fedora 14 kernel performance with ip forwarding workload
by Jesse Brandeburg
The other day I was running the stock fedora kernel on my ip
forwarding setup, to see what the performance was, and the performance
wasn't very good.
system is S5520HC dual socket 2.93GHz Xeon 5570 (Nehalem) with 3 quad
port 82580 adapters (12 ports). Traffic is bidirectional 64 byte
packets being forwarded and received on each port, basically port to
port routing. I am only using 12 flows currently.
The driver is igb, and I am using an affinity script that lines up
each pair of ports that are forwarding traffic into optimal
configurations for cache locality. I am also disabling
remote_node_defrag_ratio to stop cross node traffic.
With the fedora default kernel from F14 it appears that
CONFIG_NETFILTER=y means that I cannot unload all of netfilter even if
I stop iptables service.
perf showed netfilter being prominent, and removing it gives me much
higher throughput. Is there a reason CONFIG_NETFILTER=y ? Isn't it a
good thing to be able to disable netfilter if you want to?
Jesse
8 years, 9 months
[PATCHSET] utrace for 3.1 kernel
by Oleg Nesterov
Hello.
utrace patches for 3.1 kernel. Untested, will try to do some tests
tomorrow.
I do not want to spam you all and the lists, please look at
git://git.kernel.org/pub/scm/linux/kernel/git/oleg/misc.git utrace-3.1
Oleg Nesterov (29):
utrace: add utrace_init_task/utrace_free_task calls
tracehooks: add utrace hooks
tracehooks: reintroduce tracehook_consider_fatal_signal()
add utrace hooks into sig_ignored() and recalc_sigpending()
restore the EXEC/EXIT/CLONE utrace hooks
utrace: utrace_report_death() can use task_utrace_struct()
restore the DEATH/REAP utrace hooks
utrace: remove jobctl bits
ptrace: take ->siglock around s/TRACED/RUNNING/
introduce wake_up_quiescent()
introduce ptrace_signal_wake_up()
wait_task_inactive: treat task->state and match_state as bitmasks
introduce TASK_UTRACED state
utrace: use TASK_UTRACED instead of TASK_TRACED
reintroduce tracehook_finish_jctl() as utrace_end_stop()
teach wake_up_quiescent() to do "selective" wake_up
ptrace_stop: do not assume the task is running after wake_up_quiescent()
get_signal_to_deliver: restore/restructure utrace/ptrace signal reporting
utrace_get_signal: s/JOBCTL_STOP_PENDING/JOBCTL_PENDING_MASK/
introduce ptrace_set_syscall_trace()
introduce PT_SYSCALL_TRACE flag
utrace: don't clear TIF_SYSCALL_TRACE if it is needed by ptrace
introduce task_utrace_lock/task_utrace_unlock
teach ptrace_set_syscall_trace() to play well with utrace
introduce PT_SINGLE_STEP and PT_SINGLE_BLOCK
utrace: finish_resume_report: don't do user_xxx_step() if ptrace_wants_step()
ptrace: shift user_*_step() from ptrace_resume() to ptrace_stop()
ptrace_disable: no need to disable stepping
ptrace_report_syscall: check TIF_SYSCALL_EMU
Roland McGrath (1):
utrace core
Documentation/DocBook/Makefile | 2 +-
Documentation/DocBook/utrace.tmpl | 589 +++++++++
arch/s390/kernel/traps.c | 4 +-
arch/x86/kernel/ptrace.c | 1 -
fs/exec.c | 5 +-
fs/proc/array.c | 14 +-
include/linux/ptrace.h | 7 +
include/linux/sched.h | 25 +-
include/linux/signal.h | 2 +
include/linux/tracehook.h | 53 +-
include/linux/utrace.h | 773 ++++++++++++
init/Kconfig | 9 +
kernel/Makefile | 1 +
kernel/exit.c | 5 +
kernel/fork.c | 9 +
kernel/ptrace.c | 57 +-
kernel/sched.c | 2 +-
kernel/signal.c | 97 ++-
kernel/utrace.c | 2461 +++++++++++++++++++++++++++++++++++++
19 files changed, 4062 insertions(+), 54 deletions(-)
create mode 100644 Documentation/DocBook/utrace.tmpl
create mode 100644 include/linux/utrace.h
create mode 100644 kernel/utrace.c
12 years, 1 month
[F15 PATCH 0/6] ARM: Add support for pata_imx (for Genesi Smarttop)
by Niels de Vos
Hi there!
This patchset contains only cherry-picked patches from Linus' current
tree. Without these patches, the Genesi Smarttop does not recognize the
onboard SSD-chip. The required driver (pata_imx) is already included in
the ARM verion of the kernel, only the bindings with the board are
missing.
Instead of sending all patches in separate emails, I'm attaching them
numbered and all for easy inclusion in the F15 kernel. I have done
local testing and can happily report that the 2.6.41.6-1 kernel can use
the onboard SSD when appying these patches.
I (and other Fedora ARM owners of a Smarttop) appreciate it if these
patches can be included in a next release. Many thanks,
Niels
---
Arnaud Patard (4):
imx51: add pata device
imx51: add pata clock
imx: efika: Enable pata.
Fix pata imx resource
Fabio Estevam (1):
ARM: imx: Add PATA resources for other i.MX processors
Uwe Kleine-König (1):
ARM: mx5: fix remaining inconsistent names for irqs
arch/arm/mach-mx5/Kconfig | 1 +
arch/arm/mach-mx5/clock-mx51-mx53.c | 7 +-
arch/arm/mach-mx5/devices-imx51.h | 4 +
arch/arm/mach-mx5/devices.c | 10 +-
arch/arm/mach-mx5/mm.c | 8 +-
arch/arm/mach-mx5/mx51_efika.c | 2 +
arch/arm/plat-mxc/devices/Kconfig | 3 +
arch/arm/plat-mxc/devices/Makefile | 1 +
arch/arm/plat-mxc/devices/platform-pata_imx.c | 59 ++++++++
arch/arm/plat-mxc/include/mach/devices-common.h | 8 +
arch/arm/plat-mxc/include/mach/mx35.h | 2 +-
arch/arm/plat-mxc/include/mach/mx51.h | 162 +++++++++++-----------
12 files changed, 175 insertions(+), 92 deletions(-)
create mode 100644 arch/arm/plat-mxc/devices/platform-pata_imx.c
12 years, 2 months
Re: yum and kernel-modules-extra
by Josh Boyer
On Mon, Dec 26, 2011 at 12:32 PM, Thorsten Leemhuis
<Thorsten_Leemhuis(a)gmx.de> wrote:
> On 26.12.2011 18:01, Josh Boyer wrote:
>> On Mon, Dec 26, 2011 at 11:11 AM, Bruno Wolff III <bruno(a)wolff.to> wrote:
>>> I submitted an RFE (bug 770444) to include kernel-modules-extra in the
>>> default list of install only packages for yum.
>> The entire point of kernel-modules-extra is to get them out of the
>> default install. If there are modules in that subpackage that are
>> needed by default, then they should be moved back to the main package. [...]
>
> That's afaics not what Bruno proposes. What he afaics means is: when a
> new kernel gets released, than the package kernel-modules-extra (just
> like the kernel itself) needs to be installed (like a "rpm -i" would do)
> and not updated (rpm -U), otherwise you loose the modules for current or
> older kernels on "yum update" every time a new kernel gets released.
Ah, indeed. My apologies Bruno, I misread "install only" as
"install". Your RFE sounds good indeed.
josh
12 years, 2 months
yum and kernel-modules-extra
by Bruno Wolff III
I submitted an RFE (bug 770444) to include kernel-modules-extra in the
default list of install only packages for yum.
12 years, 2 months
grubby fatal error: unable to find a suitable template
by Steve Dickson
Hello all,
I'm getting the $SUBJECT error when I try to
install a rawhide kernel on an updated f16 vm.
# rpm -ihv kernel-3.2.0-0.rc6.git0.1.pnfsdlexp.fc16.x86_64.rpm
Preparing... ########################################### [100%]
1:kernel ########################################### [100%]
grubby fatal error: unable to find a suitable template
# uname -r
3.1.5-6.fc16.x86_64
Things like the network and NFS do not come up due the fact the modules can not
be found... so things are not being installed correctly....
Google searching the error I see a number bugzillas that are being blissfully
ignored...
Just curious if anybody else has seen this error and what they did about it...
steved.
12 years, 3 months
iwlagn probs with 3.1.5
by Genes MailLists
Has anyone else noticed wifi problems (via iwlagn) on 3.1.5?
After a wake from sleep today my lenovo laptop could not connect to
wifi. It had connected fine on a different wifi.
It got stuck in 'authenticating' phase - if the supplicant logs are a
valid indicator.
Booting back to 3.1.4 got it going again immediately.
I did not see any out of the ordinary messages.t
I know there been quite a bit of noise about iwlagn driver on LKML -
but I thought most of the patches had gone into 3.1.5 ...
gene
12 years, 3 months
Re: [kernel] This should never be =y
by Kyle McMartin
On Mon, Dec 19, 2011 at 09:48:06PM +0000, Dave Jones wrote:
> commit f76e0239ae4715d34b39397046b51825fb3fa47c
> Author: Dave Jones <davej(a)redhat.com>
> Date: Mon Dec 19 16:47:56 2011 -0500
>
> This should never be =y
>
Looks like nobody noticed when it changed from a bool to a tristate in
2006...
^1da177e (Linus Torvalds 2005-04-16 15:20:36 -0700 31) config
DEV_APPLETALK
9cac2c35 (David S. Miller 2006-07-30 20:20:54 -0700 32) tristate
"Appletalk interfaces support"
^1da177e (Linus Torvalds 2005-04-16 15:20:36 -0700 33) depends
on ATALK
^1da177e (Linus Torvalds 2005-04-16 15:20:36 -0700 34) help
^1da177e (Linus Torvalds 2005-04-16 15:20:36 -0700 35)
AppleTalk is the protocol that Apple computers can use to communicate
^1da177e (Linus Torvalds 2005-04-16 15:20:36 -0700 36) on a
network. If your Linux box is connected to such a network, and wish
^1da177e (Linus Torvalds 2005-04-16 15:20:36 -0700 37) to do
IP over it, or you have a LocalTalk card and wish to use it to
^1da177e (Linus Torvalds 2005-04-16 15:20:36 -0700 38)
connect to the AppleTalk network, say Y.
This isn't the first time this has happened...
--Kyle
12 years, 3 months
RE: Route Metric
by Alexander Bahoor
Sorry for the broadcast, but don't see a group with networking and wireless.
Can someone please recommend a list of wireless client adapters that support fedora or linux in general? I checked a few vendor, but did not see download for them.
Rgrds,
Alex
12 years, 3 months
Disable Intel IOMMU by default in kernel-3.1.5-7
by Joshua C.
I saw in the change log for the kernel-3.1.5-7 that the Intel IOMMU
has been disabled by default. Can someone explain this decision or
point me to the right discussion on the internet?
--joshua
12 years, 3 months