on some boxes we can see a high load, but i have
no idea, what the cause of this:
Mem: 27740K used, 1264K free, 0K shrd, 2160K buff, 9592K cached
CPU: 0% usr 0% sys 0% nic 97% idle 0% io 0% irq 1% sirq
Load average: 1.03 1.02 1.05 1/40 9974
PID PPID USER STAT VSZ %VSZ %CPU
There is one process/thread that is in a block state but uses no CPU. I got
similar behavior with a blocked syscall waiting for a SAN lun disk recover.
Probably your process has a blocked syscall. Is there any kernel oops dmesg?
Try to kill the remaining process. A blocked process will not die
Second look. Check if khubd keeps state D (uninterruptible sleep). Maybe a
USB device caused it.
Em 24/04/2014 04:42, Luiz Angelo Daros de Luca luizl...@gmail.com
escreveu:
There is one process/thread that is in a block state but uses no CPU. I got
similar behavior with a blocked syscall waiting
Applied, thanks. Will get into OpenWrt with next netifd-bump.
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
Adding an interface in the interface list in case of a config update will
override the node version of the old interface and thus overriding the dynamic
interface marker.
In case of config update interface_set_dynamic is done on an invalid interface
pointer as the new interface pointer has been
No need to install policy routes if the prefix address cannot be installed;
clean up of the policy routes
and the prefix address only needs to happen when the prefix assignment was
active.
Signed-off-by: Hans Dedecker dedec...@gmail.com
---
interface-ip.c |4 ++--
1 files changed, 2
Gives the user the control to select the correct WAN IPv4 address to be used by
the 6rd tunnel when mutiple WAN interfaces are configured
Signed-off-by: Hans Dedecker dedec...@gmail.com
---
package/network/ipv6/6rd/Makefile |2 +-
package/network/ipv6/6rd/files/6rd.sh |9 +++--
* Bastian Bittorf bitt...@bluebottle.com [24.04.2014 09:07]:
thank you 'Luiz Angelo Daros de Luca',
the reason was khubd / a USB-issue. i did nothing, just waited
and it was repaired automatically. according to dmesg, there
was something USB-related, but without user-action:
Ahh, recentlly I deployed Barrier Breaker on a friend's exactlly same
router model and he has said it was getting very similar behavior (high
loads without any apparently reason). We tried to revert back to AA
which I am not sure if clear the issues (I guess it would if it's a
kernel issue).
Hello,
talking about:
https://dev.openwrt.org/ticket/12181
patch has been tested @ Ninux.org on TP-Link WDR3600. We are now able
to successfully have tagged and untagged frames together on the same
port.
Here is the clean version of the patch rebased on current trunk:
On 24/04/2014 12:44, Saverio Proto wrote:
Hello,
talking about: https://dev.openwrt.org/ticket/12181
patch has been tested @ Ninux.org on TP-Link WDR3600. We are now
able to successfully have tagged and untagged frames together on
the same port.
Here is the clean version of the patch
Hello,
the patch is attached to this email.
thanks
Saverio
2014-04-24 12:47 GMT+02:00 John Crispin j...@phrozen.org:
On 24/04/2014 12:44, Saverio Proto wrote:
Hello,
talking about: https://dev.openwrt.org/ticket/12181
patch has been tested @ Ninux.org on TP-Link WDR3600. We are now
2014-04-23 21:52 GMT+02:00, Rafał Miłecki zaj...@gmail.com:
2014-04-23 13:41 GMT+02:00 José Vázquez ppvazquez...@gmail.com:
I don't know if any of the OpenWRT developers or contributors have
this router. If yes, my opinion is to add support for the board using
the patches sent by Matthew
Couldn't agree more. Very good email.
On 23/04/2014 18:33, cmsv wrote:
Although it would be good to have this hardware supported aren't we all
forgetting something such as BETTER alternatives without extra work and
headaches ?
I have seen this router costing from $279 to a typical price of
Hi John,
* John Crispin (j...@phrozen.org) wrote:
if you use ralink SoC please test mac80211 after this update to ensure
that we do not have a regression.
Looks good here on rt2880.
-Claudio
___
openwrt-devel mailing list
On 24/04/2014 17:23, Claudio Leite wrote:
Looks good here on rt2880.
-Claudio
great, thanks for the test effort !
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
Hi,
I am observing some issues by reading kernel messages and the git commit
log:
1. Kexec is not working.
kexec-mips.c looks for at least some lines in /proc/iomem for example
System RAM. This entry is missing at least on the TD-W8970.
Would it work with that entry is more missing? How is the
On 24/04/2014 22:05, joerg jungermann wrote:
Hi,
I am observing some issues by reading kernel messages and the git
commit log:
1. Kexec is not working. kexec-mips.c looks for at least some lines
in /proc/iomem for example System RAM. This entry is missing at
least on the TD-W8970.
Am Thursday, den 24 April hub John Crispin folgendes in die Tasten:
On 24/04/2014 22:05, joerg jungermann wrote:
Hi,
I am observing some issues by reading kernel messages and the git
commit log:
1. Kexec is not working. kexec-mips.c looks for at least some lines
in
the lantiq,switch; property is missing in the dts file
According to the current HEAD: target/linux/lantiq/dts/TDW8970.dts
[...] eth@E108000 { #address-cells = 1; #size-cells = 0;
compatible = lantiq,xrx200-net; reg = 0xE108000 0x3000 /*
switch */ 0xE10B100 0x70 /* mdio */ 0xE10B1D8 0x30
I'm using the latest trunk version and trying to use the dynamic cpu
frequency turbo mode. I've got the config.txt setup. I see that there
is a cpufreq-driver patch file, but it seems that its for the brcm2835
platform. Brcm2708 = Raspberry A, brcm2835 = Raspberry B ?. However all
the builds
Hi Guys,
I'm doing a robust test on a ramips 7620a chip, with a customized openwrt
firmware. I ran into this rare issue recently.
The system initialization stopped right before /etc/rc.d/ processing. Chances
are around 5%.
After some debug work, I located the key code that caused the issue.
The GIT_LOCAL_REPOSITORY is pretty useless if KERNEL_GIT_CLONE_URI
is not set. The code in include/ will actually be executed only
if it's set. So let's make it official and mark it as a dependency
in the configuration menu to avoid confusing the user.
Signed-off-by: Mathieu Olivari
This allows the selection of a specific branch in the menuconfig
when using a kernel downloaded from GIT.
Signed-off-by: Mathieu Olivari math...@qca.qualcomm.com
---
config/Config-devel.in |9 +
include/kernel-defaults.mk |4
2 files changed, 13 insertions(+)
diff --git
The way git options are managed in kernel-defaults.mk makes additions
difficult. If requires different code path for each option; it's
ok so far as we handle only one option, but if we want to make the git
clone mechanism more flexible, more option will be required, which
will become tedious.
So;
By default, OpenWrt doesn't patch the kernel if it's downloaded from
a git tree. This option will allow OpenWrt to download the kernel from
GIT but still allow the user to apply some patches, when enabled.
It can be very convenient when working on some standard kernel
maintained in Git, shared by
Hi John,
Thanks for your help. Once I got your patch, I would carry on the test this
weekend.
Hi Hua,
i recall seeing this bug before with uclibc, where the very first call of
fork() fails after a boot ... i have not seen this problem on uclibc in about
4 years ...
i will dig out the
27 matches
Mail list logo
