From: Kees Cook <keescook@chromium.org>
To: David Abdurachmanov <david.abdurachmanov@gmail.com>
Cc: Palmer Dabbelt <palmer@sifive.com>,
Albert Ou <aou@eecs.berkeley.edu>,
Andy Lutomirski <luto@amacapital.net>,
Will Drewry <wad@chromium.org>,
green.hu@gmail.com, deanbo422@gmail.com,
LKML <linux-kernel@vger.kernel.org>,
linux-riscv@lists.infradead.org
Subject: Re: [PATCH 2/2] riscv: fix syscall_{get,set}_arguments
Date: Thu, 6 Dec 2018 08:48:55 -0800 [thread overview]
Message-ID: <CAGXu5jJT4wY9wgCF8gB+tjgH+YQps4eUDVwZoGY=GVmLMpLPxw@mail.gmail.com> (raw)
In-Reply-To: <20181206150156.28210-3-david.abdurachmanov@gmail.com>
On Thu, Dec 6, 2018 at 7:02 AM David Abdurachmanov
<david.abdurachmanov@gmail.com> wrote:
>
> Testing with libseccomp master branch revealed that testcases with
> filters on syscall arguments were failing due to wrong values. Seccomp
> uses syscall_get_argumentsi() to copy syscall arguments, and there is a
> bug in pointer arithmetics in memcpy() call.
>
> Two alternative implementation were tested: the one in this patch and
> another one based on while-break loop. Both delivered the same results.
>
> This implementation is also used in arm, arm64 and nds32 arches.
Minor nit: can you make this the first patch? That way seccomp works
correctly from the point of introduction. :)
-Kees
>
> Signed-off-by: David Abdurachmanov <david.abdurachmanov@gmail.com>
> ---
> arch/riscv/include/asm/syscall.h | 42 ++++++++++++++++++++++++--------
> 1 file changed, 32 insertions(+), 10 deletions(-)
>
> diff --git a/arch/riscv/include/asm/syscall.h b/arch/riscv/include/asm/syscall.h
> index bba3da6ef157..26ceb434a433 100644
> --- a/arch/riscv/include/asm/syscall.h
> +++ b/arch/riscv/include/asm/syscall.h
> @@ -70,19 +70,32 @@ static inline void syscall_set_return_value(struct task_struct *task,
> regs->a0 = (long) error ?: val;
> }
>
> +#define SYSCALL_MAX_ARGS 6
> +
> static inline void syscall_get_arguments(struct task_struct *task,
> struct pt_regs *regs,
> unsigned int i, unsigned int n,
> unsigned long *args)
> {
> - BUG_ON(i + n > 6);
> + if (n == 0)
> + return;
> +
> + if (i + n > SYSCALL_MAX_ARGS) {
> + unsigned long *args_bad = args + SYSCALL_MAX_ARGS - i;
> + unsigned int n_bad = n + i - SYSCALL_MAX_ARGS;
> + pr_warning("%s called with max args %d, handling only %d\n",
> + __func__, i + n, SYSCALL_MAX_ARGS);
> + memset(args_bad, 0, n_bad * sizeof(args[0]));
> + }
> +
> if (i == 0) {
> args[0] = regs->orig_a0;
> args++;
> i++;
> n--;
> }
> - memcpy(args, ®s->a1 + i * sizeof(regs->a1), n * sizeof(args[0]));
> +
> + memcpy(args, ®s->a0 + i, n * sizeof(args[0]));
> }
>
> static inline void syscall_set_arguments(struct task_struct *task,
> @@ -90,14 +103,23 @@ static inline void syscall_set_arguments(struct task_struct *task,
> unsigned int i, unsigned int n,
> const unsigned long *args)
> {
> - BUG_ON(i + n > 6);
> - if (i == 0) {
> - regs->orig_a0 = args[0];
> - args++;
> - i++;
> - n--;
> - }
> - memcpy(®s->a1 + i * sizeof(regs->a1), args, n * sizeof(regs->a0));
> + if (n == 0)
> + return;
> +
> + if (i + n > SYSCALL_MAX_ARGS) {
> + pr_warning("%s called with max args %d, handling only %d\n",
> + __func__, i + n, SYSCALL_MAX_ARGS);
> + n = SYSCALL_MAX_ARGS - i;
> + }
> +
> + if (i == 0) {
> + regs->orig_a0 = args[0];
> + args++;
> + i++;
> + n--;
> + }
> +
> + memcpy(®s->a0 + i, args, n * sizeof(args[0]));
> }
>
> static inline int syscall_get_arch(void)
> --
> 2.19.2
>
--
Kees Cook
next prev parent reply other threads:[~2018-12-06 16:49 UTC|newest]
Thread overview: 13+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-12-06 15:01 [PATCH 0/2] riscv: add support for SECCOMP and SECCOMP_FILTER David Abdurachmanov
2018-12-06 15:01 ` [PATCH 1/2] riscv: add support for SECCOMP incl. filters David Abdurachmanov
2018-12-06 16:47 ` Kees Cook
2018-12-06 17:10 ` David Abdurachmanov
2018-12-06 16:51 ` Kees Cook
2018-12-06 17:11 ` David Abdurachmanov
2018-12-06 18:25 ` David Abdurachmanov
2018-12-06 18:32 ` Kees Cook
2018-12-06 17:06 ` Kees Cook
2018-12-06 17:12 ` David Abdurachmanov
2018-12-06 15:01 ` [PATCH 2/2] riscv: fix syscall_{get,set}_arguments David Abdurachmanov
2018-12-06 16:48 ` Kees Cook [this message]
2018-12-06 17:13 ` David Abdurachmanov
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='CAGXu5jJT4wY9wgCF8gB+tjgH+YQps4eUDVwZoGY=GVmLMpLPxw@mail.gmail.com' \
--to=keescook@chromium.org \
--cc=aou@eecs.berkeley.edu \
--cc=david.abdurachmanov@gmail.com \
--cc=deanbo422@gmail.com \
--cc=green.hu@gmail.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-riscv@lists.infradead.org \
--cc=luto@amacapital.net \
--cc=palmer@sifive.com \
--cc=wad@chromium.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).