[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20210421174105.GB52940@C02TD0UTHF1T.local>
Date: Wed, 21 Apr 2021 18:41:05 +0100
From: Mark Rutland <mark.rutland@....com>
To: He Zhe <zhe.he@...driver.com>
Cc: oleg@...hat.com, catalin.marinas@....com, will@...nel.org,
linux-arm-kernel@...ts.infradead.org, paul@...l-moore.com,
eparis@...hat.com, linux-audit@...hat.com,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH 2/3] arm64: syscall.h: Add sign extension handling in
syscall_get_return_value for compat
On Fri, Apr 16, 2021 at 03:55:32PM +0800, He Zhe wrote:
> Add sign extension handling in syscall_get_return_value so that it can
> handle 32-bit compatible case and can be used by for example audit, just
> like what syscall_get_error does.
If a compat syscall can ever legitimately return a non-error value with
bit 31 set, and this sign-extends it, is that ever going to reach
userspace as a 64-bit value?
IIUC things like mmap() can return pointers above 2GiB for a compat
task, so I'm a bit uneasy that we'd handle those wrong. I can't see a
way of preventing that unless we keep the upper 32 bits for errors.
Mark.
>
> Signed-off-by: He Zhe <zhe.he@...driver.com>
> ---
> arch/arm64/include/asm/syscall.h | 7 ++++++-
> 1 file changed, 6 insertions(+), 1 deletion(-)
>
> diff --git a/arch/arm64/include/asm/syscall.h b/arch/arm64/include/asm/syscall.h
> index cfc0672013f6..cd7a22787aeb 100644
> --- a/arch/arm64/include/asm/syscall.h
> +++ b/arch/arm64/include/asm/syscall.h
> @@ -44,7 +44,12 @@ static inline long syscall_get_error(struct task_struct *task,
> static inline long syscall_get_return_value(struct task_struct *task,
> struct pt_regs *regs)
> {
> - return regs->regs[0];
> + long val = regs->regs[0];
> +
> + if (is_compat_thread(task_thread_info(task)))
> + val = sign_extend64(val, 31);
> +
> + return val;
> }
>
> static inline void syscall_set_return_value(struct task_struct *task,
> --
> 2.17.1
>
Powered by blists - more mailing lists