From: James H. <jam...@im...> - 2013-05-08 08:45:58
|
Different architectures can have different sizes of struct sockaddr which they provide to the addrlen argument of the bind/connect syscalls. In particular the metag ABI rounds it up to 112 bytes instead of 110. * tests/net: accept 112 bytes as well as 110 for bind/connect tests. Signed-off-by: James Hogan <jam...@im...> --- tests/net | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/tests/net b/tests/net index 3fb10ba..02742c2 100755 --- a/tests/net +++ b/tests/net @@ -38,10 +38,10 @@ grep_log() grep_log socket '\(PF_(LOCAL|UNIX|FILE), SOCK_STREAM, 0\) += 0' grep_log socket '\(PF_(LOCAL|UNIX|FILE), SOCK_STREAM, 0\) += 1' -grep_log bind '\(0, \{sa_family=AF_(LOCAL|UNIX|FILE), sun_path="local-stream"\}, 110\) += 0' +grep_log bind '\(0, \{sa_family=AF_(LOCAL|UNIX|FILE), sun_path="local-stream"\}, (110|112)\) += 0' grep_log listen '\(0, 5\) += 0' grep_log getsockname '\(0, \{sa_family=AF_(LOCAL|UNIX|FILE), sun_path="local-stream"\}, \[15\]\) += 0' grep_log accept '\(0, \{sa_family=AF_(LOCAL|UNIX|FILE), NULL\}, \[2\]\) += 1' -grep_log connect '\(1, \{sa_family=AF_(LOCAL|UNIX|FILE), sun_path="local-stream"\}, 110\) += 0' +grep_log connect '\(1, \{sa_family=AF_(LOCAL|UNIX|FILE), sun_path="local-stream"\}, (110|112)\) += 0' exit 0 -- 1.8.1.2 |
From: Dmitry V. L. <ld...@al...> - 2013-05-08 16:04:39
|
On Wed, May 08, 2013 at 09:44:07AM +0100, James Hogan wrote: > Different architectures can have different sizes of struct sockaddr > which they provide to the addrlen argument of the bind/connect syscalls. > In particular the metag ABI rounds it up to 112 bytes instead of 110. Thanks. I've pushed a different fix. -- ldv |
From: James H. <jam...@im...> - 2013-05-08 16:11:24
|
On 08/05/13 17:04, Dmitry V. Levin wrote: > On Wed, May 08, 2013 at 09:44:07AM +0100, James Hogan wrote: >> Different architectures can have different sizes of struct sockaddr >> which they provide to the addrlen argument of the bind/connect syscalls. >> In particular the metag ABI rounds it up to 112 bytes instead of 110. > > Thanks. I've pushed a different fix. Okay, works for me, Thanks James |