diff mbox

[for-2.7] linux-user: Fix llseek with high bit of offset_low set

Message ID 1470938379-1133-1-git-send-email-peter.maydell@linaro.org (mailing list archive)
State New, archived
Headers show

Commit Message

Peter Maydell Aug. 11, 2016, 5:59 p.m. UTC
The llseek syscall takes two 32-bit arguments, offset_high
and offset_low, which must be combined to form a single
64-bit offset. Unfortunately we were combining them with
   (uint64_t)arg2 << 32) | arg3
and arg3 is a signed type; this meant that when promoting
arg3 to a 64-bit type it would be sign-extended. The effect
was that if the offset happened to have bit 31 set then
this bit would get sign-extended into all of bits 63..32.
Explicitly cast arg3 to abi_ulong to avoid the erroneous
sign extension.

Reported-by: Chanho Park <parkch98@gmail.com>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
---
Long-standing bug and we're quite close to 2.7 but the
fix is trivial so if somebody would like to review it
I think we could put it in...

 linux-user/syscall.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Comments

Chanho Park Aug. 15, 2016, 1:41 p.m. UTC | #1
It works perfectly.
Thanks.

Tested-by: Chanho Park <parkch98@gmail.com>

On Fri, Aug 12, 2016 at 2:59 AM, Peter Maydell <peter.maydell@linaro.org> wrote:
> The llseek syscall takes two 32-bit arguments, offset_high
> and offset_low, which must be combined to form a single
> 64-bit offset. Unfortunately we were combining them with
>    (uint64_t)arg2 << 32) | arg3
> and arg3 is a signed type; this meant that when promoting
> arg3 to a 64-bit type it would be sign-extended. The effect
> was that if the offset happened to have bit 31 set then
> this bit would get sign-extended into all of bits 63..32.
> Explicitly cast arg3 to abi_ulong to avoid the erroneous
> sign extension.
>
> Reported-by: Chanho Park <parkch98@gmail.com>
> Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
> ---
> Long-standing bug and we're quite close to 2.7 but the
> fix is trivial so if somebody would like to review it
> I think we could put it in...
>
>  linux-user/syscall.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/linux-user/syscall.c b/linux-user/syscall.c
> index ebdb753..b4e21d3 100644
> --- a/linux-user/syscall.c
> +++ b/linux-user/syscall.c
> @@ -9406,7 +9406,7 @@ abi_long do_syscall(void *cpu_env, int num, abi_long arg1,
>          {
>              int64_t res;
>  #if !defined(__NR_llseek)
> -            res = lseek(arg1, ((uint64_t)arg2 << 32) | arg3, arg5);
> +            res = lseek(arg1, ((uint64_t)arg2 << 32) | (abi_ulong)arg3, arg5);
>              if (res == -1) {
>                  ret = get_errno(res);
>              } else {
> --
> 2.7.4
>
Riku Voipio Aug. 16, 2016, 1:44 p.m. UTC | #2
Hi,

applied to linux-user-for upstream.

Riku

On Mon, Aug 15, 2016 at 10:41:35PM +0900, Chanho Park wrote:
> It works perfectly.
> Thanks.
> 
> Tested-by: Chanho Park <parkch98@gmail.com>
> 
> On Fri, Aug 12, 2016 at 2:59 AM, Peter Maydell <peter.maydell@linaro.org> wrote:
> > The llseek syscall takes two 32-bit arguments, offset_high
> > and offset_low, which must be combined to form a single
> > 64-bit offset. Unfortunately we were combining them with
> >    (uint64_t)arg2 << 32) | arg3
> > and arg3 is a signed type; this meant that when promoting
> > arg3 to a 64-bit type it would be sign-extended. The effect
> > was that if the offset happened to have bit 31 set then
> > this bit would get sign-extended into all of bits 63..32.
> > Explicitly cast arg3 to abi_ulong to avoid the erroneous
> > sign extension.
> >
> > Reported-by: Chanho Park <parkch98@gmail.com>
> > Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
> > ---
> > Long-standing bug and we're quite close to 2.7 but the
> > fix is trivial so if somebody would like to review it
> > I think we could put it in...
> >
> >  linux-user/syscall.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/linux-user/syscall.c b/linux-user/syscall.c
> > index ebdb753..b4e21d3 100644
> > --- a/linux-user/syscall.c
> > +++ b/linux-user/syscall.c
> > @@ -9406,7 +9406,7 @@ abi_long do_syscall(void *cpu_env, int num, abi_long arg1,
> >          {
> >              int64_t res;
> >  #if !defined(__NR_llseek)
> > -            res = lseek(arg1, ((uint64_t)arg2 << 32) | arg3, arg5);
> > +            res = lseek(arg1, ((uint64_t)arg2 << 32) | (abi_ulong)arg3, arg5);
> >              if (res == -1) {
> >                  ret = get_errno(res);
> >              } else {
> > --
> > 2.7.4
> >
> 
> 
> 
> -- 
> Best Regards,
> Chanho Park
Peter Maydell Aug. 16, 2016, 4:10 p.m. UTC | #3
On 16 August 2016 at 14:44, Riku Voipio <riku.voipio@iki.fi> wrote:
> Hi,
>
> applied to linux-user-for upstream.

Applied to master, thanks.

-- PMM
diff mbox

Patch

diff --git a/linux-user/syscall.c b/linux-user/syscall.c
index ebdb753..b4e21d3 100644
--- a/linux-user/syscall.c
+++ b/linux-user/syscall.c
@@ -9406,7 +9406,7 @@  abi_long do_syscall(void *cpu_env, int num, abi_long arg1,
         {
             int64_t res;
 #if !defined(__NR_llseek)
-            res = lseek(arg1, ((uint64_t)arg2 << 32) | arg3, arg5);
+            res = lseek(arg1, ((uint64_t)arg2 << 32) | (abi_ulong)arg3, arg5);
             if (res == -1) {
                 ret = get_errno(res);
             } else {