Re: High Network Perfomance
On Fri, Aug 05, 2011 at 08:34:19PM -0400, Outback Dingo wrote: > On Fri, Aug 5, 2011 at 7:31 PM, Luigi Rizzo wrote: > > > On Fri, Aug 05, 2011 at 04:07:22PM -0300, Victor Detoni wrote: > > > Hi Luigi, > > > > > > Thanks for reply. That's great solution :) Will be integrated with pf or > > is > > > it? > > > > unfortunately i don't have the time to do it, but as said > > in the thread it should not be terribly difficult. > > > > The demo image has a modified click is there a diff for click > itself. if i am not wrong the modifications have been imported (in an improved form) in the source tree for Click 2.0 cheers luigi ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: svn commit: r224659 - head/etc
Hi Doug, * Doug Barton , 20110806 02:07: > The other issue is that due to the way that mergemaster works by default > users who have existing /etc/nsswitch.conf files (which should be pretty > close to 100%) won't see this change because it doesn't affect the > $FreeBSD$ tag. If we decide to keep things this way then I would suggest > adding a comment to that file so that it will get picked up. You can work around this by doing a forced commit with SVN: - Add some whitespace to the file - Run svn commit - Remove the whitespace while the editor is open This will bump the $FreeBSD$ without changing its file contents. > However, I would much rather see us actually change the default file. > Users who are going to enable nis will already know that this file needs > to be tweaked, and by using more sensible defaults we'll be more > relevant to the majority of our userbase. So I propose the attached > (along with reverting your change obviously). I'm fine either way. Feel free to commit it (after re approval). :-) -- Ed Schouten WWW: http://80386.nl/ pgpfzCgwN13G4.pgp Description: PGP signature
Re: New installation script
On Fri, Aug 5, 2011 at 3:35 PM, Kevin Oberman wrote: > I have installed 9-Beta1 using the new installation tool and I am > generally happy with it. > The new dialog cause me to need a few more key-strokes because I was > so used to the > old behavior, but it really is more intuitive and I would not want to > see the old behavior > restored. I'll get used to it soon. > > I do have a couple of issues with the new installation tool, though. > 1. After completing the partition design I am presented with the > option to "Save" the > partitions. It is not at all cleared that "Save" actually creates the > partitions and newfses > the file systems. I suggest changing "Save" to "Commit" or Execute". > These are far > clearer and more frightening. "Save" sounds too safe, not like you are > about to update > basic disk structure and may be about to make any data on the disk unusable. > > 2. I was installing 9 into an existing set of partitions. (I > understand that this is NOT > typical.) First the system asks me about adding a partition. Oops! I > selected the only > option that was not clearly wrong, "Cancel". I was not at all sure > that it was what I > wanted, but it was. I have no idea how to improve this and it's > probably not worth > spending much time think about it. But the next step was confusing. > > I selected each of the existing partitions that I was going to use and > selected modify to > enter the name of the partition (/, /var, /usr, /tmp). I then quit and > selected the not > scarey "Save". I proceeded, but thought the "Save" was rather fast. > Then the install failed > because the partitions were already populated. I ended up re-booting > and then going > through each partition and deleting it and then selecting the slice > and creating it again. > While not a big deal, it seemed like the Modify to name the partitions > should have > triggered the newfs that was not done. > > I think my first point is pretty important. The second is far less so. > > The install went pretty well and I am generally very pleased with the > new installer. It's > certainly an improvement over the old one! Thanks to the folks who worked on > it. One of the things that's still a problem as well (since you're talking about it here) is that the partition editor assumes that all partitions are properly formatted, etc, when specifying just a mountpoint. One needs to trash the MBR / GPT metadata and start over from the beginning. GPT partitions have issues too with incomplete partition schemes (i.e. user deletes a GPT partition out of an existing setup, etc) because unfortunately the "boot" partition gets created improperly the 2nd+ time around and/or gets created multiple times for some whacky reason (I don't know why this happened, but it did!). Thanks, -Garrett ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: High Network Perfomance
On Fri, Aug 5, 2011 at 7:31 PM, Luigi Rizzo wrote: > On Fri, Aug 05, 2011 at 04:07:22PM -0300, Victor Detoni wrote: > > Hi Luigi, > > > > Thanks for reply. That's great solution :) Will be integrated with pf or > is > > it? > > unfortunately i don't have the time to do it, but as said > in the thread it should not be terribly difficult. > The demo image has a modified click is there a diff for click itself. > cheers > luigi > > > Thanks, > > Victor > > > > On Thu, Aug 4, 2011 at 9:27 PM, Luigi Rizzo wrote: > > > > > On Thu, Aug 04, 2011 at 08:39:17PM -0300, Victor Detoni wrote: > > > > Hi Guys, > > > > > > > > I'm trying tunning a FreeBSD 8.2 to high perfomance network with pf. > My > > > > server configuration is: > > > > > > > > Dell 1950 > > > > CPU: Intel(R) Xeon(R) CPU5130 @ 2.00GHz (1995.03-MHz > > > K8-class > > > > CPU) > > > > 4 x CPU > > > > 2 NIC ( > > > 1 NIC (em0: ) > > > > > > > > I want to reach the high processing of packets per second and use pf > as > > > > synproxy and we still processor to handle others packets or flows. > > > > > > > > I know that em drivers has MULTI_QUEUE implementation that helps high > > > > performance for Intel drivers, but I couldn't see more information > about. > > > We > > > > can reached 500k pps, but no more traffic was processed by this > > > interface. > > > > > > > > I've already enabled net.isr.direct but with Intel Drivers does not > work > > > and > > > > the most processors are in System instead of Interrupts, why? When I > > > enable > > > > net.isr.direct the processing is balanced for on CPU in system and > > > another > > > > in interrupt and I reached 1M pps, but the total perfomance is down, > the > > > > load grow up too fast. > > > > > > > > I've changed some parameter in sysctl for intel drivers, but it > doesn't > > > have > > > > effect. > > > > > > > > Someone know what I can do to reach more packets performance? I want > to > > > use > > > > this FreeBSD as a router/firewall only. > > > > > > if you feel like doing a bit of coding yourself, you could try netmap > > > > > > http://info.iet.unipi.it/~luigi/netmap/ > > > > > > cheers > > > luigi > > > > > > > Thanks, > > > > Victor > > > > ___ > > > > freebsd-current@freebsd.org mailing list > > > > http://lists.freebsd.org/mailman/listinfo/freebsd-current > > > > To unsubscribe, send any mail to " > > > freebsd-current-unsubscr...@freebsd.org" > > > > > ___ > > freebsd-current@freebsd.org mailing list > > http://lists.freebsd.org/mailman/listinfo/freebsd-current > > To unsubscribe, send any mail to " > freebsd-current-unsubscr...@freebsd.org" > ___ > freebsd-current@freebsd.org mailing list > http://lists.freebsd.org/mailman/listinfo/freebsd-current > To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org" > ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: svn commit: r224659 - head/etc
Ed,
First off, thanks for taking a look at this. I've always been
uncomfortable with our default nsswitch.conf file because most users do
not use nis, and although everything works with our (previous) default
it produces errors in the logs that are very non-obvious.
That said, I think the approach you have here needs some tweaking. You
only posted it to -current yesterday, and I hadn't even gotten past
thinking "Ok, that's a good direction to start looking" before it was
committed already. :) I know deadlines are tight prior to the release,
but I don't think they are _that_ tight.
I have 2 suggestions. The first is that if we go this way that we tweak
the sed statements to comment out the _compat entries rather than
deleting them. This will accomplish the same thing but be slightly less
confusing for users who eventually want to implement nis. This can be
accomplished like this:
-e 's/.*_compat:/# &/' -e 's/compat$/files/'
The other issue is that due to the way that mergemaster works by default
users who have existing /etc/nsswitch.conf files (which should be pretty
close to 100%) won't see this change because it doesn't affect the
$FreeBSD$ tag. If we decide to keep things this way then I would suggest
adding a comment to that file so that it will get picked up.
However, I would much rather see us actually change the default file.
Users who are going to enable nis will already know that this file needs
to be tweaked, and by using more sensible defaults we'll be more
relevant to the majority of our userbase. So I propose the attached
(along with reverting your change obviously).
hth,
Doug
On 08/05/2011 10:33, Ed Schouten wrote:
> Author: ed
> Date: Fri Aug 5 17:33:12 2011
> New Revision: 224659
> URL: http://svn.freebsd.org/changeset/base/224659
>
> Log:
> Remove NIS entries from /etc/nsswitch.conf for WITHOUT_NIS.
>
> We already modify various configuration files in /etc based on build
> configuration. This is not done for nsswitch.conf right now when setting
> WITHOUT_NIS. This breaks various utilities, including crond, that depend
> on working databases.
>
> Approved by:re (kib)
> MFC after: 1 month
>
> Modified:
> head/etc/Makefile
>
> Modified: head/etc/Makefile
> ==
> --- head/etc/Makefile Fri Aug 5 17:15:46 2011(r224658)
> +++ head/etc/Makefile Fri Aug 5 17:33:12 2011(r224659)
> @@ -256,6 +256,10 @@ distribution:
> ${DESTDIR}/boot/device.hints
> .endif
> .endif
> +.if ${MK_NIS} == "no"
> + sed -i "" -e '/_compat:/d' -e 's/compat/files/' \
> + ${DESTDIR}/etc/nsswitch.conf
> +.endif
>
> distrib-dirs:
> mtree -eU ${MTREE_FOLLOWS_SYMLINKS} -f ${.CURDIR}/mtree/BSD.root.dist
> -p ${DESTDIR}/
>
--
Nothin' ever doesn't change, but nothin' changes much.
-- OK Go
Breadth of IT experience, and depth of knowledge in the DNS.
Yours for the right price. :) http://SupersetSolutions.com/
#
# $FreeBSD: head/etc/nsswitch.conf 158266 2006-05-03 15:14:47Z ume $
#
# To enable NIS the commented entries should replace their uncommented
# equivalents. See nsswitch.conf(5) for more information.
#
group: files
#group_compat: nis
#group: compat
passwd: files
#passwd_compat: nis
#passwd: compat
services: files
#services_compat: nis
#services: compat
hosts: files dns
networks: files
shells: files
protocols: files
rpc: files
___
freebsd-current@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-current
To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Can't map a Spanish keyboard on Current but in FreeBSD 7.4-STABLE it works fin.
I'm running kde on both and 7.4 works equally well with ttys and with kde4-4.6.5. My FreeBSD 9.0-BETA1 works with ttys but not even close with kde4-4.6.5. Is this me or could it be kde or Current? There doesn't seem to be any changes in the language files spanish.iso.acc.kbd, for example. I've been tolerating this for the last week since setting it up with Beta1. Thanks, ed ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: High Network Perfomance
On Fri, Aug 05, 2011 at 04:07:22PM -0300, Victor Detoni wrote: > Hi Luigi, > > Thanks for reply. That's great solution :) Will be integrated with pf or is > it? unfortunately i don't have the time to do it, but as said in the thread it should not be terribly difficult. cheers luigi > Thanks, > Victor > > On Thu, Aug 4, 2011 at 9:27 PM, Luigi Rizzo wrote: > > > On Thu, Aug 04, 2011 at 08:39:17PM -0300, Victor Detoni wrote: > > > Hi Guys, > > > > > > I'm trying tunning a FreeBSD 8.2 to high perfomance network with pf. My > > > server configuration is: > > > > > > Dell 1950 > > > CPU: Intel(R) Xeon(R) CPU5130 @ 2.00GHz (1995.03-MHz > > K8-class > > > CPU) > > > 4 x CPU > > > 2 NIC ( > > 1 NIC (em0: ) > > > > > > I want to reach the high processing of packets per second and use pf as > > > synproxy and we still processor to handle others packets or flows. > > > > > > I know that em drivers has MULTI_QUEUE implementation that helps high > > > performance for Intel drivers, but I couldn't see more information about. > > We > > > can reached 500k pps, but no more traffic was processed by this > > interface. > > > > > > I've already enabled net.isr.direct but with Intel Drivers does not work > > and > > > the most processors are in System instead of Interrupts, why? When I > > enable > > > net.isr.direct the processing is balanced for on CPU in system and > > another > > > in interrupt and I reached 1M pps, but the total perfomance is down, the > > > load grow up too fast. > > > > > > I've changed some parameter in sysctl for intel drivers, but it doesn't > > have > > > effect. > > > > > > Someone know what I can do to reach more packets performance? I want to > > use > > > this FreeBSD as a router/firewall only. > > > > if you feel like doing a bit of coding yourself, you could try netmap > > > > http://info.iet.unipi.it/~luigi/netmap/ > > > > cheers > > luigi > > > > > Thanks, > > > Victor > > > ___ > > > freebsd-current@freebsd.org mailing list > > > http://lists.freebsd.org/mailman/listinfo/freebsd-current > > > To unsubscribe, send any mail to " > > freebsd-current-unsubscr...@freebsd.org" > > > ___ > freebsd-current@freebsd.org mailing list > http://lists.freebsd.org/mailman/listinfo/freebsd-current > To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org" ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
New installation script
I have installed 9-Beta1 using the new installation tool and I am generally happy with it. The new dialog cause me to need a few more key-strokes because I was so used to the old behavior, but it really is more intuitive and I would not want to see the old behavior restored. I'll get used to it soon. I do have a couple of issues with the new installation tool, though. 1. After completing the partition design I am presented with the option to "Save" the partitions. It is not at all cleared that "Save" actually creates the partitions and newfses the file systems. I suggest changing "Save" to "Commit" or Execute". These are far clearer and more frightening. "Save" sounds too safe, not like you are about to update basic disk structure and may be about to make any data on the disk unusable. 2. I was installing 9 into an existing set of partitions. (I understand that this is NOT typical.) First the system asks me about adding a partition. Oops! I selected the only option that was not clearly wrong, "Cancel". I was not at all sure that it was what I wanted, but it was. I have no idea how to improve this and it's probably not worth spending much time think about it. But the next step was confusing. I selected each of the existing partitions that I was going to use and selected modify to enter the name of the partition (/, /var, /usr, /tmp). I then quit and selected the not scarey "Save". I proceeded, but thought the "Save" was rather fast. Then the install failed because the partitions were already populated. I ended up re-booting and then going through each partition and deleting it and then selecting the slice and creating it again. While not a big deal, it seemed like the Modify to name the partitions should have triggered the newfs that was not done. I think my first point is pretty important. The second is far less so. The install went pretty well and I am generally very pleased with the new installer. It's certainly an improvement over the old one! Thanks to the folks who worked on it. -- R. Kevin Oberman, Network Engineer - Retired E-mail: kob6...@gmail.com ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: High Network Perfomance
Hi Luigi, Thanks for reply. That's great solution :) Will be integrated with pf or is it? Thanks, Victor On Thu, Aug 4, 2011 at 9:27 PM, Luigi Rizzo wrote: > On Thu, Aug 04, 2011 at 08:39:17PM -0300, Victor Detoni wrote: > > Hi Guys, > > > > I'm trying tunning a FreeBSD 8.2 to high perfomance network with pf. My > > server configuration is: > > > > Dell 1950 > > CPU: Intel(R) Xeon(R) CPU5130 @ 2.00GHz (1995.03-MHz > K8-class > > CPU) > > 4 x CPU > > 2 NIC ( > 1 NIC (em0: ) > > > > I want to reach the high processing of packets per second and use pf as > > synproxy and we still processor to handle others packets or flows. > > > > I know that em drivers has MULTI_QUEUE implementation that helps high > > performance for Intel drivers, but I couldn't see more information about. > We > > can reached 500k pps, but no more traffic was processed by this > interface. > > > > I've already enabled net.isr.direct but with Intel Drivers does not work > and > > the most processors are in System instead of Interrupts, why? When I > enable > > net.isr.direct the processing is balanced for on CPU in system and > another > > in interrupt and I reached 1M pps, but the total perfomance is down, the > > load grow up too fast. > > > > I've changed some parameter in sysctl for intel drivers, but it doesn't > have > > effect. > > > > Someone know what I can do to reach more packets performance? I want to > use > > this FreeBSD as a router/firewall only. > > if you feel like doing a bit of coding yourself, you could try netmap > > http://info.iet.unipi.it/~luigi/netmap/ > > cheers > luigi > > > Thanks, > > Victor > > ___ > > freebsd-current@freebsd.org mailing list > > http://lists.freebsd.org/mailman/listinfo/freebsd-current > > To unsubscribe, send any mail to " > freebsd-current-unsubscr...@freebsd.org" > ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
9-BETA1 Fatal trap 9 in __mnt_vnode_next
I am running FreeBSD 9-Beta1 (csup from July 31) and I had a Fatal trap 9 in __mnt_vnode_next: (transcribed from a picture) Fatal trap 9: general protection fault while in kernel mode cpuid = 0; apic id =00 instruction pointer = 0x20:0x8086d790 stack pointer = 0x20:0xff8340be0a40 frame pointer = 0x28:0xff8340be0a60 code segment= base 0x0, limit 0xf, type 0x1b = DPL 0, pres 1, long 1, def32 0, gran 1 processor eflags= interrupt enabled, resume, IOPL = 0 current process = 18 (syncer) [ thread pid 18 tid 100077 ] Stopped at __mnt_vnode_next+0xc0: movq %rax,0x30(%rdx) db> bt Tracing pid 18 tid 100077 td 0xfe00090378c0 __mnt_vnode_next() at __mnt_vnode_next+0xc0 vfs_msync() at vfs_msync+0x1c7 sync_fsync() at sync_fsync+0x13f VOP_FSYNC_APV() at VOP_FSYN_APV+0x49 sync_vnode() at sync_vnode+0x156 sched_sync() at sched_sync+0x270 fork_exit() at fork_exit+0x120 fork_trampoline() at fork_trampoline+0xe --- trap 0, rip = 0, rsp = 0xff8340be0d00, rbp = 0 --- db> I have compiled the world and kernel myself with /etc/malloc.conf -> aj, make.conf is: WRKDIRPREFIX=/usr/obj CFLAGS= -O2 -fno-strict-aliasing -pipe CPUTYPE?=athlon64 COPTFLAGS= -O -pipe # added by use.perl 2011-08-01 19:47:48 PERL_VERSION=5.12.4 I removed the WITNESS and INVARIANT options from the kernel config. /boot/loader.conf: geom_eli_load="YES" geom_mirror_load="YES" zfs_load="YES" vm.kmem_size=12G I have the root filesystem on ZFS, and swap in a mirrored GELI. At the moment of the crash I also had a UFS filesystem mounted. Hopefully this is useful to someone. Best, Edwin ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
ahci(4) breaks on MSI X58 Pro-E
Hi all, The day before yesterday I upgraded a server of mine at Hetzner to FreeBSD 9. It seems to use an MSI X58 Pro-E motherboard, having an ICH10 SATA-controller. It seems that under high load, I get timeouts on the SATA controller. It seems disabling MSI has no effect. Unfortunately it's a sort-of production system, so I can't do a lot of testing with it and I also don't have any hands-on console access. Is there anyone out there who has a system at home that also has an X58 motherboard and can test latest HEAD to see whether it is a recurring problem? I've attached the output of dmesg for now. Disabling the ahci driver and using ata seems to work. -- Ed Schouten WWW: http://80386.nl/ Copyright (c) 1992-2011 The FreeBSD Project. Copyright (c) 1979, 1980, 1983, 1986, 1988, 1989, 1991, 1992, 1993, 1994 The Regents of the University of California. All rights reserved. FreeBSD is a registered trademark of The FreeBSD Foundation. FreeBSD 9.0-BETA1 #0 r+4f5f78a: Wed Aug 3 10:42:02 CEST 2011 e...@hertog.hoeg.nl:/usr/obj/export/pub/src/sys/HERTOG amd64 CPU: Intel(R) Core(TM) i7 CPU 920 @ 2.67GHz (2673.36-MHz K8-class CPU) Origin = "GenuineIntel" Id = 0x106a5 Family = 6 Model = 1a Stepping = 5 Features=0xbfebfbff Features2=0x98e3bd AMD Features=0x28100800 AMD Features2=0x1 TSC: P-state invariant, performance statistics real memory = 8594128896 (8196 MB) avail memory = 8233840640 (7852 MB) Event timer "LAPIC" quality 400 ACPI APIC Table: <7522MS A7522800> FreeBSD/SMP: Multiprocessor System Detected: 8 CPUs FreeBSD/SMP: 1 package(s) x 4 core(s) x 2 SMT threads cpu0 (BSP): APIC ID: 0 cpu1 (AP): APIC ID: 1 cpu2 (AP): APIC ID: 2 cpu3 (AP): APIC ID: 3 cpu4 (AP): APIC ID: 4 cpu5 (AP): APIC ID: 5 cpu6 (AP): APIC ID: 6 cpu7 (AP): APIC ID: 7 ioapic0 irqs 0-23 on motherboard acpi0: <7522MS A7522800> on motherboard acpi0: Power Button (fixed) acpi0: reservation of 0, a (3) failed acpi0: reservation of 10, bff0 (3) failed Timecounter "ACPI-fast" frequency 3579545 Hz quality 900 acpi_timer0: <24-bit timer at 3.579545MHz> port 0x808-0x80b on acpi0 cpu0: on acpi0 ACPI Warning: Incorrect checksum in table [OEMB] - 0xD0, should be 0xCF (20110527/tbutils-282) cpu1: on acpi0 cpu2: on acpi0 cpu3: on acpi0 cpu4: on acpi0 cpu5: on acpi0 cpu6: on acpi0 cpu7: on acpi0 pcib0: port 0xcf8-0xcff on acpi0 pci0: on pcib0 pcib1: at device 1.0 on pci0 pci1: on pcib1 pcib2: at device 3.0 on pci0 pci2: on pcib2 vgapci0: port 0xcc00-0xcc7f mem 0xfa00-0xfaff,0xd000-0xdfff,0xf800-0xf9ff irq 16 at device 0.0 on pci2 pcib3: at device 7.0 on pci0 pci3: on pcib3 pci0: at device 20.0 (no driver attached) pci0: at device 20.1 (no driver attached) pci0: at device 20.2 (no driver attached) pci0: at device 20.3 (no driver attached) pci0: at device 26.0 (no driver attached) pci0: at device 26.1 (no driver attached) pci0: at device 26.2 (no driver attached) pci0: at device 26.7 (no driver attached) pcib4: irq 17 at device 28.0 on pci0 pci4: on pcib4 pcib5: irq 17 at device 28.4 on pci0 pci6: on pcib5 re0: port 0xe800-0xe8ff mem 0xfbeff000-0xfbef,0xf6ff-0xf6ff irq 16 at device 0.0 on pci6 re0: Using 1 MSI-X message re0: Chip rev. 0x3c00 re0: MAC rev. 0x0040 miibus0: on re0 rgephy0: PHY 1 on miibus0 rgephy0: none, 10baseT, 10baseT-FDX, 10baseT-FDX-flow, 100baseTX, 100baseTX-FDX, 100baseTX-FDX-flow, 1000baseT, 1000baseT-master, 1000baseT-FDX, 1000baseT-FDX-master, 1000baseT-FDX-flow, 1000baseT-FDX-flow-master, auto, auto-flow re0: Ethernet address: 40:61:86:be:cd:da pci0: at device 29.0 (no driver attached) pci0: at device 29.1 (no driver attached) pci0: at device 29.2 (no driver attached) pci0: at device 29.7 (no driver attached) pcib6: at device 30.0 on pci0 pci7: on pcib6 isab0: at device 31.0 on pci0 isa0: on isab0 ahci0: port 0xb000-0xb007,0xac00-0xac03,0xa880-0xa887,0xa800-0xa803,0xa480-0xa49f mem 0xf7ffa000-0xf7ffa7ff irq 19 at device 31.2 on pci0 ahci0: AHCI v1.20 with 6 3Gbps ports, Port Multiplier not supported ahcich0: at channel 0 on ahci0 ahcich1: at channel 1 on ahci0 ahcich2: at channel 2 on ahci0 ahcich3: at channel 3 on ahci0 ahcich4: at channel 4 on ahci0 ahcich5: at channel 5 on ahci0 pci0: at device 31.3 (no driver attached) acpi_button0: on acpi0 attimer0: port 0x40-0x43 irq 0 on acpi0 Timecounter "i8254" frequency 1193182 Hz quality 0 Event timer "i8254" frequency 1193182 Hz quality 100 atrtc0: port 0x70-0x71 irq 8 on acpi0 Event timer "RTC" frequency 32768 Hz quality 0 uart0: <16550 or compatible> port 0x3f8-0x3ff irq 4 flags 0x10 on acpi0 atkbdc0: port 0x60,0x64 irq 1 on acpi0 atkbd0: irq 1 on atkbdc0 atkbd0: [GIANT-LOCKED] hpet0: iomem 0xfed0-0xfed003ff on acpi0 Timecounter "HPET" frequency 14318180 Hz quality 950 Event timer "HPET" frequency 14318180 Hz quality 450 Event timer "HPET1" frequency 14318180 Hz quality 440 Event timer "HPET2" frequency
Re: graid fresh install
On 8/5/2011 12:01 PM, Mike Tancsa wrote: > > and mount it. However, speeds seem a bit slow, but I am not sure if > thats to be expected. > > i5# dd if=/dev/zero of=/mnt/test bs=1024k count=100 > 100+0 records in > 100+0 records out > 104857600 bytes transferred in 13.996080 secs (7491926 bytes/sec) > i5# One extra bit of oddness with graid, is that if I disable the RAID in the BIOS and boot up the disks just with plain old AHCI, graid still works. Is that a quirk of the BIOS, or does it generally work that way? ---Mike -- --- Mike Tancsa, tel +1 519 651 3400 Sentex Communications, m...@sentex.net Providing Internet services since 1994 www.sentex.net Cambridge, Ontario Canada http://www.tancsa.com/ ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: Bug: devfs is sure to have the bug.
On Fri, Aug 05, 2011 at 06:45:22PM +0300, Jaakko Heinonen wrote:
> On 2011-08-03, Kostik Belousov wrote:
> > On Wed, Aug 03, 2011 at 02:44:23PM +0900, Kohji Okuno wrote:
> > > > devfs_populate(), and the context holds only "dm->dm_lock" in
> > > > devfs_populate().
> > > >
> > > > On the other hand, "devfs_generation" is incremented in devfs_create()
> > > > and devfs_destroy() the context holds only "devmtx" in devfs_create()
> > > > and devfs_destroy().
> > > >
> > > > If a context executes devfs_create() when other context is executing
> > > > (***), then "dm->dm_generation" is updated incorrect value.
> > > > As a result, we can not open the last detected device (we receive
> > > > ENOENT).
> >
> > I think the problem you described is real, and suggested change is right.
> > Initially, I thought that we should work with devfs_generation as with
> > the atomic type due to unlocked access in the devfs_populate(), but then
> > convinced myself that this is not needed.
> >
> > But also, I think there is another half of the problem. Namely,
> > devfs_lookup() calls devfs_populate_vp(), and then does lookup with the
> > help of devfs_lookupx(). We will miss the generation update
> > happen after the drop of the dm_lock in devfs_populate_vp() to reacquire
> > the directory vnode lock.
>
> I don't understand this. devfs_generation is not protected with dm_lock
> in devfs_create() and devfs_destroy(). On the other hand if you mean
> that another thread calls devfs_populate() while we drop dm_lock in
> devfs_populate_vp(), isn't the mount point up to date when we re-lock
> dm_lock?
Yes, I was not quite exact in describing what I mean, and the reference
to dm_lock drop is both vague and not correct.
I am after the fact that we do allow the situation where it is externally
visible that new cdev node was successfully created before the lookup
returns ENOENT for the path of the node.
>
> > @@ -630,13 +630,15 @@ devfs_populate_loop(struct devfs_mount *dm, int
> > cleanup)
> > void
> > devfs_populate(struct devfs_mount *dm)
> > {
> > + unsigned gen;
> >
> > sx_assert(&dm->dm_lock, SX_XLOCKED);
> > - if (dm->dm_generation == devfs_generation)
> > + gen = devfs_generation;
> > + if (dm->dm_generation == gen)
> > return;
> > while (devfs_populate_loop(dm, 0))
> > continue;
> > - dm->dm_generation = devfs_generation;
> > + dm->dm_generation = gen;
> > }
>
> After this change dm->dm_generation may be stale although the mount
> point is up to date? This is probably harmless, though.
This must be harmless, in the worst case it will cause more calls to
the populate. In fact, this even allows the dm_generation to roll backward,
which is again harmless.
pgph7QBq5GQ8j.pgp
Description: PGP signature
Re: sata port multiplier
2011/8/5 Mike Tancsa > On 8/4/2011 9:59 PM, Nenhum_de_Nos wrote: > > hail, > > > > any info on what port multiplier I could buy to make the 4 port Sil3124 > at > > least make up to 8 ? > > speaking about PM, does anybody already tried new usb3 pm ? ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: High Network Perfomance
Oh, wow! Great news! It's much more than I expected. Unfortunately I am too far from the Bay Area. on 05/08/2011 19:12 Luigi Rizzo said the following: > Its the latter. The libpcap-over-netmap code has been there for > a while, but documentation was not completely up to date so > we are both at fault. > > I just updated the webpage with links to another paper reporting > performance of the libpcap emulation library, Click and OpenvSwitch. > > There is also a new set of slides for a talk i am going to give at > various locations in the Bay Area next week (if you are around, > email me privately for details) > > Our libpcap makes netmap immediately available for basically > all pcap clients (sure, you might need a bit of tinkering, > and some apps could have their own speed issues -- see our study > on Click and OpenvSwitch). > Click userspace now runs (on FreeBSD+netmap) as fast or better than > the in-kernel linux version. Which means that all research > and prototypes that were bound to Linux because of this > reason, now could consider switching platform. > > Note that having netmap does not prevent the existing stack from > working. I have designed the system in a way that allows > incremental improvement/replacement of the components. > > I am working on bringing outside the kernel ipfw+dummynet > (relatively straightforward, i did it already once a couple of years ago), > and then the routing tables. > > netgraph could be another candidate (though i dont know > how much work it involves; netgraph is very similar to Click, > and the latter might have a richer set of elements). > > For TCP (and high speed TCP) i don't have a clear view on > what are the bottlenecks, but with 1500-byte MTU your pps rates > are 20 times lower, so the problem is entirely different and > saving 400ns per packet does help, but not as much as when > packets arrive every 70ns. > > cheers > luigi > -+--- > Prof. Luigi RIZZO, ri...@iet.unipi.it . Dip. di Ing. dell'Informazione > http://www.iet.unipi.it/~luigi/. Universita` di Pisa > TEL +39-050-2211611 . via Diotisalvi 2 > Mobile +39-338-6809875 . 56122 PISA (Italy) > -+--- -- Andriy Gapon ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: 9.0-BETA1 installer issues
On Wed, Aug 03, 2011 at 09:03:22PM +0200, Marc Fonvieille wrote: > On Wed, Aug 03, 2011 at 08:28:34PM +0200, Marc Fonvieille wrote: > > On Tue, Aug 02, 2011 at 08:36:01AM -0500, Nathan Whitehorn wrote: > > > On 08/02/11 04:41, Bruce Cran wrote: > > > > I've been trying out 9.0-BETA1: it's a lot easier to install than > > > > previous releases with bsdinstall, but I spotted a few issues: > > > > > > Good! Thanks for checking. > > > > > > > Typo - "Resovler Configuration". > > > > If I leave the resolver window for a while it gets corrupted with: > > > > > > > > Aug 2 10:31:23 dhclient[973]: Bogus domain search list 15: lan, . > > > > > > Interesting. It looks like DHCP doesn't like your local setup... > > > > > > > In the documentation installation screen, it should say "At a > > > > minimum..." - the 'a' is missing. Also, there should perhaps be a > > > > semi-colon between "English version" and "this is the original". The > > > > menu also doesn't appear to do anything once you select "OK". > > > > > > The spelling fixes are easy to fix. The documentation issue is more > > > confusing. It should begin running pkg_add, after you press OK, assuming > > > you selected something. Do you have the installer log handy? It will be > > > in /tmp. > > > > > [...] > > > > Hmm I think it's "default" PACKAGESITE env variable pointing on > > non-existing > > ftp://ftp.freebsd.org/pub/FreeBSD/ports//packages-9-beta1/Latest/ > > > > I'm wrong, I did an install and same behavior as Bruce. > I looked in /tmp/bsdinstall_log: > > Running installation step: docsintall > pkg_add: unable to fetch > 'ftp://ftp.freebsd.org/pub/FreeBSD/ports/amd64/packages-9-current/Latest/en-freebsd-doc.tbz' > by URL ^^ I haven't looked at bsdinstall lately, but IIRC there's no dialog that offers a mirror selection? It would be nice to select a nearby server. pgpKnltL7Ojjf.pgp Description: PGP signature
Re: Bug: devfs is sure to have the bug.
On 2011-08-03, Kostik Belousov wrote:
> On Wed, Aug 03, 2011 at 02:44:23PM +0900, Kohji Okuno wrote:
> > > devfs_populate(), and the context holds only "dm->dm_lock" in
> > > devfs_populate().
> > >
> > > On the other hand, "devfs_generation" is incremented in devfs_create()
> > > and devfs_destroy() the context holds only "devmtx" in devfs_create()
> > > and devfs_destroy().
> > >
> > > If a context executes devfs_create() when other context is executing
> > > (***), then "dm->dm_generation" is updated incorrect value.
> > > As a result, we can not open the last detected device (we receive ENOENT).
>
> I think the problem you described is real, and suggested change is right.
> Initially, I thought that we should work with devfs_generation as with
> the atomic type due to unlocked access in the devfs_populate(), but then
> convinced myself that this is not needed.
>
> But also, I think there is another half of the problem. Namely,
> devfs_lookup() calls devfs_populate_vp(), and then does lookup with the
> help of devfs_lookupx(). We will miss the generation update
> happen after the drop of the dm_lock in devfs_populate_vp() to reacquire
> the directory vnode lock.
I don't understand this. devfs_generation is not protected with dm_lock
in devfs_create() and devfs_destroy(). On the other hand if you mean
that another thread calls devfs_populate() while we drop dm_lock in
devfs_populate_vp(), isn't the mount point up to date when we re-lock
dm_lock?
> @@ -630,13 +630,15 @@ devfs_populate_loop(struct devfs_mount *dm, int cleanup)
> void
> devfs_populate(struct devfs_mount *dm)
> {
> + unsigned gen;
>
> sx_assert(&dm->dm_lock, SX_XLOCKED);
> - if (dm->dm_generation == devfs_generation)
> + gen = devfs_generation;
> + if (dm->dm_generation == gen)
> return;
> while (devfs_populate_loop(dm, 0))
> continue;
> - dm->dm_generation = devfs_generation;
> + dm->dm_generation = gen;
> }
After this change dm->dm_generation may be stale although the mount
point is up to date? This is probably harmless, though.
--
Jaakko
___
freebsd-current@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-current
To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
graid fresh install
I netbooted 9.0 Beta and was going to try and install the system onto a raid1 mirror built with graid. However, neither sysinstall nor bsdinstall see it fully I have the box netbooted (Intel Matrix raid) i5# graid status -ags Intel-a5bbf622 OPTIMAL ada0 (ACTIVE (ACTIVE)) Intel-a5bbf622 OPTIMAL ada1 (ACTIVE (ACTIVE)) i5# I can newfs /dev/raid/r0 i5# ls -l /dev/raid/r0 crw-r- 1 root operator0, 98 Aug 5 10:43 /dev/raid/r0 i5# and mount it. However, speeds seem a bit slow, but I am not sure if thats to be expected. i5# dd if=/dev/zero of=/mnt/test bs=1024k count=100 100+0 records in 100+0 records out 104857600 bytes transferred in 13.996080 secs (7491926 bytes/sec) i5# i5# iostat -c 1000 ada0 ada1 ttyada0 ada1 cpu tin tout KB/t tps MB/s KB/t tps MB/s us ni sy in id 2 204 124.47 5 0.57 124.28 5 0.57 0 0 0 0 100 0 186 128.00 60 7.49 128.00 62 7.74 0 0 0 0 100 064 128.00 55 6.87 128.00 50 6.24 0 0 0 1 99 063 128.00 60 7.49 128.00 64 7.99 0 0 0 1 99 063 128.00 40 4.99 128.00 40 4.99 0 0 0 1 99 063 126.43 61 7.52 128.00 59 7.37 0 0 0 0 99 063 128.00 45 5.62 128.00 43 5.37 0 0 0 1 99 063 128.00 50 6.24 128.00 51 6.37 0 0 0 0 100 064 128.00 50 6.24 128.00 49 6.12 0 0 0 1 99 063 128.00 80 9.99 128.00 81 10.11 0 0 0 0 100 064 128.00 70 8.74 128.00 71 8.87 0 0 0 0 99 But when I run sysintall, it does not find any disks. If I run bsdinstall, it sees ada0 and ada1, but I cannot partition and label the raid partition. Am I supposed to just write to ada0 ? What is the best way to do it Guided just shows ada0 and ada1 The partition editor does show md0 4.0 MBmalloc ada0 465 GB raid/r0 465 GB ada1 465 GB raid/r0 465 GB But I cannot label any of the r0 'disks' It comes up with Invalid argument. arg0 'Intel-a5bbf622' ada0 at ahcich0 bus 0 scbus0 target 0 lun 0 ada0: ATA-8 SATA 3.x device ada0: 300.000MB/s transfers (SATA 2.x, UDMA6, PIO 8192bytes) ada0: Command Queueing enabled ada0: 476940MB (976773168 512 byte sectors: 16H 63S/T 16383C) ada0: Previously was known as ad4 ada1 at ahcich1 bus 0 scbus1 target 0 lun 0 ada1: ATA-8 SATA 3.x device ada1: 300.000MB/s transfers (SATA 2.x, UDMA6, PIO 8192bytes) ada1: Command Queueing enabled ada1: 476940MB (976773168 512 byte sectors: 16H 63S/T 16383C) ada1: Previously was known as ad6 ahci0: port 0x3068-0x306f,0x3074-0x3077,0x3060-0x3067,0x3070-0x3073,0x3020-0x303f mem 0xb262-0xb26207ff irq 18 at device 31.2 o n pci0 ahci0: AHCI v1.30 with 6 3Gbps ports, Port Multiplier not supported ahcich0: at channel 0 on ahci0 ahcich1: at channel 1 on ahci0 ahcich2: at channel 2 on ahci0 ahcich3: at channel 3 on ahci0 ahcich4: at channel 4 on ahci0 ahcich5: at channel 5 on ahci0 ---Mike -- --- Mike Tancsa, tel +1 519 651 3400 Sentex Communications, m...@sentex.net Providing Internet services since 1994 www.sentex.net Cambridge, Ontario Canada http://www.tancsa.com/ ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: High Network Perfomance
On Fri, Aug 05, 2011 at 06:34:54PM +0300, Andriy Gapon wrote: > on 05/08/2011 18:23 Vlad Galu said the following: > > Netmap's scope may be narrow, but it's a great alternative to > > proprietary implementations that are provided by only a handful of > > vendors and tied to their hardware. I don't think Luigi foresees lots > > of users either, but having it in the base system is a lot better than > > not having it at all, IMHO. > > No doubt. > > On the other hand, it was a little bit misleading of Luigi to suggest netmap > to a > user who merely asked about tuning FreeBSD network (routing/firewall) > performance :-) Sure, i took the chance to advertise the product a bit :) But the original poster essentially wanted to "go as fast as i can" and any tweaks that one can do on the system won't bring him much further than the 500Kpps he is seeing. Perhaps 1Mpps. Not too exciting. > netmap can no doubt be useful, but perhaps let's not overhype it before it > actually proves itself on practical tasks. Or has it already? I might have > missed that. Its the latter. The libpcap-over-netmap code has been there for a while, but documentation was not completely up to date so we are both at fault. I just updated the webpage with links to another paper reporting performance of the libpcap emulation library, Click and OpenvSwitch. There is also a new set of slides for a talk i am going to give at various locations in the Bay Area next week (if you are around, email me privately for details) Our libpcap makes netmap immediately available for basically all pcap clients (sure, you might need a bit of tinkering, and some apps could have their own speed issues -- see our study on Click and OpenvSwitch). Click userspace now runs (on FreeBSD+netmap) as fast or better than the in-kernel linux version. Which means that all research and prototypes that were bound to Linux because of this reason, now could consider switching platform. Note that having netmap does not prevent the existing stack from working. I have designed the system in a way that allows incremental improvement/replacement of the components. I am working on bringing outside the kernel ipfw+dummynet (relatively straightforward, i did it already once a couple of years ago), and then the routing tables. netgraph could be another candidate (though i dont know how much work it involves; netgraph is very similar to Click, and the latter might have a richer set of elements). For TCP (and high speed TCP) i don't have a clear view on what are the bottlenecks, but with 1500-byte MTU your pps rates are 20 times lower, so the problem is entirely different and saving 400ns per packet does help, but not as much as when packets arrive every 70ns. cheers luigi -+--- Prof. Luigi RIZZO, ri...@iet.unipi.it . Dip. di Ing. dell'Informazione http://www.iet.unipi.it/~luigi/. Universita` di Pisa TEL +39-050-2211611 . via Diotisalvi 2 Mobile +39-338-6809875 . 56122 PISA (Italy) -+--- ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: fdescfs mount causes hard lock
On Friday 05 August 2011 11:41 am, Jung-uk Kim wrote: > On Friday 05 August 2011 04:33 am, Ian FREISLICH wrote: > > Hi > > > > Since CURRENT Aug 2, my machine has hardlocked at boot time on > > "Mounting local filesystems". I traced it this morning to > > > > fdesc /dev/fd fdescfs rw 0 0 > > > > Not mounting this stopped the hardlock. I added it because one > > installed port requeried it, but I can't remember which port that > > was. > > > > Is this related to the panic recently reported by David > > Wolfskill? > > Yes, reverting r224614 fixed the hang for me. Alternatively, > please try the last patch posted in the thread. Actually, this patch was committed as r224655, it seems. :-) Jung-uk Kim ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: High Network Perfomance
on 05/08/2011 18:40 Vlad Galu said the following: > On Fri, Aug 5, 2011 at 5:34 PM, Andriy Gapon wrote: >> on 05/08/2011 18:23 Vlad Galu said the following: >>> Netmap's scope may be narrow, but it's a great alternative to >>> proprietary implementations that are provided by only a handful of >>> vendors and tied to their hardware. I don't think Luigi foresees lots >>> of users either, but having it in the base system is a lot better than >>> not having it at all, IMHO. >> >> No doubt. >> >> On the other hand, it was a little bit misleading of Luigi to suggest netmap >> to a >> user who merely asked about tuning FreeBSD network (routing/firewall) >> performance :-) >> >> netmap can no doubt be useful, but perhaps let's not overhype it before it >> actually proves itself on practical tasks. Or has it already? I might have >> missed that. > > It's great for IDS/IPS, which is exactly what I'm using it for :) Oh, yes, haven't thought about this. -- Andriy Gapon ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: fdescfs mount causes hard lock
On Friday 05 August 2011 04:33 am, Ian FREISLICH wrote: > Hi > > Since CURRENT Aug 2, my machine has hardlocked at boot time on > "Mounting local filesystems". I traced it this morning to > > fdesc /dev/fd fdescfs rw 0 0 > > Not mounting this stopped the hardlock. I added it because one > installed port requeried it, but I can't remember which port that > was. > > Is this related to the panic recently reported by David Wolfskill? Yes, reverting r224614 fixed the hang for me. Alternatively, please try the last patch posted in the thread. Jung-uk Kim ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: High Network Perfomance
On Fri, Aug 5, 2011 at 5:34 PM, Andriy Gapon wrote: > on 05/08/2011 18:23 Vlad Galu said the following: >> Netmap's scope may be narrow, but it's a great alternative to >> proprietary implementations that are provided by only a handful of >> vendors and tied to their hardware. I don't think Luigi foresees lots >> of users either, but having it in the base system is a lot better than >> not having it at all, IMHO. > > No doubt. > > On the other hand, it was a little bit misleading of Luigi to suggest netmap > to a > user who merely asked about tuning FreeBSD network (routing/firewall) > performance :-) > > netmap can no doubt be useful, but perhaps let's not overhype it before it > actually proves itself on practical tasks. Or has it already? I might have > missed that. It's great for IDS/IPS, which is exactly what I'm using it for :) > > -- > Andriy Gapon > -- Good, fast & cheap. Pick any two. ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: High Network Perfomance
on 05/08/2011 18:23 Vlad Galu said the following: > Netmap's scope may be narrow, but it's a great alternative to > proprietary implementations that are provided by only a handful of > vendors and tied to their hardware. I don't think Luigi foresees lots > of users either, but having it in the base system is a lot better than > not having it at all, IMHO. No doubt. On the other hand, it was a little bit misleading of Luigi to suggest netmap to a user who merely asked about tuning FreeBSD network (routing/firewall) performance :-) netmap can no doubt be useful, but perhaps let's not overhype it before it actually proves itself on practical tasks. Or has it already? I might have missed that. -- Andriy Gapon ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: High Network Perfomance
On Fri, Aug 5, 2011 at 5:03 PM, Andriy Gapon wrote: > on 05/08/2011 17:56 Luigi Rizzo said the following: >> i asked permission to re@ to integrate it in 9.0 but it was considered >> a bit premature. >> However i am not too worried because the system dependencies are >> minimal and it changes no API/ABI or internal data structure so >> it is easy to add it at a later time. > > /rant > > I think that having real, useful in practice applications that make use of the > netmap would definitely speed up its adoption. > > Maybe I am too pessimistic here, but I don't foresee too many users of netmap > as > long as it remains just a mechanism that potentially can greatly speed up > things > if you manage to write your own applications that do those things via netmap. > Full TCP/IP stack with sockets API on top of it and lots of available > applications > on top of that is one thing, an interface to a network card is a totally > different > thing on a scale of usability (especially the by the end-users). Netmap's scope may be narrow, but it's a great alternative to proprietary implementations that are provided by only a handful of vendors and tied to their hardware. I don't think Luigi foresees lots of users either, but having it in the base system is a lot better than not having it at all, IMHO. > > -- > Andriy Gapon > ___ > freebsd-current@freebsd.org mailing list > http://lists.freebsd.org/mailman/listinfo/freebsd-current > To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org" > -- Good, fast & cheap. Pick any two. ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: High Network Perfomance
On Friday, August 05, 2011 02:27:33 AM Luigi Rizzo wrote: > if you feel like doing a bit of coding yourself, you could try netmap > > http://info.iet.unipi.it/~luigi/netmap/ Out of curiosity, will this code be integrated in FreeBSD? Because it seems really useful and a lot of programs might benefit from this. Regards, Pieter ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: High Network Perfomance
on 05/08/2011 17:56 Luigi Rizzo said the following: > i asked permission to re@ to integrate it in 9.0 but it was considered > a bit premature. > However i am not too worried because the system dependencies are > minimal and it changes no API/ABI or internal data structure so > it is easy to add it at a later time. /rant I think that having real, useful in practice applications that make use of the netmap would definitely speed up its adoption. Maybe I am too pessimistic here, but I don't foresee too many users of netmap as long as it remains just a mechanism that potentially can greatly speed up things if you manage to write your own applications that do those things via netmap. Full TCP/IP stack with sockets API on top of it and lots of available applications on top of that is one thing, an interface to a network card is a totally different thing on a scale of usability (especially the by the end-users). -- Andriy Gapon ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: High Network Perfomance
On Fri, Aug 05, 2011 at 04:28:19PM +0200, Pieter de Goeje wrote: > On Friday, August 05, 2011 02:27:33 AM Luigi Rizzo wrote: > > if you feel like doing a bit of coding yourself, you could try netmap > > > > http://info.iet.unipi.it/~luigi/netmap/ > > Out of curiosity, will this code be integrated in FreeBSD? Because it seems > really useful and a lot of programs might benefit from this. i asked permission to re@ to integrate it in 9.0 but it was considered a bit premature. However i am not too worried because the system dependencies are minimal and it changes no API/ABI or internal data structure so it is easy to add it at a later time. cheers luigi ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: gptzfsboot error using HP Smart Array P410i Controller
Hello Everyone,
Despite the BIOS information about the nonexistent floppy, the zfsboot.c code
will prevent to boot from the first HDD if a floppy is given as a first
available device.
The drive 0x0 (floppy) will be probed before the code below and an error occurs:
[…]
gptzfsboot: error 1 lba 32
gptzfsboot: error 1 lba 1
[…]
The continue statement will skip the rest of the iteration because
if ((i | DRV_HARD) == *(uint8_t *)PTOV(ARGS))
is true if the drive equals 0x80. As a result we do not call probe_drive()
for this drive.
Eliminating
if ((i | DRV_HARD) == *(uint8_t *)PTOV(ARGS))
continue;
would help.
Any comments will be appreciated.
Best Regards,
Christoph
i386/zfsboot/zfsboot.c
int
main(void)
{
[…]
/*
* Probe the rest of the drives that the bios knows about. This
* will find any other available pools and it may fill in missing
* vdevs for the boot pool.
*/
for (i = 0; i < *(unsigned char *)PTOV(BIOS_NUMDRIVES); i++) {
if ((i | DRV_HARD) == *(uint8_t *)PTOV(ARGS))
continue;
if (!int13probe(i | DRV_HARD))
break;
[…]
probe_drive(dsk, NULL);
}
[…]
On Aug 5, 2011, at 1:47 PM, Christoph Hoffmann wrote:
> Hello John,
>
> No, I and not using clang.
>
> My problem persists even I apply the patch.
>
> As a workaround I have to put OS on second LUN presented by the
> P410i Controller.
>
> Regards,
>
> Christoph
>
> --
> Christoph Hoffmann
>
>
> On Aug 5, 2011, at 1:37 PM, John Baldwin wrote:
>
>> On Thursday, August 04, 2011 3:26:49 pm Christoph Hoffmann wrote:
>>> Hello Everyone,
>>>
>>> The system will successfully boot only if the OS installation is laying on
>>> the second drive or higher (0x81 and more).
>>
>> Are you using clang? If so, you should try either using GCC or using this
>> patch with clang as a workaround from the previous thread on zfsboot issues:
>>
>> Index: sys/boot/i386/zfsboot/Makefile
>> ===
>> --- sys/boot/i386/zfsboot/Makefile (revision 224653)
>> +++ sys/boot/i386/zfsboot/Makefile (working copy)
>> @@ -20,7 +20,6 @@
>> -fomit-frame-pointer \
>> -fno-unit-at-a-time \
>> -mno-align-long-strings \
>> - -mrtd \
>> -DBOOT2 \
>> -DSIOPRT=${BOOT_COMCONSOLE_PORT} \
>> -DSIOFMT=${B2SIOFMT} \
>> Index: sys/boot/i386/gptzfsboot/Makefile
>> ===
>> --- sys/boot/i386/gptzfsboot/Makefile (revision 224653)
>> +++ sys/boot/i386/gptzfsboot/Makefile (working copy)
>> @@ -22,7 +22,6 @@
>> -fomit-frame-pointer \
>> -fno-unit-at-a-time \
>> -mno-align-long-strings \
>> - -mrtd \
>> -DGPT -DBOOT2 \
>> -DSIOPRT=${BOOT_COMCONSOLE_PORT} \
>> -DSIOFMT=${B2SIOFMT} \
>>
>> --
>> John Baldwin
>> ___
>> freebsd-current@freebsd.org mailing list
>> http://lists.freebsd.org/mailman/listinfo/freebsd-current
>> To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
>
> ___
> freebsd-current@freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-current
> To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
___
freebsd-current@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-current
To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: sata port multiplier
On 8/4/2011 9:59 PM, Nenhum_de_Nos wrote: > hail, > > any info on what port multiplier I could buy to make the 4 port Sil3124 at > least make up to 8 ? I have used a number of PMs from http://www.addonics.com/. 8.x works great. eg. # camcontrol devlist at scbus0 target 0 lun 0 (pass0,ada0) at scbus0 target 1 lun 0 (pass1,ada1) at scbus0 target 2 lun 0 (pass2,ada2) at scbus0 target 3 lun 0 (pass3,ada3) at scbus0 target 15 lun 0 (pass4,pmp0) at scbus1 target 0 lun 0 (pass5,ada4) at scbus1 target 1 lun 0 (pass6,ada5) at scbus1 target 2 lun 0 (pass7,ada6) at scbus1 target 3 lun 0 (pass8,ada7) at scbus1 target 4 lun 0 (pass9,ada8) at scbus1 target 15 lun 0 (pass10,pmp1) ---Mike -- --- Mike Tancsa, tel +1 519 651 3400 Sentex Communications, m...@sentex.net Providing Internet services since 1994 www.sentex.net Cambridge, Ontario Canada http://www.tancsa.com/ ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: panic: share -> excl @r224632
On Fri, Aug 05, 2011 at 10:43:32AM +0200, Martin Matuska wrote: > Patch updated. > ... OK; after reverting kern/vfs_mount.c to its pre-patched status on my laptop, I applied this most recent patch without incident; it built & boots fine, which is now running: FreeBSD g1-222.catwhisker.org. 9.0-BETA1 FreeBSD 9.0-BETA1 #305 r224654M: Fri Aug 5 04:57:16 PDT 2011 r...@g1-222.catwhisker.org.:/usr/obj/usr/src/sys/CANARY i386 I also applied it to my build machine (which had not encountered problems with the code as originally committed); building was uneventful, as was booting; it is now running: FreeBSD freebeast.catwhisker.org 9.0-BETA1 FreeBSD 9.0-BETA1 #575 r224654M: Fri Aug 5 05:05:08 PDT 2011 r...@freebeast.catwhisker.org:/usr/obj/usr/src/sys/GENERIC i386 Peace, david -- David H. Wolfskill da...@catwhisker.org Depriving a girl or boy of an opportunity for education is evil. See http://www.catwhisker.org/~david/publickey.gpg for my public key. pgpALOZaCA83G.pgp Description: PGP signature
Re: High Network Perfomance
Victor Detoni wrote: > Hi Guys, > > I'm trying tunning a FreeBSD 8.2 to high perfomance network with pf. My > server configuration is: > > Dell 1950 > CPU: Intel(R) Xeon(R) CPU5130 @ 2.00GHz (1995.03-MHz K8-class > CPU) > 4 x CPU > 2 NIC ( 1 NIC (em0: ) > > I want to reach the high processing of packets per second and use pf as > synproxy and we still processor to handle others packets or flows. Benchmarking I did a few years ago showed a strong correlation between forwarding rate and CPU L1 cache size. As well as an inverse relationship to the number of CPUs in the system. At the time, some architectures were worse than others. Intel Pentium4/Xeons had a halving and AMD Opteron/Athlon had about a 7% reduction in forwarding rate with SMP compared to UP. I haven't had the chance to re-run these tests recently. Set net.inet.ip.fastforwarding=1 and run benchmarks to test your forwarding rates with different configurations. See for some results: http://docs.freebsd.org/cgi/getmsg.cgi?fetch=77846+0+archive/2008/freebsd-net/20080120.freebsd-net Ian -- Ian Freislich ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: gptzfsboot error using HP Smart Array P410i Controller
Hello John,
No, I and not using clang.
My problem persists even I apply the patch.
As a workaround I have to put OS on second LUN presented by the
P410i Controller.
Regards,
Christoph
--
Christoph Hoffmann
On Aug 5, 2011, at 1:37 PM, John Baldwin wrote:
> On Thursday, August 04, 2011 3:26:49 pm Christoph Hoffmann wrote:
>> Hello Everyone,
>>
>> The system will successfully boot only if the OS installation is laying on
>> the second drive or higher (0x81 and more).
>
> Are you using clang? If so, you should try either using GCC or using this
> patch with clang as a workaround from the previous thread on zfsboot issues:
>
> Index: sys/boot/i386/zfsboot/Makefile
> ===
> --- sys/boot/i386/zfsboot/Makefile (revision 224653)
> +++ sys/boot/i386/zfsboot/Makefile (working copy)
> @@ -20,7 +20,6 @@
>-fomit-frame-pointer \
>-fno-unit-at-a-time \
>-mno-align-long-strings \
> - -mrtd \
>-DBOOT2 \
>-DSIOPRT=${BOOT_COMCONSOLE_PORT} \
>-DSIOFMT=${B2SIOFMT} \
> Index: sys/boot/i386/gptzfsboot/Makefile
> ===
> --- sys/boot/i386/gptzfsboot/Makefile (revision 224653)
> +++ sys/boot/i386/gptzfsboot/Makefile (working copy)
> @@ -22,7 +22,6 @@
>-fomit-frame-pointer \
>-fno-unit-at-a-time \
>-mno-align-long-strings \
> - -mrtd \
>-DGPT -DBOOT2 \
>-DSIOPRT=${BOOT_COMCONSOLE_PORT} \
>-DSIOFMT=${B2SIOFMT} \
>
> --
> John Baldwin
> ___
> freebsd-current@freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-current
> To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
___
freebsd-current@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-current
To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: gptzfsboot error using HP Smart Array P410i Controller
On Thursday, August 04, 2011 3:26:49 pm Christoph Hoffmann wrote:
> Hello Everyone,
>
> The system will successfully boot only if the OS installation is laying on
> the second drive or higher (0x81 and more).
Are you using clang? If so, you should try either using GCC or using this
patch with clang as a workaround from the previous thread on zfsboot issues:
Index: sys/boot/i386/zfsboot/Makefile
===
--- sys/boot/i386/zfsboot/Makefile (revision 224653)
+++ sys/boot/i386/zfsboot/Makefile (working copy)
@@ -20,7 +20,6 @@
-fomit-frame-pointer \
-fno-unit-at-a-time \
-mno-align-long-strings \
- -mrtd \
-DBOOT2 \
-DSIOPRT=${BOOT_COMCONSOLE_PORT} \
-DSIOFMT=${B2SIOFMT} \
Index: sys/boot/i386/gptzfsboot/Makefile
===
--- sys/boot/i386/gptzfsboot/Makefile (revision 224653)
+++ sys/boot/i386/gptzfsboot/Makefile (working copy)
@@ -22,7 +22,6 @@
-fomit-frame-pointer \
-fno-unit-at-a-time \
-mno-align-long-strings \
- -mrtd \
-DGPT -DBOOT2 \
-DSIOPRT=${BOOT_COMCONSOLE_PORT} \
-DSIOFMT=${B2SIOFMT} \
--
John Baldwin
___
freebsd-current@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-current
To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: fdescfs mount causes hard lock
Am 05.08.2011 10:33 (UTC+1) schrieb Ian FREISLICH: Hi Since CURRENT Aug 2, my machine has hardlocked at boot time on "Mounting local filesystems". I traced it this morning to fdesc /dev/fd fdescfs rw 0 0 Not mounting this stopped the hardlock. I added it because one installed port requeried it, but I can't remember which port that was. I am observing the same since 02/08/2011. Without mounting fdescfs the system does not freeze anymore. The reason for using fdescfs was java/openjdk, if I remember right. Rainer Hurling Is this related to the panic recently reported by David Wolfskill? Ian ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: [clang] (gpt)zfsboot is broken: zfs_alloc()/zfs_free() mismatch
On 2011-08-05 02:17, Pawel Worach wrote: ... A workaround for the hang on boot and "error 1 lba X" failures is the following patch, it would be interesting if it also makes the zfs_alloc/free error go away too. ... dim@ has filed an upstream bug, http://llvm.org/bugs/show_bug.cgi?id=10591 This PR is about another issue, which I discovered during debugging of the gptzfsboot problem. The actual problem when compiling (gpt)zfsboot with clang and -mrtd seems to be that clang screws something up with the function calling conventions, causing crashes and other weird things. I have not yet determined the root cause of this. ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: [clang] (gpt)zfsboot is broken: zfs_alloc()/zfs_free() mismatch
On 2011-08-05 11:45, Olivier Smedts wrote: ... A workaround for the hang on boot and "error 1 lba X" failures is the following patch, it would be interesting if it also makes the zfs_alloc/free error go away too. Yes, it solved the "ZFS: zfs_alloc()/zfs_free() mismatch" error when gptzfsboot-ing after building with clang. I don't know if it still works with gcc and the patch applied. Yeah, that should not be a problem. The -mrtd option causes the compiled code to shrink a little, but that is not essential for the ZFS boot loaders. ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: resolvconf script overwrites entries in resolv.conf - RDNSS/DNSSL related
On Thu, 04 Aug 2011 23:52:54 -0700, Doug Barton wrote: > > On 08/04/2011 22:59, Mattia Rossi wrote: > > I've finally patched my 8.2 IPv6 gateway with the RDNSS/DNSSL patches > > The script anyhow overwrites my previous manual entries in > > /etc/resolv.conf which I need for my manual IPv4 setup... > > > Check 'man resolvconf.conf' for information on name_servers_append. It > would probably be nice if there was a _prepend equivalent. > Mattia, when you say you have the patches, which ones? To be clear, the RDNSS/DNSSL support that was committed to head was very heavily modified from that which I proposed and which is on my web site. In particular, the resolvconf(8) tool that I offered was not used at all; the version in head is the openresolv tool by Roy Marples. Doug's response is in regard to the resolvconf version in head. FWIW, the resolvconf version from my site will use the most recent nameservers received by from either dhclient-script or rtadvd but you can also add "static" entries using standard resolv.conf syntax in the /var/db/resolv.db file; see the man page with that version. I will update my RDNSS/DNSSL web page now to add a warning that my patches have been superseded and that folk should use the committed versions where possible. -jr signature.asc Description: PGP signature
Re: [clang] (gpt)zfsboot is broken: zfs_alloc()/zfs_free() mismatch
Hello Olivier,
Thank you very for your work.
Using the path I have successfully rebuild, install and boot my system.
Regards,
Christoph
On Aug 5, 2011, at 11:45 AM, Olivier Smedts wrote:
> 2011/8/5 Pawel Worach :
>> On Aug 1, 2011, at 14:24, Test Rat wrote:
>>
>>> Anyone else? I can still reproduce with trunk r136607.
>>> boot and gptboot seem to be unaffected.
>>>
>>> IIRC, with previous clang import it just stuck during boot
>>> without any error messages.
>>
>> A workaround for the hang on boot and "error 1 lba X" failures is the
>> following patch, it would be interesting if it also makes the zfs_alloc/free
>> error go away too.
>
> Yes, it solved the "ZFS: zfs_alloc()/zfs_free() mismatch" error when
> gptzfsboot-ing after building with clang. I don't know if it still
> works with gcc and the patch applied.
>
>> Index: sys/boot/i386/zfsboot/Makefile
>> ===
>> --- sys/boot/i386/zfsboot/Makefile (revision 224653)
>> +++ sys/boot/i386/zfsboot/Makefile (working copy)
>> @@ -20,7 +20,6 @@
>>-fomit-frame-pointer \
>>-fno-unit-at-a-time \
>>-mno-align-long-strings \
>> - -mrtd \
>>-DBOOT2 \
>>-DSIOPRT=${BOOT_COMCONSOLE_PORT} \
>>-DSIOFMT=${B2SIOFMT} \
>> Index: sys/boot/i386/gptzfsboot/Makefile
>> ===
>> --- sys/boot/i386/gptzfsboot/Makefile (revision 224653)
>> +++ sys/boot/i386/gptzfsboot/Makefile (working copy)
>> @@ -22,7 +22,6 @@
>>-fomit-frame-pointer \
>>-fno-unit-at-a-time \
>>-mno-align-long-strings \
>> - -mrtd \
>>-DGPT -DBOOT2 \
>>-DSIOPRT=${BOOT_COMCONSOLE_PORT} \
>>-DSIOFMT=${B2SIOFMT} \
>>
>> dim@ has filed an upstream bug, http://llvm.org/bugs/show_bug.cgi?id=10591
>>
>> Regards
>> --
>> Pawel
>>
>> ___
>> freebsd-current@freebsd.org mailing list
>> http://lists.freebsd.org/mailman/listinfo/freebsd-current
>> To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
>>
>
>
>
> --
> Olivier Smedts _
> ASCII ribbon campaign ( )
> e-mail: oliv...@gid0.org- against HTML email & vCards X
> www: http://www.gid0.org- against proprietary attachments / \
>
> "Il y a seulement 10 sortes de gens dans le monde :
> ceux qui comprennent le binaire,
> et ceux qui ne le comprennent pas."
> ___
> freebsd-current@freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-current
> To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
___
freebsd-current@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-current
To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: [clang] (gpt)zfsboot is broken: zfs_alloc()/zfs_free() mismatch
2011/8/5 Pawel Worach :
> On Aug 1, 2011, at 14:24, Test Rat wrote:
>
>> Anyone else? I can still reproduce with trunk r136607.
>> boot and gptboot seem to be unaffected.
>>
>> IIRC, with previous clang import it just stuck during boot
>> without any error messages.
>
> A workaround for the hang on boot and "error 1 lba X" failures is the
> following patch, it would be interesting if it also makes the zfs_alloc/free
> error go away too.
Yes, it solved the "ZFS: zfs_alloc()/zfs_free() mismatch" error when
gptzfsboot-ing after building with clang. I don't know if it still
works with gcc and the patch applied.
> Index: sys/boot/i386/zfsboot/Makefile
> ===
> --- sys/boot/i386/zfsboot/Makefile (revision 224653)
> +++ sys/boot/i386/zfsboot/Makefile (working copy)
> @@ -20,7 +20,6 @@
> -fomit-frame-pointer \
> -fno-unit-at-a-time \
> -mno-align-long-strings \
> - -mrtd \
> -DBOOT2 \
> -DSIOPRT=${BOOT_COMCONSOLE_PORT} \
> -DSIOFMT=${B2SIOFMT} \
> Index: sys/boot/i386/gptzfsboot/Makefile
> ===
> --- sys/boot/i386/gptzfsboot/Makefile (revision 224653)
> +++ sys/boot/i386/gptzfsboot/Makefile (working copy)
> @@ -22,7 +22,6 @@
> -fomit-frame-pointer \
> -fno-unit-at-a-time \
> -mno-align-long-strings \
> - -mrtd \
> -DGPT -DBOOT2 \
> -DSIOPRT=${BOOT_COMCONSOLE_PORT} \
> -DSIOFMT=${B2SIOFMT} \
>
> dim@ has filed an upstream bug, http://llvm.org/bugs/show_bug.cgi?id=10591
>
> Regards
> --
> Pawel
>
> ___
> freebsd-current@freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-current
> To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
>
--
Olivier Smedts _
ASCII ribbon campaign ( )
e-mail: oliv...@gid0.org - against HTML email & vCards X
www: http://www.gid0.org - against proprietary attachments / \
"Il y a seulement 10 sortes de gens dans le monde :
ceux qui comprennent le binaire,
et ceux qui ne le comprennent pas."
___
freebsd-current@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-current
To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: panic: share -> excl @r224632
Patch updated.
On 05.08.2011 10:26, Kostik Belousov wrote:
> On Fri, Aug 05, 2011 at 10:18:49AM +0200, Martin Matuska wrote:
>> I agree to Kostik's approach, but I suggest implementing it in a
>> separate function and also use for the unmount() part.
>>
>> Please review attached patch.
> Since you are moving the fragment to a function, you may somewhat reduce
> the code duplication by moving at least free() and return to the end
> of function and jumping to it.
>
> Also, after more looking at this, I think now that the check for VI_DOOMED
> is not needed, due to relookup and comparision of vp and vp1.
--
Martin Matuska
FreeBSD committer
http://blog.vx.sk
Index: sys/kern/vfs_mount.c
===
--- sys/kern/vfs_mount.c (revision 224654)
+++ sys/kern/vfs_mount.c (working copy)
@@ -362,6 +362,58 @@
}
/*
+ * Verify vnode's global path
+ */
+static int
+vfs_verify_global_path(struct thread *td, struct vnode *vp, char *fspath)
+{
+ struct nameidata nd;
+ struct vnode *vp1;
+ char *rpath, *fbuf;
+ int error;
+
+ ASSERT_VOP_ELOCKED(vp, __func__);
+
+ /* Construct global filesystem path from vp. */
+ VOP_UNLOCK(vp, 0);
+ error = vn_fullpath_global(td, vp, &rpath, &fbuf);
+ if (error != 0) {
+ vrele(vp);
+ return (error);
+ }
+ if (strlen(rpath) >= MNAMELEN) {
+ vrele(vp);
+ error = ENAMETOOLONG;
+ goto out;
+ }
+
+ /*
+ * Re-lookup the vnode by path. As a side effect, the vnode is
+ * relocked. If vnode was renamed, return ENOENT.
+ */
+ NDINIT(&nd, LOOKUP, FOLLOW | LOCKLEAF | MPSAFE | AUDITVNODE1,
+ UIO_SYSSPACE, fspath, td);
+ error = namei(&nd);
+ if (error != 0) {
+ vrele(vp);
+ goto out;
+ }
+ NDFREE(&nd, NDF_ONLY_PNBUF);
+ vp1 = nd.ni_vp;
+ vrele(vp);
+ if (vp1 != vp) {
+ vput(vp1);
+ error = ENOENT;
+ goto out;
+ }
+
+ strlcpy(fspath,rpath,MNAMELEN);
+out:
+ free(fbuf, M_TEMP);
+ return (error);
+}
+
+/*
* Mount a filesystem.
*/
int
@@ -745,6 +797,7 @@
vfs_domount_first(
struct thread *td, /* Calling thread. */
struct vfsconf *vfsp, /* File system type. */
+ char *fspath, /* Mount path. */
struct vnode *vp, /* Vnode to be covered. */
int fsflags, /* Flags common to all filesystems. */
struct vfsoptlist **optlist /* Options local to the filesystem. */
@@ -753,25 +806,12 @@
struct vattr va;
struct mount *mp;
struct vnode *newdp;
- char *fspath, *fbuf;
int error;
mtx_assert(&Giant, MA_OWNED);
ASSERT_VOP_ELOCKED(vp, __func__);
KASSERT((fsflags & MNT_UPDATE) == 0, ("MNT_UPDATE shouldn't be here"));
- /* Construct global filesystem path from vp. */
- error = vn_fullpath_global(td, vp, &fspath, &fbuf);
- if (error != 0) {
- vput(vp);
- return (error);
- }
- if (strlen(fspath) >= MNAMELEN) {
- vput(vp);
- free(fbuf, M_TEMP);
- return (ENAMETOOLONG);
- }
-
/*
* If the user is not root, ensure that they own the directory
* onto which we are attempting to mount.
@@ -793,14 +833,12 @@
}
if (error != 0) {
vput(vp);
- free(fbuf, M_TEMP);
return (error);
}
VOP_UNLOCK(vp, 0);
/* Allocate and initialize the filesystem. */
mp = vfs_mount_alloc(vp, vfsp, fspath, td->td_ucred);
- free(fbuf, M_TEMP);
/* XXXMAC: pass to vfs_mount_alloc? */
mp->mnt_optnew = *optlist;
/* Set the mount level flags. */
@@ -1083,15 +1121,15 @@
mtx_lock(&Giant);
NDFREE(&nd, NDF_ONLY_PNBUF);
vp = nd.ni_vp;
- if ((fsflags & MNT_UPDATE) == 0)
- error = vfs_domount_first(td, vfsp, vp, fsflags, optlist);
- else
+ if ((fsflags & MNT_UPDATE) == 0) {
+ error = vfs_verify_global_path(td, vp, fspath);
+ if (error == 0)
+ error = vfs_domount_first(td, vfsp, fspath, vp,
+ fsflags, optlist);
+ } else
error = vfs_domount_update(td, vp, fsflags, optlist);
mtx_unlock(&Giant);
- ASSERT_VI_UNLOCKED(vp, __func__);
- ASSERT_VOP_UNLOCKED(vp, __func__);
-
return (error);
}
@@ -1118,7 +1156,7 @@
{
struct mount *mp;
struct nameidata nd;
- char *pathbuf, *rpathbuf, *fbuf;
+ char *pathbuf;
int error, id0, id1;
AUDIT_ARG_VALUE(uap->flags);
@@ -1164,15 +1202,10 @@
UIO_SYSSPACE, pathbuf, td);
if (namei(&nd) == 0) {
NDFREE(&nd, NDF_ONLY_PNBUF);
-if (vn_fullpath_global(td, nd.ni_vp, &rpathbuf,
-&fbuf) == 0) {
- if (strlen(rpathbuf) < MNAMELEN) {
- strlcpy(pathbuf, rpathbuf,
- MNAMELEN);
- }
- free(fbuf, M_TEMP);
-}
-vput(nd.ni_vp);
+error = vfs_verify_global_path(td, nd.ni_vp,
+pathbuf);
+if (error == 0)
+ vput(nd.ni_vp);
}
}
mtx_lock(&mountlist_mtx);
___
freebsd-current@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-current
To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
fdescfs mount causes hard lock
Hi Since CURRENT Aug 2, my machine has hardlocked at boot time on "Mounting local filesystems". I traced it this morning to fdesc /dev/fd fdescfs rw 0 0 Not mounting this stopped the hardlock. I added it because one installed port requeried it, but I can't remember which port that was. Is this related to the panic recently reported by David Wolfskill? Ian -- Ian Freislich ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: panic: share -> excl @r224632
On Fri, Aug 05, 2011 at 10:18:49AM +0200, Martin Matuska wrote: > I agree to Kostik's approach, but I suggest implementing it in a > separate function and also use for the unmount() part. > > Please review attached patch. Since you are moving the fragment to a function, you may somewhat reduce the code duplication by moving at least free() and return to the end of function and jumping to it. Also, after more looking at this, I think now that the check for VI_DOOMED is not needed, due to relookup and comparision of vp and vp1. pgpnolFGBZp3Z.pgp Description: PGP signature
Re: panic: share -> excl @r224632
I agree to Kostik's approach, but I suggest implementing it in a
separate function and also use for the unmount() part.
Please review attached patch.
On 04.08.2011 15:59, Kostik Belousov wrote:
> On Thu, Aug 04, 2011 at 05:54:54AM -0700, David Wolfskill wrote:
>> I've only seen this on my laptop; my build machine doesn't exhibit the
>> panic.
>>
>> r224602 is the most recent point I've built that does not exhibit the
>> panic at all.
>>
>> The first few lines (hand-transcribed; I have no serial console on this
>> laptop -- the one shortcoming it has):
>>
>> shared lock of (lockmgr) devfs @/usr/src/sys/kern/vfs_cache.c:1116
>> while exclusively locked from /usr/src/sys/kern/vfs_subr.c:2134
>> panic: share -> excl
>>
>> The backtrace (which I wasn't patient enough to trto transcribe twice;
>> sorry) appeared to involve "nmount", and the panic occurred directly
>> after mounting the file systems during the transition from single-user
>> mode to multi-user mode.
>>
>> The output from "svn update" going from r224602 -> r224632 shows the
>> following files being updated:
>>
>> Uusr.sbin/faithd/faithd.8
>> Uusr.sbin/jail/jail.8
>> Ulib/libproc/proc_create.c
>> Usys/arm/arm/irq_dispatch.S
>> Usys/arm/sa11x0/sa11x0_irq.S
>> Usys/powerpc/booke/locore.S
>> Usys/powerpc/booke/platform_bare.c
>> Usys/powerpc/booke/pmap.c
>> Usys/nfsclient/nfsnode.h
>> Usys/nfsclient/nfs_node.c
>> Usys/kern/kern_jail.c
>> Usys/kern/vfs_mount.c
>> Usys/fs/nfsclient/nfsnode.h
>> Usys/fs/nfsclient/nfs_clnode.c
>> Usys/mips/mips/exception.S
>> Usys/dev/ahci/ahci.c
>> Usys/dev/ata/chipsets/ata-nvidia.c
>> Usys/dev/ath/ath_hal/ah_eeprom_v4k.c
>> Usys/i386/ibcs2/imgact_coff.c
>> Usys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_vnops.c
>> Updated to revision 224632.
>>
>> Updating to r224648 (this morning) has had no apparent effect on the
>> cited panic: it occurs in the same way, at the same time, accompanied by
>> the same messages (citing the same files and lines).
>>
>> I've attached dmesg.boot from r224602.
>>
>> I will see if I can find a commit that affected at least one of the
>> affected files in the above list that I can revert to avoid the panic,
>> but I'm a bit slow for a while yet, so I figured I'd finally get around
>> to posting this, in the hope that someone cleverer than me will spot
>> the problem and suggest a circumvention to try.
>>
>> And I'm quite willing to try such things.
>>
>> Note: this is FreeBSD/i386; nothing special: no jails; running on real
>> hardware; no attempts to use ZFS
> I am sure that this is caused by r224614.
> I forgot that vn_fullpath cannot operate on the locked vnode.
>
> Try this.
>
> diff --git a/sys/kern/vfs_mount.c b/sys/kern/vfs_mount.c
> index d601c56..54f123c 100644
> --- a/sys/kern/vfs_mount.c
> +++ b/sys/kern/vfs_mount.c
> @@ -746,13 +746,15 @@ vfs_domount_first(
> struct thread *td, /* Calling thread. */
> struct vfsconf *vfsp, /* File system type. */
> struct vnode *vp, /* Vnode to be covered. */
> + char *ufspath,
> int fsflags,/* Flags common to all filesystems. */
> struct vfsoptlist **optlist /* Options local to the filesystem. */
> )
> {
> struct vattr va;
> + struct nameidata nd;
> struct mount *mp;
> - struct vnode *newdp;
> + struct vnode *newdp, *vp1;
> char *fspath, *fbuf;
> int error;
>
> @@ -761,16 +763,45 @@ vfs_domount_first(
> KASSERT((fsflags & MNT_UPDATE) == 0, ("MNT_UPDATE shouldn't be here"));
>
> /* Construct global filesystem path from vp. */
> + VOP_UNLOCK(vp, 0);
> error = vn_fullpath_global(td, vp, &fspath, &fbuf);
> if (error != 0) {
> - vput(vp);
> + vrele(vp);
> return (error);
> }
> if (strlen(fspath) >= MNAMELEN) {
> - vput(vp);
> + vrele(vp);
> free(fbuf, M_TEMP);
> return (ENAMETOOLONG);
> }
> + if ((vp->v_iflag & VI_DOOMED) != 0) {
> + vrele(vp);
> + free(fbuf, M_TEMP);
> + return (EBADF);
> + }
> +
> + /*
> + * Re-lookup the vnode by path. As a side effect, the vnode is
> + * relocked. If vnode was renamed, return ENOENT.
> + */
> + NDINIT(&nd, LOOKUP, FOLLOW | LOCKLEAF | MPSAFE | AUDITVNODE1,
> + UIO_SYSSPACE, ufspath, td);
> + error = namei(&nd);
> + if (error != 0) {
> + vrele(vp);
> + free(fbuf, M_TEMP);
> + return (error);
> + }
> + if (NDHASGIANT(&nd))
> + mtx_unlock(&Giant);
> + NDFREE(&nd, NDF_ONLY_PNBUF);
> + vp1 = nd.ni_vp;
> + vrele(vp);
> + if (vp1 != vp) {
> + vput(vp1);
> + free(fbuf, M_TEMP);
> + return (ENOENT);
> + }
>
> /*
>* If the user is not root, ensure that they
