On Sun, Dec 03, 2006 at 12:33:39AM +0300, Alexey Dobriyan wrote:
David Binderman's icc logs:
net/rose/rose_route.c(399): remark #593: variable err was set but never used
Which is more of a non-fix because all the callers ignore the return
value ...
Ralf
-
To unsubscribe from this list: send
From: Ralf Baechle [EMAIL PROTECTED]
Date: Sun, 3 Dec 2006 08:52:10 +
On Sun, Dec 03, 2006 at 12:33:39AM +0300, Alexey Dobriyan wrote:
David Binderman's icc logs:
net/rose/rose_route.c(399): remark #593: variable err was set but never
used
Which is more of a non-fix because all
Stephen,
On 12/2/06, Stephen Hemminger [EMAIL PROTECTED] wrote:
If transmit lock is contended on, then push return code back
and retry at higher level.
Looking at qdisc_restart, it seems to me that the NETDEV_TX_LOCKED
return code must only be used if the device features LLTX. With your
On Fri, 2006-10-06 at 17:20 +, Joseph S. Myers wrote:
The kernel headers installed by Linux 2.6.19-rc1 make
headers_install do not work for building glibc, because glibc expects
linux/rtnetlink.h to provide various definitions, some of which have
been moved to linux/if_addr.h and some of
On Wed, 2006-14-11 at 16:17 +0100, Daniel Lezcano wrote:
The attached document describes the network isolation at the layer 2
and at the layer 3 ..
Daniel,
I apologize for taking this long to get back to you. The document (I
hope) made it clear to me at least the difference between the two
On Mon, Sep 04, 2006 at 05:44:58PM +0800, Wong Edison wrote:
pageexec report an oops for tcp_lp_owd_calculator(). This is due to
tcp_lp_remote_hz_estimator can return 0.
This patch fix the handling of lp-flag, so will set lp-flag as FALSE
if rhz = 0
Signed-off-by: Wong Hoi Sing Edison
On Sat, 2006-12-02 at 16:49 -0600, Steve Wise wrote:
+
+static struct ib_ah *iwch_ah_create(struct ib_pd *pd,
+ struct ib_ah_attr *ah_attr)
+{
+ return ERR_PTR(-ENOSYS);
+}
-ENOSYS is just about ALWAYS a bug in that it's guaranteed to be the
wrong error
I have removed the Re: just to add some freshness to the discussion
So i read quickly the rest of the discussions. I was almost suprised to
find that i agree with Eric on a lot of opinions (we also agree that
vindaloo is good for you i guess);-
The two issues that stood out for me (in addition to
This improves dumping performance of xfrm policies. I started getting
bothered when i noticed it took upto a minute dumping 60K policies from
the kernel.
I have another one i am testing that uses the same approach for SAs.
Will send in the next hour.
Against net-2.6.20.
cheers,
jamal
[XFRM]
Ok. Just a quick summary of where I see the discussion.
We all agree that L2 isolation is needed at some point.
The approaches discussed for L2 and L3 are sufficiently orthogonal
that we can implement then in either order. You would need to
unshare L3 to unshare L2, but if we think of them as
Ok, heres is the SA version and the numbers.
Against net-2.6.20
cheers,
jamal
[XFRM] Optimize SA dumping
Same comments as in [XFRM] Optimize policy dumping
Again, the same issue as in the policy optimization; we could do better
only if we didnt have to do bending backwards to make pfkey
On Sun, Dec 03, 2006 at 07:26:02AM -0500, jamal wrote:
On Wed, 2006-14-11 at 16:17 +0100, Daniel Lezcano wrote:
The attached document describes the network isolation at the layer 2
and at the layer 3 ..
Daniel,
I apologize for taking this long to get back to you. The document (I
hope)
On Sun, 2006-03-12 at 17:37 +0100, Herbert Poetzl wrote:
On Sun, Dec 03, 2006 at 07:26:02AM -0500, jamal wrote:
To use an extreme example: if i picked apache as a
binary compiled 10 years ago, it will run on the L2 approach but not on
the L3 approach. Is this understanding correct? I find
Maxime Austruy wrote:
Found a hang while playing with wpa_supplicant and a zd1211 usb adapter
using linux 2.6.19. The problem is that ieee80211softmac_wx_set_mlme
forgets to release a mutex when mlme-cmd is IW_MLME_DEAUTH. The fix
below allows me to kill wpa_supplicant and to restart it
Hi,
Some questions,suggestions,:
+cxgb3_cpl_handler_func t3c_handlers[NUM_CPL_CMDS];
Can it be static'ified? (I suppose not.)
+struct cxgb3_client t3c_client = {
+ .name = iw_cxgb3,
+ .add = open_rnic_dev,
+ .remove = close_rnic_dev,
+ .handlers = t3c_handlers,
+
Add new rt2500usb and rt73usb device id numbers.
Signed-off-by Ivo van Doorn [EMAIL PROTECTED]
---
diff -rU3 wireless-dev-base/drivers/net/wireless/d80211/rt2x00/rt2500usb.c
wireless-dev-base-rt2x00/drivers/net/wireless/d80211/rt2x00/rt2500usb.c
---
Always use kzalloc instead of kmalloc.
Remove duplicate init functions.
And destroy the workqueue before freeing
resources, otherwise a thread on the queue might
still want to access that resource.
Signed-off-by Ivo van Doorn [EMAIL PROTECTED]
---
diff -rU3
Correctly initialize the channel_change_time.
Make sure that channel is reset afterwards,
otherwise the channel is not correctly initialized
and rx/tx will fail.
Signed-off-by Ivo van Doorn [EMAIL PROTECTED]
---
diff -rU3 wireless-dev-mac/drivers/net/wireless/d80211/rt2x00/rt2400pci.c
Move the crc handling of rt61pci and rt73usb into a seperate module.
This will create the crc-itu-t module inside the rt2x00 folder.
Signed-off-by Ivo van Doorn [EMAIL PROTECTED]
---
diff -rNU3 wireless-dev-base/drivers/net/wireless/d80211/rt2x00/Kconfig
rt61pci and rt73usb have the WMM ring priorities
backwards. RING_AC_VO is the most important ring
while RING_AC_BK the least important ring.
Lets reorder the ring handling.
(And fix some small typos in the comments regarding
the rings)
Signed-off-by Ivo van Doorn [EMAIL PROTECTED]
---
diff -rU3
Each received packet has a signal field, this field
can be translated into the rate with which the frame
has been received.
Create a seperate function for this since the
conversion is equal for all drivers.
Signed-off-by Ivo van Doorn [EMAIL PROTECTED]
---
diff -rU3
Correctly let the non-monitor and monitor interfaces
excist peacefully together. Make sure the configuration
is always accurate and allows the correct packets to
come through, let the interface enable the radio at
the correct time etc. etc.
Signed-off-by Ivo van Doorn [EMAIL PROTECTED]
---
Overhaul the byteordering mechanism.
All byteordering happens at the reading and writing
of the register/eeprom/descriptor instead of the
get/set_field functions.
This makes sparse very happy and reduces the errors
significantly.
Signed-off-by Ivo van Doorn [EMAIL PROTECTED]
---
Put the link tuning in a workqueue, this prevents
the interrupthandlers from being busy for a too long
period and blocking new inetrrupt handling.
To do this correctly we add a link structure containing
all information regarding the link status.
Signed-off-by Ivo van Doorn [EMAIL PROTECTED]
---
As usual, when I make a large patch series,
I overlook important bits...
This will fix all issues that have arisen from this
patch series.
Signed-off-by Ivo van Doorn [EMAIL PROTECTED]
---
diff -rU3 wireless-dev-misc/drivers/net/wireless/d80211/rt2x00/rt2500pci.c
The length of a packet that needs to be send over a USB device,
needs to have an even length. odd lengths will cause problems.
At the moment there is no solution for the enabling of promisc
mode on usb devices. The function is called from interrupt
context which means the driver cannot access a
Correctly detect the maxssi settings from the EEPROM
where available, and correct the dummy values that
had been added in the initial patch to support
noise and signal measurement.
MAX_RX_SSI is a value that is different on each chipset,
so it should be a driver specific define.
Signed-off-by Ivo
Latest d80211 stack no longer provides any ethtool support.
At the moment there is no quick replacement possible for
the ethtool features (debugfs is under investigation, but
requires more work).
Signed-off-by Ivo van Doorn [EMAIL PROTECTED]
---
diff -rU3
Always call ieee80211_wake_queue if the ring
is not full after the txrun. The ieee80211_wake_queue
is responsible for chacking if the queue was stopped
or not.
The current implementation of checking the ring_full
before the txdone run was flawed and race conditions
could occur that blocked all tx
Make sure all statistics the d80211 stack requires.
Some of this requires values to be read during
interrupt process (add a new function to handle this).
And other fields can be read from the registers at
request time.
Note that rt61pci and rt73usb had an invalid registername
the legacy drivers
No set_field commands are required for the mac registers.
This was previously done for the byteordering.
But since the MAC is already read in the correct byteorder
this had never had to happen at all anyway.
Signed-off-by Ivo van Doorn [EMAIL PROTECTED]
---
diff -rU3
rt2400pci, rt2500pci and rt61pci share exactly the
same code for the eeprom reading. The only difference
is that rt61pci has a slightly different register reading
approach. In any case we have a lot of duplicate code.
Create a new module eeprom_93cx6 inside the rt2x00 folder
and make rt2x00 use
Remove the rt2x00usb_vendor_request from the rt2x00usb.h
header, and place it into the rt2x00_vendor_request.
This means that the rt2x00_vendor_request function
needs a timeout value especially for commands that
require a lot of time (i.e. Firmware writing).
Signed-off-by Ivo van Doorn [EMAIL
rt61pci and rt73usb legacy drivers have hinted that there are
race conditions with the bbp register handling. This must be
fixed by doing a busy check before the actual read command.
At the same time we can remove duplicate code by putting that
busy check into a seperate function.
Signed-off-by
Call activity_led() function after each rxdone run
to make the led blink.
Signed-off-by Ivo van Doorn [EMAIL PROTECTED]
---
diff -rU3 wireless-dev-channel/drivers/net/wireless/d80211/rt2x00/rt2400pci.c
wireless-dev-led/drivers/net/wireless/d80211/rt2x00/rt2400pci.c
---
Correctly handle rts frames in txdone, by freeing the
packet. Also use the is_rts_frame to detect if the
frame was rts. This was done incorrectly previously.
Signed-off-by Ivo van Doorn [EMAIL PROTECTED]
---
diff -rU3 wireless-dev-timing/drivers/net/wireless/d80211/rt2x00/rt2400pci.c
Hi,
This patch series will bring rt2x00 completely up to date.
This includes the compile fixes after last d80211 update,
and brings so many fixes it is hard to summarize them.
Basically what it comes down to is that with this version
several people have managed to get the drivers working
and even
We work with the EEPROM by using the word number
as offset. Fix USB drivers to use the correct offset.
Signed-off-by Ivo van Doorn [EMAIL PROTECTED]
---
diff -rU3 wireless-dev-rssi/drivers/net/wireless/d80211/rt2x00/rt2500usb.h
Fix rt2x00 compilation problems due to the d80211 update.
Signed-off-by Ivo van Doorn [EMAIL PROTECTED]
---
http://www.mendiosus.nl/rt2x00/06_d80211.diff
-
To unsubscribe from this list: send the line unsubscribe netdev in
the body of a message to [EMAIL PROTECTED]
More majordomo info at
Misc fixes.
* Correctly set the RFCSR value using the setfield function.
* Remove the DISABLE_RX register setting during initialization.
* Changing the durationid should not add but overwrite. (sparse fix)
* Prevent false warnings about ignoring ring initialization,
by only mentioning which
Remove txd and rxd descriptor structures.
All access to the descriptors should be done
by treating the descriptor as a register.
For this we do add a desc structure containing
a array, and several methods for reading and writing
to that array.
This is prepares rt2x00 for the overhaul of the
Introduce new defines for the SIFS, PIFS, EIFS, DIFS,
and make use of it in the drivers.
Signed-off-by Ivo van Doorn [EMAIL PROTECTED]
---
diff -rU3 wireless-dev-stats/drivers/net/wireless/d80211/rt2x00/rt2400pci.c
wireless-dev-timing/drivers/net/wireless/d80211/rt2x00/rt2400pci.c
---
On Sunday 03 December 2006 13:19, Ivo van Doorn wrote:
rt2400pci, rt2500pci and rt61pci share exactly the
same code for the eeprom reading. The only difference
is that rt61pci has a slightly different register reading
approach. In any case we have a lot of duplicate code.
Create a new module
On Sunday 03 December 2006 19:39, Michael Wu wrote:
On Sunday 03 December 2006 13:19, Ivo van Doorn wrote:
rt2400pci, rt2500pci and rt61pci share exactly the
same code for the eeprom reading. The only difference
is that rt61pci has a slightly different register reading
approach. In any
From: Randy Dunlap [EMAIL PROTECTED]
Fix help text typos.
Fix help text width to be 80 columns: better for editors and for
menuconfig in 80-column terms.
Signed-off-by: Randy Dunlap [EMAIL PROTECTED]
---
drivers/net/phy/Kconfig |7 ---
1 file changed, 4 insertions(+), 3 deletions(-)
From: Maxime Austruy [EMAIL PROTECTED]
Routine ieee80211softmac_wx_set_mlme has one return that fails
to release a mutex acquired at entry.
Signed-off-by: Maxime Austruy [EMAIL PROTECTED]
Signed-off-by: Larry Finger [EMAIL PROTECTED]
---
John and Chris,
This error was introduced in the
Hi,
This patch is a resend of a patch I has been send to the linux kernel
and netdev list earlier. The most recent version of a few weeks back
didn't compile since I missed 1 line in my patch that changed
include/linux/input.h.
This patch will offer the handling of radiokeys on a laptop.
Such
On Sun, 2006-12-03 at 19:36 +0100, Ivo van Doorn wrote:
+
+ down(master-key_sem);
+
Hi,
this one seems to be used as a mutex only, please consider using a mutex
instead, as is the default for new code since 2.6.16 or so
Greetings,
Arjan van de Ven
--
if you want to mail me at
On Sun, 2006-12-03 at 19:36 +0100, Ivo van Doorn wrote:
+static int rfkill_open(struct input_dev *dev)
+{
+ ((struct rfkill_key*)(dev-private))-open_count++;
+ return 0;
+}
Hi,
this open_count thing smells fishy to me; what are the locking rules for
it? What guarantees that the
On Sun, 2006-12-03 at 19:36 +0100, Ivo van Doorn wrote:
Hi,
This patch is a resend of a patch I has been send to the linux kernel
and netdev list earlier. The most recent version of a few weeks back
didn't compile since I missed 1 line in my patch that changed
include/linux/input.h.
This
On Sunday 03 December 2006 20:18, Arjan van de Ven wrote:
On Sun, 2006-12-03 at 19:36 +0100, Ivo van Doorn wrote:
+
+ down(master-key_sem);
+
Hi,
this one seems to be used as a mutex only, please consider using a mutex
instead, as is the default for new code since 2.6.16 or
On Sunday 03 December 2006 20:20, Arjan van de Ven wrote:
this open_count thing smells fishy to me; what are the locking rules for
it? What guarantees that the readers of it don't get the value changed
underneath them between looking at the value and doing whatever action
depends on it's value
Hi,
This patch is a resend of a patch I has been send to the linux kernel
and netdev list earlier. The most recent version of a few weeks back
didn't compile since I missed 1 line in my patch that changed
include/linux/input.h.
This patch will offer the handling of radiokeys on a
rfkill with the fixes as suggested by Arjan.
- instead of a semaphore a mutex is now being used.
- open_count changing is now locked by the mutex.
Signed-off-by Ivo van Doorn [EMAIL PROTECTED]
---
diff --git a/drivers/input/misc/Kconfig b/drivers/input/misc/Kconfig
index ba0e88c..6986d59
Please pull from tag 'r8169-upstream-20061204-00' in repository
git://electric-eye.fr.zoreil.com/home/romieu/linux-2.6.git#r8169-upstream-20061204-00
to get the changes below.
Distance from 'upstream'
-
a0e21b002db5bce80aac82c21cd996423487a44c
The 8110SB based n2100 board signals a lot of what ought to be
PCI data parity errors durint operation of the 8169 as target.
Experiment proved that the driver can ignore the error and
process the packet as if nothing had happened.
Let's add an ad-hoc knob to enable users to fix their system
Mostly taken from Realtek's driver.
It's a bit yucky but the original is even worse.
Signed-off-by: Francois Romieu [EMAIL PROTECTED]
Signed-off-by: Darren Salt [EMAIL PROTECTED]
---
drivers/net/r8169.c | 58 +++
1 files changed, 44
This is commented out in Realtek's driver as well.
Signed-off-by: Francois Romieu [EMAIL PROTECTED]
---
drivers/net/r8169.c |5 -
1 files changed, 0 insertions(+), 5 deletions(-)
diff --git a/drivers/net/r8169.c b/drivers/net/r8169.c
index 7438049..0b57050 100644
---
Ok, I can not count up to 5 and the serie really contains 4 patches.
Btw the patch a0e21b002db5bce80aac82c21cd996423487a44c (top of the
pile) contained a reference to bug 7617 instead of 7517. It has been
replaced and tag r8169-upstream-20061204-00 now points to
Kazunori, a bug from the changes I did apply:
[ 761.318131] kernel BUG at net/key/af_key.c:1925!
[ 761.318285] Caller[1010c0e0]: pfkey_send_policy_notify+0x6c/0x250
[af_key]
[ 761.318296] Caller[0062f6dc]: km_policy_notify+0x28/0x50
[ 761.318309] Caller[1014a94c]:
From: David Miller [EMAIL PROTECTED]
Date: Sun, 03 Dec 2006 17:58:47 -0800 (PST)
Kazunori, a bug from the changes I did apply:
[ 761.318131] kernel BUG at net/key/af_key.c:1925!
I found the problem, it's because of the xfrm_user.c change where
we clobber the xp-family value with ut-family.
Hello David,
Thank you for your tracing the bug.
I understood the issue.
Mmm, if we can not use ut-family, can we use
ut-id.family instead?
Or is it also uninitialized?
David Miller wrote:
From: David Miller [EMAIL PROTECTED]
Date: Sun, 03 Dec 2006 17:58:47 -0800 (PST)
Kazunori, a bug from
From: Kazunori MIYAZAWA [EMAIL PROTECTED]
Date: Mon, 04 Dec 2006 11:50:09 +0900
Thank you for your tracing the bug.
I understood the issue.
Mmm, if we can not use ut-family, can we use
ut-id.family instead?
Or is it also uninitialized?
struct xfrm_id does not have a family field
struct
Kernel = 2.6.19 SMP
bcmwl5 firmware = 4.10.40.0
Devscape bcm43xx_d80211 driver = Latest git pull
Wireless card = Broadcom Corporation Dell Wireless 1390 WLAN Mini-PCI
Card (rev 01) (PCI-E)
I seem to be having a problem associating with my AP--everything appears
fine and I can bring my
On Sun, Dec 03, 2006 at 07:12:08PM -0800, David Miller wrote:
For the time being I'm thinking of using the following
patch. I removed the xp-family clobbering, the AF_KEY
changes don't do this, so there is no reason for the
xfrm_user side to do it either.
This patch looks fine to me.
I am
David Miller wrote:
From: Kazunori MIYAZAWA [EMAIL PROTECTED]
Date: Mon, 04 Dec 2006 11:50:09 +0900
Thank you for your tracing the bug.
I understood the issue.
Mmm, if we can not use ut-family, can we use
ut-id.family instead?
Or is it also uninitialized?
struct xfrm_id does not have a
Hi all:
I am absolutely new to kernel module programming. I have to modify the
hostap -code such that whenever a network device is registered, it is
assigned an ip address.
can i some how use the devinet.c functions such as
inetdev_init,inet_insert_ifa,etc fucntions even though they are not
I am afraid that the method used in the patch is not native, it breaks
on IP fragments.
IPVS is a kind of layer-4 switching, it routes packet by checking
layer-4 information
such as address and port number. ip_vs_in() is hooked at
NF_IP_LOCAL_IN, so
that all the packets received by
From: Kazunori MIYAZAWA [EMAIL PROTECTED]
Date: Mon, 04 Dec 2006 13:26:29 +0900
If uninitialized ut-family is AF_INET or AF_INET6 by chance
and the family of outer addresses (ut-saddr) is differnt
ut-family, it results some garbage in the kernel as you know.
I think it does not results any
Hi All:
When I tested linux kernel 2.6.18.3, and find that kernel statistics
about IPSTATS_MIB_OUTNOROUTES which exsits in file /proc/net/snmp
doesn't increase correctly. The criteria conform to RFC2011:
ipOutNoRoutes OBJECT-TYPE
SYNTAX Counter32
MAX-ACCESS read-only
Hi,
I've just ported 8 patches that were in the wireless-2.6 version of
zd1211rw that weren't in the wireless-dev version. All of them, except for
the cleanups patch, applied with just an offset, and all of the broken hunks
in the cleanups patch were just thrown away since the code they
71 matches
Mail list logo
