On Mon, Apr 17, 2023 at 05:53:28PM +0200, Alexander Bluhm wrote:
> On Mon, Apr 17, 2023 at 04:32:13PM +0200, Alexander Bluhm wrote:
> > On Mon, Apr 17, 2023 at 02:36:57AM +0300, Vitaliy Makkoveev wrote:
> > > It seems rt_setsource() needs some attention, but sysctl_source() could
> > > be called with shared netlock just now.
> >
> > I think rtable_setsource() is not MP safe.  It is documented as [K]
> > kernel lock.  But that is not true and makes no sense.  It should
> > be exclusive netlock.  in_pcbselsrc() calls rtable_getsource() with
> > netlock.  We should rename source to ar_source so we can grep for
> > its users.
> 
> Perhaps something like this.  I will run it through regress.

Rebased after your rename commit.  Regress test did not trigger
lock assertions.  ar_source is a pointer to an ifa_addr, so net
lock should be safe.

ok?

bluhm

Index: net/art.h
===================================================================
RCS file: /cvs/src/sys/net/art.h,v
retrieving revision 1.22
diff -u -p -r1.22 art.h
--- net/art.h   18 Apr 2023 10:19:16 -0000      1.22
+++ net/art.h   18 Apr 2023 13:32:46 -0000
@@ -41,7 +41,7 @@ struct art_root {
        uint8_t                  ar_nlvl;       /* [I] Number of levels */
        uint8_t                  ar_alen;       /* [I] Address length in bits */
        uint8_t                  ar_off;        /* [I] Offset of key in bytes */
-       struct sockaddr         *ar_source;     /* [K] optional src addr to use 
*/
+       struct sockaddr         *ar_source;     /* [N] use optional src addr */
 };
 
 #define ISLEAF(e)      (((unsigned long)(e) & 1) == 0)
Index: net/rtable.c
===================================================================
RCS file: /cvs/src/sys/net/rtable.c,v
retrieving revision 1.81
diff -u -p -r1.81 rtable.c
--- net/rtable.c        18 Apr 2023 10:19:16 -0000      1.81
+++ net/rtable.c        18 Apr 2023 13:32:46 -0000
@@ -376,6 +376,8 @@ rtable_setsource(unsigned int rtableid, 
 {
        struct art_root         *ar;
 
+       NET_ASSERT_LOCKED_EXCLUSIVE();
+
        if ((ar = rtable_get(rtableid, af)) == NULL)
                return (EAFNOSUPPORT);
 
@@ -388,6 +390,8 @@ struct sockaddr *
 rtable_getsource(unsigned int rtableid, int af)
 {
        struct art_root         *ar;
+
+       NET_ASSERT_LOCKED();
 
        ar = rtable_get(rtableid, af);
        if (ar == NULL)
Index: net/rtsock.c
===================================================================
RCS file: /cvs/src/sys/net/rtsock.c,v
retrieving revision 1.362
diff -u -p -r1.362 rtsock.c
--- net/rtsock.c        18 Apr 2023 09:56:54 -0000      1.362
+++ net/rtsock.c        18 Apr 2023 13:32:46 -0000
@@ -853,8 +853,10 @@ route_output(struct mbuf *m, struct sock
                        error = EINVAL;
                        goto fail;
                }
-               if ((error =
-                   rt_setsource(tableid, info.rti_info[RTAX_IFA])) != 0)
+               NET_LOCK();
+               error = rt_setsource(tableid, info.rti_info[RTAX_IFA]);
+               NET_UNLOCK();
+               if (error)
                        goto fail;
        } else {
                error = rtm_output(rtm, &rt, &info, prio, tableid);
@@ -862,7 +864,9 @@ route_output(struct mbuf *m, struct sock
                        type = rtm->rtm_type;
                        seq = rtm->rtm_seq;
                        free(rtm, M_RTABLE, len);
+                       NET_LOCK_SHARED();
                        rtm = rtm_report(rt, type, seq, tableid);
+                       NET_UNLOCK_SHARED();
                        len = rtm->rtm_msglen;
                }
        }

Reply via email to