| Rusty Russell |
| Mar 15, 3:39 pm 2009 |
| Stephen Hemminger | Re: [PATCH] bonding: allow bond in mode balance-alb to w ...
On Sat, 14 Mar 2009 10:49:11 +0100
Something like this:
--- a/net/core/dev.c 2009-03-15 15:55:02.098126056 -0700
+++ b/net/core/dev.c 2009-03-15 16:02:43.999251305 -0700
@@ -3830,6 +3830,7 @@ int dev_set_mac_address(struct net_devic
{
const struct net_device_ops *ops = dev->netdev_ops;
int err;
+ char save_addr[MAX_ADDR_LEN];
if (!ops->ndo_set_mac_address)
return -EOPNOTSUPP;
@@ -3837,9 +3838,17 @@ int dev_set_mac_address(struct net_devic
return -EINVAL;
if ...
| Mar 15, 4:12 pm 2009 |
| Ben Hutchings | Re: build bug on kfree(struct sk_buff*)
[...]
There are many types that should not be allocated and freed using the
general-purpose heap functions. Why provide this check only for struct
skbuff?
Perhaps you should try to find a general way of detecting this, such as
a type annotation for sparse.
Ben.
--
Ben Hutchings, Senior Software Engineer, Solarflare Communications
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
--
| Mar 15, 12:33 pm 2009 |
| Roel Kluin | Re: build bug on kfree(struct sk_buff*)
I noticed that my initial attempt was invalid. The one below appears to be
correct, although I am not sure whether there is a better implementation.
this patch also fixes a kfree(skb) in net/core/dev.c, what do you think?
Roel
Signed-off-by: Roel Kluin <roel.kluin@gmail.com>
---
include/linux/kernel.h | 4 ++++
include/linux/slab.h | 3 +++
mm/slab.c | 2 ++
mm/slob.c | 2 ++
mm/slub.c | 2 ++
net/core/dev.c | 2 +-
6 ...
| Mar 15, 9:42 am 2009 |
| Jeff Kirsher |
| Mar 14, 6:39 pm 2009 |
| Ramkrishna Vepa |
| Mar 14, 6:06 pm 2009 |
| Bill Fink |
| Mar 15, 4:29 pm 2009 |
| Ramkrishna Vepa |
| Mar 15, 4:32 pm 2009 |
| Ben Hutchings | Re: [net-2.6 PATCH 1/10] Neterion: New driver: Driver he ...
On Sat, 2009-03-14 at 00:20 -0800, Ramkrishna Vepa wrote:
Same for most of this.
NAPI should be enabled all the time. There is no need to make it an
option.
This should be controllable through ethtool ETHTOOL_{G,S}RXFH
operations.
What's an RxD block? Some fixed number of RX descriptors? I think this
Just like any other module.
[...]
Doesn't apply when using a current kernel (if that even works in SUSE
These don't apply to in-tree modules.
Not sure this ...
| Mar 15, 12:04 pm 2009 |
| Ben Hutchings | Re: [net-2.6 PATCH 3/10] Neterion: New driver: Kconfig a ...
[...]
These should really come last in the series so they don't refer to
nonexistent files. This may not matter for a new driver as Dave can
squash them together into one patch.
Ben.
--
Ben Hutchings, Senior Software Engineer, Solarflare Communications
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
--
| Mar 15, 12:09 pm 2009 |
| Ramkrishna Vepa |
| Mar 15, 3:42 pm 2009 |
| Ben Hutchings | Re: [net-next PATCH 11/11] ixgbe: Add documentation for ...
[...]
This is the same as for any network driver, so there is no point
including it in driver-specific documentation.
Ben.
--
Ben Hutchings, Senior Software Engineer, Solarflare Communications
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
--
| Mar 15, 12:25 pm 2009 |
| Francois Romieu |
| Mar 15, 3:09 am 2009 |
| Jarek Poplawski | [PATCH v2] pkt_sched: Change misleading code in class delete.
A few words were missing...
----------------> take 2
While looking for a possible reason of bugzilla report on HTB oops:
http://bugzilla.kernel.org/show_bug.cgi?id=12858
I found the code in htb_delete calling htb_destroy_class on zero
refcount is very misleading: it can suggest this is a common path, and
destroy is called under sch_tree_lock. Actually, this can never happen
like this because before deletion cops->get() is done, and after
delete a class is still used by tclass_notify. The class ...
| Mar 15, 5:35 am 2009 |
| Eric Dumazet | Re: IGMP Join dropping multicast packets
I could not reproduce the problem on my machines (bnx2 adapter), even if changing
NUMSOCK from 55 to 200 in joiner.c
Is your network a 100Mb one or Gigabit ?
Try to slow down your joiner ?
(Could be a flood of IGMP messages your router/switch cannot cope with)
Please describe your "rock solid" network setup (kind of network adapters you have, kind of router...)
Each time an address is added, NIC driver have to reprogram mcfilter of
the device. Maybe some NIC can drop some packets at this ...
| Mar 14, 7:37 pm 2009 |
| Rafael J. Wysocki |
| Mar 15, 3:41 am 2009 |
| Jeff Chua | Re: 2.6.29-rc8: Reported regressions from 2.6.28
Still not working in linux-2.6.29-rc8. Broken after the commit below.
There were many changes to wireless after this commit, and simply
reverting this commit will break compiling.
Jeff.
commit 41bb73eeac5ff5fb217257ba33b654747b3abf11
Author: Johannes Berg <johannes@sipsolutions.net>
Date: Wed Oct 29 01:09:37 2008 +0100
mac80211: remove SSID driver code
--
| Mar 14, 7:58 pm 2009 |
| Johannes Berg |
| Mar 15, 11:11 am 2009 |
| Jeff Chua | Re: 2.6.29-rc8: Reported regressions from 2.6.28
The commit below is causing problem with associating with the hidden AP as well.
Thanks,
Jeff.
71c11fb57b924c160297ccd9e1761db598d00ac2 is first bad commit
commit 71c11fb57b924c160297ccd9e1761db598d00ac2
Author: Johannes Berg <johannes@sipsolutions.net>
Date: Tue Oct 28 18:29:48 2008 +0100
b43/legacy: remove SSID code
The SSID programmed into the device is used by the ucode only
to reply to probe requests, a functionality we disable anyway
because it doesn't fit ...
| Mar 14, 8:06 pm 2009 |
| Ingo Molnar | Re: 2.6.29-rc8: Reported regressions from 2.6.28
Even if so (bisection is very hard and error-prone) why do you
shape your reaction to it as a personal attack? Why do you say
"Jeff is totally unbelievable" - why dont you say something more
amicable like:
" Hm, that's weird - that commit does not even seem to affect
the driver you are working with. Could you please re-check
the final bits of the bisection to make sure you got the
right commit ID? "
Instead of this irritated-sounding attack tone you are using.
It's not ...
| Mar 15, 1:26 pm 2009 |
| Linus Torvalds | Re: 2.6.29-rc8: Reported regressions from 2.6.28
Jeff - can you test the kernels before-and-after this commit (with _no_
other changes) and descibe the differences?
Johannes - "not believable" is simply not an argument. If Jeff can show a
difference, then your disbelief is totally irrelevant, and clearly shows
that you are basing your beliefs on incorrect assumptions (like some
specific version of firmware that isn't the whole story).
Linus
--
| Mar 15, 11:44 am 2009 |
| Johannes Berg |
| Mar 15, 12:01 pm 2009 |
| Ilpo Järvinen | [PATCH 2/7] tcp: kill dead end_seq variable in clean_rtx_queue
I've already forgotten what for this was necessary, anyway
it's no longer used (if it ever was).
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
---
net/ipv4/tcp_input.c | 3 ---
1 files changed, 0 insertions(+), 3 deletions(-)
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index cd39d1d..f527a16 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -3201,7 +3201,6 @@ static int tcp_clean_rtx_queue(struct sock *sk, int prior_fackets,
while ((skb = ...
| Mar 14, 5:23 pm 2009 |
| Ilpo Järvinen | [PATCH 6/7] tcp: cache result of earlier divides when ms ...
The results is very unlikely change every so often so we
hardly need to divide again after doing that once for a
connection. Yet, if divide still becomes necessary we
detect that and do the right thing and again settle for
non-divide state. Takes the u16 space which was previously
taken by the plain xmit_size_goal.
This should take care part of the tso vs non-tso difference
we found earlier.
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Cc: Evgeniy Polyakov <zbr@ioremap.net>
Cc: ...
| Mar 14, 5:23 pm 2009 |
| Ilpo Järvinen | [PATCH 4/7] tcp: don't check mtu probe completion in the loop
It seems that no variables clash such that we couldn't do
the check just once later on. Therefore move it.
Also kill dead obvious comment, dead argument and add
unlikely since this mtu probe does not happen too often.
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
---
net/ipv4/tcp_input.c | 13 ++++++-------
1 files changed, 6 insertions(+), 7 deletions(-)
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index b7d02c5..311c30f 100644
--- a/net/ipv4/tcp_input.c
+++ ...
| Mar 14, 5:23 pm 2009 |
| Ilpo Järvinen | [PATCH 5/7] tcp: simplify tcp_current_mss
There's very little need for most of the callsites to get
tp->xmit_goal_size updated. That will cost us divide as is,
so slice the function in two. Also, the only users of the
tp->xmit_goal_size are directly behind tcp_current_mss(),
so there's no need to store that variable into tcp_sock
at all! The drop of xmit_goal_size currently leaves 16-bit
hole and some reorganization would again be necessary to
change that (but I'm aiming to fill that hole with u16
xmit_goal_size_segs to cache the ...
| Mar 14, 5:23 pm 2009 |
| Ilpo Järvinen | [PATCH 7/7] tcp: make sure xmit goal size never becomes zero
It's not too likely to happen, would basically require crafted
packets (must hit the max guard in tcp_bound_to_half_wnd()).
It seems that nothing that bad would happen as there's tcp_mems
and congestion window that prevent runaway at some point from
hurting all too much (I'm not that sure what all those zero
sized segments we would generate do though in write queue).
Preventing it regardless is certainly the best way to go.
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Cc: Evgeniy ...
| Mar 14, 5:23 pm 2009 |
| Ilpo Järvinen | [PATCH 1/7] tcp: remove pointless .dsack/.num_sacks code
In the pure assignment case, the earlier zeroing is
still in effect.
David S. Miller raised concerns if the ifs are there to avoid
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
---
My apologies for the duplicated patches, I succeeded once again
to make mess out of it with git-send-email and 8-bit chars in
header which vger consistently is known to reject... Thus resending
to get a copy to appear on netdev too.
Grr, 3rd attempt, now with --no-signed-off-by-cc though it ...
| Mar 14, 5:23 pm 2009 |
| Ilpo Järvinen | [PATCH 3/7] tcp: consolidate paws check
Wow, it was quite tricky to merge that stream of negations
but I think I finally got it right:
check & replace_ts_recent:
(s32)(rcv_tsval - ts_recent) >= 0 => 0
(s32)(ts_recent - rcv_tsval) <= 0 => 0
discard:
(s32)(ts_recent - rcv_tsval) > TCP_PAWS_WINDOW => 1
(s32)(ts_recent - rcv_tsval) <= TCP_PAWS_WINDOW => 0
I toggled the return values of tcp_paws_check around since
the old encoding added yet-another negation making tracking
of truth-values ...
| Mar 14, 5:23 pm 2009 |
| Jeff Kirsher | [net-2.6 PATCH] igb: remove ASPM L0s workaround
From: Alexander Duyck <alexander.h.duyck@intel.com>
The L0s workaround should be moved into a pci quirk and so it is not
necessary in the driver. This update removes the L0s workaround from the
igb driver.
This was the second half of the PCI quirk patch that Matthew Wilcox did
not pick up when he picked up the quirk patch.
This patch applies to the net-2.6 tree. :)
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Signed-off-by: Jeff Kirsher ...
| Mar 14, 6:38 pm 2009 |
| David Miller |
| Mar 14, 10:26 pm 2009 |
| Kyle McMartin | [PATCH] tulip: fix crash on iface up with shirq debug
From: Kyle McMartin <kyle@redhat.com>
Tulip is currently doing request_irq before it has done its
initialization. This is usually not a problem because it hasn't
enable interrupts yet, but with DEBUG_SHIRQ on, we call the irq handler
when registering the interrupt as a sanity check.
This can result in a NULL ptr dereference, so call tulip_init_ring
before request_irq, and add a free_ring function to do the freeing
now shared with tulip_close.
Tested with a shell loop running ifup, ifdown ...
| Mar 14, 10:35 pm 2009 |
| Ilpo Järvinen |
| Mar 15, 1:45 am 2009 |
| Evgeniy Polyakov |
| Mar 15, 2:08 am 2009 |
| Evgeniy Polyakov |
| Mar 15, 1:36 am 2009 |
| Francois Romieu | [PATCH 1/2] r8169: use hardware auto-padding.
It shortens the code and fixes the current pci_unmap leak with
padded skb reported by Dave Jones.
Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
---
drivers/net/r8169.c | 8 --------
1 files changed, 0 insertions(+), 8 deletions(-)
diff --git a/drivers/net/r8169.c b/drivers/net/r8169.c
index b347340..352da2a 100644
--- a/drivers/net/r8169.c
+++ b/drivers/net/r8169.c
@@ -3363,13 +3363,6 @@ static int rtl8169_start_xmit(struct sk_buff *skb, struct net_device *dev)
opts1 |= ...
| Mar 15, 4:09 am 2009 |
| Francois Romieu | [PATCH 2/2] r8169: revert "r8169: read MAC address from ...
It fails on the following systems:
- RTL8169sc/8110sc (XID 18000000)
reported by Tim Durack <tdurack@gmail.com> (x86)
- RTL8169sb/8110sb (XID 10000000)
reported by Mikael Pettersson <mikpe@it.uu.se> (ARM)
The patch appeared to work on x86 for the following systems:
RTL8169sb/8110sb 10000000 PCI (EXT)
RTL8110s 04000000 PCI (EXT)
RTL8102e 24a00000 PCI-E (LOM)
RTL8168c/8111c 3c2000c0 PCI-E (LOM)
RTL8168b/8111b 38000000 PCI-E (LOM)
RTL8168b/8111b 38000000 PCI-E ...
| Mar 15, 4:10 am 2009 |
| Pelle Svensson | Intel 82567LM e1000e DHCP problem/question
Hi,
Quick question.
Is there any functionality inside the driver for DHCP request.
I see in the source that there is some kind of DHCP filter f the tx stream.
Can the drive issue DHCP request inside it self when the Cable is unplugged/plugged.
I'm tracing a DHCP fault.
When the net is brought down, ifdown eth0, the dhclient is killed.
Then when I unplugged/plugged the cable I get a DHCP request even that
dhclient is down.
Also the Wireshark format of the DHCP request are ...
| Mar 15, 7:04 am 2009 |
| Jiri Pirko | [PATCH] bonding: select current active slave when enslav ...
Hi.
I've hit an issue on my system when I've been using RealTek RTL8139D cards in
bonding interface in mode balancing-alb. When I enslave a card, the current
active slave (bond->curr_active_slave) is not set and the link is therefore
not functional.
----
# cat /proc/net/bonding/bond0
Ethernet Channel Bonding Driver: v3.5.0 (November 4, 2008)
Bonding Mode: adaptive load balancing
Primary Slave: None
Currently Active Slave: None
MII Status: up
MII Polling Interval (ms): 100
Up Delay ...
| Mar 15, 9:02 am 2009 |
| Andrew Morton |
| Mar 15, 2:30 pm 2009 |
| Andrew Morton |
| Mar 15, 2:32 pm 2009 |
