Source-Changes-HG archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

[src/netbsd-7-0]: src/sys/kern Pull up following revision(s) (re...



details:   https://anonhg.NetBSD.org/src/rev/88320156577d
branches:  netbsd-7-0
changeset: 319069:88320156577d
user:      martin <martin%NetBSD.org@localhost>
date:      Tue May 15 04:48:16 2018 +0000
description:
Pull up following revision(s) (requested by maxv in ticket #1602):

        sys/kern/uipc_mbuf.c: revision 1.211 (via patch)

Modify m_defrag, so that it never frees the first mbuf of the chain. While
here use the given 'flags' argument, and not M_DONTWAIT.

We have a problem with several drivers: they poll an mbuf chain from their
queues and call m_defrag on them, but m_defrag could update the mbuf
pointer, so the mbuf in the queue is no longer valid. It is not easy to
fix each driver, because doing pop+push will reorder the queue, and we
don't really want that to happen.

This problem was independently spotted by me, Kengo, Masanobu, and other
people too it seems (perhaps PR/53218).

Now m_defrag leaves the first mbuf in place, and compresses the chain
only starting from the second mbuf in the chain.

It is important not to compress the first mbuf with hacks, because the
storage of this first mbuf may be shared with other mbufs.

diffstat:

 sys/kern/uipc_mbuf.c |  32 +++++++++++++++++++-------------
 1 files changed, 19 insertions(+), 13 deletions(-)

diffs (84 lines):

diff -r 041681d4a684 -r 88320156577d sys/kern/uipc_mbuf.c
--- a/sys/kern/uipc_mbuf.c      Mon May 14 19:03:48 2018 +0000
+++ b/sys/kern/uipc_mbuf.c      Tue May 15 04:48:16 2018 +0000
@@ -1,4 +1,4 @@
-/*     $NetBSD: uipc_mbuf.c,v 1.158.4.1.2.2 2018/04/17 08:30:08 martin Exp $   */
+/*     $NetBSD: uipc_mbuf.c,v 1.158.4.1.2.3 2018/05/15 04:48:16 martin Exp $   */
 
 /*-
  * Copyright (c) 1999, 2001 The NetBSD Foundation, Inc.
@@ -62,7 +62,7 @@
  */
 
 #include <sys/cdefs.h>
-__KERNEL_RCSID(0, "$NetBSD: uipc_mbuf.c,v 1.158.4.1.2.2 2018/04/17 08:30:08 martin Exp $");
+__KERNEL_RCSID(0, "$NetBSD: uipc_mbuf.c,v 1.158.4.1.2.3 2018/05/15 04:48:16 martin Exp $");
 
 #include "opt_mbuftrace.h"
 #include "opt_nmbclusters.h"
@@ -1376,30 +1376,35 @@
 }
 
 /*
- * Copy the mbuf chain to a new mbuf chain that is as short as possible.
- * Return the new mbuf chain on success, NULL on failure.  On success,
- * free the old mbuf chain.
+ * Compress the mbuf chain. Return the new mbuf chain on success, NULL on
+ * failure. The first mbuf is preserved, and on success the pointer returned
+ * is the same as the one passed.
  */
 struct mbuf *
 m_defrag(struct mbuf *mold, int flags)
 {
        struct mbuf *m0, *mn, *n;
-       size_t sz = mold->m_pkthdr.len;
+       int sz;
 
 #ifdef DIAGNOSTIC
        if ((mold->m_flags & M_PKTHDR) == 0)
                panic("m_defrag: not a mbuf chain header");
 #endif
 
-       m0 = m_gethdr(flags, MT_DATA);
+       if (mold->m_next == NULL)
+               return mold;
+
+       m0 = m_get(flags, MT_DATA);
        if (m0 == NULL)
                return NULL;
-       M_COPY_PKTHDR(m0, mold);
        mn = m0;
 
+       sz = mold->m_pkthdr.len - mold->m_len;
+       KASSERT(sz >= 0);
+
        do {
-               if (sz > MHLEN) {
-                       MCLGET(mn, M_DONTWAIT);
+               if (sz > MLEN) {
+                       MCLGET(mn, flags);
                        if ((mn->m_flags & M_EXT) == 0) {
                                m_freem(m0);
                                return NULL;
@@ -1415,7 +1420,7 @@
 
                if (sz > 0) {
                        /* need more mbufs */
-                       n = m_get(M_NOWAIT, MT_DATA);
+                       n = m_get(flags, MT_DATA);
                        if (n == NULL) {
                                m_freem(m0);
                                return NULL;
@@ -1426,9 +1431,10 @@
                }
        } while (sz > 0);
 
-       m_freem(mold);
+       m_freem(mold->m_next);
+       mold->m_next = m0;
 
-       return m0;
+       return mold;
 }
 
 int



Home | Main Index | Thread Index | Old Index