 4.2BSD 1024 8192 16 # (Cyl. 1 - 1449)
>  >>   c: 2967552 2048 unused 0 0 # (Cyl. 1 - 1449)
>  >>   d: 2969600 0 unused 0 0 # (Cyl. 0 - 1449)
>  Hmm.. According to this: https://nxr.netbsd.org/xref/src/distrib/amd64/installimage/Makefile#8 , your image has the correct number of sectors as it should be, which means, there was some problem in generating my image file. Can you cross check by building 8.0_BETA? Meanwhile, I'll just check the image for any errors.
>  >
>  > Which file (and what variables) decide the size of partition d? (and thus a and c, as they are calculated from it.)
>  >>  The ffs file system in partition a: fsck's cleanly and is mountable/usable.
>  >>
>  >>  Of course (my umass stick is slightly bigger than the image) I a few logs like:
>  >>
>  >>  sd0: 5304320 trailing sectors not covered by disklabel
>  >>
>  >>  but that is harmless.
>  >>
>  >>  Martin

Hey! https://github.com/NetBSD/src/commit/c20c0c7bce5a6b954d5fea0a3932bfa569523a57 shows that netbsd-8 still uses 1400 MB as size while netbsd-current uses 1450 MB which was bumped 6 months ago https://github.com/NetBSD/src/commit/48053fec549ddfcb2412db8778f0cbf7e6bc9f63 . So my number of sectors is correct too. Could 'not bumping' the size for netbsd-8 be a problem? should it also be bumped