Patchwork savevm.c: Fix compilation error on 32bit host.

login
register
mail settings
Submitter Evgeny Voevodin
Date Aug. 15, 2012, 9:10 a.m.
Message ID <1345021819-892-1-git-send-email-e.voevodin@samsung.com>
Download mbox | patch
Permalink /patch/177581/
State New
Headers show

Comments

Evgeny Voevodin - Aug. 15, 2012, 9:10 a.m.
Casting of 0x0101010101010101ULL to long will truncate it to 32
bits on 32bit hosts, and won't truncate on 64bit hosts.

Signed-off-by: Evgeny Voevodin <e.voevodin@samsung.com>
---
 savevm.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
Peter Maydell - Aug. 15, 2012, 9:30 a.m.
On 15 August 2012 10:10, Evgeny Voevodin <e.voevodin@samsung.com> wrote:
> Casting of 0x0101010101010101ULL to long will truncate it to 32
> bits on 32bit hosts, and won't truncate on 64bit hosts.
>
> Signed-off-by: Evgeny Voevodin <e.voevodin@samsung.com>

Dup of http://patchwork.ozlabs.org/patch/177217/ I'm afraid.

-- PMM
Evgeny Voevodin - Aug. 15, 2012, 9:34 a.m.
On 08/15/2012 01:30 PM, Peter Maydell wrote:
> On 15 August 2012 10:10, Evgeny Voevodin <e.voevodin@samsung.com> wrote:
>> Casting of 0x0101010101010101ULL to long will truncate it to 32
>> bits on 32bit hosts, and won't truncate on 64bit hosts.
>>
>> Signed-off-by: Evgeny Voevodin <e.voevodin@samsung.com>
>
> Dup of http://patchwork.ozlabs.org/patch/177217/ I'm afraid.
>
> -- PMM
>
>

Don't be afraid, it's true. I didn't see it in the mailing list and
didn't know that the bug is already fixed there.
Michael Tokarev - Aug. 15, 2012, 11:07 a.m.
On 15.08.2012 13:10, Evgeny Voevodin wrote:
> Casting of 0x0101010101010101ULL to long will truncate it to 32
> bits on 32bit hosts, and won't truncate on 64bit hosts.
> 
> Signed-off-by: Evgeny Voevodin <e.voevodin@samsung.com>
> ---
>  savevm.c |    2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/savevm.c b/savevm.c
> index 0ea10c9..9ab4d83 100644
> --- a/savevm.c
> +++ b/savevm.c
> @@ -2473,7 +2473,7 @@ int xbzrle_encode_buffer(uint8_t *old_buf, uint8_t *new_buf, int slen,
>          /* word at a time for speed, use of 32-bit long okay */
>          if (!res) {
>              /* truncation to 32-bit long okay */
> -            long mask = 0x0101010101010101ULL;
> +            long mask = (long)0x0101010101010101ULL;
>              while (i < slen) {
>                  xor = *(long *)(old_buf + i) ^ *(long *)(new_buf + i);
>                  if ((xor - mask) & ~xor & (mask << 7)) {

Um, how it is ugly... Can't we use unsigned types for all this stuff?
Including slen too - the function parameter...

Thanks,

/mjt

Patch

diff --git a/savevm.c b/savevm.c
index 0ea10c9..9ab4d83 100644
--- a/savevm.c
+++ b/savevm.c
@@ -2473,7 +2473,7 @@  int xbzrle_encode_buffer(uint8_t *old_buf, uint8_t *new_buf, int slen,
         /* word at a time for speed, use of 32-bit long okay */
         if (!res) {
             /* truncation to 32-bit long okay */
-            long mask = 0x0101010101010101ULL;
+            long mask = (long)0x0101010101010101ULL;
             while (i < slen) {
                 xor = *(long *)(old_buf + i) ^ *(long *)(new_buf + i);
                 if ((xor - mask) & ~xor & (mask << 7)) {