From: Albert ARIBAUD <albert.u.boot@aribaud.net>
To: u-boot@lists.denx.de
Subject: [U-Boot] [PATCH] disk:efi: avoid unaligned access on efi partition
Date: Mon, 14 Oct 2013 13:46:26 +0200 [thread overview]
Message-ID: <20131014134626.7e691591@lilith> (raw)
In-Reply-To: <yw1xd2n83z3x.fsf@unicorn.mansr.com>
Hi M?ns,
On Mon, 14 Oct 2013 11:50:42 +0100, M?ns Rullg?rd <mans@mansr.com>
wrote:
> Piotr Wilczek <p.wilczek@samsung.com> writes:
>
> >> -----Original Message-----
> >> From: M?ns Rullg?rd [mailto:mans at mansr.com]
> >> Sent: Saturday, October 12, 2013 1:29 AM
> >> To: Piotr Wilczek
> >> Cc: u-boot at lists.denx.de; Tom Rini; Kyungmin Park
> >> Subject: Re: [PATCH] disk:efi: avoid unaligned access on efi partition
> >>
> >> Piotr Wilczek <p.wilczek@samsung.com> writes:
> >>
> >> > In this patch static variable and memcpy instead of an assignment are
> >> > used to avoid unaligned access exception on some ARM platforms.
> >> >
> >> > Signed-off-by: Piotr Wilczek <p.wilczek@samsung.com>
> >> > Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
> >> > CC: Tom Rini <trini@ti.com>
> >> > ---
> >> > disk/part_efi.c | 6 ++++--
> >> > 1 file changed, 4 insertions(+), 2 deletions(-)
> >> >
> >> > diff --git a/disk/part_efi.c b/disk/part_efi.c index b7524d6..303b8af
> >> > 100644
> >> > --- a/disk/part_efi.c
> >> > +++ b/disk/part_efi.c
> >> > @@ -224,7 +224,8 @@ static int set_protective_mbr(block_dev_desc_t
> >> *dev_desc)
> >> > p_mbr->signature = MSDOS_MBR_SIGNATURE;
> >> > p_mbr->partition_record[0].sys_ind = EFI_PMBR_OSTYPE_EFI_GPT;
> >> > p_mbr->partition_record[0].start_sect = 1;
> >> > - p_mbr->partition_record[0].nr_sects = (u32) dev_desc->lba;
> >> > + memcpy(&p_mbr->partition_record[0].nr_sects, &dev_desc->lba,
> >> > + sizeof(dev_desc->lba));
> >>
> >> Why is this assignment problematic? Note that the compiler may
> >> optimise the memcpy() call into a plain assignment including any
> >> alignment assumptions it was making in the original code.
> >>
> >> The correct fix is either to ensure that pointers are properly aligned
> >> or that things are annotated as potentially unaligned, whichever is
> >> more appropriate.
> >>
> > Problem is that the legacy_mbr structure consists 'le16 unknown' field
> > before 'partition_record' filed and the structure is packed. As a result the
> > address of 'partition_record' filed is halfword aligned. The compiler uses
> > str/ldr instructions (address must be 4-byte aligned) to copy u32 'lba' data
> > thus causing unaligned access exception.
>
> If the struct has __attribute__((packed)), gcc should do the right thing
> already. Note that on ARMv6 and later ldr/str support unaligned
> addresses unless this is explicitly disabled in the system control
> register. If you do this, you _MUST_ compile with -mno-unaligned-access.
> Otherwise you will get problems.
Please do not advise using native unaligned accesses on code that is
not strictly used by ARMv6+ architectures: the present code, for
instance, might be run on pre-ARMv6 or non-ARM platforms, and thus,
should never assume ability to perform unaligned accesses natively.
Amicalement,
--
Albert.
next prev parent reply other threads:[~2013-10-14 11:46 UTC|newest]
Thread overview: 27+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-10-11 13:31 [U-Boot] [PATCH] disk:efi: avoid unaligned access on efi partition Piotr Wilczek
2013-10-11 16:00 ` Albert ARIBAUD
2013-10-11 23:28 ` Måns Rullgård
2013-10-14 8:30 ` Piotr Wilczek
2013-10-14 10:50 ` Måns Rullgård
2013-10-14 11:46 ` Albert ARIBAUD [this message]
2013-10-14 12:19 ` Måns Rullgård
2013-10-14 13:00 ` Albert ARIBAUD
2013-10-14 13:05 ` Måns Rullgård
2013-10-14 13:48 ` Albert ARIBAUD
2013-10-14 14:09 ` Måns Rullgård
2013-10-14 15:18 ` Albert ARIBAUD
2013-10-14 15:59 ` Måns Rullgård
2013-10-14 17:26 ` Albert ARIBAUD
2013-10-15 15:23 ` Måns Rullgård
2013-10-15 16:21 ` Albert ARIBAUD
2013-10-15 16:29 ` Måns Rullgård
2013-10-15 17:07 ` Albert ARIBAUD
2013-10-14 13:49 ` Piotr Wilczek
2013-10-14 14:22 ` Albert ARIBAUD
2014-01-28 12:46 ` [U-Boot] [PATCH V2] " Piotr Wilczek
2014-01-29 21:48 ` Tom Rini
2014-02-19 14:56 ` Hector Palacios
2014-02-19 15:03 ` Tom Rini
2014-02-19 15:23 ` Tom Rini
2014-02-24 15:56 ` Lukasz Majewski
2014-02-24 16:23 ` Tom Rini
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20131014134626.7e691591@lilith \
--to=albert.u.boot@aribaud.net \
--cc=u-boot@lists.denx.de \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox