On Monday 22 January 2007 10:01, Andre Oppermann wrote:
> John Baldwin wrote:
> > Also, you've introduced another regression in that if the m_get2m() fails
it
> > should be returning ENOBUFS and not EFAULT to userland. The comments in
> > sosend_*() about 'EFAULT being the only possible error' are obviously
> > wrong. :)
>
> We're always calling with M_WAITOK because we're coming from userland and
> may sleep forever.
Bah, ok. I was confused because I guess TRYWAIT can't return NULL anymore,
but sosend_copyin() still has:
MGETHDR(m, M_TRYWAIT, MT_DATA);
if (m == NULL) {
error = ENOBUFS;
goto out;
}
although it would seem that the ZERO_COPY_SOCKETS code just needs to be
updated. Maybe like so:
Index: uipc_socket.c
===================================================================
RCS file: /usr/cvs/src/sys/kern/uipc_socket.c,v
retrieving revision 1.288
diff -u -r1.288 uipc_socket.c
--- uipc_socket.c 8 Jan 2007 17:49:59 -0000 1.288
+++ uipc_socket.c 22 Jan 2007 15:52:06 -0000
@@ -857,20 +857,11 @@
if (resid >= MINCLSIZE) {
#ifdef ZERO_COPY_SOCKETS
if (top == NULL) {
- MGETHDR(m, M_TRYWAIT, MT_DATA);
- if (m == NULL) {
- error = ENOBUFS;
- goto out;
- }
+ m = m_gethdr(M_TRYWAIT, MT_DATA);
m->m_pkthdr.len = 0;
m->m_pkthdr.rcvif = NULL;
- } else {
- MGET(m, M_TRYWAIT, MT_DATA);
- if (m == NULL) {
- error = ENOBUFS;
- goto out;
- }
- }
+ } else
+ m = m_get(M_TRYWAIT, MT_DATA);
if (so_zero_copy_send &&
resid>=PAGE_SIZE &&
*space>=PAGE_SIZE &&
@@ -881,14 +872,8 @@
len = cow_send;
}
if (!cow_send) {
- MCLGET(m, M_TRYWAIT);
- if ((m->m_flags & M_EXT) == 0) {
- m_free(m);
- m = NULL;
- } else {
- len = min(min(MCLBYTES, resid),
- *space);
- }
+ m_clget(m, M_TRYWAIT);
+ len = min(min(MCLBYTES, resid), *space);
}
#else /* ZERO_COPY_SOCKETS */
if (top == NULL) {
--
John Baldwin
_______________________________________________
[email protected] mailing list
http://lists.freebsd.org/mailman/listinfo/cvs-all
To unsubscribe, send any mail to "[EMAIL PROTECTED]"
