Date: Wed, 21 Jan 2015 22:33:58 -0700 From: Warner Losh <imp@bsdimp.com> To: Jia-Shiun Li <jiashiun@gmail.com> Cc: freebsd-arm <freebsd-arm@freebsd.org> Subject: Re: Raspberry Pi stability improved Message-ID: <54A211A3-3B38-4CD3-BE18-2579E1FB6EBE@bsdimp.com> In-Reply-To: <CAHNYxxPrNB842wmGc2jREZZZ2QoO=Rk7vJsPx3yYie4sgO0meA@mail.gmail.com> References: <DE1783C0-C19F-448D-92EF-A51826D7FE2B@freebsd.org> <B1DA56CBB1774BB4ACE66D69BB9A662F@ad.peach.ne.jp> <DB7CB7F1-79A7-4F0D-8C99-8A62AFC74D6A@freebsd.org> <DBE5557E-81E3-42E2-8F1F-A0865E9AC477@bsdimp.com> <CAHNYxxPrNB842wmGc2jREZZZ2QoO=Rk7vJsPx3yYie4sgO0meA@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
> On Jan 21, 2015, at 10:02 PM, Jia-Shiun Li <jiashiun@gmail.com> wrote: >=20 > On Mon, Jan 19, 2015 at 5:41 AM, Warner Losh <imp@bsdimp.com> wrote: >=20 > If you reside with gpart, you may hit a frustratingly stupid gpart = bug. >=20 > We tell a lie about the CHS of the SD cards. This lie works its way = into the alignment of the partitions. > The md and da devices also lie about the CHS of SD cards. They tell = different lies (and often times > different USB adapters tell different lies) so that screws up gpart. = It assumes that the CHS is The Truth > and The One True Way to create partitions is to have them cylinder = group sized and aligned, even when > resizing an existing partition. However, since the CHS is a lie, and = nobody can get their story straight, > gpart expresses its displeasure by starting the partition at a new = location, effectively killing your > filesystem which now starts at the wrong place on the disk... >=20 > That=E2=80=99s what=E2=80=99s killing the autoresize. :( Sadly, I have = no fix for this, but it is a problem I hit about 6-9 months > ago=E2=80=A6 I talked about the problem then, but I can=E2=80=99t = recall if it was on a public mailing list or just private > email to part=E2=80=99s author. >=20 >=20 > for md, is it sufficient to give -x -y arguments when creating md = images and partitioning on them? Not sure about official release and = snapshot images, but crochet-build just give -x 63 -y 255 = unconditionally. Aligning to a more modern value, say 4K or 1M, should = be reasonable. Better if matched that on physical devices to be written = to later. Aligning to 1M is easy and likely catches most of the cases we care = about. However, there is no way to know the final values in crochet. > as for CHS, is it still relevant beyond FAT family partitions, after = 20 years of LBA? If not probably CHS alignment should only be followed = when handling FAT partitions. I understand gpart wants to keep it = generalized but in reality FAT/CHS is becoming more and more a specific = case comparing with others. gpart can choose to do this rounding or not. CHS alignment hasn=E2=80=99t = been required for FAT partitions in about 25 maybe 30 years. Warner > =20
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?54A211A3-3B38-4CD3-BE18-2579E1FB6EBE>