| Andy Fleming | [RFC] Change mii_bus id field to a string
Having the id field be an int was making more complex bus topologies
excessively difficult. For now, just convert it to a string, and
change all "bus->id = val" to snprintf(id, MII_BUS_ID_LEN, "%x", val).
---
At the very least this needs to be tested by the various driver maintainers.
arch/powerpc/platforms/82xx/ep8248e.c | 2 +-
arch/powerpc/platforms/pasemi/gpio_mdio.c | 2 +-
arch/powerpc/sysdev/fsl_soc.c | 5 ++-
drivers/net/au1000_eth.c | 6 ...
| Apr 2, 6:49 pm 2008 |
| Chuck Ebbert |
| Apr 2, 6:46 pm 2008 |
| Sebastian Siewior | [PATCH 2/5] m68knommu: fec: small coding style cleanup
this is basically
- lident
- a few returns in set_multicast_list() in order to get less ident levels
because there is no else
- remove __inline__
Signed-off-by: Sebastian Siewior <sebastian@linutronix.de>
--- a/drivers/net/fec.c
+++ b/drivers/net/fec.c
@@ -84,15 +84,15 @@ static unsigned int fec_hw[] = {
#elif defined(CONFIG_M523x) || defined(CONFIG_M528x)
(MCF_MBAR + 0x1000),
#elif defined(CONFIG_M520x)
- (MCF_MBAR+0x30000),
+ (MCF_MBAR + 0x30000),
#elif defined(CONFIG_M532x)
...
| Apr 2, 4:58 pm 2008 |
| Sebastian Siewior | [PATCH 1/5] fec: kill warnings
linux-2.6-mk68/drivers/net/fec.c: In function 'fec_enet_module_init':
linux-2.6-mk68/drivers/net/fec.c:2627: warning: unused variable 'j'
linux-2.6-mk68/drivers/net/fec.c: At top level:
linux-2.6-mk68/drivers/net/fec.c:2136: warning: 'mii_link_interrupt' defined but not used
Signed-off-by: Sebastian Siewior <bigeasy@linutronix.de>
--- a/drivers/net/fec.c
+++ b/drivers/net/fec.c
@@ -67,6 +67,11 @@
#define FEC_MAX_PORTS 1
#endif
+#if defined(CONFIG_M523x) || defined(CONFIG_M527x) || def...
| Apr 2, 4:58 pm 2008 |
| Sebastian Siewior | [PATCH 3/5] m68knommu: fec typedef a function
Make a typedef for a function that is refered as a function pointer.
Also move the conditial call of the callback function in fec_enet_mii()
of the area that is protected with spinlock beacause the callbacked
function might grab the same lock.
Signed-off-by: Sebastian Siewior <bigeasy@linutronix.de>
--- a/drivers/net/fec.c
+++ b/drivers/net/fec.c
@@ -116,10 +116,10 @@ static unsigned char fec_mac_default[] =
/* Forward declarations of some structures to support different PHYs
*/
-
+t...
| Apr 2, 4:58 pm 2008 |
| Sebastian Siewior | [PATCH 5/5] m68knommu: dont allocate unused interrupts
After fixing (hopefully) fixing locking I run into this:
|irq 88: nobody cared (try booting with the "irqpoll" option)
|Stack from 00214b74:
| 00214b88 00047940 00179b88 00046c36 00179b88 00214bc8 00047b20 00000058
| 00179b88 00000000 00000058 00179bb8 002cdf40 001415ba 00242000 0014163e
| 00179b88 00046c36 0014163e 000479a6 00179b88 00214bfc 00046da8 00000058
| 00179b88 00000000 00000000 00241c00 00247484 40001000 002473e0 002cf080
| 00247000 002334ac 00214c0c 0002...
| Apr 2, 4:58 pm 2008 |
| Sebastian Siewior | [PATCH 4/5] m68knommu: fec fixup locking
Now there are two spinlocks:
- one for HW access like the RX/TX buffer & friends (but the MAC address)
- one for the mii lists.
Theoretically those two spinlocks could become one, but that way it looks
better.
The locks are taken before the first access to the HW (like retrieving the
current pointer to the ring buffer) and released after we finished. This
fixes some races that were covered / fixed by 0a504779d.
Signed-off-by: Sebastian Siewior <bigeasy@linutronix.de>
--- a/drivers/net...
| Apr 2, 4:58 pm 2008 |
| Sebastian Siewior | [PATCH 0/5] fixup locking on m68knommu fec
This series basically fixes my locking problem. We have here:
#1 the old one which removes some undefined vars.
#2 cleans the file with lident. I had trouble reading the code...
#3 type def a function pointer so we don't have the long defines
and move the callback after releasing the lock.
#4 This one fixes the locking I guess. My mcf5235 did not race anymore
after that one. I saw 0a504779d while forming the patch so I guess
we could revert that one, Greg?
#5 removes the registration of int...
| Apr 2, 4:58 pm 2008 |
| Auke Kok | [PATCH 1/5] e1000: remove PCI Express device IDs
We do not want to prolong the situation much longer that e1000
and e1000e support these devices at the same time. As a result,
take out the bandage that was added for the interim period
and remove all the PCI Express device IDs from e1000.
Signed-off-by: Auke Kok <auke-jan.h.kok@intel.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Jeff Garzik <jeff@garzik.org>
---
drivers/net/Kconfig | 3 ---
drivers/net/e1000/e1000_main.c | 33 -------------------...
| Apr 2, 4:48 pm 2008 |
| Auke Kok | [PATCH 5/5] e1000e: reformat register test code, fix some mi...
From: Bruce Allan <bruce.w.allan@intel.com>
The register tests should be run with all the proper flags enabled
to maximize the test coverage code and make sure we are as close
as we can get to testing regular traffic.
Reformat the code for readability. Minor cleanups in the descriptor
ring setup.
Signed-off-by: Bruce Allan <bruce.w.allan@intel.com>
Signed-off-by: Auke Kok <auke-jan.h.kok@intel.com>
---
drivers/net/e1000e/ethtool.c | 123 ++++++++++++++++++++++------------...
| Apr 2, 4:48 pm 2008 |
| Auke Kok | [PATCH 4/5] e1000e: rename a few functions
From: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Several minor cosmetic function renames.
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: Auke Kok <auke-jan.h.kok@intel.com>
---
drivers/net/e1000e/82571.c | 8 ++++----
drivers/net/e1000e/e1000.h | 4 ++--
drivers/net/e1000e/es2lan.c | 4 ++--
drivers/net/e1000e/ich8lan.c | 6 +++---
drivers/net/e1000e/lib.c | 6 +++---
drivers/net/e1000e/netdev.c | 8 ++++----
6 files c...
| Apr 2, 4:48 pm 2008 |
| Auke Kok | [PATCH 3/5] e1000e: Make arrays out of these Rx/Tx registers
From: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
With multiple queues coming into the code these base control
registers need to be made into arrays.
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: Auke Kok <auke-jan.h.kok@intel.com>
---
drivers/net/e1000e/82571.c | 24 ++++++++++++------------
drivers/net/e1000e/es2lan.c | 24 ++++++++++++------------
drivers/net/e1000e/hw.h | 10 ++++++----
drivers/net/e1000e/ich8lan.c | 24 ++++++++...
| Apr 2, 4:48 pm 2008 |
| Auke Kok | [PATCH 2/5] e1000e: limit EEPROM size accesses
From: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: Auke Kok <auke-jan.h.kok@intel.com>
---
drivers/net/e1000e/82571.c | 4 ++++
drivers/net/e1000e/es2lan.c | 4 ++++
2 files changed, 8 insertions(+), 0 deletions(-)
diff --git a/drivers/net/e1000e/82571.c b/drivers/net/e1000e/82571.c
index f7e1619..9b46a00 100644
--- a/drivers/net/e1000e/82571.c
+++ b/drivers/net/e1000e/82571.c
@@ -171,6 +171,10 ...
| Apr 2, 4:48 pm 2008 |
| Jeff Garzik | Re: GFP_ATOMIC page allocation failures.
Do you people hear what you're saying???
I respectfully but strongly disagree with this.
We do __not__ need a whitelist (__GFP_NOWARN) of drivers that handle
allocation failures properly. That's a long list, a maintenance
nightmare, and it is punishing good behavior.
It has been true for over a decade that allocations should be checked
for NULL, and GFP_ATOMIC allocations MUST be checked for NULL.
Let's not crap all over good drivers, because a few bad apples don't
have the proper c...
| Apr 2, 1:21 pm 2008 |
| Andrew Morton | Re: GFP_ATOMIC page allocation failures.
After you've read Nick's comments (which I pray you have not), and after
you've convinced us and yourself of their wrongness, you might like to
consider adding a __GFP_NOWARN to netdev_alloc_skb().
--
| Apr 2, 1:33 pm 2008 |
| Jeff Garzik | Re: GFP_ATOMIC page allocation failures.
Already done so. Adding __GFP_NOWARN to netdev_alloc_skb() is wrong
for several reasons.
It doesn't change the underlying conditions.
It doesn't fix the desire to stamp other drivers in this manner.
And most importantly, it is not even correct: the handling of the
allocation failure remains delegated to the netdev_alloc_skb() users,
which may or may not be properly handling allocation failures.
Put simply, you don't know if the caller is stupid or smart. And there
are a _lot_ of call...
| Apr 2, 2:18 pm 2008 |
| Kok, Auke | Re: GFP_ATOMIC page allocation failures.
FYI e1000 and family already count various levels of alloc failures resulting from
this:
alloc_rx_buff_failed - page alloc failure (might be harmless)
rx_no_buffer_count - no buffer available for HW to use (harmless, hw will retry)
rx_missed_errors - hw dropped a packet because of above failures
still I personally think the page alloc warnings are a good thing and we've had
several issues resolve quickly because of them.
shutting them up completely moves the focus to our driver which ...
| Apr 2, 2:37 pm 2008 |
| Stephen Hemminger | [PATCH] sky2: fix suspend/resume races
There are a couple of possible races on suspend/resume reported by Dell.
First the driver needs to block new packets from being queued for Tx.
The other less likely problem is the watchdog timer going off
during resume.
Note: there is no need to explicitly restart the watchdog timer
since it gets restarted by link up on resume.
Signed-off-by: Stephen Hemminger <shemminger@linux-foundation.org>
--- a/drivers/net/sky2.c 2008-04-01 11:42:56.000000000 -0700
+++ b/drivers/net/sky2.c 2008-04-0...
| Apr 2, 1:16 pm 2008 |
| Dan Noe | [PATCH (fixed)] cxgb3: Fix __must_check warning with dev_dbg.
Fix the warning:
drivers/net/cxgb3/cxgb3_main.c: In function ‘offload_open’:
drivers/net/cxgb3/cxgb3_main.c:936: warning: ignoring return value of
‘sysfs_create_group’, declared with attribute warn_unused_result
Now the return value is checked; if sysfs_create_group() returns failure,
a warning is printed using dev_dbg, and the code continues as before. Use
of dev_dbg ensures printk is not needlessly included unless desired for
debugging.
Signed-off-by: Dan Noe <dpn@isomerica.net>...
| Apr 2, 1:11 pm 2008 |
| Andrew Morton |
| Apr 2, 12:14 pm 2008 |
| Stephen Hemminger | Re: [Bugme-new] [Bug 10380] New: sc92031 does not create a d...
Add missing sysfs device association. Compile tested only -ENOHW.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
---
Patch against 2.6.25-rc8, ignore previous patch.
--- a/drivers/net/sc92031.c 2008-04-02 09:34:35.000000000 -0700
+++ b/drivers/net/sc92031.c 2008-04-02 09:47:41.000000000 -0700
@@ -1446,6 +1446,7 @@ static int __devinit sc92031_probe(struc
}
pci_set_drvdata(pdev, dev);
+ SET_NETDEV_DEV(dev, &pdev->dev);
#if SC92031_USE_BAR == 0
dev->mem_...
| Apr 2, 12:49 pm 2008 |
| Thomas Scheiblauer |
| Apr 2, 1:08 pm 2008 |
| Stephen Hemminger | [PATCH 3/3] sc92031: use netdev_alloc_skb
Use netdev_alloc_skb since it handles any NUMA node memory localtion issues
and sets skb->dev. Since device driver was not setting skb->dev, I bet
filter rules based on device would not work.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
--- a/drivers/net/sc92031.c 2008-04-02 10:03:00.000000000 -0700
+++ b/drivers/net/sc92031.c 2008-04-02 10:03:04.000000000 -0700
@@ -796,7 +796,7 @@ static void _sc92031_rx_tasklet(struct n
rx_len -= rx_size_align + 4;
- skb = dev...
| Apr 2, 1:13 pm 2008 |
| Stephen Hemminger | [PATCH 2/3] sc92031: start transmit return value bugfix
Any negative return value from start_xmit is interpreted as NETDEV_TX_LOCK
which is not what this driver wants. It should return 0 (NETDEV_TX_OK)
when it consumes a packet.
Also, use skb_padto() as the generic way to pad small frames.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
--- a/drivers/net/sc92031.c 2008-04-02 09:50:22.000000000 -0700
+++ b/drivers/net/sc92031.c 2008-04-02 09:53:15.000000000 -0700
@@ -947,16 +947,16 @@ static struct net_device_stats *sc92031_
stati...
| Apr 2, 1:11 pm 2008 |
| Stephen Hemminger | [PATCH 1/3] sc92031: use net_device stats
Statistics structure is available for use in net_device structure.
Compile tested only.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
---
Patch against 2.6.25-rc8 with previous SET_NETDEV_DEV patch
applied.
--- a/drivers/net/sc92031.c 2008-04-02 09:47:41.000000000 -0700
+++ b/drivers/net/sc92031.c 2008-04-02 09:49:34.000000000 -0700
@@ -311,7 +311,6 @@ struct sc92031_priv {
/* for dev->get_stats */
long rx_value;
- struct net_device_stats stats;
};
/* I don...
| Apr 2, 1:11 pm 2008 |
| Stephen Hemminger | [PATCH 1/3] sc92031: use net_device stats
Statistics structure is available for use in net_device structure.
Compile tested only.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
---
Patch against 2.6.25-rc8 with previous SET_NETDEV_DEV patch
applied.
--- a/drivers/net/sc92031.c 2008-04-02 09:47:41.000000000 -0700
+++ b/drivers/net/sc92031.c 2008-04-02 09:49:34.000000000 -0700
@@ -311,7 +311,6 @@ struct sc92031_priv {
/* for dev->get_stats */
long rx_value;
- struct net_device_stats stats;
};
/* I don...
| Apr 2, 1:08 pm 2008 |
| Stephen Hemminger | [PATCH] sc92031: sysfs link missing
Add missing sysfs device association. Compile tested only -ENOHW.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
--- a/drivers/net/sc92031.c 2008-04-02 09:34:35.000000000 -0700
+++ b/drivers/net/sc92031.c 2008-04-02 09:36:53.000000000 -0700
@@ -1446,6 +1446,7 @@ static int __devinit sc92031_probe(struc
}
pci_set_drvdata(pdev, dev);
+ SET_NETDEV_DEV(dev, &hw->pdev->dev);
#if SC92031_USE_BAR == 0
dev->mem_start = pci_resource_start(pdev, SC92031_USE_BAR...
| Apr 2, 12:39 pm 2008 |
| Andrew Morton | Re: GFP_ATOMIC page allocation failures.
Yes, it's useful debugging. It tells us when we mucked up the page
allocator.
It also tells us when we mucked up the net driver - I doubt if we (or at
least, I) would have discovered that e1000 does a 32k allocation for a 5k(?)
Yes, I suppose so.
hm, tricky.
--
| Apr 2, 11:54 am 2008 |
| Pavel Emelyanov | [PATCH net-2.6.26 3/3][TUN][NETNS]: Allow to register tun de...
This is basically means that a net is set for a new device, but
actually this involves two more steps:
1. mark the tun device as "local", i.e. do not allow for it to
move across namespaces.
This is done so, since tun device is most often associated to some
file (and thus to some process) and moving the device alone is not
valid while keeping the file and the process outside.
2. get the tun device's net when tun becomes attached and put one
when it becomes detached.
This is needed to h...
| Apr 2, 10:01 am 2008 |
| Max Krasnyanskiy | Re: [PATCH net-2.6.26 3/3][TUN][NETNS]: Allow to register tu...
Hi Pavel,
All three patches look perfectly fine to me.
Feel free to add Acked-by: Max Krasnyansky <maxk@qualcomm.com>
I'm not sure what you mean "by file lives on" here. I believe you're talking
about persistent tun devices. ie Those that exist in detached state and are
not attached to any file descriptors.
"net" refcounting logic there looks fine to me.
Max
--
| Apr 2, 6:54 pm 2008 |
| Pavel Emelyanov | [PATCH net-2.6.26 2/3][TUN][NETNS]: Actually make the tun_de...
Remove the static tun_dev_list and replace its occurrences in
driver with per-net one.
It is used in two places - in tun_set_iff and tun_cleanup. In
the first case it's legal to use current net_ns. In the cleanup
call - move the loop, that unregisters all devices in net exit
hook.
Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
---
drivers/net/tun.c | 35 +++++++++++++++++------------------
1 files changed, 17 insertions(+), 18 deletions(-)
diff --git a/drivers/net/tun.c b/driv...
| Apr 2, 9:55 am 2008 |
| Pavel Emelyanov | [PATCH net-2.6.26 1/3][TUN][NETNS]: Introduce the tun_net st...
This is the first step in making tuntap devices work in net
namespaces. The structure mentioned is pointed by struct net
and tun driver will fill this one on load. It will contain only
the tun devices list.
So declare this structure and introduce net init and exit hooks.
Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
---
drivers/net/tun.c | 43 ++++++++++++++++++++++++++++++++++++++++++-
include/net/net_namespace.h | 4 ++++
2 files changed, 46 insertions(+), 1 del...
| Apr 2, 9:50 am 2008 |
| Dan Noe | [PATCH] cxgb3: Fix __must_check warning with dev_dbg.
Fix the warning:
drivers/net/cxgb3/cxgb3_main.c: In function ‘offload_open’:
drivers/net/cxgb3/cxgb3_main.c:936: warning: ignoring return value of
‘sysfs_create_group’, declared with attribute warn_unused_result
Now the return value is checked; if sysfs_create_group() returns failure,
a warning is printed using dev_dbg, and the code continues as before. Use
of dev_dbg ensures printk is not needlessly included unless desired for
debugging.
Signed-off-by: Dan Noe <dpn@isomerica.net>...
| Apr 2, 7:50 am 2008 |
| Chris Snook |
| Apr 2, 11:49 am 2008 |
| Joakim Tjernlund | Quick patch for ucc_geth, add copybreak and move TX processi...
This is a quick hack to make ucc_geth handle a
ping -f -l 10 gracefully. Without it, the CPU will lockup during the ping flood.
I don't have time ATM to clean it up, but if anyone wants to he is welcome :)
Commenst on the geleral approach is wanted though, why do I need to move TX
processing to soft IRQ to make the system response under heavy load?
Kernel 2.6.23
Jocke
---
drivers/net/ucc_geth.c | 85 +++++++++++++++++++++++++++++++++--------------
1 files changed, 59 insertions(+), 26 del...
| Apr 2, 7:48 am 2008 |
| Matvejchikov Ilya |
| Apr 2, 5:50 am 2008 |
| Kamalesh Babulal | [BUILD_FAILURE] 2.6.25-rc8-mm1 build failure on x86_64 with ...
Hi Andrew,
The 2.6.25-rc8-mm1 kernel build fails on x86_64, when compiled with randconfig option
In file included from include/net/dst.h:15,
from include/net/sock.h:57,
from include/linux/if_pppox.h:145,
from fs/compat_ioctl.c:39:
include/net/neighbour.h: In function
| Apr 2, 5:02 am 2008 |
| KOSAKI Motohiro | Re: GFP_ATOMIC page allocation failures.
Hi
if network guys hope known-good driver should call
__netdev_alloc_skb(dev, length, GFP_ATOMIC|__GFP_NOWARN) instead netdev_alloc_skb,
I think we should make netdev_alloc_skb_nowarn.
static inline void *netdev_alloc_skb_nowarn(struct net_device *dev, unsigned int length)
{
return __netdev_alloc_skb(dev, length, GFP_ATOMIC|__GFP_NOWARN);
}
if not, we make unnecessary various alloc_skb
--
| Apr 2, 4:17 am 2008 |
| David Miller | Re: GFP_ATOMIC page allocation failures.
From: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Giving it a proper name like this takes away the indication that this
situation is very special.
Two leading underscores to an interface means "something special and
unusual requiring more careful consideration than usual is occuring
here." netdev_alloc_skb_nowarn() on the other hand, does not
convey this meaning.
And we will have very few drivers that use this construct, thus it
really is best to handle things the way Andrew has.
...
| Apr 2, 4:24 am 2008 |
| Peter Zijlstra |
| Apr 2, 7:04 am 2008 |
| David Miller |
| Apr 2, 2:45 pm 2008 |
| Jeff Garzik | Re: GFP_ATOMIC page allocation failures.
Correct...
In addition, I remain worried about potential edge cases in drivers that
do "burst refill" style allocations, rather than the [IMO safer]
as-you-go style that tg3 employs.
I haven't done a serious audit, but I have nagging doubts about the
behavior of burst-refill drivers when faced with a burst of failed
allocations... do they stop DMA operation, or do they permit hardware
to wander into part of the DMA ring where, in the previous "cycle"
through the ring, valid DMA addresse...
| Apr 2, 3:06 pm 2008 |
| KOSAKI Motohiro | Re: GFP_ATOMIC page allocation failures.
Hmmm,
at least, tg3 driver often output the same message on my test environment.
(because I often run stress test ;)
I doubt almost driver has the same problem under heavy workload.
and I think tg3 is known-good driver too.
because its alloc failure doesn't cause any bad thing at all.
Do you favorite the following patch?
---------------------------------------------------------------------
We get rather a lot of reports of page allocation warnings coming out of
tg3. But this driver is ...
| Apr 2, 6:00 am 2008 |
| KOSAKI Motohiro |
| Apr 2, 6:56 am 2008 |
| David Miller | Re: GFP_ATOMIC page allocation failures.
From: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Yes, I know about it. It will occur for any device
But modern chips are going to support segmentation of
receive packets into multiple individual pages and
therefore not have this order>0 page allocation
requirement.
That is my point.
--
| Apr 2, 2:44 pm 2008 |
| Michael Chan |
| Apr 2, 4:12 pm 2008 |
| David Miller |
| Apr 2, 4:53 pm 2008 |
| Andrew Morton |
| Apr 2, 4:43 am 2008 |
| YOSHIFUJI Hideaki / | [PATCH 2.6.26] [IPV6] RAW: Remove ancient comment.
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
---
net/ipv6/raw.c | 5 -----
1 files changed, 0 insertions(+), 5 deletions(-)
diff --git a/net/ipv6/raw.c b/net/ipv6/raw.c
index c1e495f..aae6ced 100644
--- a/net/ipv6/raw.c
+++ b/net/ipv6/raw.c
@@ -168,11 +168,6 @@ static int ipv6_raw_deliver(struct sk_buff *skb, int nexthdr)
read_lock(&raw_v6_hashinfo.lock);
sk = sk_head(&raw_v6_hashinfo.ht[hash]);
- /*
- * The first socket found will be delivered after
- ...
| Apr 2, 2:36 am 2008 |
| David Miller |
| Apr 2, 2:57 am 2008 |
