[Qemu-devel] [PATCH] migration: Don't use *_to_cpup()

[Qemu-devel] [PATCH] migration: Don't use *_to_cpup()

[Peter Maydell]

The *_to_cpup() functions just compose a pointer dereference with
a *_to_cpu() byteswap. Instead use ld*_p(), which handles potential
pointer misaligment and avoids the need to cast the pointer.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
---
The motivation here is that I'd like to get rid of _to_cpup()
entirely: we don't have many places that use it.
---
 migration/migration.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/migration/migration.c b/migration/migration.c
index 7ecbade..551da0a 100644
--- a/migration/migration.c
+++ b/migration/migration.c
@@ -1381,7 +1381,7 @@ static void *source_return_path_thread(void *opaque)
         /* OK, we have the message and the data */
         switch (header_type) {
         case MIG_RP_MSG_SHUT:
-            sibling_error = be32_to_cpup((uint32_t *)buf);
+            sibling_error = ldl_be_p(buf);
             trace_source_return_path_thread_shut(sibling_error);
             if (sibling_error) {
                 error_report("RP: Sibling indicated error %d", sibling_error);
@@ -1395,13 +1395,13 @@ static void *source_return_path_thread(void *opaque)
             goto out;
 
         case MIG_RP_MSG_PONG:
-            tmp32 = be32_to_cpup((uint32_t *)buf);
+            tmp32 = ldl_be_p(buf);
             trace_source_return_path_thread_pong(tmp32);
             break;
 
         case MIG_RP_MSG_REQ_PAGES:
-            start = be64_to_cpup((uint64_t *)buf);
-            len = be32_to_cpup((uint32_t *)(buf + 8));
+            start = ldq_be_p(buf);
+            len = ldl_be_p(buf + 8);
             migrate_handle_rp_req_pages(ms, NULL, start, len);
             break;
 
@@ -1409,8 +1409,8 @@ static void *source_return_path_thread(void *opaque)
             expected_len = 12 + 1; /* header + termination */
 
             if (header_len >= expected_len) {
-                start = be64_to_cpup((uint64_t *)buf);
-                len = be32_to_cpup((uint32_t *)(buf + 8));
+                start = ldq_be_p(buf);
+                len = ldl_be_p(buf + 8);
                 /* Now we expect an idstr */
                 tmp32 = buf[12]; /* Length of the following idstr */
                 buf[13 + tmp32] = '\0';
-- 
1.9.1

Re: [Qemu-devel] [PATCH] migration: Don't use *_to_cpup()

[Eric Blake]

[-- Attachment #1: Type: text/plain, Size: 735 bytes --]

On 06/10/2016 09:01 AM, Peter Maydell wrote:
> The *_to_cpup() functions just compose a pointer dereference with
> a *_to_cpu() byteswap. Instead use ld*_p(), which handles potential
> pointer misaligment and avoids the need to cast the pointer.

s/misaligment/misalignment/

> 
> Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
> ---
> The motivation here is that I'd like to get rid of _to_cpup()
> entirely: we don't have many places that use it.
> ---
>  migration/migration.c | 12 ++++++------
>  1 file changed, 6 insertions(+), 6 deletions(-)
> 

Reviewed-by: Eric Blake <eblake@redhat.com>

-- 
Eric Blake   eblake redhat com    +1-919-301-3266
Libvirt virtualization library http://libvirt.org


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 604 bytes --]

Re: [Qemu-devel] [PATCH] migration: Don't use *_to_cpup()

[Peter Maydell]

On 10 June 2016 at 16:05, Eric Blake <eblake@redhat.com> wrote:
> On 06/10/2016 09:01 AM, Peter Maydell wrote:
>> The *_to_cpup() functions just compose a pointer dereference with
>> a *_to_cpu() byteswap. Instead use ld*_p(), which handles potential
>> pointer misaligment and avoids the need to cast the pointer.
>
> s/misaligment/misalignment/
>
>>
>> Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
>> ---
>> The motivation here is that I'd like to get rid of _to_cpup()
>> entirely: we don't have many places that use it.
>> ---
>>  migration/migration.c | 12 ++++++------
>>  1 file changed, 6 insertions(+), 6 deletions(-)
>>
>
> Reviewed-by: Eric Blake <eblake@redhat.com>

Thanks. I just noticed that I missed the use of cpu_to_be64w()
for loads, so I'll send a v2.

-- PMM