* [PATCH 2.6.20] Corrects small word-size dependency in exposed kernel structure
@ 2007-05-11 10:01 Marc A. Pelletier
0 siblings, 0 replies; only message in thread
From: Marc A. Pelletier @ 2007-05-11 10:01 UTC (permalink / raw)
To: netdev
Corrects a small word-size dependency in an exposed kernel structure
so that mmaped packet sockets will work correctly from 32-bit userspace
on a 64-bit kernel.
Signed-off-by: Marc A. Pelletier <marc@abovesecurity.com>
---
As an aside, I've considered changing the setsockopt() for that feature
so that it returns critical values back to userspace (stride between
frames, actual offsets of substructures) rather than needing userspace
to guess them, but that can't be done without
(a) breaking current users of PACKET_RX_RING; or
(b) adding a new setsockopt.
Which is best?
Also, I have a couple of sanity fixes for the ring packet implementation
I'd like to discuss. Anyone specific maintaining that bit of code?
-- Marc A. Pelletier
--- orig/include/linux/if_packet.h 2007-05-10 14:27:49 -0400
+++ new/include/linux/if_packet.h 2007-05-10 14:27:50 -0400
@@ -50,7 +50,7 @@
struct tpacket_hdr
{
- unsigned long tp_status;
+ unsigned int tp_status;
#define TP_STATUS_KERNEL 0
#define TP_STATUS_USER 1
#define TP_STATUS_COPY 2
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2007-05-11 14:27 UTC | newest]
Thread overview: (only message) (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2007-05-11 10:01 [PATCH 2.6.20] Corrects small word-size dependency in exposed kernel structure Marc A. Pelletier
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).