[hail patch 2/3] fix 32/64 wire interoperability

[hail patch 2/3] fix 32/64 wire interoperability

[Pete Zaitcev]

Testing found that tabled and chunkd running on CPUs with different
word length cannot talk to each other.

The bug was introduced by commit ea5d20bc22aeed077312c9c1824e84651af17a16.

The fix is to add named padding that takes the place of the invisible
padding, thus making the layout platform-neutral.

Signed-off-by: Pete Zaitcev <zaitcev@redhat.com>

---
 include/chunk_msg.h |    1 +
 1 file changed, 1 insertion(+)

diff --git a/include/chunk_msg.h b/include/chunk_msg.h
index a34fc21..4c170e4 100644
--- a/include/chunk_msg.h
+++ b/include/chunk_msg.h
@@ -91,6 +91,7 @@ struct chunksrv_resp {
 	uint32_t		nonce;	/* txn id, copied from request */
 	uint64_t		data_len;		/* len of addn'l data */
 	unsigned char		hash[CHD_CSUM_SZ];	/* SHA1 checksum */
+	unsigned char		rsv2[4];		/* pad for 64 bits */
 };
 
 struct chunksrv_resp_get {

Re: [hail patch 2/3] fix 32/64 wire interoperability

[Jeff Garzik]

On 08/04/2010 07:16 PM, Pete Zaitcev wrote:
> Testing found that tabled and chunkd running on CPUs with different
> word length cannot talk to each other.
>
> The bug was introduced by commit ea5d20bc22aeed077312c9c1824e84651af17a16.
>
> The fix is to add named padding that takes the place of the invisible
> padding, thus making the layout platform-neutral.
>
> Signed-off-by: Pete Zaitcev<zaitcev@redhat.com>
>
> ---
>   include/chunk_msg.h |    1 +
>   1 file changed, 1 insertion(+)
>
> diff --git a/include/chunk_msg.h b/include/chunk_msg.h
> index a34fc21..4c170e4 100644
> --- a/include/chunk_msg.h
> +++ b/include/chunk_msg.h
> @@ -91,6 +91,7 @@ struct chunksrv_resp {
>   	uint32_t		nonce;	/* txn id, copied from request */
>   	uint64_t		data_len;		/* len of addn'l data */
>   	unsigned char		hash[CHD_CSUM_SZ];	/* SHA1 checksum */
> +	unsigned char		rsv2[4];		/* pad for 64 bits */
>   };

good catch.  applied 1-3, and pushed out.

I wonder if we shouldn't switch to attribute(packed) for safety, though.

Re: [hail patch 2/3] fix 32/64 wire interoperability

[Pete Zaitcev]

On Thu, 05 Aug 2010 17:25:30 -0400
Jeff Garzik <jeff@garzik.org> wrote:

> I wonder if we shouldn't switch to attribute(packed) for safety, though.

I dunno, as long as we test this...  I irrationally hate the
gcc attributes, although probably any viable competitor will have
to support them. Talk about embrace-and-extend. Fine when FSF does
it, not fine when Microsoft does, huh.

-- Pete