5.1 and Sun X4100M2
Just installed 5.1/amd64 on a Sun X4100M2. Booting shows the following: ... nfe0 at pci0 dev 10 function 0 NVIDIA CK804 LAN rev 0xf3: apic 15 int 15, address 00:14:4f:7d:0f:0c eephy0 at nfe0 phy 1: 88E Gigabit PHY, rev. 2 ... nfe1 at pci6 dev 10 function 0 NVIDIA CK804 LAN rev 0xf3: apic 15 int 7, address 00:14:4f:7d:0f:0d nfe1: no PHY found! ... Seems like the two Intel onboard NICs are found successfully: ... em0 at pci12 dev 1 function 0 Intel PRO/1000MT (82546EB) rev 0x03: apic 17 int 0, address 00:14:4f:7d:0f:0e em1 at pci12 dev 1 function 1 Intel PRO/1000MT (82546EB) rev 0x03: apic 17 int 1, address 00:14:4f:7d:0f:0f ... This server also has an Intel I340-T4 quad-port ethernet card installed. After we installed 5.1, it was recognized: ... em0 at pci4 dev 0 function 0 Intel I340-T4 (82580) rev 0x01: apic 15 int 7, address 90:e2:ba:05:26:bc em1 at pci4 dev 0 function 1 Intel I340-T4 (82580) rev 0x01: apic 15 int 10, address 90:e2:ba:05:26:bd em2 at pci4 dev 0 function 2 Intel I340-T4 (82580) rev 0x01: apic 15 int 6, address 90:e2:ba:05:26:be em3 at pci4 dev 0 function 3 Intel I340-T4 (82580) rev 0x01: apic 15 int 5, address 90:e2:ba:05:26:bf ... em4 at pci12 dev 1 function 0 Intel PRO/1000MT (82546EB) rev 0x03: apic 17 int 0, address 00:14:4f:7d:0f:0e em5 at pci12 dev 1 function 1 Intel PRO/1000MT (82546EB) rev 0x03: apic 17 int 1, address 00:14:4f:7d:0f:0f ... However, with no changes to the hardware, after a reboot, this card is no longer recognized. Anyone else running 5.1 on an X4100M2? -- albert chin (ch...@thewrittenword.com)
Re: intermittent crash
On 2012/07/21 21:52, Mike Belopuhov wrote:
On Sat, Jul 21, 2012 at 19:25 +0100, Stuart Henderson wrote:
On 2012/07/21 18:49, Mike Belopuhov wrote:
this sppp_clear_ip_addrs_args dance looks totally unneeded if i read
the diff correctly. why don't you just pass sp as an argument to
the workq_add_task?
yes, that makes sense. I think he did it that way because it was
modelled on sppp_set_ip_addrs which needed to pass in two addresses
as well as the pointer.
i see. we can change that other call to workq_add_task to take sp
and args parameters because workq callback takes two and currently
the second one is ignored.
makes sense, but I think that one can wait until after unlock.
Index: if_spppsubr.c
===
RCS file: /cvs/src/sys/net/if_spppsubr.c,v
retrieving revision 1.96
diff -u -p -r1.96 if_spppsubr.c
--- if_spppsubr.c 28 Jan 2012 12:14:45 - 1.96
+++ if_spppsubr.c 21 Jul 2012 18:24:59 -
@@ -398,7 +398,7 @@ HIDE void sppp_qflush(struct ifqueue *if
int sppp_update_gw_walker(struct radix_node *rn, void *arg, u_int);
void sppp_update_gw(struct ifnet *ifp);
HIDE void sppp_set_ip_addrs(void *, void *);
-HIDE void sppp_clear_ip_addrs(struct sppp *sp);
+HIDE void sppp_clear_ip_addrs(struct sppp *, void *);
HIDE void sppp_set_phase(struct sppp *sp);
/* our control protocol descriptors */
@@ -3096,9 +3096,15 @@ sppp_ipcp_tls(struct sppp *sp)
HIDE void
sppp_ipcp_tlf(struct sppp *sp)
{
+ struct ifnet *ifp = sp-pp_if;
+
if (sp-ipcp.flags (IPCP_MYADDR_DYN|IPCP_HISADDR_DYN))
/* Some address was dynamic, clear it again. */
- sppp_clear_ip_addrs(sp);
+ if (workq_add_task(NULL, 0,
+ (void *)sppp_clear_ip_addrs, sp, NULL)) {
this cast is wrong. sppp_clear_ip_addrs should take two void *
arguments.
oops yes, I put it on the wrong argument.
+ printf(%s: workq_add_task failed, cannot clear
+ addresses\n, ifp-if_xname);
+ }
/* we no longer need LCP */
sp-lcp.protos = ~(1 IDX_IPCP);
@@ -4760,16 +4766,22 @@ sppp_set_ip_addrs(void *arg1, void *arg2
}
/*
- * Clear IP addresses. Must be called at splnet.
+ * Work queue task clearing addresses from process context.
+ * Clear IP addresses.
*/
HIDE void
-sppp_clear_ip_addrs(struct sppp *sp)
+sppp_clear_ip_addrs(struct sppp *sp, void *arg2)
{
it should still have a void sppp_clear_ip_addrs(void *, void *)
signature.
better like this?
Index: if_spppsubr.c
===
RCS file: /cvs/src/sys/net/if_spppsubr.c,v
retrieving revision 1.96
diff -u -p -r1.96 if_spppsubr.c
--- if_spppsubr.c 28 Jan 2012 12:14:45 - 1.96
+++ if_spppsubr.c 22 Jul 2012 10:00:47 -
@@ -398,7 +398,7 @@ HIDE void sppp_qflush(struct ifqueue *if
int sppp_update_gw_walker(struct radix_node *rn, void *arg, u_int);
void sppp_update_gw(struct ifnet *ifp);
HIDE void sppp_set_ip_addrs(void *, void *);
-HIDE void sppp_clear_ip_addrs(struct sppp *sp);
+HIDE void sppp_clear_ip_addrs(void *, void *);
HIDE void sppp_set_phase(struct sppp *sp);
/* our control protocol descriptors */
@@ -3096,9 +3096,15 @@ sppp_ipcp_tls(struct sppp *sp)
HIDE void
sppp_ipcp_tlf(struct sppp *sp)
{
+ struct ifnet *ifp = sp-pp_if;
+
if (sp-ipcp.flags (IPCP_MYADDR_DYN|IPCP_HISADDR_DYN))
/* Some address was dynamic, clear it again. */
- sppp_clear_ip_addrs(sp);
+ if (workq_add_task(NULL, 0,
+ sppp_clear_ip_addrs, (void *)sp, NULL)) {
+ printf(%s: workq_add_task failed, cannot clear
+ addresses\n, ifp-if_xname);
+ }
/* we no longer need LCP */
sp-lcp.protos = ~(1 IDX_IPCP);
@@ -4760,16 +4766,23 @@ sppp_set_ip_addrs(void *arg1, void *arg2
}
/*
- * Clear IP addresses. Must be called at splnet.
+ * Work queue task clearing addresses from process context.
+ * Clear IP addresses.
*/
HIDE void
-sppp_clear_ip_addrs(struct sppp *sp)
+sppp_clear_ip_addrs(void *arg1, void *arg2)
{
+ struct sppp *sp = (struct sppp *)arg1;
struct ifnet *ifp = sp-pp_if;
+ int debug = ifp-if_flags IFF_DEBUG;
struct ifaddr *ifa;
struct sockaddr_in *si;
struct sockaddr_in *dest;
+ int s;
+
+ s = splsoftnet();
+
u_int32_t remote;
if (sp-ipcp.flags IPCP_HISADDR_DYN)
remote = sp-ipcp.saved_hisaddr;
@@ -4792,6 +4805,7 @@ sppp_clear_ip_addrs(struct sppp *sp)
}
if (ifa si) {
+ int error;
struct sockaddr_in new_sin = *si;
in_ifscrub(ifp, ifatoia(ifa));
@@ -4800,10 +4814,17 @@ sppp_clear_ip_addrs(struct sppp *sp)
if
Aprovecha las promos de Belleza
Si no podes visualizar este mail, ingresa a: http://news1.bonuscupon.com.ar/r.html?uid=1.22.295h.10c.58kirnlfh9
Re: intermittent crash
On Sun, Jul 22, 2012 at 11:03 +0100, Stuart Henderson wrote:
On 2012/07/21 21:52, Mike Belopuhov wrote:
On Sat, Jul 21, 2012 at 19:25 +0100, Stuart Henderson wrote:
On 2012/07/21 18:49, Mike Belopuhov wrote:
this sppp_clear_ip_addrs_args dance looks totally unneeded if i read
the diff correctly. why don't you just pass sp as an argument to
the workq_add_task?
yes, that makes sense. I think he did it that way because it was
modelled on sppp_set_ip_addrs which needed to pass in two addresses
as well as the pointer.
i see. we can change that other call to workq_add_task to take sp
and args parameters because workq callback takes two and currently
the second one is ignored.
makes sense, but I think that one can wait until after unlock.
sure.
better like this?
i'm OK with the diff, albeit a tiny nit below.
Index: if_spppsubr.c
===
RCS file: /cvs/src/sys/net/if_spppsubr.c,v
retrieving revision 1.96
diff -u -p -r1.96 if_spppsubr.c
--- if_spppsubr.c 28 Jan 2012 12:14:45 - 1.96
+++ if_spppsubr.c 22 Jul 2012 10:00:47 -
@@ -398,7 +398,7 @@ HIDE void sppp_qflush(struct ifqueue *if
int sppp_update_gw_walker(struct radix_node *rn, void *arg, u_int);
void sppp_update_gw(struct ifnet *ifp);
HIDE void sppp_set_ip_addrs(void *, void *);
-HIDE void sppp_clear_ip_addrs(struct sppp *sp);
+HIDE void sppp_clear_ip_addrs(void *, void *);
HIDE void sppp_set_phase(struct sppp *sp);
/* our control protocol descriptors */
@@ -3096,9 +3096,15 @@ sppp_ipcp_tls(struct sppp *sp)
HIDE void
sppp_ipcp_tlf(struct sppp *sp)
{
+ struct ifnet *ifp = sp-pp_if;
+
if (sp-ipcp.flags (IPCP_MYADDR_DYN|IPCP_HISADDR_DYN))
/* Some address was dynamic, clear it again. */
- sppp_clear_ip_addrs(sp);
+ if (workq_add_task(NULL, 0,
+ sppp_clear_ip_addrs, (void *)sp, NULL)) {
+ printf(%s: workq_add_task failed, cannot clear
+ addresses\n, ifp-if_xname);
+ }
/* we no longer need LCP */
sp-lcp.protos = ~(1 IDX_IPCP);
@@ -4760,16 +4766,23 @@ sppp_set_ip_addrs(void *arg1, void *arg2
}
/*
- * Clear IP addresses. Must be called at splnet.
+ * Work queue task clearing addresses from process context.
+ * Clear IP addresses.
*/
HIDE void
-sppp_clear_ip_addrs(struct sppp *sp)
+sppp_clear_ip_addrs(void *arg1, void *arg2)
{
+ struct sppp *sp = (struct sppp *)arg1;
struct ifnet *ifp = sp-pp_if;
+ int debug = ifp-if_flags IFF_DEBUG;
struct ifaddr *ifa;
struct sockaddr_in *si;
struct sockaddr_in *dest;
+ int s;
you don't need to add extra empty line here.
+
+ s = splsoftnet();
+
u_int32_t remote;
if (sp-ipcp.flags IPCP_HISADDR_DYN)
remote = sp-ipcp.saved_hisaddr;
@@ -4792,6 +4805,7 @@ sppp_clear_ip_addrs(struct sppp *sp)
}
if (ifa si) {
+ int error;
struct sockaddr_in new_sin = *si;
in_ifscrub(ifp, ifatoia(ifa));
@@ -4800,10 +4814,17 @@ sppp_clear_ip_addrs(struct sppp *sp)
if (sp-ipcp.flags IPCP_HISADDR_DYN)
/* replace peer addr in place */
dest-sin_addr.s_addr = sp-ipcp.saved_hisaddr;
- if (!in_ifinit(ifp, ifatoia(ifa), new_sin, 0, 0))
+ if (!(error = in_ifinit(ifp, ifatoia(ifa), new_sin, 0, 0)))
dohooks(ifp-if_addrhooks, 0);
+ if (debug error) {
+ log(LOG_DEBUG, SPP_FMT sppp_clear_ip_addrs: in_ifinit
+ failed, error=%d\n, SPP_ARGS(ifp), error);
+ splx(s);
+ return;
+ }
sppp_update_gw(ifp);
}
+ splx(s);
}
Re: fix for ClickPad
You want this in now... and it has been tested on how many machines to see if any new regressions show up? Come on guys. Don't have me OK this. Instead, have me OK it based on saying something like This must go in. You take the risk. I don't want to.
Re: fix for ClickPad
On Sun, Jul 22, 2012 at 10:38:50AM -0600, Theo de Raadt wrote: You want this in now... and it has been tested on how many machines to see if any new regressions show up? Come on guys. Don't have me OK this. Instead, have me OK it based on saying something like This must go in. You take the risk. I don't want to. looking closer at my mail history, the patch is not as important as I believed it was. (I thought it was part of the fix for all ALPS touchpads that was discussend and committed while I was busy with work and not looking too closely at patches, but this part has been committed separatly). Anyways, It fixes issues seen on synaptics clickpads (as found on the X220 for example) for at least gilles@ and me. Other people with clickpads have often de-activated them in the BIOS because of these issues. Having that working in 5.2 would have been nice. I've tested that patch on all my machines with pms (X40, X220, toshiba, msi, samsung laptops, plus one desktop with a regular ps/2 mouse), but I it doesn't count. So it can wait after release, since your comments seem to indicate that is this your preference. Alexandr, just don't forget about it... -- Matthieu Herrb
Le rollup Eco est maintenant à 39 euros
Le rollup Eco à 39* ht avec impression: Impression HD et housse de transport incluses Idéal pour vos actions one shot Format 85x200 Montage ultra rapide Disponible en 4 jours * Hors frais de traitement de visuel (10/visuel différent) Simple et rapide: commandez online en 5 minutes sur vedi-express GROUND BOARD DISPLAY - NOUVEAU - Display double face pour votre communication extérieure - Piquets pour ancrage en sol meuble - Elastiques fournis pour fixer la bâche - Assemblage aisé: 5 minutes à 1 personne - Housse de transport incluse - Image vendue séparément VOS BACHES SUR MESURE - Impression HD - 8 finitions au choix - 4 supports d'impression - Disponible en 3 jours STOP TROTTOIRS CHEVALETS - Formats A1, B1, A0 et A2. - Assure une communication recto verso. - Cadres v-clic en aluminium anodisé. - Installation aisée - Affiches vendues séparément. VEDI express © 2002-2012 Rollup, Pop Up, bâches, stands parapluie aux meilleurs prix et dans les meilleurs délais. Powered by VEDI ne plus recevoir nos newsletters
Mention RFC 6106 in rtadvd.conf(5) SEE ALSO section
Mention RFC 6106 in the rtadvd.conf(5) man pages SEE ALSO section. Index: rtadvd.conf.5 === RCS file: /home/cvs/src/usr.sbin/rtadvd/rtadvd.conf.5,v retrieving revision 1.28 diff -u -p -r1.28 rtadvd.conf.5 --- rtadvd.conf.5 8 Jul 2012 10:46:00 - 1.28 +++ rtadvd.conf.5 23 Jul 2012 01:27:20 - @@ -345,6 +345,12 @@ Thomas Narten, Erik Nordmark and W. A. S Neighbor Discovery for IP version 6 (IPv6) .Dc , RFC 2461 +.Pp +J. Jeong, S. Park, L. Beloeil, and S. Madanapalli, +.Do +IPv6 Router Advertisement Options for DNS Configuration +.Dc , +RFC 6106 .Sh HISTORY The .Xr rtadvd 8 -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean.
