date:20150902

Re: svn commit: r287366 - head/sys/kern

2015-09-02 Thread Baptiste Daroussin

On Tue, Sep 01, 2015 at 02:05:30PM +, Konstantin Belousov wrote:
> Author: kib
> Date: Tue Sep  1 14:05:29 2015
> New Revision: 287366
> URL: https://svnweb.freebsd.org/changeset/base/287366
> 
> Log:
>   Exit notification for EVFILT_PROC removes knote from the knlist.  In
>   particular, this invalidates the knote kn_link linkage, making the
>   SLIST_FOREACH() loop accessing undefined values (e.g. trashed by
>   QUEUE_MACRO_DEBUG).  If the knote is freed by other thread when kq
>   lock is released or when influx is cleared, e.g. by knote_scan() for
>   kqueue owning the knote, the iteration step would access freed memory.
>   
>   Use SLIST_FOREACH_SAFE() to fix iteration.
>   
>   Diagnosed by:   avg
>   Tested by:  avg, lstewart, pawel
>   Sponsored by:   The FreeBSD Foundation
>   MFC after:  2 weeks
> 
Thank you! that makes my package builders stable again!

Best regards,
Bapt


pgpYdGuiPvQ3J.pgp
Description: PGP signature

svn commit: r287404 - head/lib/libc/net

2015-09-02 Thread Hiroki Sato

Author: hrs
Date: Wed Sep  2 16:50:49 2015
New Revision: 287404
URL: https://svnweb.freebsd.org/changeset/base/287404

Log:
  - snprintf() returns at most size-1 of the chars printed into
the buffer.  (n == hostlen) also means the buffer length was
too short.
  
  - Use sdl->sdl_data only when (sdl->sdl_nlen > 0 && sdl->sdl_alen == 0)
to prevent redundant output.

Modified:
  head/lib/libc/net/getnameinfo.c

Modified: head/lib/libc/net/getnameinfo.c
==
--- head/lib/libc/net/getnameinfo.c Wed Sep  2 16:48:03 2015
(r287403)
+++ head/lib/libc/net/getnameinfo.c Wed Sep  2 16:50:49 2015
(r287404)
@@ -394,26 +394,22 @@ getnameinfo_link(const struct sockaddr *
 
if (sdl->sdl_nlen == 0 && sdl->sdl_alen == 0 && sdl->sdl_slen == 0) {
n = snprintf(host, hostlen, "link#%d", sdl->sdl_index);
-   if (n > hostlen) {
+   if (n >= hostlen) {
*host = '\0';
return (EAI_MEMORY);
}
return (0);
}
 
-   if (sdl->sdl_nlen > 0) {
-   if (sdl->sdl_nlen + 1 > hostlen) {
+   if (sdl->sdl_nlen > 0 && sdl->sdl_alen == 0) {
+   n = sdl->sdl_nlen;
+   if (n >= hostlen) {
*host = '\0';
return (EAI_MEMORY);
}
memcpy(host, sdl->sdl_data, sdl->sdl_nlen);
-   n = sdl->sdl_nlen;
-   host += n;
-   if (sdl->sdl_alen > 0) {
-   *host++ = ':';
-   n++;
-   }
-   hostlen -= n;
+   host[n] = '\0';
+   return (0);
}
 
switch (sdl->sdl_type) {
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"

svn commit: r287400 - head

2015-09-02 Thread Gleb Smirnoff

Author: glebius
Date: Wed Sep  2 15:42:14 2015
New Revision: 287400
URL: https://svnweb.freebsd.org/changeset/base/287400

Log:
  The ${BUILDKERNELS:[2..-1]} appears to produce a non zero result for
  a one word variable, which is quite unexpected from documentation.
  So, to avoid double installation of a single kernel, protect the extra
  kernels loop with ${BUILDKERNELS:[#]} > 1 conditional.
  
  Sponsored by: Netflix
  Sponsored by: Nginx, Inc.

Modified:
  head/Makefile.inc1

Modified: head/Makefile.inc1
==
--- head/Makefile.inc1  Wed Sep  2 15:23:51 2015(r287399)
+++ head/Makefile.inc1  Wed Sep  2 15:42:14 2015(r287400)
@@ -1127,6 +1127,7 @@ reinstallkernel reinstallkernel.debug: _
cd ${KRNLOBJDIR}/${INSTALLKERNEL}; \
${CROSSENV} PATH=${TMPPATH} \
${MAKE} ${IMAKE_INSTALL} KERNEL=${INSTKERNNAME} 
${.TARGET:S/kernel//}
+.if ${BUILDKERNELS:[#]} > 1
 .for _kernel in ${BUILDKERNELS:[2..-1]}
@echo "--"
@echo ">>> Installing kernel ${_kernel}"
@@ -1135,6 +1136,7 @@ reinstallkernel reinstallkernel.debug: _
${CROSSENV} PATH=${TMPPATH} \
${MAKE} ${IMAKE_INSTALL} KERNEL=${INSTKERNNAME}.${_kernel} 
${.TARGET:S/kernel//}
 .endfor
+.endif
 
 distributekernel distributekernel.debug:
 .if empty(INSTALLKERNEL)
@@ -1154,6 +1156,7 @@ distributekernel distributekernel.debug:
sed -e 's|^./kernel|.|' ${DESTDIR}/${DISTDIR}/kernel.premeta > \
${DESTDIR}/${DISTDIR}/kernel.meta
 .endif
+.if ${BUILDKERNELS:[#]} > 1
 .for _kernel in ${BUILDKERNELS:[2..-1]}
 .if defined(NO_ROOT)
echo "#${MTREE_MAGIC}" > ${DESTDIR}/${DISTDIR}/kernel.${_kernel}.premeta
@@ -1170,27 +1173,32 @@ distributekernel distributekernel.debug:
${DESTDIR}/${DISTDIR}/kernel.${_kernel}.meta
 .endif
 .endfor
+.endif
 
 packagekernel:
 .if defined(NO_ROOT)
cd ${DESTDIR}/${DISTDIR}/kernel; \
tar cvf - @${DESTDIR}/${DISTDIR}/kernel.meta | \
${XZ_CMD} > ${DESTDIR}/${DISTDIR}/kernel.txz
+.if ${BUILDKERNELS:[#]} > 1
 .for _kernel in ${BUILDKERNELS:[2..-1]}
cd ${DESTDIR}/${DISTDIR}/kernel.${_kernel}; \
tar cvf - @${DESTDIR}/${DISTDIR}/kernel.${_kernel}.meta | \
${XZ_CMD} > ${DESTDIR}/${DISTDIR}/kernel.${_kernel}.txz
 .endfor
+.endif
 .else
cd ${DESTDIR}/${DISTDIR}/kernel; \
tar cvf - . | \
${XZ_CMD} > ${DESTDIR}/${DISTDIR}/kernel.txz
+.if ${BUILDKERNELS:[#]} > 1
 .for _kernel in ${BUILDKERNELS:[2..-1]}
cd ${DESTDIR}/${DISTDIR}/kernel.${_kernel}; \
tar cvf - . | \
${XZ_CMD} > ${DESTDIR}/${DISTDIR}/kernel.${_kernel}.txz
 .endfor
 .endif
+.endif
 
 #
 # doxygen
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"

svn commit: r287402 - head/sys/net

2015-09-02 Thread Hiroki Sato

Author: hrs
Date: Wed Sep  2 16:30:45 2015
New Revision: 287402
URL: https://svnweb.freebsd.org/changeset/base/287402

Log:
  Fix a panic which was reproducible by an infinite loop of
  "ifconfig epair0 create && ifconfig epair0a destroy".
  
  This was caused by an uninitialized function pointer in
  softc->media.

Modified:
  head/sys/net/if_epair.c

Modified: head/sys/net/if_epair.c
==
--- head/sys/net/if_epair.c Wed Sep  2 16:06:25 2015(r287401)
+++ head/sys/net/if_epair.c Wed Sep  2 16:30:45 2015(r287402)
@@ -809,6 +809,14 @@ epair_clone_create(struct if_clone *ifc,
netisr_get_cpuid(sca->ifp->if_index % netisr_get_cpucount());
scb->cpuid =
netisr_get_cpuid(scb->ifp->if_index % netisr_get_cpucount());
+
+   /* Initialise pseudo media types. */
+   ifmedia_init(>media, 0, epair_media_change, epair_media_status);
+   ifmedia_add(>media, IFM_ETHER | IFM_10G_T, 0, NULL);
+   ifmedia_set(>media, IFM_ETHER | IFM_10G_T);
+   ifmedia_init(>media, 0, epair_media_change, epair_media_status);
+   ifmedia_add(>media, IFM_ETHER | IFM_10G_T, 0, NULL);
+   ifmedia_set(>media, IFM_ETHER | IFM_10G_T);

/* Finish initialization of interface a. */
ifp = sca->ifp;
@@ -867,14 +875,6 @@ epair_clone_create(struct if_clone *ifc,
strlcpy(name, sca->ifp->if_xname, len);
DPRINTF("name='%s/%db' created sca=%p scb=%p\n", name, unit, sca, scb);
 
-   /* Initialise pseudo media types. */
-   ifmedia_init(>media, 0, epair_media_change, epair_media_status);
-   ifmedia_add(>media, IFM_ETHER | IFM_10G_T, 0, NULL);
-   ifmedia_set(>media, IFM_ETHER | IFM_10G_T);
-   ifmedia_init(>media, 0, epair_media_change, epair_media_status);
-   ifmedia_add(>media, IFM_ETHER | IFM_10G_T, 0, NULL);
-   ifmedia_set(>media, IFM_ETHER | IFM_10G_T);
-
/* Tell the world, that we are ready to rock. */
sca->ifp->if_drv_flags |= IFF_DRV_RUNNING;
scb->ifp->if_drv_flags |= IFF_DRV_RUNNING;
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"

svn commit: r287403 - head/sys/dev/ioat

2015-09-02 Thread Conrad E. Meyer

Author: cem
Date: Wed Sep  2 16:48:03 2015
New Revision: 287403
URL: https://svnweb.freebsd.org/changeset/base/287403

Log:
  ioat: re-initialize interrupts after resetting hw on BDXDE
  
  Resetting some generations of the I/OAT hardware (just BDXDE for now)
  resets the corresponding MSI-X registers.  So, teardown and
  re-initialize interrupts after resetting the hardware.
  
  Reviewed by:  jimharris
  Approved by:  markj (mentor)
  Sponsored by: EMC / Isilon Storage Division
  Differential Revision:https://reviews.freebsd.org/D3549

Modified:
  head/sys/dev/ioat/ioat.c

Modified: head/sys/dev/ioat/ioat.c
==
--- head/sys/dev/ioat/ioat.cWed Sep  2 16:30:45 2015(r287402)
+++ head/sys/dev/ioat/ioat.cWed Sep  2 16:48:03 2015(r287403)
@@ -53,12 +53,14 @@ __FBSDID("$FreeBSD$");
 static int ioat_probe(device_t device);
 static int ioat_attach(device_t device);
 static int ioat_detach(device_t device);
+static int ioat_setup_intr(struct ioat_softc *ioat);
+static int ioat_teardown_intr(struct ioat_softc *ioat);
 static int ioat3_attach(device_t device);
 static int ioat_map_pci_bar(struct ioat_softc *ioat);
 static void ioat_dmamap_cb(void *arg, bus_dma_segment_t *segs, int nseg,
 int error);
-static int ioat_interrupt_setup(struct ioat_softc *ioat);
 static void ioat_interrupt_handler(void *arg);
+static boolean_t ioat_is_bdxde(struct ioat_softc *ioat);
 static void ioat_process_events(struct ioat_softc *ioat);
 static inline uint32_t ioat_get_active(struct ioat_softc *ioat);
 static inline uint32_t ioat_get_ring_space(struct ioat_softc *ioat);
@@ -220,13 +222,15 @@ ioat_attach(device_t device)
goto err;
 
ioat->version = ioat_read_cbver(ioat);
-   ioat_interrupt_setup(ioat);
-
if (ioat->version < IOAT_VER_3_0) {
error = ENODEV;
goto err;
}
 
+   error = ioat_setup_intr(ioat);
+   if (error != 0)
+   return (error);
+
error = ioat3_attach(device);
if (error != 0)
goto err;
@@ -273,15 +277,23 @@ ioat_detach(device_t device)
 
bus_dma_tag_destroy(ioat->hw_desc_tag);
 
+   ioat_teardown_intr(ioat);
+
+   return (0);
+}
+
+static int
+ioat_teardown_intr(struct ioat_softc *ioat)
+{
+
if (ioat->tag != NULL)
-   bus_teardown_intr(device, ioat->res, ioat->tag);
+   bus_teardown_intr(ioat->device, ioat->res, ioat->tag);
 
if (ioat->res != NULL)
-   bus_release_resource(device, SYS_RES_IRQ,
+   bus_release_resource(ioat->device, SYS_RES_IRQ,
rman_get_rid(ioat->res), ioat->res);
 
-   pci_release_msi(device);
-
+   pci_release_msi(ioat->device);
return (0);
 }
 
@@ -455,7 +467,7 @@ ioat_dmamap_cb(void *arg, bus_dma_segmen
  * Interrupt setup and handlers
  */
 static int
-ioat_interrupt_setup(struct ioat_softc *ioat)
+ioat_setup_intr(struct ioat_softc *ioat)
 {
uint32_t num_vectors;
int error;
@@ -498,6 +510,23 @@ ioat_interrupt_setup(struct ioat_softc *
return (0);
 }
 
+static boolean_t
+ioat_is_bdxde(struct ioat_softc *ioat)
+{
+   u_int32_t pciid;
+
+   pciid = pci_get_devid(ioat->device);
+   switch (pciid) {
+   case 0x6f508086:
+   case 0x6f518086:
+   case 0x6f528086:
+   case 0x6f538086:
+   return (TRUE);
+   }
+
+   return (FALSE);
+}
+
 static void
 ioat_interrupt_handler(void *arg)
 {
@@ -918,7 +947,7 @@ ioat_reset_hw(struct ioat_softc *ioat)
 {
uint64_t status;
uint32_t chanerr;
-   int timeout;
+   int timeout, error;
 
status = ioat_get_chansts(ioat);
if (is_ioat_active(status) || is_ioat_idle(status))
@@ -953,6 +982,20 @@ ioat_reset_hw(struct ioat_softc *ioat)
if (timeout == 20)
return (ETIMEDOUT);
 
+   /*
+* BDXDE models reset MSI-X registers on device reset.  We must
+* teardown and re-setup interrupts.
+*/
+   if (ioat_is_bdxde(ioat)) {
+   error = ioat_teardown_intr(ioat);
+   if (error)
+   return (error);
+
+   error = ioat_setup_intr(ioat);
+   if (error)
+   return (error);
+   }
+
return (0);
 }
 
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"

svn commit: r287399 - head/sys/dev/iwn

2015-09-02 Thread Gleb Smirnoff

Author: glebius
Date: Wed Sep  2 15:23:51 2015
New Revision: 287399
URL: https://svnweb.freebsd.org/changeset/base/287399

Log:
  Remove the software queue, which is a remnant of ifnet ifqueue.
  
  Reviewed by:  adrian
  Sponsored by: Netflix
  Sponsored by: Nginx, Inc.

Modified:
  head/sys/dev/iwn/if_iwn.c
  head/sys/dev/iwn/if_iwnvar.h

Modified: head/sys/dev/iwn/if_iwn.c
==
--- head/sys/dev/iwn/if_iwn.c   Wed Sep  2 14:38:16 2015(r287398)
+++ head/sys/dev/iwn/if_iwn.c   Wed Sep  2 15:23:51 2015(r287399)
@@ -231,7 +231,6 @@ static void iwn_xmit_task(void *arg0, in
 static int iwn_raw_xmit(struct ieee80211_node *, struct mbuf *,
const struct ieee80211_bpf_params *);
 static int iwn_transmit(struct ieee80211com *, struct mbuf *);
-static voidiwn_start_locked(struct iwn_softc *);
 static voidiwn_watchdog(void *);
 static int iwn_ioctl(struct ieee80211com *, u_long , void *);
 static voidiwn_parent(struct ieee80211com *);
@@ -471,7 +470,6 @@ iwn_attach(device_t dev)
}
 
IWN_LOCK_INIT(sc);
-   mbufq_init(>sc_snd, ifqmaxlen);
 
/* Read hardware revision and attach. */
sc->hw_type = (IWN_READ(sc, IWN_HW_REV) >> IWN_HW_REV_TYPE_SHIFT)
@@ -1409,8 +1407,6 @@ iwn_detach(device_t dev)
ieee80211_ifdetach(>sc_ic);
}
 
-   mbufq_drain(>sc_snd);
-
/* Uninstall interrupt handler. */
if (sc->irq != NULL) {
bus_teardown_intr(dev, sc->irq, sc->sc_ih);
@@ -3597,14 +3593,10 @@ iwn_tx_done(struct iwn_softc *sc, struct
(status & IWN_TX_FAIL) != 0);
 
sc->sc_tx_timer = 0;
-   if (--ring->queued < IWN_TX_RING_LOMARK) {
+   if (--ring->queued < IWN_TX_RING_LOMARK)
sc->qfullmsk &= ~(1 << ring->qid);
-   if (sc->qfullmsk == 0)
-   iwn_start_locked(sc);
-   }
 
DPRINTF(sc, IWN_DEBUG_TRACE, "->%s: end\n",__func__);
-
 }
 
 /*
@@ -3781,14 +3773,10 @@ iwn_ampdu_tx_done(struct iwn_softc *sc, 
}
 
sc->sc_tx_timer = 0;
-   if (ring->queued < IWN_TX_RING_LOMARK) {
+   if (ring->queued < IWN_TX_RING_LOMARK)
sc->qfullmsk &= ~(1 << ring->qid);
-   if (sc->qfullmsk == 0)
-   iwn_start_locked(sc);
-   }
 
DPRINTF(sc, IWN_DEBUG_TRACE, "->%s: end\n",__func__);
-
 }
 
 /*
@@ -4948,54 +4936,30 @@ iwn_raw_xmit(struct ieee80211_node *ni, 
 static int
 iwn_transmit(struct ieee80211com *ic, struct mbuf *m)
 {
-   struct iwn_softc *sc;
+   struct iwn_softc *sc = ic->ic_softc;
+   struct ieee80211_node *ni;
int error;
 
-   sc = ic->ic_softc;
-
IWN_LOCK(sc);
-   if ((sc->sc_flags & IWN_FLAG_RUNNING) == 0) {
+   if ((sc->sc_flags & IWN_FLAG_RUNNING) == 0 || sc->sc_beacon_wait) {
IWN_UNLOCK(sc);
return (ENXIO);
}
-   error = mbufq_enqueue(>sc_snd, m);
-   if (error) {
-   IWN_UNLOCK(sc);
-   return (error);
-   }
-   iwn_start_locked(sc);
-   IWN_UNLOCK(sc);
-   return (0);
-}
-
-static void
-iwn_start_locked(struct iwn_softc *sc)
-{
-   struct ieee80211_node *ni;
-   struct mbuf *m;
 
-   IWN_LOCK_ASSERT(sc);
-
-   /*
-* If we're waiting for a beacon, we can just exit out here
-* and wait for the taskqueue to be kicked.
-*/
-   if (sc->sc_beacon_wait) {
-   return;
+   if (sc->qfullmsk) {
+   IWN_UNLOCK(sc);
+   return (ENOBUFS);
}
 
-   DPRINTF(sc, IWN_DEBUG_XMIT, "%s: called\n", __func__);
-   while (sc->qfullmsk == 0 && 
-   (m = mbufq_dequeue(>sc_snd)) != NULL) {
-   ni = (struct ieee80211_node *)m->m_pkthdr.rcvif;
-   if (iwn_tx_data(sc, m, ni) != 0) {
-   if_inc_counter(ni->ni_vap->iv_ifp,
-   IFCOUNTER_OERRORS, 1);
-   ieee80211_free_node(ni);
-   } else
-   sc->sc_tx_timer = 5;
-   }
-   DPRINTF(sc, IWN_DEBUG_XMIT, "%s: done\n", __func__);
+   ni = (struct ieee80211_node *)m->m_pkthdr.rcvif;
+   error = iwn_tx_data(sc, m, ni);
+   if (error) {
+   if_inc_counter(ni->ni_vap->iv_ifp, IFCOUNTER_OERRORS, 1);
+   ieee80211_free_node(ni);
+   } else
+   sc->sc_tx_timer = 5;
+   IWN_UNLOCK(sc);
+   return (error);
 }
 
 static void
@@ -8731,9 +8695,6 @@ iwn_panicked(void *arg0, int pending)
"%s: could not move to run state\n", __func__);
}
 
-   /* Only run start once the NIC is in a useful state, like associated */
-   iwn_start_locked(sc);
-
IWN_UNLOCK(sc);
 }
 

Modified: head/sys/dev/iwn/if_iwnvar.h
==
---

Re: svn commit: r287394 - head/etc

2015-09-02 Thread Warner Losh

On Wed, Sep 2, 2015 at 6:46 AM, Gleb Smirnoff  wrote:

> +   set wifi-driver-regex
> +
>  "(ath|bwi|bwn|ipw|iwi|iwn|malo|mwl|ral|rsu|rum|run|uath|upgt|\
> +   ural|urtw|urtwn|wi|wpi|wtap|zyd)[0-9]+";
>

This moves us in the wrong direction. We've been looking at ways of
getting rid of the NIC regex since it went in. Now there's another one
to fight with :(

Warner
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"

svn commit: r287405 - head/sys/geom

2015-09-02 Thread Warner Losh

Author: imp
Date: Wed Sep  2 17:29:30 2015
New Revision: 287405
URL: https://svnweb.freebsd.org/changeset/base/287405

Log:
  After the introduction of direct dispatch, the pacing code in g_down()
  broke in two ways. One, the pacing variable was accessed in multiple
  threads in an unsafe way. Two, since large numbers of I/O could come
  down from the buf layer at one time, large numbers of allocation
  failures could happen all at once, resulting in a huge pace value that
  would limit I/Os to 10 IOPS for minutes (or even hours) at a
  time. While a real solution to these problems requires substantial
  work (to go to a no-allocation after the first model, or to have some
  way to wait for more memory with some kind of reserve for pager and
  swapper requests), it is relatively easy to make this simplistic
  pacing less pathological.
  
  Move to using a volatile variable with loads and stores. While this is
  a little racy, losing the race is safe: either you get memory and
  proceed, or you don't and queue. Second, sleep for 1ms (or one tick, whichever
  is larger) instead of 100ms. This removes the artificial 10 IOPS limit
  while still easing up on new I/Os during memory shortages. Remove
  tying the amount of time we do this to the number of failed requests
  and do it only as long as we keep failing requests.
  
  Finally, to avoid needless recursion when memory is tight (start ->
  g_io_deliver() -> g_io_request() -> start -> ... until we use 1/2 the
  stack), don't do direct dispatch while pacing. This should be a rare
  event (not steady state) so the performance hit here is worth the
  extra safety of not starving g_down() with directly dispatched I/O.
  
  Differential Review: https://reviews.freebsd.org/D3546

Modified:
  head/sys/geom/geom_io.c

Modified: head/sys/geom/geom_io.c
==
--- head/sys/geom/geom_io.c Wed Sep  2 16:50:49 2015(r287404)
+++ head/sys/geom/geom_io.c Wed Sep  2 17:29:30 2015(r287405)
@@ -71,7 +71,17 @@ static struct g_bioq g_bio_run_down;
 static struct g_bioq g_bio_run_up;
 static struct g_bioq g_bio_run_task;
 
-static u_int pace;
+/*
+ * Pace is a hint that we've had some trouble recently allocating
+ * bios, so we should back off trying to send I/O down the stack
+ * a bit to let the problem resolve. When pacing, we also turn
+ * off direct dispatch to also reduce memory pressure from I/Os
+ * there, at the expxense of some added latency while the memory
+ * pressures exist. See g_io_schedule_down() for more details
+ * and limitations.
+ */
+static volatile u_int pace;
+
 static uma_zone_t  biozone;
 
 /*
@@ -521,7 +531,8 @@ g_io_request(struct bio *bp, struct g_co
(pp->flags & G_PF_DIRECT_RECEIVE) != 0 &&
!g_is_geom_thread(curthread) &&
((pp->flags & G_PF_ACCEPT_UNMAPPED) != 0 ||
-   (bp->bio_flags & BIO_UNMAPPED) == 0 || THREAD_CAN_SLEEP());
+   (bp->bio_flags & BIO_UNMAPPED) == 0 || THREAD_CAN_SLEEP()) &&
+   pace == 0;
if (direct) {
/* Block direct execution if less then half of stack left. */
size_t  st, su;
@@ -688,7 +699,7 @@ g_io_deliver(struct bio *bp, int error)
bp->bio_driver2 = NULL;
bp->bio_pflags = 0;
g_io_request(bp, cp);
-   pace++;
+   pace = 1;
return;
 }
 
@@ -777,10 +788,33 @@ g_io_schedule_down(struct thread *tp __u
}
CTR0(KTR_GEOM, "g_down has work to do");
g_bioq_unlock(_bio_run_down);
-   if (pace > 0) {
-   CTR1(KTR_GEOM, "g_down pacing self (pace %d)", pace);
-   pause("g_down", hz/10);
-   pace--;
+   if (pace != 0) {
+   /*
+* There has been at least one memory allocation
+* failure since the last I/O completed. Pause 1ms to
+* give the system a chance to free up memory. We only
+* do this once because a large number of allocations
+* can fail in the direct dispatch case and there's no
+* relationship between the number of these failures and
+* the length of the outage. If there's still an outage,
+* we'll pause again and again until it's
+* resolved. Older versions paused longer and once per
+* allocation failure. This was OK for a single threaded
+* g_down, but with direct dispatch would lead to max of
+* 10 IOPs for minutes at a time when transient memory
+* issues prevented allocation for a batch of requests
+* from the upper layers.
+*
+* XXX This pacing is really

Re: svn commit: r287403 - head/sys/dev/ioat

2015-09-02 Thread John Baldwin

On Wednesday, September 02, 2015 04:48:04 PM Conrad E. Meyer wrote:
> Author: cem
> Date: Wed Sep  2 16:48:03 2015
> New Revision: 287403
> URL: https://svnweb.freebsd.org/changeset/base/287403
> 
> Log:
>   ioat: re-initialize interrupts after resetting hw on BDXDE
>   
>   Resetting some generations of the I/OAT hardware (just BDXDE for now)
>   resets the corresponding MSI-X registers.  So, teardown and
>   re-initialize interrupts after resetting the hardware.
>   
>   Reviewed by:jimharris
>   Approved by:markj (mentor)
>   Sponsored by:   EMC / Isilon Storage Division
>   Differential Revision:  https://reviews.freebsd.org/D3549

Alternatively you could use pci_restore_state() (before reset) and
pci_save_state() (after reset) to restore standard PCI config registers
(including MSI/MSI-X) after a reset.

This might be more scalable if you want to ensure other PCI config
registers (e.g. PCI-e capabilities) are restored after a reset.

-- 
John Baldwin
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"

svn commit: r287417 - head/usr.sbin/makefs

2015-09-02 Thread Xin LI

Author: delphij
Date: Thu Sep  3 01:15:23 2015
New Revision: 287417
URL: https://svnweb.freebsd.org/changeset/base/287417

Log:
  Don't leak 'var'.
  
  Reported by:  clang static analyzer

Modified:
  head/usr.sbin/makefs/mtree.c

Modified: head/usr.sbin/makefs/mtree.c
==
--- head/usr.sbin/makefs/mtree.cWed Sep  2 23:14:39 2015
(r287416)
+++ head/usr.sbin/makefs/mtree.cThu Sep  3 01:15:23 2015
(r287417)
@@ -181,7 +181,7 @@ static char *
 mtree_resolve(const char *spec, int *istemp)
 {
struct sbuf *sb;
-   char *res, *var;
+   char *res, *var = NULL;
const char *base, *p, *v;
size_t len;
int c, error, quoted, subst;
@@ -284,8 +284,10 @@ mtree_resolve(const char *spec, int *ist
free(res);
}
free(var);
+   var = NULL;
}
 
+   free(var);
sbuf_finish(sb);
res = (error == 0) ? strdup(sbuf_data(sb)) : NULL;
sbuf_delete(sb);
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"

svn commit: r287418 - in head/sys/powerpc: include powerpc

2015-09-02 Thread Justin Hibbits

Author: jhibbits
Date: Thu Sep  3 01:38:15 2015
New Revision: 287418
URL: https://svnweb.freebsd.org/changeset/base/287418

Log:
  pmap_mapdev_attr() also takes a vm_paddr_t.
  
  This was missed in r235936.  With recent work for 36-bit paddr, this is now
  needed.

Modified:
  head/sys/powerpc/include/pmap.h
  head/sys/powerpc/powerpc/pmap_dispatch.c

Modified: head/sys/powerpc/include/pmap.h
==
--- head/sys/powerpc/include/pmap.h Thu Sep  3 01:15:23 2015
(r287417)
+++ head/sys/powerpc/include/pmap.h Thu Sep  3 01:38:15 2015
(r287418)
@@ -235,7 +235,7 @@ voidpmap_kenter(vm_offset_t va, vm_pad
 void   pmap_kenter_attr(vm_offset_t va, vm_offset_t pa, vm_memattr_t);
 void   pmap_kremove(vm_offset_t);
 void   *pmap_mapdev(vm_paddr_t, vm_size_t);
-void   *pmap_mapdev_attr(vm_offset_t, vm_size_t, vm_memattr_t);
+void   *pmap_mapdev_attr(vm_paddr_t, vm_size_t, vm_memattr_t);
 void   pmap_unmapdev(vm_offset_t, vm_size_t);
 void   pmap_page_set_memattr(vm_page_t, vm_memattr_t);
 void   pmap_deactivate(struct thread *);

Modified: head/sys/powerpc/powerpc/pmap_dispatch.c
==
--- head/sys/powerpc/powerpc/pmap_dispatch.cThu Sep  3 01:15:23 2015
(r287417)
+++ head/sys/powerpc/powerpc/pmap_dispatch.cThu Sep  3 01:38:15 2015
(r287418)
@@ -463,7 +463,7 @@ pmap_mapdev(vm_paddr_t pa, vm_size_t siz
 }
 
 void *
-pmap_mapdev_attr(vm_offset_t pa, vm_size_t size, vm_memattr_t attr)
+pmap_mapdev_attr(vm_paddr_t pa, vm_size_t size, vm_memattr_t attr)
 {
 
CTR4(KTR_PMAP, "%s(%#x, %#x, %#x)", __func__, pa, size, attr);
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"

svn commit: r287419 - head/sys/boot/fdt/dts/arm

2015-09-02 Thread Oleksandr Tymoshenko

Author: gonzo
Date: Thu Sep  3 02:28:18 2015
New Revision: 287419
URL: https://svnweb.freebsd.org/changeset/base/287419

Log:
  Enable both i2c1 and i2c2. These devices are disabled in TI's DTS
  so they were disabled during DTS transition. Though there are
  no standard devices/drivers on them people might use iic(4) userland
  interface to access these buses.

Modified:
  head/sys/boot/fdt/dts/arm/beaglebone-black.dts

Modified: head/sys/boot/fdt/dts/arm/beaglebone-black.dts
==
--- head/sys/boot/fdt/dts/arm/beaglebone-black.dts  Thu Sep  3 01:38:15 
2015(r287418)
+++ head/sys/boot/fdt/dts/arm/beaglebone-black.dts  Thu Sep  3 02:28:18 
2015(r287419)
@@ -30,6 +30,22 @@
 #include "am335x-boneblack.dts"
 #include "beaglebone-common.dtsi"
 
+_pinmux {
+   i2c1_pins: pinmux_i2c1_pins {
+   pinctrl-single,pins = <
+   0x158 (PIN_INPUT_PULLUP | MUX_MODE2)/* 
spi0_d1.i2c1_sda */
+   0x15c (PIN_INPUT_PULLUP | MUX_MODE2)/* 
spi0_cs0.i2c1_scl */
+   >;
+   };
+
+   i2c2_pins: pinmux_i2c2_pins {
+   pinctrl-single,pins = <
+   0x178 (PIN_INPUT_PULLUP | MUX_MODE3)/* 
uart1_ctsn.i2c2_sda */
+   0x17c (PIN_INPUT_PULLUP | MUX_MODE3)/* 
uart1_rtsn.i2c2_scl */
+   >;
+   };
+};
+
  {
tda998x: hdmi-encoder {
compatible = "nxp,tda998x";
@@ -42,6 +58,20 @@
};
 };
 
+ {
+   pinctrl-names = "default";
+   pinctrl-0 = <_pins>;
+
+   status = "okay";
+};
+
+ {
+   pinctrl-names = "default";
+   pinctrl-0 = <_pins>;
+
+   status = "okay";
+};
+
  {
hdmi = <>;
 };
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"

svn commit: r287421 - head/usr.sbin/sysrc

2015-09-02 Thread Devin Teske

Author: dteske
Date: Thu Sep  3 03:58:59 2015
New Revision: 287421
URL: https://svnweb.freebsd.org/changeset/base/287421

Log:
  Fix mandoc(1) "WARNING: end of line whitespace"
  
  MFC after:3 days
  X-MFC-to: stable/10

Modified:
  head/usr.sbin/sysrc/sysrc.8

Modified: head/usr.sbin/sysrc/sysrc.8
==
--- head/usr.sbin/sysrc/sysrc.8 Thu Sep  3 03:29:44 2015(r287420)
+++ head/usr.sbin/sysrc/sysrc.8 Thu Sep  3 03:58:59 2015(r287421)
@@ -199,8 +199,8 @@ syntax to add items to existing values,
 the first character of the value is taken as the delimiter separating items
 .Pq usually Qo \  Qc or Qo , Qc .
 For example, in the following statement:
-.Bl -tag -width indent+
-.It \ 
+.Bl -item -offset indent
+.It
 .Nm
 cloned_interfaces+=" gif0"
 .El
@@ -223,25 +223,25 @@ For example, the above and below stateme
 .Dq gif0
 starts with an alpha-numeric character
 .Pq the letter Li g :
-.Bl -tag -width indent+
-.It \ 
+.Bl -item -offset indent
+.It
 .Nm
 cloned_interfaces+=gif0
 .El
 .Pp
 Take the following sequence for example:
-.Bl -tag -width indent+
-.It \ 
+.Bl -item -offset indent
+.It
 .Nm
 cloned_interfaces= # start with NULL
-.It \ 
+.It
 .Nm
 cloned_interfaces+=gif0
 .Dl # NULL -> `gif0' Pq NB: no preceding delimiter
-.It \ 
+.It
 .Nm
 cloned_interfaces+=gif0 # no change
-.It \ 
+.It
 .Nm
 cloned_interfaces+="tun0 gif0"
 .Dl # `gif0' -> `gif0 tun0' Pq NB: no duplication
@@ -277,21 +277,21 @@ For example, the above and below stateme
 .Dq gif0
 starts with an alpha-numeric character
 .Pq the letter Li g :
-.Bl -tag -width indent+
-.It \ 
+.Bl -item -offset indent
+.It
 .Nm
 cloned_interfaces-=gif0
 .El
 .Pp
 Take the following sequence for example:
-.Bl -tag -width indent+
-.It \ 
+.Bl -item -offset indent
+.It
 .Nm
 foo="bar baz" # start
-.It \ 
+.It
 .Nm
 foo-=bar # `bar baz' -> `baz'
-.It \ 
+.It
 .Nm
 foo-=baz # `baz' -> NULL
 .El
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"

svn commit: r287420 - head/usr.sbin/sysrc

2015-09-02 Thread Devin Teske

Author: dteske
Date: Thu Sep  3 03:29:44 2015
New Revision: 287420
URL: https://svnweb.freebsd.org/changeset/base/287420

Log:
  Remove non-functional examples.
  
  MFC after:3 days
  X-MFC-to: stable/10

Modified:
  head/usr.sbin/sysrc/sysrc.8

Modified: head/usr.sbin/sysrc/sysrc.8
==
--- head/usr.sbin/sysrc/sysrc.8 Thu Sep  3 02:28:18 2015(r287419)
+++ head/usr.sbin/sysrc/sysrc.8 Thu Sep  3 03:29:44 2015(r287420)
@@ -24,7 +24,7 @@
 .\"
 .\" $FreeBSD$
 .\"
-.Dd March 4, 2015
+.Dd September 2, 2015
 .Dt SYSRC 8
 .Os
 .Sh NAME
@@ -394,27 +394,6 @@ usbd_flags-"default"
 .Nm
 cloned_interfaces+"alternate"
 .Dl returns "alternate" if $cloned_interfaces is set .
-.Pp
-.Nm
-\&'#kern_securelevel'
-.Dl returns length in characters of $kern_securelevel .
-.Pp
-.Nm
-\&'hostname?'
-.Dl returns NULL and error status 2 if $hostname is unset Pq or if set, 
returns the value of $hostname with no error status .
-.Pp
-.Nm
-\&'hostname:?'
-.Dl returns NULL and error status 2 if $hostname is unset or NULL Pq or if set 
and non-NULL, returns value without error status .
-.Sh LIMITATIONS
-The
-.Nm
-utility presently does not support the
-.Ql rc.conf.d
-collection of system configuration files
-.Pq which requires a service name to be known during execution .
-.Pp
-This will be corrected by a future enhancement.
 .Sh SEE ALSO
 .Xr jls 1 ,
 .Xr rc.conf 5 ,
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"

svn commit: r287422 - head/sys/boot/efi/loader/arch/amd64

2015-09-02 Thread Marcel Moolenaar

Author: marcel
Date: Thu Sep  3 04:35:17 2015
New Revision: 287422
URL: https://svnweb.freebsd.org/changeset/base/287422

Log:
  For UGA, the frame buffer address obtained by scanning the
  PCI BARs does not necessarily correspond to the upper-left
  most pixel. Scan the frame buffer for which byte changed
  when changing the pixel at (0,0).
  
  Use the same technique to determine the stride. Except for
  changing the pixel at (0,0), we change the pixel at (0,1).
  
  PR:   202730
  Tested by:hartzell (at) alerce.com

Modified:
  head/sys/boot/efi/loader/arch/amd64/framebuffer.c

Modified: head/sys/boot/efi/loader/arch/amd64/framebuffer.c
==
--- head/sys/boot/efi/loader/arch/amd64/framebuffer.c   Thu Sep  3 03:58:59 
2015(r287421)
+++ head/sys/boot/efi/loader/arch/amd64/framebuffer.c   Thu Sep  3 04:35:17 
2015(r287422)
@@ -43,6 +43,21 @@ static EFI_GUID gop_guid = EFI_GRAPHICS_
 static EFI_GUID pciio_guid = EFI_PCI_IO_PROTOCOL_GUID;
 static EFI_GUID uga_guid = EFI_UGA_DRAW_PROTOCOL_GUID;
 
+static u_int
+efifb_color_depth(struct efi_fb *efifb)
+{
+   uint32_t mask;
+   u_int depth;
+
+   mask = efifb->fb_mask_red | efifb->fb_mask_green |
+   efifb->fb_mask_blue | efifb->fb_mask_reserved;
+   if (mask == 0)
+   return (0);
+   for (depth = 1; mask != 1; depth++)
+   mask >>= 1;
+   return (depth);
+}
+
 static int
 efifb_mask_from_pixfmt(struct efi_fb *efifb, EFI_GRAPHICS_PIXEL_FORMAT pixfmt,
 EFI_PIXEL_BITMASK *pixinfo)
@@ -92,83 +107,198 @@ efifb_from_gop(struct efi_fb *efifb, EFI
return (result);
 }
 
-static int
-efifb_from_uga(struct efi_fb *efifb, EFI_UGA_DRAW_PROTOCOL *uga)
+static ssize_t
+efifb_uga_find_pixel(EFI_UGA_DRAW_PROTOCOL *uga, u_int line,
+EFI_PCI_IO_PROTOCOL *pciio, uint64_t addr, uint64_t size)
+{
+   EFI_UGA_PIXEL pix0, pix1;
+   uint8_t *data1, *data2;
+   size_t count, maxcount = 1024;
+   ssize_t ofs;
+   EFI_STATUS status;
+   u_int idx;
+
+   status = uga->Blt(uga, , EfiUgaVideoToBltBuffer,
+   0, line, 0, 0, 1, 1, 0);
+   if (EFI_ERROR(status)) {
+   printf("UGA BLT operation failed (video->buffer)");
+   return (-1);
+   }
+   pix1.Red = ~pix0.Red;
+   pix1.Green = ~pix0.Green;
+   pix1.Blue = ~pix0.Blue;
+   pix1.Reserved = 0;
+
+   data1 = calloc(maxcount, 2);
+   if (data1 == NULL) {
+   printf("Unable to allocate memory");
+   return (-1);
+   }
+   data2 = data1 + maxcount;
+
+   ofs = 0;
+   while (size > 0) {
+   count = min(size, maxcount);
+
+   status = pciio->Mem.Read(pciio, EfiPciIoWidthUint32,
+   EFI_PCI_IO_PASS_THROUGH_BAR, addr + ofs, count >> 2,
+   data1);
+   if (EFI_ERROR(status)) {
+   printf("Error reading frame buffer (before)");
+   goto fail;
+   }
+   status = uga->Blt(uga, , EfiUgaBltBufferToVideo,
+   0, 0, 0, line, 1, 1, 0);
+   if (EFI_ERROR(status)) {
+   printf("UGA BLT operation failed (modify)");
+   goto fail;
+   }
+   status = pciio->Mem.Read(pciio, EfiPciIoWidthUint32,
+   EFI_PCI_IO_PASS_THROUGH_BAR, addr + ofs, count >> 2,
+   data2);
+   if (EFI_ERROR(status)) {
+   printf("Error reading frame buffer (after)");
+   goto fail;
+   }
+   status = uga->Blt(uga, , EfiUgaBltBufferToVideo,
+   0, 0, 0, line, 1, 1, 0);
+   if (EFI_ERROR(status)) {
+   printf("UGA BLT operation failed (restore)");
+   goto fail;
+   }
+   for (idx = 0; idx < count; idx++) {
+   if (data1[idx] != data2[idx]) {
+   free(data1);
+   return (ofs + (idx & ~3));
+   }
+   }
+   ofs += count;
+   size -= count;
+   }
+   printf("Couldn't find the pixel");
+
+ fail:
+   printf(" -- error %lu\n", status & ~EFI_ERROR_MASK);
+   free(data1);
+   return (-1);
+}
+
+static EFI_STATUS
+efifb_uga_detect_framebuffer(EFI_UGA_DRAW_PROTOCOL *uga,
+EFI_PCI_IO_PROTOCOL **pciiop, uint64_t *addrp, uint64_t *sizep)
 {
-   uint8_t *buf;
EFI_PCI_IO_PROTOCOL *pciio;
-   EFI_HANDLE handle;
+   EFI_HANDLE *buf, *hp;
+   uint8_t *resattr;
+   uint64_t a, addr, s, size;
+   ssize_t ofs;
EFI_STATUS status;
-   UINTN bufofs, bufsz;
-   uint64_t address, length;
-   uint32_t horiz, vert, depth, refresh;
+   UINTN bufsz;
u_int bar;
 
-   status =

svn commit: r287408 - in head/bin/sh: . tests/parser

2015-09-02 Thread Jilles Tjoelker

Author: jilles
Date: Wed Sep  2 19:49:55 2015
New Revision: 287408
URL: https://svnweb.freebsd.org/changeset/base/287408

Log:
  sh: Allow empty << EOF markers.

Added:
  head/bin/sh/tests/parser/heredoc13.0   (contents, props changed)
Modified:
  head/bin/sh/parser.c
  head/bin/sh/tests/parser/Makefile

Modified: head/bin/sh/parser.c
==
--- head/bin/sh/parser.cWed Sep  2 18:51:36 2015(r287407)
+++ head/bin/sh/parser.cWed Sep  2 19:49:55 2015(r287408)
@@ -106,6 +106,8 @@ static int startlinno;  /* line # where 
 static int funclinno;  /* line # where the current function started */
 static struct parser_temp *parser_temp;
 
+#define NOEOFMARK ((const char *))
+
 
 static union node *list(int);
 static union node *andor(void);
@@ -972,6 +974,10 @@ checkend(int c, const char *eofmark, int
pungetc();
pushstring(eofmark + 1, q - (eofmark + 1), NULL);
}
+   } else if (c == '\n' && *eofmark == '\0') {
+   c = PEOF;
+   plinno++;
+   needprompt = doprompt;
}
return (c);
 }
@@ -1383,7 +1389,7 @@ readtoken1(int firstc, char const *initi
 
STARTSTACKSTR(out);
loop: { /* for each line, until end of word */
-   if (eofmark)
+   if (eofmark && eofmark != NOEOFMARK)
/* set c to PEOF if at end of here document */
c = checkend(c, eofmark, striptabs);
for (;;) {  /* until end of line or end of word */
@@ -2046,7 +2052,7 @@ expandstr(const char *ps)
parser_temp = NULL;
setinputstring(ps, 1);
doprompt = 0;
-   readtoken1(pgetc(), DQSYNTAX, "", 0);
+   readtoken1(pgetc(), DQSYNTAX, NOEOFMARK, 0);
if (backquotelist != NULL)
error("Command substitution not allowed here");
 

Modified: head/bin/sh/tests/parser/Makefile
==
--- head/bin/sh/tests/parser/Makefile   Wed Sep  2 18:51:36 2015
(r287407)
+++ head/bin/sh/tests/parser/Makefile   Wed Sep  2 19:49:55 2015
(r287408)
@@ -57,6 +57,7 @@ FILES+=   heredoc9.0
 FILES+=heredoc10.0
 FILES+=heredoc11.0
 FILES+=heredoc12.0
+FILES+=heredoc13.0
 FILES+=line-cont1.0
 FILES+=line-cont2.0
 FILES+=line-cont3.0

Added: head/bin/sh/tests/parser/heredoc13.0
==
--- /dev/null   00:00:00 1970   (empty, because file is newly added)
+++ head/bin/sh/tests/parser/heredoc13.0Wed Sep  2 19:49:55 2015
(r287408)
@@ -0,0 +1,21 @@
+# $FreeBSD$
+
+failures=0
+
+check() {
+   if ! eval "[ $* ]"; then
+   echo "Failed: $*"
+   : $((failures += 1))
+   fi
+}
+
+check '"$(cat <<""
+
+echo yes)" = "yes"'
+
+check '"$(cat <<""
+yes
+
+)" = "yes"'
+
+exit $((failures != 0))
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"

svn commit: r287407 - head/usr.bin/netstat

2015-09-02 Thread Hiroki Sato

Author: hrs
Date: Wed Sep  2 18:51:36 2015
New Revision: 287407
URL: https://svnweb.freebsd.org/changeset/base/287407

Log:
  Simplify kvm symbol resolution and error handling.  The symbol table
  nl_symbols will eventually be organized into several modules depending
  on MK_* variables.

Added:
  head/usr.bin/netstat/nlist_symbols   (contents, props changed)
Modified:
  head/usr.bin/netstat/Makefile
  head/usr.bin/netstat/main.c
  head/usr.bin/netstat/mroute.c
  head/usr.bin/netstat/netgraph.c
  head/usr.bin/netstat/netisr.c
  head/usr.bin/netstat/netstat.h
  head/usr.bin/netstat/route.c

Modified: head/usr.bin/netstat/Makefile
==
--- head/usr.bin/netstat/Makefile   Wed Sep  2 18:42:35 2015
(r287406)
+++ head/usr.bin/netstat/Makefile   Wed Sep  2 18:51:36 2015
(r287407)
@@ -4,9 +4,32 @@
 .include 
 
 PROG=  netstat
-SRCS=  if.c inet.c main.c mbuf.c mroute.c netisr.c route.c \
+SRCS=  if.c inet.c main.c mbuf.c mroute.c netisr.c nl_symbols.c route.c \
unix.c mroute6.c ipsec.c bpf.c pfkey.c sctp.c \
flowtable.c
+DPSRCS=nl_defs.h
+
+nl_symbols.c: nlist_symbols
+   awk '\
+   BEGIN { \
+   print "#include "; \
+   print "#include "; \
+   print "struct nlist nl[] = {"; \
+   } \
+   !/^\#/ { printf("\t{ .n_name = \"%s\" },\n", $$2); } \
+   END { print "\t{ .n_name = NULL },\n};" } \
+   ' < ${.ALLSRC} > ${.TARGET} || rm -f ${.TARGET}
+nl_defs.h: nlist_symbols
+   awk '\
+   BEGIN { \
+   print "#include "; \
+   print "extern struct nlist nl[];"; \
+   i = 0; \
+   } \
+   !/^\#/ { printf("\#define\tN%s\t%s\n", toupper($$2), i++); }' \
+   < ${.ALLSRC} > ${.TARGET} || rm -f ${.TARGET}
+CLEANFILES+=   nl_symbols.c nl_defs.h
+CFLAGS+=   -I${.OBJDIR}
 
 WARNS?=3
 CFLAGS+=-fno-strict-aliasing

Modified: head/usr.bin/netstat/main.c
==
--- head/usr.bin/netstat/main.c Wed Sep  2 18:42:35 2015(r287406)
+++ head/usr.bin/netstat/main.c Wed Sep  2 18:51:36 2015(r287407)
@@ -28,7 +28,7 @@
  */
 
 #ifndef lint
-char const copyright[] =
+static char const copyright[] =
 "@(#) Copyright (c) 1983, 1988, 1993\n\
Regents of the University of California.  All rights reserved.\n";
 #endif /* not lint */
@@ -69,95 +69,10 @@ __FBSDID("$FreeBSD$");
 #include 
 #include 
 #include "netstat.h"
+#include "nl_defs.h"
 #include 
 
-static struct nlist nl[] = {
-#defineN_RTSTAT0
-   { .n_name = "_rtstat" },
-#defineN_RTREE 1
-   { .n_name = "_rt_tables"},
-#defineN_MRTSTAT   2
-   { .n_name = "_mrtstat" },
-#defineN_MFCHASHTBL3
-   { .n_name = "_mfchashtbl" },
-#defineN_VIFTABLE  4
-   { .n_name = "_viftable" },
-#defineN_NGSOCKS   5
-   { .n_name = "_ngsocklist"},
-#defineN_IP6STAT   6
-   { .n_name = "_ip6stat" },
-#defineN_ICMP6STAT 7
-   { .n_name = "_icmp6stat" },
-#defineN_IPSECSTAT 8
-   { .n_name = "_ipsec4stat" },
-#defineN_IPSEC6STAT9
-   { .n_name = "_ipsec6stat" },
-#defineN_PIM6STAT  10
-   { .n_name = "_pim6stat" },
-#defineN_MRT6STAT  11
-   { .n_name = "_mrt6stat" },
-#defineN_MF6CTABLE 12
-   { .n_name = "_mf6ctable" },
-#defineN_MIF6TABLE 13
-   { .n_name = "_mif6table" },
-#defineN_PFKEYSTAT 14
-   { .n_name = "_pfkeystat" },
-#defineN_RTTRASH   15
-   { .n_name = "_rttrash" },
-#defineN_CARPSTAT  16
-   { .n_name = "_carpstats" },
-#defineN_PFSYNCSTAT17
-   { .n_name = "_pfsyncstats" },
-#defineN_AHSTAT18
-   { .n_name = "_ahstat" },
-#defineN_ESPSTAT   19
-   { .n_name = "_espstat" },
-#defineN_IPCOMPSTAT20
-   { .n_name = "_ipcompstat" },
-#defineN_TCPSTAT   21
-   { .n_name = "_tcpstat" },
-#defineN_UDPSTAT   22
-   { .n_name = "_udpstat" },
-#defineN_IPSTAT23
-   { .n_name = "_ipstat" },
-#defineN_ICMPSTAT  24
-   { .n_name = "_icmpstat" },
-#defineN_IGMPSTAT  25
-   { .n_name = "_igmpstat" },
-#defineN_PIMSTAT   26
-   { .n_name = "_pimstat" },
-#defineN_TCBINFO   27
-   { .n_name = "_tcbinfo" },
-#defineN_UDBINFO   28
-   { .n_name = "_udbinfo" },
-#defineN_DIVCBINFO 29
-   { .n_name = "_divcbinfo" },
-#defineN_RIPCBINFO 30
-   { .n_name = "_ripcbinfo" },
-#defineN_UNP_COUNT 31
-   { .n_name = "_unp_count" },
-#defineN_UNP_GENCNT32
-   { .n_name = "_unp_gencnt" },
-#defineN_UNP_DHEAD

Re: svn commit: r287403 - head/sys/dev/ioat

2015-09-02 Thread Conrad Meyer

On Wed, Sep 2, 2015 at 10:26 AM, John Baldwin  wrote:
> Alternatively you could use pci_restore_state() (before reset) and
> pci_save_state() (after reset) to restore standard PCI config registers
> (including MSI/MSI-X) after a reset.
>
> This might be more scalable if you want to ensure other PCI config
> registers (e.g. PCI-e capabilities) are restored after a reset.


Hi John,

Do you mean pci_save_state() before reset and pci_restore_state()
afterwards? Seems reasonable to me.

Thanks,
Conrad
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"

Re: svn commit: r287403 - head/sys/dev/ioat

2015-09-02 Thread Conrad Meyer

https://reviews.freebsd.org/D3552

On Wed, Sep 2, 2015 at 12:15 PM, Conrad Meyer  wrote:
> On Wed, Sep 2, 2015 at 10:26 AM, John Baldwin  wrote:
>> Alternatively you could use pci_restore_state() (before reset) and
>> pci_save_state() (after reset) to restore standard PCI config registers
>> (including MSI/MSI-X) after a reset.
>>
>> This might be more scalable if you want to ensure other PCI config
>> registers (e.g. PCI-e capabilities) are restored after a reset.
>
>
> Hi John,
>
> Do you mean pci_save_state() before reset and pci_restore_state()
> afterwards? Seems reasonable to me.
>
> Thanks,
> Conrad
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"

svn commit: r287406 - head/usr.bin/netstat

2015-09-02 Thread Hiroki Sato

Author: hrs
Date: Wed Sep  2 18:42:35 2015
New Revision: 287406
URL: https://svnweb.freebsd.org/changeset/base/287406

Log:
  Divide statistics in the number of packets with 1000 instead of 1024
  in human-readable form.
  
  PR:   183598

Modified:
  head/usr.bin/netstat/if.c

Modified: head/usr.bin/netstat/if.c
==
--- head/usr.bin/netstat/if.c   Wed Sep  2 17:29:30 2015(r287405)
+++ head/usr.bin/netstat/if.c   Wed Sep  2 18:42:35 2015(r287406)
@@ -201,7 +201,7 @@ pfsync_stats(u_long off, const char *nam
  */
 static void
 show_stat(const char *fmt, int width, const char *name,
-u_long value, short showvalue)
+u_long value, short showvalue, int div1000)
 {
const char *lsep, *rsep;
char newfmt[64];
@@ -238,7 +238,8 @@ show_stat(const char *fmt, int width, co
 
/* Format in human readable form. */
humanize_number(buf, sizeof(buf), (int64_t)value, "",
-   HN_AUTOSCALE, HN_NOSPACE | HN_DECIMAL);
+   HN_AUTOSCALE, HN_NOSPACE | HN_DECIMAL | \
+   ((div1000) ? HN_DIVISOR_1000 : 0));
maybe_pad(lsep);
snprintf(newfmt, sizeof(newfmt), "{:%s/%%%ds}", name, width);
xo_emit(newfmt, buf);
@@ -353,7 +354,7 @@ intpr(void (*pfunc)(char *), int af)
name, ifa->ifa_flags, xname);
 
 #define IFA_MTU(ifa)   (((struct if_data *)(ifa)->ifa_data)->ifi_mtu)
-   show_stat("lu", 6, "mtu", IFA_MTU(ifa), IFA_MTU(ifa));
+   show_stat("lu", 6, "mtu", IFA_MTU(ifa), IFA_MTU(ifa), 0);
 #undef IFA_MTU
 
switch (ifa->ifa_addr->sa_family) {
@@ -416,22 +417,25 @@ intpr(void (*pfunc)(char *), int af)
 
 #defineIFA_STAT(s) (((struct if_data *)ifa->ifa_data)->ifi_ ## s)
show_stat("lu", 8, "received-packets", IFA_STAT(ipackets),
-   link|network);
-   show_stat("lu", 5, "received-errors", IFA_STAT(ierrors), link);
-   show_stat("lu", 5, "dropped-packets", IFA_STAT(iqdrops), link);
+   link|network, 1);
+   show_stat("lu", 5, "received-errors", IFA_STAT(ierrors),
+   link, 1);
+   show_stat("lu", 5, "dropped-packets", IFA_STAT(iqdrops),
+   link, 1);
if (bflag)
show_stat("lu", 10, "received-bytes", IFA_STAT(ibytes),
-   link|network);
+   link|network, 0);
show_stat("lu", 8, "sent-packets", IFA_STAT(opackets),
-   link|network);
-   show_stat("lu", 5, "send-errors", IFA_STAT(oerrors), link);
+   link|network, 1);
+   show_stat("lu", 5, "send-errors", IFA_STAT(oerrors), link, 1);
if (bflag)
show_stat("lu", 10, "sent-bytes", IFA_STAT(obytes),
-   link|network);
-   show_stat("NRSlu", 5, "collisions", IFA_STAT(collisions), link);
+   link|network, 0);
+   show_stat("NRSlu", 5, "collisions", IFA_STAT(collisions),
+   link, 1);
if (dflag)
show_stat("LSlu", 5, "dropped-packets",
-   IFA_STAT(oqdrops), link);
+   IFA_STAT(oqdrops), link, 1);
xo_emit("\n");
 
if (!aflag) {
@@ -616,24 +620,24 @@ loop:
 
xo_open_instance("stats");
show_stat("lu", 10, "received-packets",
-   new->ift_ip - old->ift_ip, 1);
+   new->ift_ip - old->ift_ip, 1, 1);
show_stat("lu", 5, "received-errors",
-   new->ift_ie - old->ift_ie, 1);
+   new->ift_ie - old->ift_ie, 1, 1);
show_stat("lu", 5, "dropped-packets",
-   new->ift_id - old->ift_id, 1);
+   new->ift_id - old->ift_id, 1, 1);
show_stat("lu", 10, "received-bytes",
-   new->ift_ib - old->ift_ib, 1);
+   new->ift_ib - old->ift_ib, 1, 0);
show_stat("lu", 10, "sent-packets",
-   new->ift_op - old->ift_op, 1);
+   new->ift_op - old->ift_op, 1, 1);
show_stat("lu", 5, "send-errors",
-   new->ift_oe - old->ift_oe, 1);
+   new->ift_oe - old->ift_oe, 1, 1);
show_stat("lu", 10, "sent-bytes",
-   new->ift_ob - old->ift_ob, 1);
+   new->ift_ob - old->ift_ob, 1, 0);
show_stat("NRSlu", 5, "collisions",
-   new->ift_co - old->ift_co, 1);
+   new->ift_co - old->ift_co, 1, 1);
if (dflag)
show_stat("LSlu", 5, "dropped-packets",
-   new->ift_od - old->ift_od, 1);
+   new->ift_od - old->ift_od, 1, 1);
xo_close_instance("stats");
xo_emit("\n");
xo_flush();
___
svn-src-head@freebsd.org

Re: svn commit: r284598 - head/share/mk

2015-09-02 Thread Simon J. Gerraty

Bryan Drewery  wrote:
> My concern is that checked in 'local' files should not be changed by
> FreeBSD. I should not have to fight conflicts of _my customizations_
> against _FreeBSD customizations (against bmake upstream)_. There is so
> much logic in these local.* files, they seem more aptly named
> 'freebsd.*' as they seem to be intended to be customizations for
> FreeBSD, rather than optional customizations for a developer or other
> downstream consumer of FreeBSD.

Yes that's true though src.* is what freebsd is using afaik
It will take some time - to optimize this, cleanly separating
infrastructure from freebsd specifics.
It is not unreasonable for some logic to exist in the checked in
local.*.mk to serve as examples.

FWIW in re-working our internal build to accommodate what is now in
FreeBSD, we now have many of the local.*.mk's include a juniper.*.mk to
hold our local customizations with minimal risk of conflicts

___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"

Notice to appear in Court #000560940

2015-09-02 Thread District Court

Notice to Appear,

You have to appear in the Court on the September 10.
Please, prepare all the documents relating to the case and bring them to Court 
on the specified date.
Note: The case will be heard by the judge in your absence if you do not come.

You can find the Court Notice is in the attachment.

Kind regards,
Leonard Lamb,
District Clerk.

___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"

svn commit: r287416 - head/sys/kern

2015-09-02 Thread Mateusz Guzik

Author: mjg
Date: Wed Sep  2 23:14:39 2015
New Revision: 287416
URL: https://svnweb.freebsd.org/changeset/base/287416

Log:
  fd: remove UMA_ZONE_ZINIT argument from Files zone
  
  Originally it was added in order to prevent trashing of objects with
  INVARIANTS enabled. The same effect is now provided with mere UMA_ZONE_NOFREE.
  
  This reverts r286921.
  
  Discussed with:   kib

Modified:
  head/sys/kern/kern_descrip.c

Modified: head/sys/kern/kern_descrip.c
==
--- head/sys/kern/kern_descrip.cWed Sep  2 23:09:01 2015
(r287415)
+++ head/sys/kern/kern_descrip.cWed Sep  2 23:14:39 2015
(r287416)
@@ -3833,7 +3833,7 @@ filelistinit(void *dummy)
 {
 
file_zone = uma_zcreate("Files", sizeof(struct file), NULL, NULL,
-   NULL, NULL, UMA_ALIGN_PTR, UMA_ZONE_NOFREE | UMA_ZONE_ZINIT);
+   NULL, NULL, UMA_ALIGN_PTR, UMA_ZONE_NOFREE);
filedesc0_zone = uma_zcreate("filedesc0", sizeof(struct filedesc0),
NULL, NULL, NULL, NULL, UMA_ALIGN_PTR, 0);
mtx_init(_lock, "sigio lock", NULL, MTX_DEF);
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"

svn commit: r287415 - head/sys/vm

2015-09-02 Thread Mateusz Guzik

Author: mjg
Date: Wed Sep  2 23:09:01 2015
New Revision: 287415
URL: https://svnweb.freebsd.org/changeset/base/287415

Log:
  Don't trash memory from UMA_ZONE_NOFREE zones.
  
  Objects obtained from such zones are supposed to retain type stability,
  which was violated by aforementioned trashing.
  
  This is a follow-up to r284861.
  
  Discussed with:   kib

Modified:
  head/sys/vm/uma_core.c

Modified: head/sys/vm/uma_core.c
==
--- head/sys/vm/uma_core.c  Wed Sep  2 22:48:41 2015(r287414)
+++ head/sys/vm/uma_core.c  Wed Sep  2 23:09:01 2015(r287415)
@@ -1946,8 +1946,8 @@ uma_zcreate(const char *name, size_t siz
 * destructor, pass UMA constructor/destructor which checks for
 * memory use after free.
 */
-   if ((!(flags & UMA_ZONE_ZINIT)) && ctor == NULL && dtor == NULL &&
-   uminit == NULL && fini == NULL) {
+   if ((!(flags & (UMA_ZONE_ZINIT | UMA_ZONE_NOFREE))) &&
+   ctor == NULL && dtor == NULL && uminit == NULL && fini == NULL) {
args.ctor = trash_ctor;
args.dtor = trash_dtor;
args.uminit = trash_init;
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"

svn commit: r287413 - head/usr.sbin/sysrc

2015-09-02 Thread Devin Teske

Author: dteske
Date: Wed Sep  2 21:53:48 2015
New Revision: 287413
URL: https://svnweb.freebsd.org/changeset/base/287413

Log:
  Minor code cleanups (no functional changes).
  
  MFC after:3 days
  X-MFC-to: stable/10

Modified:
  head/usr.sbin/sysrc/sysrc

Modified: head/usr.sbin/sysrc/sysrc
==
--- head/usr.sbin/sysrc/sysrc   Wed Sep  2 21:13:06 2015(r287412)
+++ head/usr.sbin/sysrc/sysrc   Wed Sep  2 21:53:48 2015(r287413)
@@ -257,14 +257,14 @@ while getopts aAcdDef:Fhij:nNqR:vxX flag
F) SHOW_FILE=1 ;;
h) usage ;; # NOTREACHED
i) IGNORE_UNKNOWNS=1 ;;
-   j) [ "$OPTARG" ] || die \
-   "%s: Missing or null argument to \`-j' flag" "$pgm"
+   j) [ "$OPTARG" ] ||
+   die "%s: Missing or null argument to \`-j' flag" "$pgm"
   JAIL="$OPTARG" ;;
n) SHOW_NAME= ;;
N) SHOW_VALUE= ;;
q) QUIET=1 VERBOSE= ;;
-   R) [ "$OPTARG" ] || die \
-   "%s: Missing or null argument to \`-R' flag" "$pgm"
+   R) [ "$OPTARG" ] ||
+   die "%s: Missing or null argument to \`-R' flag" "$pgm"
   ROOTDIR="$OPTARG" ;;
v) VERBOSE=1 QUIET= ;;
x) DELETE=${DELETE:-1} ;;
@@ -384,13 +384,12 @@ if [ "$JAIL" -o "$ROOTDIR" ]; then
# jls(1) or jexec(8) utilities are missing.
#
if f_have jexec && f_have jls; then
-   jid="`jls jid path | \
-   (
+   jid=$( jls jid path |
while read JID JROOT; do
[ "$JROOT" = "$ROOTDIR" ] || continue
echo $JID
done
-   )`"
+   )
 
#
# If multiple running jails match the specified root
@@ -472,8 +471,8 @@ if [ "$SHOW_ALL" ]; then
# If passed `-a' (rather than `-A'), re-purge the
# environment, removing the rc.conf(5) defaults.
#
-   [ "$SHOW_ALL" = "1" ] \
-   && f_clean_env --except rc_conf_files $EXCEPT
+   [ "$SHOW_ALL" = "1" ] &&
+   f_clean_env --except rc_conf_files $EXCEPT
 
#
# If `-f file' was passed, set $rc_conf_files to an
@@ -491,8 +490,7 @@ if [ "$SHOW_ALL" ]; then
#
[ "$SHOW_ALL" = "1" -a \
  "$( f_sysrc_find rc_conf_files )" = "$RC_DEFAULTS" \
-   ] \
-   && unset rc_conf_files
+   ] && unset rc_conf_files
fi
 
for NAME in $( set |
@@ -524,7 +522,7 @@ if [ "$SHOW_ALL" ]; then
continue
fi
 
-   [ "$VERBOSE" ] && \
+   [ "$VERBOSE" ] &&
echo -n "$( f_sysrc_find "$NAME" ): "
 
#
@@ -560,7 +558,7 @@ while [ $# -gt 0 ]; do
 *) mode=ASSIGN
esac
 
-   [ "$DESCRIBE" ] && \
+   [ "$DESCRIBE" ] &&
echo "$NAME: $( f_sysrc_desc "$NAME" )"
 
case "$1" in
@@ -573,7 +571,7 @@ while [ $# -gt 0 ]; do
# If verbose, prefix line with where the directive lives
if [ "$VERBOSE" -a ! "$CHECK_ONLY" ]; then
file=$( f_sysrc_find "$NAME" )
-   [ "$file" = "$RC_DEFAULTS" -o ! "$file" ] && \
+   [ "$file" = "$RC_DEFAULTS" -o ! "$file" ] &&
file=$( f_sysrc_get 'rc_conf_files%%[$IFS]*' )
if [ "$SHOW_EQUALS" ]; then
echo -n ": $file; "
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"

svn commit: r287414 - head/sys/dev/ioat

2015-09-02 Thread Conrad E. Meyer

Author: cem
Date: Wed Sep  2 22:48:41 2015
New Revision: 287414
URL: https://svnweb.freebsd.org/changeset/base/287414

Log:
  ioat(4): pci_save/restore_state to persist MSI-X registers over BDXDE reset
  
  Also for BWD devices, per jimharris@.
  
  Reviewed by:  jhb
  Approved by:  markj (mentor)
  Sponsored by: EMC / Isilon Storage Division
  Differential Revision:https://reviews.freebsd.org/D3552

Modified:
  head/sys/dev/ioat/ioat.c

Modified: head/sys/dev/ioat/ioat.c
==
--- head/sys/dev/ioat/ioat.cWed Sep  2 21:53:48 2015(r287413)
+++ head/sys/dev/ioat/ioat.cWed Sep  2 22:48:41 2015(r287414)
@@ -60,7 +60,7 @@ static int ioat_map_pci_bar(struct ioat_
 static void ioat_dmamap_cb(void *arg, bus_dma_segment_t *segs, int nseg,
 int error);
 static void ioat_interrupt_handler(void *arg);
-static boolean_t ioat_is_bdxde(struct ioat_softc *ioat);
+static boolean_t ioat_model_resets_msix(struct ioat_softc *ioat);
 static void ioat_process_events(struct ioat_softc *ioat);
 static inline uint32_t ioat_get_active(struct ioat_softc *ioat);
 static inline uint32_t ioat_get_ring_space(struct ioat_softc *ioat);
@@ -511,12 +511,18 @@ ioat_setup_intr(struct ioat_softc *ioat)
 }
 
 static boolean_t
-ioat_is_bdxde(struct ioat_softc *ioat)
+ioat_model_resets_msix(struct ioat_softc *ioat)
 {
u_int32_t pciid;
 
pciid = pci_get_devid(ioat->device);
switch (pciid) {
+   /* BWD: */
+   case 0x0c508086:
+   case 0x0c518086:
+   case 0x0c528086:
+   case 0x0c538086:
+   /* BDXDE: */
case 0x6f508086:
case 0x6f518086:
case 0x6f528086:
@@ -947,7 +953,7 @@ ioat_reset_hw(struct ioat_softc *ioat)
 {
uint64_t status;
uint32_t chanerr;
-   int timeout, error;
+   int timeout;
 
status = ioat_get_chansts(ioat);
if (is_ioat_active(status) || is_ioat_idle(status))
@@ -974,6 +980,13 @@ ioat_reset_hw(struct ioat_softc *ioat)
chanerr = pci_read_config(ioat->device, IOAT_CFG_CHANERR_INT_OFFSET, 4);
pci_write_config(ioat->device, IOAT_CFG_CHANERR_INT_OFFSET, chanerr, 4);
 
+   /*
+* BDXDE and BWD models reset MSI-X registers on device reset.
+* Save/restore their contents manually.
+*/
+   if (ioat_model_resets_msix(ioat))
+   pci_save_state(ioat->device);
+
ioat_reset(ioat);
 
/* Wait at most 20 ms */
@@ -982,19 +995,8 @@ ioat_reset_hw(struct ioat_softc *ioat)
if (timeout == 20)
return (ETIMEDOUT);
 
-   /*
-* BDXDE models reset MSI-X registers on device reset.  We must
-* teardown and re-setup interrupts.
-*/
-   if (ioat_is_bdxde(ioat)) {
-   error = ioat_teardown_intr(ioat);
-   if (error)
-   return (error);
-
-   error = ioat_setup_intr(ioat);
-   if (error)
-   return (error);
-   }
+   if (ioat_model_resets_msix(ioat))
+   pci_restore_state(ioat->device);
 
return (0);
 }
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"

Re: svn commit: r287358 - head

2015-09-02 Thread Alexander Kabaev

On Wed, 2 Sep 2015 16:20:34 +0300
Gleb Smirnoff  wrote:

> On Wed, Sep 02, 2015 at 09:08:25AM -0400, Alexander Kabaev wrote:
> A> > 
> A> > Does that happen to you, or do you just suppose that this is
> A> > going to happen looking at the code?
> A> > 
> A> > -- 
> A> > Totus tuus, Glebius.
> A> 
> A> It does happen to me, I wouldn't bother you otherwise.
> A> 
> A> 'make installworld KERNCONF=KANBSD' now results in kernel installed
> A> twice, once as /boot/kernel and once as /boot/kernel.KANBSD.
> A> Re-running it once more, I even get to
> A> have /boot/kernel.KANBSD.old.
> A> 
> A> As a side point, 'make buildworld' does not suffer from same issue.
> 
> installworld? Now I'm really confused. Was that typo?
> 
> -- 
> Totus tuus, Glebius.

Just to make sure information gets through un-munged this time, the
output of the
sudo make installkernel KERNCONF=KANBSD 2>&1 | tee installkernel.log is
available at 
https://people.freebsd.org/~kan/installkernel.log

-- 
Alexander Kabaev


pgpU8_6PwQLZj.pgp
Description: OpenPGP digital signature

Re: svn commit: r287358 - head

2015-09-02 Thread Alexander Kabaev

On Wed, 2 Sep 2015 16:20:34 +0300
Gleb Smirnoff  wrote:

> On Wed, Sep 02, 2015 at 09:08:25AM -0400, Alexander Kabaev wrote:
> A> > A> > Modified: head/Makefile.inc1
> A> > A> > 
> ==
> A> > A> > --- head/Makefile.inc1  Tue Sep  1 11:46:13 2015
> A> > A> > (r287357) +++ head/Makefile.inc1Tue Sep  1 11:59:12
> A> > A> > 2015(r287358) @@ -1127,6 +1127,14 @@ reinstallkernel
> A> > A> > reinstallkernel.debug: _ cd ${KRNLOBJDIR}/${INSTALLKERNEL};
> A> > A> > \ ${CROSSENV} PATH=${TMPPATH} \
> A> > A> > ${MAKE} ${IMAKE_INSTALL} KERNEL=${INSTKERNNAME}
> A> > A> > ${.TARGET:S/kernel//} +.for _kernel in
> A> > A> > ${BUILDKERNELS:[2..-1]}
> A> > A> > +   @echo
> A> > A> > "--"
> A> > A> > +   @echo ">>> Installing kernel ${_kernel}"
> A> > A> > +   @echo
> A> > A> > "--"
> A> > A> > +   cd ${KRNLOBJDIR}/${_kernel}; \
> A> > A> > +   ${CROSSENV} PATH=${TMPPATH} \
> A> > A> > +   ${MAKE} ${IMAKE_INSTALL}
> A> > A> > KERNEL=${INSTKERNNAME}.${_kernel} ${.TARGET:S/kernel//}
> A> > A> > +.endfor 
> A> > A> >  distributekernel distributekernel.debug:
> A> > A> >  .if empty(INSTALLKERNEL)
> A> > A> 
> A> > A> This is probably not doing what you think it is doing. Now,
> A> > A> when installing my kernel, built from config file , I
> A> > A> get it installed into /boot/kernel, which is good and normal,
> A> > A> and I get an extra and highly undesired copy in
> A> > A> boot/kernel.. That one I didn't ask for and would
> A> > A> rather see it gone. 
> A> > 
> A> > Does that happen to you, or do you just suppose that this is
> A> > going to happen looking at the code?
> A> > 
> A> > -- 
> A> > Totus tuus, Glebius.
> A> 
> A> It does happen to me, I wouldn't bother you otherwise.
> A> 
> A> 'make installworld KERNCONF=KANBSD' now results in kernel installed
> A> twice, once as /boot/kernel and once as /boot/kernel.KANBSD.
> A> Re-running it once more, I even get to
> A> have /boot/kernel.KANBSD.old.
> A> 
> A> As a side point, 'make buildworld' does not suffer from same issue.
> 
> installworld? Now I'm really confused. Was that typo?
> 
> -- 
> Totus tuus, Glebius.

Sorry, read them as installkernel and buildkernel respectively.
-- 
Alexander Kabaev


pgpq2PO2g3oK_.pgp
Description: OpenPGP digital signature

Re: svn commit: r287358 - head

2015-09-02 Thread Alexander Motin

On 02.09.2015 16:36, Alexander Kabaev wrote:
> On Wed, 2 Sep 2015 16:20:34 +0300
> Gleb Smirnoff  wrote:
> 
>> On Wed, Sep 02, 2015 at 09:08:25AM -0400, Alexander Kabaev wrote:
>> A> > 
>> A> > Does that happen to you, or do you just suppose that this is
>> A> > going to happen looking at the code?
>> A> > 
>> A> > -- 
>> A> > Totus tuus, Glebius.
>> A> 
>> A> It does happen to me, I wouldn't bother you otherwise.
>> A> 
>> A> 'make installworld KERNCONF=KANBSD' now results in kernel installed
>> A> twice, once as /boot/kernel and once as /boot/kernel.KANBSD.
>> A> Re-running it once more, I even get to
>> A> have /boot/kernel.KANBSD.old.
>> A> 
>> A> As a side point, 'make buildworld' does not suffer from same issue.
>>
>> installworld? Now I'm really confused. Was that typo?
>>
>> -- 
>> Totus tuus, Glebius.
> 
> Just to make sure information gets through un-munged this time, the
> output of the
> sudo make installkernel KERNCONF=KANBSD 2>&1 | tee installkernel.log is
> available at 
> https://people.freebsd.org/~kan/installkernel.log

Sorry if it is documented somewhere, but can this be disabled somehow?
On my development server installkernel for some reason already took more
time then buildkernel -DKERNFAST. Now I suppose it will take twice more.
I am doing it hundred times a day, and I am not happy. :-\

-- 
Alexander Motin
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"

svn commit: r287396 - in head: sbin/bsdlabel sbin/dumpfs sbin/fdisk sbin/ffsinfo sbin/mdconfig sbin/newfs sbin/newfs_msdos sbin/newfs_nandfs sbin/reboot share/man/man4 share/man/man7 share/man/man8...

2015-09-02 Thread Edward Tomasz Napierala

Author: trasz
Date: Wed Sep  2 14:08:43 2015
New Revision: 287396
URL: https://svnweb.freebsd.org/changeset/base/287396

Log:
  It's 2015, and some people are still trying to use fdisk and then
  go asking what debug flags to set for GEOM to make it work.  Advice
  them to use gpart(8) instead.
  
  Something similar should probably done with disklabel,
  but I need to rewrite the disklabel examples first.
  
  Reviewed by:  wblock@
  MFC after:1 month
  Sponsored by: The FreeBSD Foundation
  Differential Revision:https://reviews.freebsd.org/D3315

Modified:
  head/sbin/bsdlabel/bsdlabel.8
  head/sbin/dumpfs/dumpfs.8
  head/sbin/fdisk/fdisk.8
  head/sbin/ffsinfo/ffsinfo.8
  head/sbin/mdconfig/mdconfig.8
  head/sbin/newfs/newfs.8
  head/sbin/newfs_msdos/newfs_msdos.8
  head/sbin/newfs_nandfs/newfs_nandfs.8
  head/sbin/reboot/boot_i386.8
  head/share/man/man4/da.4
  head/share/man/man4/md.4
  head/share/man/man4/vpo.4
  head/share/man/man7/tuning.7
  head/share/man/man8/picobsd.8
  head/usr.sbin/boot0cfg/boot0cfg.8

Modified: head/sbin/bsdlabel/bsdlabel.8
==
--- head/sbin/bsdlabel/bsdlabel.8   Wed Sep  2 14:04:13 2015
(r287395)
+++ head/sbin/bsdlabel/bsdlabel.8   Wed Sep  2 14:08:43 2015
(r287396)
@@ -445,7 +445,10 @@ to properly
 recognize the disk:
 .Bd -literal -offset indent
 dd if=/dev/zero of=/dev/da0 bs=512 count=32
-fdisk -BI da0
+gpart create -s MBR da0
+gpart add -t freebsd da0
+gpart set -a active -i 1 da0
+gpart bootcode -b /boot/mbr da0
 dd if=/dev/zero of=/dev/da0s1 bs=512 count=32
 bsdlabel -w -B da0s1
 bsdlabel -e da0s1
@@ -495,6 +498,5 @@ are not generally compatible.
 .Xr md 4 ,
 .Xr disktab 5 ,
 .Xr boot0cfg 8 ,
-.Xr fdisk 8 ,
 .Xr gpart 8 ,
 .Xr newfs 8

Modified: head/sbin/dumpfs/dumpfs.8
==
--- head/sbin/dumpfs/dumpfs.8   Wed Sep  2 14:04:13 2015(r287395)
+++ head/sbin/dumpfs/dumpfs.8   Wed Sep  2 14:08:43 2015(r287396)
@@ -100,8 +100,8 @@ flag is needed if the filesystem uses
 .Sh SEE ALSO
 .Xr disktab 5 ,
 .Xr fs 5 ,
-.Xr disklabel 8 ,
 .Xr fsck 8 ,
+.Xr gpart 8 ,
 .Xr newfs 8 ,
 .Xr tunefs 8
 .Sh HISTORY

Modified: head/sbin/fdisk/fdisk.8
==
--- head/sbin/fdisk/fdisk.8 Wed Sep  2 14:04:13 2015(r287395)
+++ head/sbin/fdisk/fdisk.8 Wed Sep  2 14:08:43 2015(r287396)
@@ -39,6 +39,13 @@ The
 utility can be used to divide space on the disk into slices and set one
 active.
 .Sh DESCRIPTION
+.Bf -symbolic
+This command is obsolete.
+Users are advised to use
+.Xr gpart 8
+instead.
+.Ef
+.Pp
 The
 .Fx
 utility,

Modified: head/sbin/ffsinfo/ffsinfo.8
==
--- head/sbin/ffsinfo/ffsinfo.8 Wed Sep  2 14:04:13 2015(r287395)
+++ head/sbin/ffsinfo/ffsinfo.8 Wed Sep  2 14:08:43 2015(r287396)
@@ -121,9 +121,9 @@ to
 .Pa /var/tmp/ffsinfo
 with all available information.
 .Sh SEE ALSO
-.Xr disklabel 8 ,
 .Xr dumpfs 8 ,
 .Xr fsck 8 ,
+.Xr gpart 8 ,
 .Xr growfs 8 ,
 .Xr gvinum 8 ,
 .Xr newfs 8 ,

Modified: head/sbin/mdconfig/mdconfig.8
==
--- head/sbin/mdconfig/mdconfig.8   Wed Sep  2 14:04:13 2015
(r287395)
+++ head/sbin/mdconfig/mdconfig.8   Wed Sep  2 14:08:43 2015
(r287396)
@@ -300,8 +300,7 @@ mount /dev/md1.nop /mnt
 .Sh SEE ALSO
 .Xr md 4 ,
 .Xr ffs 7 ,
-.Xr bsdlabel 8 ,
-.Xr fdisk 8 ,
+.Xr gpart 8 ,
 .Xr mdmfs 8 ,
 .Xr malloc 9
 .Sh HISTORY

Modified: head/sbin/newfs/newfs.8
==
--- head/sbin/newfs/newfs.8 Wed Sep  2 14:04:13 2015(r287395)
+++ head/sbin/newfs/newfs.8 Wed Sep  2 14:08:43 2015(r287396)
@@ -303,11 +303,11 @@ on file systems that contain many small 
 .Xr geom 4 ,
 .Xr disktab 5 ,
 .Xr fs 5 ,
-.Xr bsdlabel 8 ,
 .Xr camcontrol 8 ,
 .Xr dump 8 ,
 .Xr dumpfs 8 ,
 .Xr fsck 8 ,
+.Xr gpart 8 ,
 .Xr gjournal 8 ,
 .Xr growfs 8 ,
 .Xr gvinum 8 ,

Modified: head/sbin/newfs_msdos/newfs_msdos.8
==
--- head/sbin/newfs_msdos/newfs_msdos.8 Wed Sep  2 14:04:13 2015
(r287395)
+++ head/sbin/newfs_msdos/newfs_msdos.8 Wed Sep  2 14:08:43 2015
(r287396)
@@ -228,9 +228,7 @@ Create a 30MB image file, with the FAT p
 newfs_msdos -C 30M -@63s ./somefile
 .Ed
 .Sh SEE ALSO
-.Xr disktab 5 ,
-.Xr disklabel 8 ,
-.Xr fdisk 8 ,
+.Xr gpart 8 ,
 .Xr newfs 8
 .Sh HISTORY
 The

Modified: head/sbin/newfs_nandfs/newfs_nandfs.8
==
--- head/sbin/newfs_nandfs/newfs_nandfs.8   Wed Sep  2 14:04:13 2015
(r287395)
+++ head/sbin/newfs_nandfs/newfs_nandfs.8   Wed Sep  2

svn commit: r287395 - head/sys/compat/linux

2015-09-02 Thread Edward Tomasz Napierala

Author: trasz
Date: Wed Sep  2 14:04:13 2015
New Revision: 287395
URL: https://svnweb.freebsd.org/changeset/base/287395

Log:
  Fixes a panic triggered by threaded Linux applications when running
  with RACCT/RCTL enabled.
  
  Reviewed by:  ngie@, ed@
  Tested by:Larry Rosenman 
  MFC after:1 month
  Sponsored by: The FreeBSD Foundation
  Differential Revision:https://reviews.freebsd.org/D3470

Modified:
  head/sys/compat/linux/linux_fork.c

Modified: head/sys/compat/linux/linux_fork.c
==
--- head/sys/compat/linux/linux_fork.c  Wed Sep  2 12:46:42 2015
(r287394)
+++ head/sys/compat/linux/linux_fork.c  Wed Sep  2 14:04:13 2015
(r287395)
@@ -285,10 +285,20 @@ linux_clone_thread(struct thread *td, st
 
p = td->td_proc;
 
+#ifdef RACCT
+   if (racct_enable) {
+   PROC_LOCK(p);
+   error = racct_add(p, RACCT_NTHR, 1);
+   PROC_UNLOCK(p);
+   if (error != 0)
+   return (EPROCLIM);
+   }
+#endif
+
/* Initialize our td */
error = kern_thr_alloc(p, 0, );
if (error)
-   return (error);
+   goto fail;


cpu_set_upcall(newtd, td);
 
@@ -369,6 +379,16 @@ linux_clone_thread(struct thread *td, st
td->td_retval[0] = newtd->td_tid;
 
return (0);
+
+fail:
+#ifdef RACCT
+   if (racct_enable) {
+   PROC_LOCK(p);
+   racct_sub(p, RACCT_NTHR, 1);
+   PROC_UNLOCK(p);
+   }
+#endif
+   return (error);
 }
 
 int
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"

Re: svn commit: r287358 - head

2015-09-02 Thread Alexander Kabaev

On Wed, 02 Sep 2015 16:57:01 +0300
Alexander Motin  wrote:

> On 02.09.2015 16:36, Alexander Kabaev wrote:
> > On Wed, 2 Sep 2015 16:20:34 +0300
> > Gleb Smirnoff  wrote:
> > 
> >> On Wed, Sep 02, 2015 at 09:08:25AM -0400, Alexander Kabaev wrote:
> >> A> > 
> >> A> > Does that happen to you, or do you just suppose that this is
> >> A> > going to happen looking at the code?
> >> A> > 
> >> A> > -- 
> >> A> > Totus tuus, Glebius.
> >> A> 
> >> A> It does happen to me, I wouldn't bother you otherwise.
> >> A> 
> >> A> 'make installworld KERNCONF=KANBSD' now results in kernel
> >> A> installed twice, once as /boot/kernel and once
> >> A> as /boot/kernel.KANBSD. Re-running it once more, I even get to
> >> A> have /boot/kernel.KANBSD.old.
> >> A> 
> >> A> As a side point, 'make buildworld' does not suffer from same
> >> A> issue.
> >>
> >> installworld? Now I'm really confused. Was that typo?
> >>
> >> -- 
> >> Totus tuus, Glebius.
> > 
> > Just to make sure information gets through un-munged this time, the
> > output of the
> > sudo make installkernel KERNCONF=KANBSD 2>&1 | tee
> > installkernel.log is available at 
> > https://people.freebsd.org/~kan/installkernel.log
> 
> Sorry if it is documented somewhere, but can this be disabled somehow?
> On my development server installkernel for some reason already took
> more time then buildkernel -DKERNFAST. Now I suppose it will take
> twice more. I am doing it hundred times a day, and I am not happy. :-\
> 

Well, I do not think this this was the intended behaviour, so it is
pretty safe to assume it will get fixed. Makefile.inc1 uses documented
'[2..-1]' subscript to prevent same kernel installed more than once but
somehow that does not work.

-- 
Alexander Kabaev


pgp6R2PLhZDx1.pgp
Description: OpenPGP digital signature

svn commit: r287397 - head/sbin/mount_unionfs

2015-09-02 Thread Edward Tomasz Napierala

Author: trasz
Date: Wed Sep  2 14:10:09 2015
New Revision: 287397
URL: https://svnweb.freebsd.org/changeset/base/287397

Log:
  Fix markup.
  
  MFC after:1 month
  Sponsored by: The FreeBSD Foundation

Modified:
  head/sbin/mount_unionfs/mount_unionfs.8

Modified: head/sbin/mount_unionfs/mount_unionfs.8
==
--- head/sbin/mount_unionfs/mount_unionfs.8 Wed Sep  2 14:08:43 2015
(r287396)
+++ head/sbin/mount_unionfs/mount_unionfs.8 Wed Sep  2 14:10:09 2015
(r287397)
@@ -83,7 +83,7 @@ becomes the upper layer.
 However,
 .Ar uniondir
 remains the mount point.
-.It Sm Cm copymode No = Cm traditional | transparent | masquerade Sm
+.It Cm copymode No = Cm traditional | transparent | masquerade
 Specifies the way to create a file or a directory in the upper layer
 automatically when needed.
 The
@@ -98,7 +98,7 @@ For behavior of the
 mode, see
 .Sx MASQUERADE MODE
 below.
-.It Sm Cm whiteout No = Cm always | whenneeded Sm
+.It Cm whiteout No = Cm always | whenneeded
 Specifies whether whiteouts should always be made in the upper layer
 when removing a file or directory or only when it already exists in the
 lower layer.
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"

Re: svn commit: r287358 - head

2015-09-02 Thread Gleb Smirnoff

  Alexander,

On Tue, Sep 01, 2015 at 10:44:38PM -0400, Alexander Kabaev wrote:
A> On Tue, 1 Sep 2015 11:59:12 + (UTC)
A> > Log:
A> >   Not only build with buildworld, but also install with installworld
A> > all alternative kernels.
A> >   
A> >   Sponsored by:Netflix
A> >   Sponsored by:Nginx, Inc.
A> > 
A> > Modified:
A> >   head/Makefile.inc1
A> > 
A> > Modified: head/Makefile.inc1
A> > 
==
A> > --- head/Makefile.inc1 Tue Sep  1 11:46:13 2015
A> > (r287357) +++ head/Makefile.inc1   Tue Sep  1 11:59:12
A> > 2015   (r287358) @@ -1127,6 +1127,14 @@ reinstallkernel
A> > reinstallkernel.debug: _ cd ${KRNLOBJDIR}/${INSTALLKERNEL}; \
A> >${CROSSENV} PATH=${TMPPATH} \
A> >${MAKE} ${IMAKE_INSTALL} KERNEL=${INSTKERNNAME}
A> > ${.TARGET:S/kernel//} +.for _kernel in ${BUILDKERNELS:[2..-1]}
A> > +  @echo
A> > "--"
A> > +  @echo ">>> Installing kernel ${_kernel}"
A> > +  @echo
A> > "--"
A> > +  cd ${KRNLOBJDIR}/${_kernel}; \
A> > +  ${CROSSENV} PATH=${TMPPATH} \
A> > +  ${MAKE} ${IMAKE_INSTALL}
A> > KERNEL=${INSTKERNNAME}.${_kernel} ${.TARGET:S/kernel//} +.endfor
A> >  
A> >  distributekernel distributekernel.debug:
A> >  .if empty(INSTALLKERNEL)
A> 
A> This is probably not doing what you think it is doing. Now, when
A> installing my kernel, built from config file , I get it installed
A> into /boot/kernel, which is good and normal, and I get an extra and
A> highly undesired copy in boot/kernel.. That one I didn't ask for
A> and would rather see it gone. 

Does that happen to you, or do you just suppose that this is going to
happen looking at the code?

-- 
Totus tuus, Glebius.
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"

Re: svn commit: r287358 - head

2015-09-02 Thread Alexander Kabaev

On Wed, 2 Sep 2015 15:50:24 +0300
Gleb Smirnoff  wrote:

>   Alexander,
> 
> On Tue, Sep 01, 2015 at 10:44:38PM -0400, Alexander Kabaev wrote:
> A> On Tue, 1 Sep 2015 11:59:12 + (UTC)
> A> > Log:
> A> >   Not only build with buildworld, but also install with
> A> > installworld all alternative kernels.
> A> >   
> A> >   Sponsored by:  Netflix
> A> >   Sponsored by:  Nginx, Inc.
> A> > 
> A> > Modified:
> A> >   head/Makefile.inc1
> A> > 
> A> > Modified: head/Makefile.inc1
> A> > 
> ==
> A> > --- head/Makefile.inc1   Tue Sep  1 11:46:13 2015
> A> > (r287357) +++ head/Makefile.inc1 Tue Sep  1 11:59:12
> A> > 2015 (r287358) @@ -1127,6 +1127,14 @@ reinstallkernel
> A> > reinstallkernel.debug: _ cd ${KRNLOBJDIR}/${INSTALLKERNEL}; \
> A> >  ${CROSSENV} PATH=${TMPPATH} \
> A> >  ${MAKE} ${IMAKE_INSTALL} KERNEL=${INSTKERNNAME}
> A> > ${.TARGET:S/kernel//} +.for _kernel in ${BUILDKERNELS:[2..-1]}
> A> > +@echo
> A> > "--"
> A> > +@echo ">>> Installing kernel ${_kernel}"
> A> > +@echo
> A> > "--"
> A> > +cd ${KRNLOBJDIR}/${_kernel}; \
> A> > +${CROSSENV} PATH=${TMPPATH} \
> A> > +${MAKE} ${IMAKE_INSTALL}
> A> > KERNEL=${INSTKERNNAME}.${_kernel} ${.TARGET:S/kernel//} +.endfor
> A> >  
> A> >  distributekernel distributekernel.debug:
> A> >  .if empty(INSTALLKERNEL)
> A> 
> A> This is probably not doing what you think it is doing. Now, when
> A> installing my kernel, built from config file , I get it
> A> installed into /boot/kernel, which is good and normal, and I get
> A> an extra and highly undesired copy in boot/kernel.. That one
> A> I didn't ask for and would rather see it gone. 
> 
> Does that happen to you, or do you just suppose that this is going to
> happen looking at the code?
> 
> -- 
> Totus tuus, Glebius.

It does happen to me, I wouldn't bother you otherwise.

'make installworld KERNCONF=KANBSD' now results in kernel installed
twice, once as /boot/kernel and once as /boot/kernel.KANBSD. Re-running
it once more, I even get to have /boot/kernel.KANBSD.old.

As a side point, 'make buildworld' does not suffer from same issue.

-- 
Alexander Kabaev


pgp2OOXjm5T36.pgp
Description: OpenPGP digital signature

Re: svn commit: r287358 - head

2015-09-02 Thread Gleb Smirnoff

On Wed, Sep 02, 2015 at 09:08:25AM -0400, Alexander Kabaev wrote:
A> > A> > Modified: head/Makefile.inc1
A> > A> > 
==
A> > A> > --- head/Makefile.inc1Tue Sep  1 11:46:13 2015
A> > A> > (r287357) +++ head/Makefile.inc1  Tue Sep  1 11:59:12
A> > A> > 2015  (r287358) @@ -1127,6 +1127,14 @@ reinstallkernel
A> > A> > reinstallkernel.debug: _ cd ${KRNLOBJDIR}/${INSTALLKERNEL}; \
A> > A> >   ${CROSSENV} PATH=${TMPPATH} \
A> > A> >   ${MAKE} ${IMAKE_INSTALL} KERNEL=${INSTKERNNAME}
A> > A> > ${.TARGET:S/kernel//} +.for _kernel in ${BUILDKERNELS:[2..-1]}
A> > A> > + @echo
A> > A> > "--"
A> > A> > + @echo ">>> Installing kernel ${_kernel}"
A> > A> > + @echo
A> > A> > "--"
A> > A> > + cd ${KRNLOBJDIR}/${_kernel}; \
A> > A> > + ${CROSSENV} PATH=${TMPPATH} \
A> > A> > + ${MAKE} ${IMAKE_INSTALL}
A> > A> > KERNEL=${INSTKERNNAME}.${_kernel} ${.TARGET:S/kernel//} +.endfor
A> > A> >  
A> > A> >  distributekernel distributekernel.debug:
A> > A> >  .if empty(INSTALLKERNEL)
A> > A> 
A> > A> This is probably not doing what you think it is doing. Now, when
A> > A> installing my kernel, built from config file , I get it
A> > A> installed into /boot/kernel, which is good and normal, and I get
A> > A> an extra and highly undesired copy in boot/kernel.. That one
A> > A> I didn't ask for and would rather see it gone. 
A> > 
A> > Does that happen to you, or do you just suppose that this is going to
A> > happen looking at the code?
A> > 
A> > -- 
A> > Totus tuus, Glebius.
A> 
A> It does happen to me, I wouldn't bother you otherwise.
A> 
A> 'make installworld KERNCONF=KANBSD' now results in kernel installed
A> twice, once as /boot/kernel and once as /boot/kernel.KANBSD. Re-running
A> it once more, I even get to have /boot/kernel.KANBSD.old.
A> 
A> As a side point, 'make buildworld' does not suffer from same issue.

installworld? Now I'm really confused. Was that typo?

-- 
Totus tuus, Glebius.
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"

svn commit: r287394 - head/etc

2015-09-02 Thread Gleb Smirnoff

Author: glebius
Date: Wed Sep  2 12:46:42 2015
New Revision: 287394
URL: https://svnweb.freebsd.org/changeset/base/287394

Log:
  Fix dynamic attach/detach of 802.11 devices after r287197:
  
  o In pccard_ether add code to start children of a 802.11
device, that are configured in rc.conf.
  o In devd.conf provide a regex matching all 802.11 devices,
and on match run pccard_ether to spawn children.
  
  PR:   202784
  Submitted by: 
  In collaboration with:"Oleg V. Nauman" 

Modified:
  head/etc/devd.conf
  head/etc/pccard_ether

Modified: head/etc/devd.conf
==
--- head/etc/devd.conf  Wed Sep  2 05:55:57 2015(r287393)
+++ head/etc/devd.conf  Wed Sep  2 12:46:42 2015(r287394)
@@ -22,6 +22,9 @@ options {
"(aac|adv|adw|aha|ahb|ahc|ahd|aic|amd|amr|asr|bt|ciss|ct|dpt|\
esp|ida|iir|ips|isp|mlx|mly|mpt|ncr|ncv|nsp|stg|sym|trm|wds)\
[0-9]+";
+   set wifi-driver-regex
+   "(ath|bwi|bwn|ipw|iwi|iwn|malo|mwl|ral|rsu|rum|run|uath|upgt|\
+   ural|urtw|urtwn|wi|wpi|wtap|zyd)[0-9]+";
 };
 
 # Note that the attach/detach with the highest value wins, so that one can
@@ -57,17 +60,16 @@ notify 0 {
 };
 
 #
-# Like Ethernet devices, but separate because
-# they have a different media type.  We may want
-# to exploit this later.
+# Like Ethernet devices, but separate because 802.11 require spawning
+# wlan(4) interface.
 #
-detach 0 {
-   media-type "802.11";
-   action "/etc/pccard_ether $device-name stop";
-};
 attach 0 {
-   media-type "802.11";
-   action "/etc/pccard_ether $device-name start";
+   device-name "$wifi-driver-regex";
+   action "/etc/pccard_ether $device-name startchildren";
+};
+detach 0 {
+   device-name "$wifi-driver-regex";
+   action "/etc/pccard_ether $device-name stopchildren";
 };
 notify 0 {
match "system"  "IFNET";

Modified: head/etc/pccard_ether
==
--- head/etc/pccard_ether   Wed Sep  2 05:55:57 2015(r287393)
+++ head/etc/pccard_ether   Wed Sep  2 12:46:42 2015(r287394)
@@ -17,6 +17,9 @@ stop_precmd="checkauto"
 stop_cmd="pccard_ether_stop"
 restart_precmd="checkauto"
 restart_cmd="pccard_ether_restart"
+startchildren_cmd="pccard_ether_startchildren"
+stopchildren_cmd="pccard_ether_stopchildren"
+extra_commands="startchildren stopchildren"
 
 setup_routes()
 {
@@ -114,6 +117,20 @@ pccard_ether_restart()
pccard_ether_start
 }
 
+pccard_ether_startchildren()
+{
+   for child in `get_if_var $ifn wlans_IF`; do
+   /etc/rc.d/netif quietstart $child
+   done
+}
+
+pccard_ether_stopchildren()
+{
+   for child in `get_if_var $ifn wlans_IF`; do
+   /etc/rc.d/netif quietstop $child
+   done
+}
+
 ifn=$1
 shift
 if [ -z "$*" ]; then
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"

Re: svn commit: r287358 - head

2015-09-02 Thread Olivier Cochard-Labbé

On Wed, Sep 2, 2015 at 3:57 PM, Alexander Motin  wrote:

>
>
> Sorry if it is documented somewhere, but can this be disabled somehow?
> On my development server installkernel for some reason already took more
> time then buildkernel -DKERNFAST. Now I suppose it will take twice more.
> I am doing it hundred times a day, and I am not happy. :-\
>
>
This create problem for generating nanobsd image too: It install kernel
twice, then the resulting /boot is too big for being copied into the small
OS partition.

Regards,

Olivier
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"

Re: svn commit: r287358 - head

2015-09-02 Thread Oliver Pinter

On 9/2/15, Olivier Cochard-Labbé  wrote:
> On Wed, Sep 2, 2015 at 3:57 PM, Alexander Motin  wrote:
>
>>
>>
>> Sorry if it is documented somewhere, but can this be disabled somehow?
>> On my development server installkernel for some reason already took more
>> time then buildkernel -DKERNFAST. Now I suppose it will take twice more.
>> I am doing it hundred times a day, and I am not happy. :-\
>>
>>
> This create problem for generating nanobsd image too: It install kernel
> twice, then the resulting /boot is too big for being copied into the small
> OS partition.

We have problems with this change too, because our kernel added twice
to installer mediums.

>
> Regards,
>
> Olivier
> ___
> svn-src-head@freebsd.org mailing list
> https://lists.freebsd.org/mailman/listinfo/svn-src-head
> To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"

Re: svn commit: r287394 - head/etc

2015-09-02 Thread Shawn Webb

Hey Gleb,

On Wed, 2015-09-02 at 12:46 +, Gleb Smirnoff wrote:
> Author: glebius
> Date: Wed Sep  2 12:46:42 2015
> New Revision: 287394
> URL: https://svnweb.freebsd.org/changeset/base/287394
> 
> Log:
>   Fix dynamic attach/detach of 802.11 devices after r287197:
>   
>   o In pccard_ether add code to start children of a 802.11
> device, that are configured in rc.conf.
>   o In devd.conf provide a regex matching all 802.11 devices,
> and on match run pccard_ether to spawn children.
>   
>   PR: 202784
>   Submitted by:   
>   In collaboration with:  "Oleg V. Nauman" 
> 
> Modified:
>   head/etc/devd.conf
>   head/etc/pccard_ether
> 
> Modified: head/etc/devd.conf
> ==
> --- head/etc/devd.confWed Sep  2 05:55:57 2015(r287393)
> +++ head/etc/devd.confWed Sep  2 12:46:42 2015(r287394)
> @@ -22,6 +22,9 @@ options {
>   "(aac|adv|adw|aha|ahb|ahc|ahd|aic|amd|amr|asr|bt|ciss|ct|dpt|\
>   esp|ida|iir|ips|isp|mlx|mly|mpt|ncr|ncv|nsp|stg|sym|trm|wds)\
>   [0-9]+";
> + set wifi-driver-regex
> + "(ath|bwi|bwn|ipw|iwi|iwn|malo|mwl|ral|rsu|rum|run|uath|upgt|\
> + ural|urtw|urtwn|wi|wpi|wtap|zyd)[0-9]+";
>  };

Should the new iwm driver also be added?

>  
>  # Note that the attach/detach with the highest value wins, so that one can
> @@ -57,17 +60,16 @@ notify 0 {
>  };
>  
>  #
> -# Like Ethernet devices, but separate because
> -# they have a different media type.  We may want
> -# to exploit this later.
> +# Like Ethernet devices, but separate because 802.11 require spawning
> +# wlan(4) interface.
>  #
> -detach 0 {
> - media-type "802.11";
> - action "/etc/pccard_ether $device-name stop";
> -};
>  attach 0 {
> - media-type "802.11";
> - action "/etc/pccard_ether $device-name start";
> + device-name "$wifi-driver-regex";
> + action "/etc/pccard_ether $device-name startchildren";
> +};
> +detach 0 {
> + device-name "$wifi-driver-regex";
> + action "/etc/pccard_ether $device-name stopchildren";
>  };
>  notify 0 {
>   match "system"  "IFNET";
> 
> Modified: head/etc/pccard_ether
> ==
> --- head/etc/pccard_ether Wed Sep  2 05:55:57 2015(r287393)
> +++ head/etc/pccard_ether Wed Sep  2 12:46:42 2015(r287394)
> @@ -17,6 +17,9 @@ stop_precmd="checkauto"
>  stop_cmd="pccard_ether_stop"
>  restart_precmd="checkauto"
>  restart_cmd="pccard_ether_restart"
> +startchildren_cmd="pccard_ether_startchildren"
> +stopchildren_cmd="pccard_ether_stopchildren"
> +extra_commands="startchildren stopchildren"
>  
>  setup_routes()
>  {
> @@ -114,6 +117,20 @@ pccard_ether_restart()
>   pccard_ether_start
>  }
>  
> +pccard_ether_startchildren()
> +{
> + for child in `get_if_var $ifn wlans_IF`; do
> + /etc/rc.d/netif quietstart $child
> + done
> +}
> +
> +pccard_ether_stopchildren()
> +{
> + for child in `get_if_var $ifn wlans_IF`; do
> + /etc/rc.d/netif quietstop $child
> + done
> +}
> +
>  ifn=$1
>  shift
>  if [ -z "$*" ]; then
> ___
> svn-src-...@freebsd.org mailing list
> https://lists.freebsd.org/mailman/listinfo/svn-src-all
> To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

-- 
Shawn Webb
HardenedBSD

GPG Key ID:  0x6A84658F52456EEE
GPG Key Fingerprint: 2ABA B6BD EF6A F486 BE89  3D9E 6A84 658F 5245 6EEE


signature.asc
Description: This is a digitally signed message part

svn commit: r287398 - head/etc

2015-09-02 Thread Gleb Smirnoff

Author: glebius
Date: Wed Sep  2 14:38:16 2015
New Revision: 287398
URL: https://svnweb.freebsd.org/changeset/base/287398

Log:
  Add iwm(4), that was missing in r287394.
  
  Submitted by: Shawn Webb

Modified:
  head/etc/devd.conf

Modified: head/etc/devd.conf
==
--- head/etc/devd.conf  Wed Sep  2 14:10:09 2015(r287397)
+++ head/etc/devd.conf  Wed Sep  2 14:38:16 2015(r287398)
@@ -23,8 +23,8 @@ options {
esp|ida|iir|ips|isp|mlx|mly|mpt|ncr|ncv|nsp|stg|sym|trm|wds)\
[0-9]+";
set wifi-driver-regex
-   "(ath|bwi|bwn|ipw|iwi|iwn|malo|mwl|ral|rsu|rum|run|uath|upgt|\
-   ural|urtw|urtwn|wi|wpi|wtap|zyd)[0-9]+";
+   "(ath|bwi|bwn|ipw|iwi|iwm|iwn|malo|mwl|ral|rsu|rum|run|uath|\
+   upgt|ural|urtw|urtwn|wi|wpi|wtap|zyd)[0-9]+";
 };
 
 # Note that the attach/detach with the highest value wins, so that one can
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"

Re: svn commit: r287396 - in head: sbin/bsdlabel sbin/dumpfs sbin/fdisk sbin/ffsinfo sbin/mdconfig sbin/newfs sbin/newfs_msdos sbin/newfs_nandfs sbin/reboot share/man/man4 share/man/man7 share/man/man

2015-09-02 Thread Gleb Smirnoff

On Wed, Sep 02, 2015 at 02:08:43PM +, Edward Tomasz Napierala wrote:
E> Author: trasz
E> Date: Wed Sep  2 14:08:43 2015
E> New Revision: 287396
E> URL: https://svnweb.freebsd.org/changeset/base/287396
E> 
E> Log:
E>   It's 2015, and some people are still trying to use fdisk and then
E>   go asking what debug flags to set for GEOM to make it work.  Advice
E>   them to use gpart(8) instead.
E>   
E>   Something similar should probably done with disklabel,
E>   but I need to rewrite the disklabel examples first.

Thanks!

Do we still have functionality of fdisk/bsdlabel that isn't covered
by gpart? Can we simply remove the tools?

-- 
Totus tuus, Glebius.
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"

42 matches

Mail list logo