| For You |
| Dec 20, 3:49 pm 2010 |
| evaluacion crediticia | Aprenda a Otorgar Creditos Sanos, Prevenga y Recupere su ...
CURSO TALLER
Aprenda a Otorgar Creditos Sanos, Prevenga y Recupere su Cartera Vencida
Experto instructor: Lic. Juan Manuel Lopez Vitela )
Duracion: (1 dia) 10 hrs.
Inversion: $4,200 pesos mas IVA / 420 dlls.
Objetivo: Capacitar al participante para que identifique los pasos
necesarios en el otorgamiento del cridito, desde el llenado de la
solicitud hasta el desarrollo de un modelo ...
| Dec 20, 2:44 pm 2010 |
| Miod Vallat | msleep vs IPL_NONE mutexes
I just noticed this morning that msleep() will not behave correctly on
some platforms, when passed a mutex associated to the IPL_NONE level
(i.e. plain mutex, without an associated spl semantic).
msleep() needs to run at splsched(), and fiddles with the mutex fields
to make sure mtx_leave() will not lower spl below IPL_SCHED, and fiddles
later on, after taking the mutex back, to make the next mtx_leave()
return to the level the first mtx_enter() was run at.
This works on amd64, i386, powerpc ...
| Dec 20, 11:36 am 2010 |
| Mark Kettenis | Re: msleep vs IPL_NONE mutexes
I think the solution is to make sure splx() isn't expensive if it
doesn't actually lower the priority level. If the guts of splx()
really are expensive, we should check whether we are already at the
desired priority level and skip the expensive bits if we are. They
should be a no-op in that case. Now that check can still be somewhat
expensive if the current priority level is kept in a hardware register
on some slow bus far away. But then we could consider keeping a
software copy in struct ...
| Dec 20, 1:06 pm 2010 |
| Top Shop |
| Dec 20, 9:24 am 2010 |
| Mark Kettenis |
| Dec 20, 9:46 am 2010 |
| Claudio Jeker | mclusters and pool
With the change in rev. 1.99 of subr_pool() changing the limit
will no longer set the high watermark as well.
This causes i386 to stop forwading packets on very busy systems because
the pool thread is not fast enough to replenish the pool. So reintroduce
the pool_sethiwat() in this particular case as a workaround.
Index: uipc_mbuf.c
===================================================================
RCS file: /cvs/src/sys/kern/uipc_mbuf.c,v
retrieving revision 1.147
diff -u -p -r1.147 ...
| Dec 20, 8:12 am 2010 |
| Mike Belopuhov | Re: mclusters and pool
On Mon, Dec 20, 2010 at 5:46 PM, Mark Kettenis <mark.kettenis@xs4all.nl>
no, that change is correct in any case :)
| Dec 20, 9:51 am 2010 |
| Andreas Bartelt |
| Dec 20, 9:22 am 2010 |
| Mike Belopuhov | Re: mclusters and pool
On Mon, Dec 20, 2010 at 5:11 PM, Mark Kettenis <mark.kettenis@xs4all.nl>
i think we should restore the behavior to what it was before until
we'll have drivers fixed.
| Dec 20, 9:29 am 2010 |
| Mark Kettenis | Re: mclusters and pool
Sorry, claudio, but I think you need to fix the driver(s) that have
this issue. Even with adjusting the pool limits you may still end up
with an empty ring, and the driver should be able to restart itself if
mbufs come available again. A potential way to fix this is to start a
timeout if you end up with an empty ring because of mbug allocation
failures, and attempt to refill the ring when the timout fires.
| Dec 20, 9:11 am 2010 |
| TBX | Ventiladores para el hogar y la industria
$ 235 Pesos Ventilador 3 en 1 Ventilador
3 en 1 / Ventilador de pie, apoyado en el
piso o colgado / 3 velocidades / 1.7 mts de
altura. 500mm de d&iacute;ametro / Paletas
de metal. 18" / Movimiento
panor&aacute;mico / Fabricado bajo las
normas ISO / 150Wats $ 470 Pesos ...
| Dec 20, 4:43 am 2010 |
| Novedades |
| Dec 19, 5:56 pm 2010 |
| dalildevil486 | 龄房产信息○
SP-------5X-----9z------#(Fa#&#Piao#)?IMb?* #,WIQ/#:137 171 387 03#, QQ#: "a"a"Z KD>E6~ R;NeR;
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com
| Dec 19, 6:45 pm 2010 |
| Dear Internet Friend |
| Dec 19, 5:57 pm 2010 |
| Kenneth R Westerback |
| Dec 20, 10:26 am 2010 |
| Kenneth R Westerback | Re: softraid(4) disks are of wrong size
The same issue ('c' being set by kernel) will apply to softraid
devices too. You seem to have discovered a bug, so your test process
was good from that standpoint. :-). But you shouldn't use 'c' for
a filesystem or RAID.
.... Ken
| Dec 20, 6:09 am 2010 |
| David Gwynne |
| Dec 19, 11:26 pm 2010 |
| David Gwynne | Re: softraid(4) disks are of wrong size
compiles. can someone test?
Index: softraid.c
===================================================================
RCS file: /cvs/src/sys/dev/softraid.c,v
retrieving revision 1.216
diff -u -p -r1.216 softraid.c
--- softraid.c 6 Nov 2010 23:01:56 -0000 1.216
+++ softraid.c 20 Dec 2010 08:21:33 -0000
@@ -3507,21 +3507,22 @@ sr_raid_read_cap(struct sr_workunit *wu)
struct scsi_read_cap_data rcd;
struct scsi_read_cap_data_16 rcd16;
int rv = 1;
+ int64_t size = ...
| Dec 20, 1:24 am 2010 |
| Andreas Bartelt | Re: softraid(4) disks are of wrong size
On 12/20/10 18:26, Kenneth R Westerback wrote:
Debugger
panic
sr_crypto_finish_io
sr_crypto_intr
sdstrategy
spec_strategy
VOP_STRATEGY
sr_startwu_callback
sd.c:572 calls bounds_check_with_label (in kern/subr_disk.c) which
returns '-1' in line 698. My assumption was that this eventually leads
to the panic.
| Dec 20, 10:59 am 2010 |
| Andreas Bartelt | Re: softraid(4) disks are of wrong size
On 12/19/10 22:28, Kenneth R Westerback wrote:
I've used the native disk sd4c as RAID only for testing, since this
improved my understanding of the problem.
Is there also a problem when newfs is used on the virtual softraid(4)
disk 'c' partition (rsd5c in my example)?
I originally discovered the problem when I tried newfs(8) on rsd5c where
the native disk was sd4a. As I wrote in my initial mail, this results in
a kernel panic. The kernel panics in sr_crypto_finish_io(), and before ...
| Dec 20, 3:50 am 2010 |
| Kenneth R Westerback | Re: softraid(4) disks are of wrong size
If anyone can test some rebuilds with this or dlg@'s diff, at least
some RAID 1 rebuilds, that would be very appreciated confirmation
that this doesn't affect any internal logic of softraid.
.... Ken
| Dec 20, 8:02 am 2010 |
| Kenneth R Westerback | Re: softraid(4) disks are of wrong size
Pretty much identical to the one I sent directly out to marco, Andreas
and tedu. Andreas has reported it worked for him. Mine is below for
completeness. I like my variable name better, but int64_t may be the
better type to use.
.... Ken
Index: softraid.c
===================================================================
RCS file: /cvs/src/sys/dev/softraid.c,v
retrieving revision 1.216
diff -u -p -r1.216 softraid.c
--- softraid.c 6 Nov 2010 23:01:56 -0000 1.216
+++ softraid.c 19 Dec ...
| Dec 20, 6:06 am 2010 |
| Marco Peereboom | Re: softraid(4) disks are of wrong size
I am working on this bug and it is a side effect of wrapping VOP_*
calls. krw meanwhile commited a fix for the bug that exposed the panic
and should make your test case happy. The underlying bug needs more
work though but I am working on it.
| Dec 20, 12:18 pm 2010 |
| Claudio Jeker | Re: convert netisr to real softinterrupt
Well this was not "quick" but I finally finsihed the testing.
Tested with packet generator set to a packet size of 512B.
All test were done on the same machine using the dual port ix(4) as
interface. Forwarding performance measured with pf disabled, no
POOL_DEBUG and a pool fix for the mcl2k pool for i386.
No other fiddling was done. dmesg of the box is below.
The numbers are the top forwarding speeds as measured by the packet
generator right before livelock detection would kick in (which ...
| Dec 20, 7:45 am 2010 |
| Kenneth R Westerback |
| Dec 20, 10:18 am 2010 |
| Claudio Jeker | kill an unused IPv6 statistic
ip6s_exthdrtoolong is not used at all so why bother keeping it around?
I guess this was a pre-m_pulldown() counter but is now unused.
Index: sys/netinet6/ip6_var.h
===================================================================
RCS file: /cvs/src/sys/netinet6/ip6_var.h,v
retrieving revision 1.37
diff -u -p -r1.37 ip6_var.h
--- sys/netinet6/ip6_var.h 8 Jul 2010 19:42:46 -0000 1.37
+++ sys/netinet6/ip6_var.h 29 Nov 2010 17:26:21 -0000
@@ -196,7 +196,6 @@ struct ip6stat {
u_int64_t ...
| Dec 20, 8:44 am 2010 |
