From: will.deacon@arm.com (Will Deacon)
To: linux-arm-kernel@lists.infradead.org
Subject: linux-4.4-rc8/arch/arm64/kernel/module.c:78: 32/64 bit problem ?
Date: Mon, 4 Jan 2016 15:28:41 +0000 [thread overview]
Message-ID: <20160104152841.GG1616@arm.com> (raw)
In-Reply-To: <CAKv+Gu9Ve6=1evFD4V2jtOCHOEqOzhmud=guioVR5NMrc=HPBw@mail.gmail.com>
On Mon, Jan 04, 2016 at 04:24:49PM +0100, Ard Biesheuvel wrote:
> On 4 January 2016 at 15:32, Ard Biesheuvel <ard.biesheuvel@linaro.org> wrote:
> > On 4 January 2016 at 15:16, Will Deacon <will.deacon@arm.com> wrote:
> >> On Mon, Jan 04, 2016 at 08:25:30AM +0000, David Binderman wrote:
> >>> Hello there,
> >>
> >> Hi David,
> >>
> >>> [linux-4.4-rc8/arch/arm64/kernel/module.c:78] ->
> >>> [linux-4.4-rc8/arch/arm64/kernel/module.c:88]: (warning) Shifting 32-bit
> >>> value by 64 bits is undefined behaviour. See condition at line 88.
> >>
> >> Curious, but how are you seeing this warning? GCC is silent for me...
> >>
> >>> Source code is
> >>>
> >>> u64 imm_mask = (1 << len) - 1;
> >>> s64 sval = do_reloc(op, place, val);
> >>>
> >>> switch (len) {
> >>> case 16:
> >>> *(s16 *)place = sval;
> >>> break;
> >>> case 32:
> >>> *(s32 *)place = sval;
> >>> break;
> >>> case 64:
> >>>
> >>> So it seems that len can be 64. Suggest new code
> >>>
> >>> u64 imm_mask = (1UL << len) - 1;
> >>
> >> That still ends up shifting by the width of the type when len == 64,
> >> which is potentially still broken. We're better off using GENMASK.
> >>
> >
> > Can't we simply return from the function rather than break from the
> > switch statement if len == 64?
> > The range check does not make any sense in that case anyway.
> >
>
> Or perhaps:
>
> ------------8<-----------------
> diff --git a/arch/arm64/kernel/module.c b/arch/arm64/kernel/module.c
> index f4bc779e62e8..fd1f4e678655 100644
> --- a/arch/arm64/kernel/module.c
> +++ b/arch/arm64/kernel/module.c
> @@ -75,14 +75,17 @@ static u64 do_reloc(enum aarch64_reloc_op
> reloc_op, void *place, u64 val)
>
> static int reloc_data(enum aarch64_reloc_op op, void *place, u64 val, int len)
> {
> - u64 imm_mask = (1 << len) - 1;
> s64 sval = do_reloc(op, place, val);
>
> switch (len) {
> case 16:
> + if (sval < S16_MIN || sval > U16_MAX)
> + return -ERANGE;
> *(s16 *)place = sval;
Doesn't this break ABS relocs, which are allowed to overflow?
Will
next prev parent reply other threads:[~2016-01-04 15:28 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-01-04 8:25 linux-4.4-rc8/arch/arm64/kernel/module.c:78: 32/64 bit problem ? David Binderman
2016-01-04 14:16 ` Will Deacon
2016-01-04 14:32 ` Ard Biesheuvel
2016-01-04 15:24 ` Ard Biesheuvel
2016-01-04 15:28 ` Will Deacon [this message]
2016-01-04 15:30 ` Will Deacon
2016-01-04 15:14 ` David Binderman
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=20160104152841.GG1616@arm.com \
--to=will.deacon@arm.com \
--cc=linux-arm-kernel@lists.infradead.org \
/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;
as well as URLs for NNTP newsgroup(s).