firefly-linux-kernel-4.4.55.git
16 years agoipv6: result of csum_fold() is already 16bit, no need to cast
Al Viro [Sun, 27 Apr 2008 05:28:58 +0000 (22:28 -0700)]
ipv6: result of csum_fold() is already 16bit, no need to cast

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[XFRM] AUDIT: Fix flowlabel text format ambibuity.
YOSHIFUJI Hideaki [Sun, 27 Apr 2008 05:24:10 +0000 (22:24 -0700)]
[XFRM] AUDIT: Fix flowlabel text format ambibuity.

Flowlabel text format was not correct and thus ambiguous.
For example, 0x00123 or 0x01203 are formatted as 0x123.
This is not what audit tools want.

Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agonet: Fix wrong interpretation of some copy_to_user() results.
Pavel Emelyanov [Fri, 25 Apr 2008 08:49:48 +0000 (01:49 -0700)]
net: Fix wrong interpretation of some copy_to_user() results.

I found some places, that erroneously return the value obtained from
the copy_to_user() call: if some amount of bytes were not able to get
to the user (this is what this one returns) the proper behavior is to
return the -EFAULT error, not that number itself.

Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agoMerge branch 'upstream-davem' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik...
David S. Miller [Fri, 25 Apr 2008 07:31:07 +0000 (00:31 -0700)]
Merge branch 'upstream-davem' of /linux/kernel/git/jgarzik/netdev-2.6

16 years agoxfrm: alg_key_len & alg_icv_len should be unsigned
Eric Dumazet [Fri, 25 Apr 2008 07:29:00 +0000 (00:29 -0700)]
xfrm: alg_key_len & alg_icv_len should be unsigned

In commit ba749ae98d5aa9d2ce9a7facde0deed454f92230 ([XFRM]: alg_key_len
should be unsigned to avoid integer divides
<http://git2.kernel.org/?p=linux/kernel/git/davem/net-2.6.git;a=commitdiff;h=ba749ae98d5aa9d2ce9a7facde0deed454f92230>)
alg_key_len field of struct xfrm_algo was converted to unsigned int to
avoid integer divides.

Then Herbert in commit 1a6509d991225ad210de54c63314fd9542922095
([IPSEC]: Add support for combined mode algorithms) added a new
structure xfrm_algo_aead, that resurrected a signed int for alg_key_len
and re-introduce integer divides.

This patch avoids these divides and saves 64 bytes of text on i386.

Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Acked-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[netdrvr] tehuti: move ioctl perm check closer to function start
Jeff Garzik [Fri, 25 Apr 2008 07:11:31 +0000 (03:11 -0400)]
[netdrvr] tehuti: move ioctl perm check closer to function start

Noticed by davem.

Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoipv6: Fix typo in net/ipv6/Kconfig
Michael Beasley [Fri, 25 Apr 2008 06:50:30 +0000 (23:50 -0700)]
ipv6: Fix typo in net/ipv6/Kconfig

Two is used in the wrong context here, as you are connecting to an
IPv6 network over IPv4; not connecting two IPv6 networks to an IPv4
one.

Signed-off-by: Michael Beasley <youvegotmoxie@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agovia-velocity: fix vlan receipt
Francois Romieu [Thu, 24 Apr 2008 21:32:33 +0000 (23:32 +0200)]
via-velocity: fix vlan receipt

- vlans were using a single CAM register (see mac_set_vlan_cam)
- setting the address filtering registers for vlans is not
  needed when there is no vlan

The non-tagged interface is filtered out as soon as a tagged
(!= 0) interface is created. Its traffic appears again when an
zero-tagged interface is created.

Tested on Via Epia SN (VT6130 chipset) with several vlans whose
tag was above or beyond 255.

Signed-off-by: Séguier Régis <rseguier@e-teleport.net>
Acked-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agotg3: sparse cleanup
Eric Dumazet [Fri, 25 Apr 2008 06:33:06 +0000 (23:33 -0700)]
tg3: sparse cleanup

Fix the following sparse warning :

drivers/net/tg3.c:4025:3: warning: context imbalance in 'tg3_restart_hw'
- unexpected unlock

Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agoforcedeth: realtek phy crossover detection
Ayaz Abdulla [Wed, 23 Apr 2008 18:37:30 +0000 (14:37 -0400)]
forcedeth: realtek phy crossover detection

This patch fixes an issue seen with the realtek 8201 phy. This phy has a
problem with crossover detection and it needs to be disabled. The
problem only arises on certain switches. Therefore, a module parameter
has been added to allow enabling crossover detection if needed. The
default will be set to disabled.

Signed-off-by: Ayaz Abdulla <aabdulla@nvidia.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoibm_newemac: Increase MDIO timeouts
Benjamin Herrenschmidt [Wed, 23 Apr 2008 05:17:14 +0000 (15:17 +1000)]
ibm_newemac: Increase MDIO timeouts

This patch doubles the MDIO timeouts in EMAC as there are field
cases where they are two short to communicate with some PHYs.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agogianfar: Fix skb allocation strategy
Andy Fleming [Tue, 22 Apr 2008 22:18:29 +0000 (17:18 -0500)]
gianfar: Fix skb allocation strategy

gianfar was unable to handle failed skb allocation for rx buffers, so
we were spinning until it succeeded.  Actually, it was worse--we were
spinning for a long time, and then silently failing.  Instead, we take
Stephen Hemminger's suggestion to try the allocation earlier, and drop the
packet if it failed.

We also make a couple of tweaks to how buffer descriptors are set up.

Signed-off-by: Andy Fleming <afleming@freescale.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agonetxen: reduce stack usage of netxen_nic_flash_print
Harvey Harrison [Tue, 22 Apr 2008 18:48:35 +0000 (11:48 -0700)]
netxen: reduce stack usage of netxen_nic_flash_print

Don't need to keep a struct netxen_new_user_info on the stack
when we only are interested in printing the serial_num.  Change
to only reading the serial_num.

Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agosmc911x: test after postfix decrement fails in smc911x_{reset,drop_pkt}
Roel Kluin [Fri, 18 Apr 2008 20:50:48 +0000 (13:50 -0700)]
smc911x: test after postfix decrement fails in smc911x_{reset,drop_pkt}

When timeout reaches 0 the postfix decrement still subtracts, so the test
fails.

[akpm@linux-foundation.org: coding-style fixes]
Signed-off-by: Roel Kluin <12o3l@tiscali.nl>
Acked-by: Peter Korsgaard <jacmet@sunsite.dk>
Cc: Jeff Garzik <jeff@garzik.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agonet drivers: fix platform driver hotplug/coldplug
Kay Sievers [Fri, 18 Apr 2008 20:50:44 +0000 (13:50 -0700)]
net drivers: fix platform driver hotplug/coldplug

Since 43cc71eed1250755986da4c0f9898f9a635cb3bf, the platform modalias is
prefixed with "platform:".  Add MODULE_ALIAS() to the hotpluggable network
platform drivers, to re-enable auto loading.

NOTE: didn't change drivers/net/fs_enet/fs_enet-main.c "old binding" support.
That looks problematic in the first place (it even uses the ancient "struct
device_driver" binding scheme for platform_bus!) and I suspect it will vanish
soonish when arch/powerpc rules the world.  Also, drivers/net/ne.c would have
needed more thought to sort out.

[akpm@linux-foundation.org: fix sgiseeq.c]
[dbrownell@users.sourceforge.net: more drivers, registration fixes]
Signed-off-by: Kay Sievers <kay.sievers@vrfy.org>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Cc: Scott Wood <scottwood@freescale.com>
Cc: Vitaly Bordug <vitb@kernel.crashing.org>
Cc: Dale Farnsworth <dale@farnsworth.org>
Cc: Ben Dooks <ben-linux@fluff.org>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Andrew Victor <andrew@sanpeople.com>
Cc: Bryan Wu <bryan.wu@analog.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoforcedeth: new backoff implementation
Ayaz Abdulla [Fri, 18 Apr 2008 20:50:43 +0000 (13:50 -0700)]
forcedeth: new backoff implementation

This patch adds support for a new backoff algorithm for half duplex supported
in newer hardware.  The old method is will be designated as legacy mode.

Re-seeding random values for the backoff algorithms are performed when a
transmit has failed due to a maximum retry count (1 to 15, where max is
considered the wraparound case of 0).

Signed-off-by: Ayaz Abdulla <aabdulla@nvidia.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoehea: make things static
Andrew Morton [Fri, 18 Apr 2008 20:50:39 +0000 (13:50 -0700)]
ehea: make things static

ehea_flush_sq() and ehea_purge_sq() should be static.

Cc: Jeff Garzik <jeff@garzik.org>
Cc: Thomas Klein <osstklei@de.ibm.com>
Cc: Thomas Klein <tklein@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agophylib: Add support for board-level PHY fixups
Andy Fleming [Fri, 18 Apr 2008 22:29:54 +0000 (17:29 -0500)]
phylib: Add support for board-level PHY fixups

Sometimes the specific interaction between the platform and the PHY
requires special handling.  For instance, to change where the PHY's
clock input is, or to add a delay to account for latency issues in the
data path.  We add a mechanism for registering a callback with the PHY
Lib to be called on matching PHYs when they are brought up, or reset.

Signed-off-by: Andy Fleming <afleming@freescale.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years ago[netdrvr] atlx: code movement: move atl1 parameter parsing
Chris Snook [Sat, 19 Apr 2008 01:51:53 +0000 (21:51 -0400)]
[netdrvr] atlx: code movement: move atl1 parameter parsing

Move some code from atlx.c to atl1.c to prevent build conflict with
the upcoming atl2 code.  No changes, just movement.

Signed-off-by: Chris Snook <csnook@redhat.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoatlx: remove flash vendor parameter
Chris Snook [Sat, 19 Apr 2008 01:47:41 +0000 (21:47 -0400)]
atlx: remove flash vendor parameter

There's no good reason to manually set the flash vendor in a module
parameter, outside of an Atheros hardware lab.  Remove it, so nobody
accidentally bricks their board using it incorrectly.

Signed-off-by: Chris Snook <csnook@redhat.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agokorina: misc cleanup
Francois Romieu [Sun, 20 Apr 2008 16:06:13 +0000 (18:06 +0200)]
korina: misc cleanup

- useless initialization (korina_ope / korina_restart)
- use a single variable for the status code in korina_probe
  and propagate the error status code from below
- useless checks in korina_remove : the variables are
  necessarily set when korina_probe succeeds

Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agokorina: fix misplaced return statement
Francois Romieu [Sun, 20 Apr 2008 16:05:31 +0000 (18:05 +0200)]
korina: fix misplaced return statement

The driver takes the error unwind path without condition.

Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoWAN: Fix confusing insmod error code for C101 too.
Krzysztof Halasa [Sun, 20 Apr 2008 17:10:56 +0000 (19:10 +0200)]
WAN: Fix confusing insmod error code for C101 too.

Signed-off-by: Krzysztof Hałasa <khc@pm.waw.pl>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoRemoving dead code in drivers/net/wan/hdlc_fr.c
Cédric Augonnet [Sun, 20 Apr 2008 17:15:51 +0000 (19:15 +0200)]
Removing dead code in drivers/net/wan/hdlc_fr.c

The local variable "prefix" is never used anymore, and the content of
this string appears a bit later, directly in a call to "alloc_netdev"
after doing exactly the same if/else test. So there seems to be no
point keeping those 4 lines anymore.

Signed-off-by: Cédric Augonnet <cedric.augonnet@ens-lyon.org>
Signed-off-by: Krzysztof Hałasa <khc@pm.waw.pl>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agotehuti: check register size
Francois Romieu [Sun, 20 Apr 2008 17:32:34 +0000 (19:32 +0200)]
tehuti: check register size

Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years ago[netdrvr] typhoon: typhoon_resume - remove call to start_queue
Grant Grundler [Mon, 21 Apr 2008 04:44:15 +0000 (22:44 -0600)]
[netdrvr] typhoon: typhoon_resume - remove call to start_queue

While trying to fix http://bugzilla.kernel.org/show_bug.cgi?id=8952
I looked at a few other drivers to figure out what drivers _should_
be doing for suspend/resume. I noticed typhoon driver is likely doing
more than it needs to.  Patch below is untested since I don't have the HW.

Suspend/resume code across NIC drivers is fairly inconsistent.
And I couldn't find any documentation on what the canonical sequence
NICs need to do for suspend or resume.  Is there any?

Barring contrary advice, I'm going model the tulip suspend/resume
fixes after tg3.c since a number of "modern" (< 5 years old) laptops
have that and I'm silly enough to assume it works.

Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoibm_newemac: PowerPC 440EP/440GR EMAC PHY clock workaround
Valentine Barshak [Tue, 22 Apr 2008 00:46:48 +0000 (10:46 +1000)]
ibm_newemac: PowerPC 440EP/440GR EMAC PHY clock workaround

This patch adds ibm_newemac PHY clock workaround for 440EP/440GR EMAC
attached to a PHY which doesn't generate RX clock if there is no link.
The code is based on the previous ibm_emac driver stuff. The 440EP/440GR
allows controlling each EMAC clock separately as opposed to global clock
selection for 440GX.

BenH: Made that #ifdef CONFIG_PPC_DCR_NATIVE for now as dcri_* stuff doesn't
exist for MMIO type DCRs like Cell. Some future rework & improvements of the
DCR infrastructure will make that cleaner but for now, this makes it work.

Signed-off-by: Valentine Barshak <vbarshak@ru.mvista.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoibm_newemac: PowerPC 440GX EMAC PHY clock workaround
Valentine Barshak [Tue, 22 Apr 2008 00:46:46 +0000 (10:46 +1000)]
ibm_newemac: PowerPC 440GX EMAC PHY clock workaround

The PowerPC 440GX Taishan board fails to reset EMAC3 (reset timeout
error) if there's no link. Because of that it fails to find PHY
chip. The older ibm_emac driver had a workaround for that: the
EMAC_CLK_INTERNAL/EMAC_CLK_EXTERNAL macros, which toggle the Ethernet
Clock Select bit in the SDR0_MFR register. This patch does the same for
"ibm,emac-440gx" compatible chips. The workaround forces clock on -all-
EMACs, so we select clock under global emac_phy_map_lock.

BenH: Made that #ifdef CONFIG_PPC_DCR_NATIVE for now as dcri_* stuff
doesn't exist for MMIO type DCRs like Cell. Some future rework &
improvements of the DCR infrastructure will make that cleaner but
for now, this makes it work.

Signed-off-by: Valentine Barshak <vbarshak@ru.mvista.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoibm_newemac Use status property for unused/unwired EMACs
Josh Boyer [Tue, 22 Apr 2008 00:46:44 +0000 (10:46 +1000)]
ibm_newemac Use status property for unused/unwired EMACs

Convert ibm_newemac to use the of_device_is_available function when checking
for unused/unwired EMACs.  We leave the current check for an "unused" property
to maintain backwards compatibility for older device trees.  Newer device
trees should simply use the standard "status" property in the EMAC node.

Signed-off-by: Josh Boyer <jwboyer@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoibm_newemac: Fix section mismatch warnings
Josh Boyer [Tue, 22 Apr 2008 00:46:43 +0000 (10:46 +1000)]
ibm_newemac: Fix section mismatch warnings

This patch fixes several section mismatch warnings in the
ibm_newemac driver similar to:

WARNING: vmlinux.o(.devinit.text+0x3a04): Section mismatch in reference from the function emac_probe() to the function .devexit.text:tah_detach()
The function __devinit emac_probe() references
a function __devexit tah_detach().

Signed-off-by: Josh Boyer <jwboyer@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoibm_newemac: Add support for 460EX/GT-type MAL rx-channel handling
Stefan Roese [Tue, 22 Apr 2008 00:46:42 +0000 (10:46 +1000)]
ibm_newemac: Add support for 460EX/GT-type MAL rx-channel handling

On some 4xx PPC's (e.g. 460EX/GT), the rx channel number is a multiple
of 8 (e.g. 8 for EMAC1, 16 for EMAC2), but enabling in MAL_RXCASR needs
the divided by 8 value for the bitmask.

Signed-off-by: Stefan Roese <sr@denx.de>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoibm_newemac: Fix problem with jumbo frame support and EMAC V4.patch
Stefan Roese [Tue, 22 Apr 2008 00:46:42 +0000 (10:46 +1000)]
ibm_newemac: Fix problem with jumbo frame support and EMAC V4.patch

This fixes the jumbo frame support on EMAC V4 systems. Now the correct
bit is set depending on the EMAC version configured.

Tested on Kilauea (405EX) and Canyonlands (460EX).

Signed-off-by: Stefan Roese <sr@denx.de>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoe1000e: Increment version to 0.2.1
Jesse Brandeburg [Wed, 23 Apr 2008 18:09:39 +0000 (11:09 -0700)]
e1000e: Increment version to 0.2.1

Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Auke Kok <auke-jan.h.kok@intel.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoigb: save and restore pcie/msi state to support EEH recovery
Auke Kok [Wed, 23 Apr 2008 18:09:34 +0000 (11:09 -0700)]
igb: save and restore pcie/msi state to support EEH recovery

To enable EEH support for pci-express network adapters, pcie/msi state
needs to be saved and restored for that adapter.

[after similar patches for ixgbe and e1000e from Wendy Xiong]

Signed-off-by: Auke Kok <auke-jan.h.kok@intel.com>
Cc: Wendy Xiong <wendyx@us.ibm.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoe1000e: save and restore pcie/msi state to support EEH recovery
Wendy Xiong [Wed, 23 Apr 2008 18:09:29 +0000 (11:09 -0700)]
e1000e: save and restore pcie/msi state to support EEH recovery

To enable EEH support for pci-express network adapters, pcie/msi state
needs to be saved and restored for that adapter.

Tested this EEH patch with 2ports and 4ports pci-express e1000e
adapters.

Signed-off-by: Wendy Xiong <wendyx@us.ibm.com>
Signed-off-by: Auke Kok <auke-jan.h.kok@intel.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoixgbe: save and restore pcie/msi state to support EEH recovery
Wendy Xiong [Wed, 23 Apr 2008 18:09:24 +0000 (11:09 -0700)]
ixgbe: save and restore pcie/msi state to support EEH recovery

To enable EEH support for pci-express network adapters, pcie/msi state
needs to be saved and restored for that adapter.

Tested this EEH patch with Intel 10G pci-express ixgbe adapter.

Signed-off-by: Wendy Xiong <wendyx@us.ibm.com>
Signed-off-by: Auke Kok <auke-jan.h.kok@intel.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoe1000e: lower ring minimum size to 64
Auke Kok [Wed, 23 Apr 2008 18:09:19 +0000 (11:09 -0700)]
e1000e: lower ring minimum size to 64

The lower limit of 80 descriptors in the ring is only valid for
one older 8254x chipset. All e1000e devices can use as low as
64 descriptors.

Signed-off-by: Auke Kok <auke-jan.h.kok@intel.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoe1000e: Fix HW Error on es2lan, ARP capture issue by BMC
David Graham [Wed, 23 Apr 2008 18:09:14 +0000 (11:09 -0700)]
e1000e: Fix HW Error on es2lan, ARP capture issue by BMC

Several components to this complex fix. The es2lan cards occasionally
gave a "HW Error" especially when forcing speed. Some users also
reported that the BMC stole ARP packets.

The fixes include setting the proper SW_FW bits to tell the BMC
that we're active and not do any un-initialization at all, so the
setup routine is largely changed.

Signed-off-by: David Graham <david.graham@intel.com>
Signed-off-by: Auke Kok <auke-jan.h.kok@intel.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoe1000e: Add interrupt moderation run-time ethtool interface
Auke Kok [Wed, 23 Apr 2008 18:09:08 +0000 (11:09 -0700)]
e1000e: Add interrupt moderation run-time ethtool interface

The ethtool -c / -C interface can now be used to modify the
irq moderation algorithm. This change does not require an
adapter reset and can thus be used at all times. The adapter
only supports changing/reading rx-usecs which has special
values for 0, 1 and 3:

0 - no irq moderation whatsoever
1 - normal moderation favoring regular mixed traffic (default)
3 - best attempt at low latency possible at cost of CPU

For values between 10 and 10000 the rx-usecs defines "the minimum
time between successive irqs" in usec, unlike the module parameter.

Signed-off-by: Auke Kok <auke-jan.h.kok@intel.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoe1000e: cleanup several stats issues
Bruce Allan [Wed, 23 Apr 2008 18:09:00 +0000 (11:09 -0700)]
e1000e: cleanup several stats issues

Several stats registers are completely unused and we just waste pci
bus time reading them. We also omit using the high 32 bits of the GORC/
GOTC counters. We can just read clear them and only read the low registers.

Mii-tool can also break es2lan if it executes a MII PHY register
ioctl while the device is in autonegotiation. Unfortunately it seems
that several applications and installations still perform this ioctl
call periodically and especially in this crucial startup time. We
can fool the ioctl by providing fail safe information that mimics
the "down" link state and only perform the dangerous PHY reads once
after link comes up to fill in the real values. As long as link
stays up the information will not change.

Signed-off-by: Bruce Allan <bruce.w.allan@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Auke Kok <auke-jan.h.kok@intel.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoS2io: Removed rx_lock and put_lock
Sreenivasa Honnur [Wed, 23 Apr 2008 17:31:33 +0000 (13:31 -0400)]
S2io: Removed rx_lock and put_lock

- Removed rx_lock and put_lock as the buffer replenishment and
  receive completion is handled serially.

Signed-off-by: Surjit Reang <surjit.reang@neterion.com>
Signed-off-by: Sreenivasa Honnur <sreenivasa.honnur@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoS2io: Removed receive buffer replenishment tasklet
Sreenivasa Honnur [Wed, 23 Apr 2008 17:29:42 +0000 (13:29 -0400)]
S2io: Removed receive buffer replenishment tasklet

- Removed receive buffer replenishment tasklet s2io_tasklet and instead
  allocating the receive buffers in either the interrupt handler (no napi)
  or the napi handler (napi enabled).

Signed-off-by: Surjit Reang <surjit.reang@neterion.com>
Signed-off-by: Sreenivasa Honnur <sreenivasa.honnur@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoS2io: Version update for memory leak fix during free_tx_buffers
Sreenivasa Honnur [Wed, 23 Apr 2008 17:28:58 +0000 (13:28 -0400)]
S2io: Version update for memory leak fix during free_tx_buffers

- Updated version number.

Signed-off-by: Santosh Rastapur <santosh.rastapur@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoS2io: Fix memory leak during free_tx_buffers
Sreenivasa Honnur [Wed, 23 Apr 2008 17:28:08 +0000 (13:28 -0400)]
S2io: Fix memory leak during free_tx_buffers

- Fix the memory leak during free_tx_buffers.

Signed-off-by: Santosh Rastapur <santosh.rastapur@neterion.com>
Signed-off-by: Ramkrishna Vepa <ram.vepa@neterion.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoucc_geth: Fix sneaky merge conflict regarding bus_id
Andy Fleming [Wed, 23 Apr 2008 21:56:17 +0000 (16:56 -0500)]
ucc_geth: Fix sneaky merge conflict regarding bus_id

The patch that changed mdio_bus to a string didn't conflict strongly enough
with the patch that added fixed PHY support to UCC.  Gather it back into
the fold.

Fixes this error:
...
 CC      drivers/net/ucc_geth.o
 'ucc_geth_probe':
 /home/bunk/linux/kernel-2.6/git/linux-2.6/drivers/net/ucc_geth.c:3935: error:
 incompatible types in assignment
 make[3]: *** [drivers/net/ucc_geth.o] Error 1

Signed-off-by: Andy Fleming <afleming@freescale.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
16 years agoipv6 RAW: Disallow IPPROTO_IPV6-level IPV6_CHECKSUM socket option on ICMPv6 sockets.
YOSHIFUJI Hideaki [Fri, 25 Apr 2008 04:30:38 +0000 (21:30 -0700)]
ipv6 RAW: Disallow IPPROTO_IPV6-level IPV6_CHECKSUM socket option on ICMPv6 sockets.

RFC3542 tells that IPV6_CHECKSUM socket option in the IPPROTO_IPV6
level is not allowed on ICMPv6 sockets.  IPPROTO_RAW level
IPV6_CHECKSUM socket option (a Linux extension) is still allowed.

Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agotcp: tcp_probe buffer overflow and incorrect return value
Tom Quetchenbach [Fri, 25 Apr 2008 04:11:58 +0000 (21:11 -0700)]
tcp: tcp_probe buffer overflow and incorrect return value

tcp_probe has a bounds-checking bug that causes many programs (less,
python) to crash reading /proc/net/tcp_probe. When it outputs a log
line to the reader, it only checks if that line alone will fit in the
reader's buffer, rather than that line and all the previous lines it
has already written.

tcpprobe_read also returns the wrong value if copy_to_user fails--it
just passes on the return value of copy_to_user (number of bytes not
copied), which makes a failure look like a success.

This patch fixes the buffer overflow and sets the return value to
-EFAULT if copy_to_user fails.

Patch is against latest net-2.6; tested briefly and seems to fix the
crashes in less and python.

Signed-off-by: Tom Quetchenbach <virtualphtn@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agoniu: Add support for Neptune FEM/NEM cards for C10 server blades
Matheos Worku [Fri, 25 Apr 2008 04:09:20 +0000 (21:09 -0700)]
niu: Add support for Neptune FEM/NEM cards for C10 server blades

[ Minor coding style and whitespace corrections, also bump
  driver version and release date. -DaveM ]

Signed-off-by: Matheos Worku <matheos.worku@sun.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agoniu: Determine the # of ports from the card's VPD data
Matheos Worku [Fri, 25 Apr 2008 04:02:37 +0000 (21:02 -0700)]
niu: Determine the # of ports from the card's VPD data

[ Fix minor whitespace and coding style stuff... -DaveM ]

Signed-off-by: Matheos Worku <matheos.worku@sun.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agoethtool: EEPROM dump no longer works for tg3 and natsemi
Mandeep Singh Baines [Fri, 25 Apr 2008 03:55:56 +0000 (20:55 -0700)]
ethtool: EEPROM dump no longer works for tg3 and natsemi

In the ethtool user-space application, tg3 and natsemi over-ride the
default implementation of dump_eeprom(). In both tg3_dump_eeprom() and
natsemi_dump_eeprom(), there is a magic number check which is not
present in the default implementation.

Commit b131dd5d ("[ETHTOOL]: Add support for large eeproms") snipped
the code which copied the ethtool_eeprom structure back to
user-space. tg3 and natsemi are over-writing the magic number field
and then checking it in user-space. With the ethtool_eeprom copy
removed, the check is failing.

The fix is simple. Add the ethtool_eeprom copy back.

Signed-off-by: Mandeep Singh Baines <msb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agoaf_key: Fix af_key.c compiler warning
Brian Haley [Fri, 25 Apr 2008 03:38:31 +0000 (20:38 -0700)]
af_key: Fix af_key.c compiler warning

net/key/af_key.c: In function ‘pfkey_spddelete’:
net/key/af_key.c:2359: warning: ‘pol_ctx’ may be used uninitialized in
this function

When CONFIG_SECURITY_NETWORK_XFRM isn't set,
security_xfrm_policy_alloc() is an inline that doesn't set pol_ctx, so
this seemed like the easiest fix short of using *uninitialized_var(pol_ctx).

Signed-off-by: Brian Haley <brian.haley@hp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agotun: Multicast handling in tun_chr_ioctl() needs proper locking.
David S. Miller [Thu, 24 Apr 2008 10:44:43 +0000 (03:44 -0700)]
tun: Multicast handling in tun_chr_ioctl() needs proper locking.

Since these operations don't go through the normal
device calls, we have to ensure we synchronize with
those paths.

Noticed by Alan Cox.

Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[NET]: Fix heavy stack usage in seq_file output routines.
Pavel Emelyanov [Thu, 24 Apr 2008 08:02:16 +0000 (01:02 -0700)]
[NET]: Fix heavy stack usage in seq_file output routines.

Plan C: we can follow the Al Viro's proposal about %n like in this patch.
The same applies to udp, fib (the /proc/net/route file), rt_cache and
sctp debug. This is minus ~150-200 bytes for each.

Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[AF_UNIX] Initialise UNIX sockets before general device initcalls
David Woodhouse [Thu, 24 Apr 2008 07:59:25 +0000 (00:59 -0700)]
[AF_UNIX] Initialise UNIX sockets before general device initcalls

When drivers call request_module(), it tries to do something with UNIX
sockets and triggers a 'runaway loop modprobe net-pf-1' warning. Avoid
this by initialising AF_UNIX support earlier.

Signed-off-by: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[RTNETLINK]: Fix bogus ASSERT_RTNL warning
Patrick McHardy [Thu, 24 Apr 2008 05:10:48 +0000 (22:10 -0700)]
[RTNETLINK]: Fix bogus ASSERT_RTNL warning

ASSERT_RTNL uses mutex_trylock to test whether the rtnl_mutex is
held. This bogus warnings when running in atomic context, which
f.e. happens when adding secondary unicast addresses through
macvlan or vlan or when synchronizing multicast addresses from
wireless devices.

Mid-term we might want to consider moving all address updates
to process context since the locking seems overly complicated,
for now just fix the bogus warning by changing ASSERT_RTNL to
use mutex_is_locked().

Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agoMerge branch 'master' of master.kernel.org:/pub/scm/linux/kernel/git/linville/wireles...
David S. Miller [Thu, 24 Apr 2008 02:44:58 +0000 (19:44 -0700)]
Merge branch 'master' of /linux/kernel/git/linville/wireless-2.6

16 years agoiwlwifi: Fix built-in compilation of iwlcore (part 2)
Zhu Yi [Thu, 24 Apr 2008 02:44:13 +0000 (19:44 -0700)]
iwlwifi: Fix built-in compilation of iwlcore (part 2)

On Wed, 2008-04-23 at 13:38 +0300, Tomas Winkler wrote:
> This patch fixes problem in Makefile that prevented
> built-in compilation of iwlcore

Here is the second part. Without this,
drivers/net/wireless/iwlwifi/build-in.o will not be linked into vmlinux.

Signed-off-by: Zhu Yi <yi.zhu@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agotun: Fix minor race in TUNSETLINK ioctl handling.
David S. Miller [Thu, 24 Apr 2008 02:37:58 +0000 (19:37 -0700)]
tun: Fix minor race in TUNSETLINK ioctl handling.

Noticed by Alan Cox.

The IFF_UP test is a bit racey, because other entities
outside of this driver's ioctl handler can modify that
state, even though this ioctl handler runs under
lock_kernel().

Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agoppp_generic: use stats from net_device structure
Paulius Zaleckas [Thu, 24 Apr 2008 01:54:01 +0000 (18:54 -0700)]
ppp_generic: use stats from net_device structure

Use stats which now is in the net_device instead of one declared in
ppp structure.
Kill ppp_net_stats function, because by default it is used identical
internal_stats function from net/core/dev.c

Signed-of-by: Paulius Zaleckas <paulius.zaleckas@teltonika.lt>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agoiwlwifi: Don't unlock priv->mutex if it isn't locked
Roland Dreier [Wed, 23 Apr 2008 23:22:13 +0000 (16:22 -0700)]
iwlwifi: Don't unlock priv->mutex if it isn't locked

Commit b716bb91 ("iwlwifi: Cancel scanning upon association") moved the
test of priv->vif in iwl{3945,4964}_mac_config_interface() outside of
where priv->mutex is held, but still tries to do mutex_unlock() on
return.  This is clearly wrong and triggers a nasty lockdep warning when
this codepath is triggered.  Fix this by removing the mutex_unlock().

Signed-off-by: Roland Dreier <rolandd@cisco.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
16 years agowireless: rndis_wlan: modparam_workaround_interval is never below 0.
Roel Kluin [Wed, 23 Apr 2008 20:10:29 +0000 (22:10 +0200)]
wireless: rndis_wlan: modparam_workaround_interval is never below 0.

priv->param_workaround_interval is unsigned, modparam_workaround_interval not.
the former is never < 0.

Signed-off-by: Roel Kluin <12o3l@tiscali.nl>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
16 years agoprism54: prism54_get_encode() test below 0 on unsigned index
Roel Kluin [Wed, 23 Apr 2008 19:56:49 +0000 (21:56 +0200)]
prism54: prism54_get_encode() test below 0 on unsigned index

previously in this function:

u32 index = (dwrq->flags & IW_ENCODE_INDEX) - 1;

index is unsigned, so if -1, the original test (below) didn't work.

Signed-off-by: Roel Kluin <12o3l@tiscali.nl>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
16 years agomac80211: update mesh EID values
Luis Carlos Cobo [Wed, 23 Apr 2008 19:34:59 +0000 (12:34 -0700)]
mac80211: update mesh EID values

This patch updates mesh EID values, some of which where conflicting with
already-approved 11h EIDs (pointed out by Tomas Winkler). I wanted to use the
values suggested in the last available 802.11 draft (2.0) but it assigns 50 to
MESH_CONFIG, the same value than EXT_SUPP_RATES. Using the values proposed in
the draft incremented by one.

Signed-off-by: Luis Carlos Cobo <luisca@cozybit.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
16 years agob43: Workaround DMA quirks
Michael Buesch [Wed, 23 Apr 2008 17:13:01 +0000 (19:13 +0200)]
b43: Workaround DMA quirks

Some mainboards/CPUs don't allow DMA masks bigger than a certain limit.
Some VIA crap^h^h^h^hdevices have an upper limit of 0xFFFFFFFF. So in this
case a 64-bit b43 device would always fail to acquire the mask.
Implement a workaround to fallback to lower DMA mask, as we can always
also support a lower mask.

Signed-off-by: Michael Buesch <mb@bu3sch.de>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
16 years agomac80211: fix use before check of Qdisc length
Ron Rindjunsky [Wed, 23 Apr 2008 10:45:12 +0000 (13:45 +0300)]
mac80211: fix use before check of Qdisc length

This patch fixes use of Qdisc length in requeue function, before we checked
the reference is valid. (Adrian Bunk's catch)

Signed-off-by: Ron Rindjunsky <ron.rindjunsky@intel.com>
Signed-off-by: Adrian Bunk <bunk@kernel.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
16 years agonet/mac80211/rx.c: fix off-by-one
Adrian Bunk [Wed, 23 Apr 2008 09:51:28 +0000 (12:51 +0300)]
net/mac80211/rx.c: fix off-by-one

This patch fixes an off-by-one in net/mac80211/rx.c introduced by
commit 8318d78a44d49ac1edf2bdec7299de3617c4232e
(cfg80211 API for channels/bitrates, mac80211 and driver conversion)
and spotted by the Coverity checker.

Signed-off-by: Adrian Bunk <bunk@kernel.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
16 years agomac80211: Fix race between ieee80211_rx_bss_put and lookup routines.
Pavel Emelyanov [Wed, 23 Apr 2008 07:47:15 +0000 (11:47 +0400)]
mac80211: Fix race between ieee80211_rx_bss_put and lookup routines.

The put routine first decrements the users counter and then
(if it is zero) locks the sta_bss_lock and removes one from
the list and the hash.

Thus, any of ieee80211_sta_config_auth, ieee80211_rx_bss_get
or ieee80211_rx_mesh_bss_get can race with it by finding a
bss that is about to get kfree-ed.

Using atomic_dec_and_lock in ieee80211_rx_bss_put takes care
of this race.

Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
16 years agoath5k: Fix radio identification on AR5424/2424
Nick Kossifidis [Mon, 21 Apr 2008 18:28:24 +0000 (21:28 +0300)]
ath5k: Fix radio identification on AR5424/2424

 *Fix radio chip identification on AR5424/2424 during ath5k_hw_attach
 *Try to assign an RF2413 radio on AR2424 for testing

 Changes-licensed-under: ISC

Signed-off-by: Nick Kossifidis <mickflemm@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
16 years agossb: Fix all-ones boardflags
Larry Finger [Sat, 19 Apr 2008 15:52:12 +0000 (17:52 +0200)]
ssb: Fix all-ones boardflags

In the SSB SPROM a field set to all ones means the value
is not defined in the SPROM.
In case of the boardflags, we need to set them to zero
to avoid confusing drivers. Drivers will only check the
flags by ANDing.

Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: Gabor Stefanik <netrolller.3d@gmail.com>
Signed-off-by: Michael Buesch <mb@bu3sch.de>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
16 years agob43: Add more btcoexist workarounds
Michael Buesch [Sat, 19 Apr 2008 14:53:00 +0000 (16:53 +0200)]
b43: Add more btcoexist workarounds

This adds more workarounds for devices with broken BT bits.

Signed-off-by: Michael Buesch <mb@bu3sch.de>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
16 years agob43: Fix HostFlags data types
Michael Buesch [Fri, 18 Apr 2008 19:06:37 +0000 (21:06 +0200)]
b43: Fix HostFlags data types

The HostFlags are a bitmask of 48bit. So we must use an u64 datatype
to hold all bits.

Signed-off-by: Michael Buesch <mb@bu3sch.de>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
16 years agob43: Workaround invalid bluetooth settings
Michael Buesch [Fri, 18 Apr 2008 18:51:41 +0000 (20:51 +0200)]
b43: Workaround invalid bluetooth settings

This adds a workaround for invalid bluetooth SPROM settings
on ASUS PCI cards.
This will stop the microcode from poking with the BT GPIO line.
This fixes data transmission on this device, as the BT GPIO line
is used for something TX related on this device
(probably the power amplifier or the radio).
This also adds a modparam knob to help debugging this in the future,
as more devices with this bug may show up.

Signed-off-by: Michael Buesch <mb@bu3sch.de>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
16 years agomac80211: MAINTAINERS update
Johannes Berg [Thu, 17 Apr 2008 17:36:28 +0000 (19:36 +0200)]
mac80211: MAINTAINERS update

This updates the mac80211 maintainers entry to
 1) remove Jiri
 2) put me first

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Cc: Michael Wu <flamingice@sourmilk.net>
Cc: Jiri Benc <jbenc@suse.cz>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
16 years agomac80211: Fix n-band association problem
Abhijeet Kolekar [Wed, 16 Apr 2008 21:02:04 +0000 (14:02 -0700)]
mac80211: Fix n-band association problem

There are two structures named wmm_info and wmm_param, they are used while
parsing the beacon frame. (Check the function ieee802_11_parse_elems).
Certain APs like D-link does not set the fifth bit in WMM IE.
While sending the association request to n-only ap it checks for wmm_ie.
If it is set then only ieee80211_ht_cap is sent during association request.
So n-only association fails.
And this patch fixes this problem by copying the wmm_info to wmm_ie,
which enables the "wmm" flag in iee80211_send_assoc.

Signed-off-by: Abhijeet Kolekar <abhijeet.kolekar@intel.com>
Acked-by: Ron Rindjunsky <ron.rindjunsky@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
16 years agoiwlwifi: Fix built-in compilation of iwlcore
Tomas Winkler [Wed, 23 Apr 2008 10:48:57 +0000 (03:48 -0700)]
iwlwifi: Fix built-in compilation of iwlcore

This patch fixes problem in Makefile that prevented
built-in compilation of iwlcore

Commit that caused this problem: eadd3c4b ("iwlwifi: make Makefile
more concise")

Signed-off-by: Tomas Winkler <tomas.winkler@intel.com>
Signed-off-by: Yi Zhu <yi.zhu@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agonet: Unexport move_addr_to_{kernel,user}
Adrian Bunk [Wed, 23 Apr 2008 10:37:49 +0000 (03:37 -0700)]
net: Unexport move_addr_to_{kernel,user}

After the removal of the Solaris binary emulation the exports of
move_addr_to_{kernel,user} are no longer used.

Signed-off-by: Adrian Bunk <bunk@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agort2x00: Select LEDS_CLASS.
David S. Miller [Wed, 23 Apr 2008 10:34:50 +0000 (03:34 -0700)]
rt2x00: Select LEDS_CLASS.

Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agoiwlwifi: Select LEDS_CLASS.
David S. Miller [Wed, 23 Apr 2008 10:34:31 +0000 (03:34 -0700)]
iwlwifi: Select LEDS_CLASS.

Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agoleds: Do not guard NEW_LEDS with HAS_IOMEM
David S. Miller [Wed, 23 Apr 2008 10:33:32 +0000 (03:33 -0700)]
leds: Do not guard NEW_LEDS with HAS_IOMEM

The LEDS infrastructure itself does not require anything
that a platform dependant upon HAS_IOMEM.

The individual drivers do, but they are properly guarded
with the necessary platform dependencies.

One can even imagine a hypervisor based LED driver that
a platform without HAS_IOMEM might have.

Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[IPSEC]: Fix catch-22 with algorithm IDs above 31
Herbert Xu [Tue, 22 Apr 2008 07:46:42 +0000 (00:46 -0700)]
[IPSEC]: Fix catch-22 with algorithm IDs above 31

As it stands it's impossible to use any authentication algorithms
with an ID above 31 portably.  It just happens to work on x86 but
fails miserably on ppc64.

The reason is that we're using a bit mask to check the algorithm
ID but the mask is only 32 bits wide.

After looking at how this is used in the field, I have concluded
that in the long term we should phase out state matching by IDs
because this is made superfluous by the reqid feature.  For current
applications, the best solution IMHO is to allow all algorithms when
the bit masks are all ~0.

The following patch does exactly that.

This bug was identified by IBM when testing on the ppc64 platform
using the NULL authentication algorithm which has an ID of 251.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agotime: Export set_normalized_timespec.
YOSHIFUJI Hideaki [Tue, 22 Apr 2008 02:45:12 +0000 (19:45 -0700)]
time: Export set_normalized_timespec.

Sorry I have just realized set_normalized_timespec() (used in
timespec_sub()) is not exported, and link will fail because of it...

Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agotcp: Make use of before macro in tcp_input.c
Arnd Hannemann [Mon, 21 Apr 2008 21:46:22 +0000 (14:46 -0700)]
tcp: Make use of before macro in tcp_input.c

Make use of tcp before macro.

Signed-off-by: Arnd Hannemann <hannemann@nets.rwth-aachen.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agohamradio: Remove unneeded and deprecated cli()/sti() calls in dmascc.c
Mark Asselstine [Mon, 21 Apr 2008 21:44:16 +0000 (14:44 -0700)]
hamradio: Remove unneeded and deprecated cli()/sti() calls in dmascc.c

These cli()/sti() calls are made in start_timer() and are therefor
redundant since the register_lock is now used to protect register
io from within scc_isr() and write_scc() (where all calls to
start_timer() originate).

Signed-off-by: Mark Asselstine <mark.asselstine@windriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[NETNS]: Remove empty ->init callback.
Pavel Emelyanov [Mon, 21 Apr 2008 21:33:16 +0000 (14:33 -0700)]
[NETNS]: Remove empty ->init callback.

The netns start-stop engine can happily live with any of
init or exit callbacks set to NULL.

Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[DCCP]: Convert do_gettimeofday() to getnstimeofday().
YOSHIFUJI Hideaki [Mon, 21 Apr 2008 21:28:45 +0000 (14:28 -0700)]
[DCCP]: Convert do_gettimeofday() to getnstimeofday().

What do_gettimeofday() does is to call getnstimeofday() and
to convert the result from timespec{} to timeval{}.
We do not always need timeval{} and we can convert timespec{}
when we really need (to print).

Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Acked-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[NETNS]: Don't initialize err variable twice.
Pavel Emelyanov [Mon, 21 Apr 2008 21:25:23 +0000 (14:25 -0700)]
[NETNS]: Don't initialize err variable twice.

The ip6_route_net_init() performs some unneeded actions.

Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[NETNS]: The ip6_fib_timer can work with garbage on net namespace stop.
Pavel Emelyanov [Mon, 21 Apr 2008 21:23:03 +0000 (14:23 -0700)]
[NETNS]: The ip6_fib_timer can work with garbage on net namespace stop.

The del_timer() function doesn't guarantee, that the timer callback
is not active by the time it exits.

Thus, the fib6_net_exit() may kfree() all the data, that is required
by the fib6_run_gc(). The race window is tiny, but slab poisoning can
trigger this bug.

Using del_timer_sync() will cure this.

Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[IPV4]: Convert do_gettimeofday() to getnstimeofday().
YOSHIFUJI Hideaki [Mon, 21 Apr 2008 09:34:08 +0000 (02:34 -0700)]
[IPV4]: Convert do_gettimeofday() to getnstimeofday().

What do_gettimeofday() does is to call getnstimeofday() and
to convert the result from timespec{} to timeval{}.
After that, these callers convert the result again to msec.
Use getnstimeofday() and convert the units at once.

Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[IPV4]: Make icmp_sk_init() static.
Adrian Bunk [Mon, 21 Apr 2008 09:31:23 +0000 (02:31 -0700)]
[IPV4]: Make icmp_sk_init() static.

This patch makes the needlessly global icmp_sk_init() static.

Signed-off-by: Adrian Bunk <bunk@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[IPV6]: Make struct ip6_prohibit_entry_template static.
Adrian Bunk [Mon, 21 Apr 2008 09:29:32 +0000 (02:29 -0700)]
[IPV6]: Make struct ip6_prohibit_entry_template static.

This patch makes the needlessly global struct
ip6_prohibit_entry_template static.

Signed-off-by: Adrian Bunk <bunk@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agotcp: Trivial fix to correct function name in a comment in net/ipv4/tcp.c
Satoru SATOH [Mon, 21 Apr 2008 09:27:58 +0000 (02:27 -0700)]
tcp: Trivial fix to correct function name in a comment in net/ipv4/tcp.c

This is a trivial fix to correct function name in a comment in
net/ipv4/tcp.c.

Signed-off-by: Satoru SATOH <satoru.satoh@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[NET]: Expose netdevice dev_id through sysfs
David Woodhouse [Sun, 20 Apr 2008 23:07:43 +0000 (16:07 -0700)]
[NET]: Expose netdevice dev_id through sysfs

Expose dev_id to userspace, because it helps to disambiguate between
interfaces where the MAC address is unique.

This should allow us to simplify the handling of persistent naming for
S390 network devices in udev -- because it can depend on a simple
attribute of the device like the other match criteria, rather than
having a special case for SUBSYSTEMS=="ccwgroup".

Signed-off-by: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agoskbuff: fix missing kernel-doc notation
Randy Dunlap [Sun, 20 Apr 2008 23:06:22 +0000 (16:06 -0700)]
skbuff: fix missing kernel-doc notation

Add kernel-doc notation for ndisc_nodetype:

Warning(linux-2.6.25-git2//include/linux/skbuff.h:340): No description found for parameter 'ndisc_nodetype'

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[ROSE]: Fix soft lockup wrt. rose_node_list_lock
Bernard Pidoux [Sun, 20 Apr 2008 22:58:07 +0000 (15:58 -0700)]
[ROSE]: Fix soft lockup wrt. rose_node_list_lock

[ INFO: possible recursive locking detected ]
2.6.25 #3
---------------------------------------------
ax25ipd/3811 is trying to acquire lock:
  (rose_node_list_lock){-+..}, at: [<f8d31f1a>] rose_get_neigh+0x1a/0xa0
[rose]

but task is already holding lock:
  (rose_node_list_lock){-+..}, at: [<f8d31fed>]
rose_route_frame+0x4d/0x620 [rose]

other info that might help us debug this:
6 locks held by ax25ipd/3811:
  #0:  (&tty->atomic_write_lock){--..}, at: [<c0259a1c>]
tty_write_lock+0x1c/0x50
  #1:  (rcu_read_lock){..--}, at: [<c02aea36>] net_rx_action+0x96/0x230
  #2:  (rcu_read_lock){..--}, at: [<c02ac5c0>] netif_receive_skb+0x100/0x2f0
  #3:  (rose_node_list_lock){-+..}, at: [<f8d31fed>]
rose_route_frame+0x4d/0x620 [rose]
  #4:  (rose_neigh_list_lock){-+..}, at: [<f8d31ff7>]
rose_route_frame+0x57/0x620 [rose]
  #5:  (rose_route_list_lock){-+..}, at: [<f8d32001>]
rose_route_frame+0x61/0x620 [rose]

stack backtrace:
Pid: 3811, comm: ax25ipd Not tainted 2.6.25 #3
  [<c0147e27>] print_deadlock_bug+0xc7/0xd0
  [<c0147eca>] check_deadlock+0x9a/0xb0
  [<c0149cd2>] validate_chain+0x1e2/0x310
  [<c0149b95>] ? validate_chain+0xa5/0x310
  [<c010a7d8>] ? native_sched_clock+0x88/0xc0
  [<c0149fa1>] __lock_acquire+0x1a1/0x750
  [<c014a5d1>] lock_acquire+0x81/0xa0
  [<f8d31f1a>] ? rose_get_neigh+0x1a/0xa0 [rose]
  [<c03201a3>] _spin_lock_bh+0x33/0x60
  [<f8d31f1a>] ? rose_get_neigh+0x1a/0xa0 [rose]
  [<f8d31f1a>] rose_get_neigh+0x1a/0xa0 [rose]
  [<f8d32404>] rose_route_frame+0x464/0x620 [rose]
  [<c031ffdd>] ? _read_unlock+0x1d/0x20
  [<f8d31fa0>] ? rose_route_frame+0x0/0x620 [rose]
  [<f8d1c396>] ax25_rx_iframe+0x66/0x3b0 [ax25]
  [<f8d1f42f>] ? ax25_start_t3timer+0x1f/0x40 [ax25]
  [<f8d1e65b>] ax25_std_frame_in+0x7fb/0x890 [ax25]
  [<c0320005>] ? _spin_unlock_bh+0x25/0x30
  [<f8d1bdf6>] ax25_kiss_rcv+0x2c6/0x800 [ax25]
  [<c02a4769>] ? sock_def_readable+0x59/0x80
  [<c014a8a7>] ? __lock_release+0x47/0x70
  [<c02a4769>] ? sock_def_readable+0x59/0x80
  [<c031ffdd>] ? _read_unlock+0x1d/0x20
  [<c02a4769>] ? sock_def_readable+0x59/0x80
  [<c02a4d3a>] ? sock_queue_rcv_skb+0x13a/0x1d0
  [<c02a4c45>] ? sock_queue_rcv_skb+0x45/0x1d0
  [<f8d1bb30>] ? ax25_kiss_rcv+0x0/0x800 [ax25]
  [<c02ac715>] netif_receive_skb+0x255/0x2f0
  [<c02ac5c0>] ? netif_receive_skb+0x100/0x2f0
  [<c02af05c>] process_backlog+0x7c/0xf0
  [<c02aeb0c>] net_rx_action+0x16c/0x230
  [<c02aea36>] ? net_rx_action+0x96/0x230
  [<c012bd53>] __do_softirq+0x93/0x120
  [<f8d2a68a>] ? mkiss_receive_buf+0x33a/0x3f0 [mkiss]
  [<c012be37>] do_softirq+0x57/0x60
  [<c012c265>] local_bh_enable_ip+0xa5/0xe0
  [<c0320005>] _spin_unlock_bh+0x25/0x30
  [<f8d2a68a>] mkiss_receive_buf+0x33a/0x3f0 [mkiss]
  [<c025ea37>] pty_write+0x47/0x60
  [<c025c620>] write_chan+0x1b0/0x220
  [<c0259a1c>] ? tty_write_lock+0x1c/0x50
  [<c011fec0>] ? default_wake_function+0x0/0x10
  [<c0259bea>] tty_write+0x12a/0x1c0
  [<c025c470>] ? write_chan+0x0/0x220
  [<c018bbc6>] vfs_write+0x96/0x130
  [<c0259ac0>] ? tty_write+0x0/0x1c0
  [<c018c24d>] sys_write+0x3d/0x70
  [<c0104d1e>] sysenter_past_esp+0x5f/0xa5
  =======================
BUG: soft lockup - CPU#0 stuck for 61s! [ax25ipd:3811]

Pid: 3811, comm: ax25ipd Not tainted (2.6.25 #3)
EIP: 0060:[<c010a9db>] EFLAGS: 00000246 CPU: 0
EIP is at native_read_tsc+0xb/0x20
EAX: b404aa2c EBX: b404a9c9 ECX: 017f1000 EDX: 0000076b
ESI: 00000001 EDI: 00000000 EBP: ecc83afc ESP: ecc83afc
  DS: 007b ES: 007b FS: 00d8 GS: 0033 SS: 0068
CR0: 8005003b CR2: b7f5f000 CR3: 2cd8e000 CR4: 000006f0
DR0: 00000000 DR1: 00000000 DR2: 00000000 DR3: 00000000
DR6: ffff0ff0 DR7: 00000400
  [<c0204937>] delay_tsc+0x17/0x30
  [<c02048e9>] __delay+0x9/0x10
  [<c02127f6>] __spin_lock_debug+0x76/0xf0
  [<c0212618>] ? spin_bug+0x18/0x100
  [<c0147923>] ? __lock_contended+0xa3/0x110
  [<c0212998>] _raw_spin_lock+0x68/0x90
  [<c03201bf>] _spin_lock_bh+0x4f/0x60
  [<f8d31f1a>] ? rose_get_neigh+0x1a/0xa0 [rose]
  [<f8d31f1a>] rose_get_neigh+0x1a/0xa0 [rose]
  [<f8d32404>] rose_route_frame+0x464/0x620 [rose]
  [<c031ffdd>] ? _read_unlock+0x1d/0x20
  [<f8d31fa0>] ? rose_route_frame+0x0/0x620 [rose]
  [<f8d1c396>] ax25_rx_iframe+0x66/0x3b0 [ax25]
  [<f8d1f42f>] ? ax25_start_t3timer+0x1f/0x40 [ax25]
  [<f8d1e65b>] ax25_std_frame_in+0x7fb/0x890 [ax25]
  [<c0320005>] ? _spin_unlock_bh+0x25/0x30
  [<f8d1bdf6>] ax25_kiss_rcv+0x2c6/0x800 [ax25]
  [<c02a4769>] ? sock_def_readable+0x59/0x80
  [<c014a8a7>] ? __lock_release+0x47/0x70
  [<c02a4769>] ? sock_def_readable+0x59/0x80
  [<c031ffdd>] ? _read_unlock+0x1d/0x20
  [<c02a4769>] ? sock_def_readable+0x59/0x80
  [<c02a4d3a>] ? sock_queue_rcv_skb+0x13a/0x1d0
  [<c02a4c45>] ? sock_queue_rcv_skb+0x45/0x1d0
  [<f8d1bb30>] ? ax25_kiss_rcv+0x0/0x800 [ax25]
  [<c02ac715>] netif_receive_skb+0x255/0x2f0
  [<c02ac5c0>] ? netif_receive_skb+0x100/0x2f0
  [<c02af05c>] process_backlog+0x7c/0xf0
  [<c02aeb0c>] net_rx_action+0x16c/0x230
  [<c02aea36>] ? net_rx_action+0x96/0x230
  [<c012bd53>] __do_softirq+0x93/0x120
  [<f8d2a68a>] ? mkiss_receive_buf+0x33a/0x3f0 [mkiss]
  [<c012be37>] do_softirq+0x57/0x60
  [<c012c265>] local_bh_enable_ip+0xa5/0xe0
  [<c0320005>] _spin_unlock_bh+0x25/0x30
  [<f8d2a68a>] mkiss_receive_buf+0x33a/0x3f0 [mkiss]
  [<c025ea37>] pty_write+0x47/0x60
  [<c025c620>] write_chan+0x1b0/0x220
  [<c0259a1c>] ? tty_write_lock+0x1c/0x50
  [<c011fec0>] ? default_wake_function+0x0/0x10
  [<c0259bea>] tty_write+0x12a/0x1c0
  [<c025c470>] ? write_chan+0x0/0x220
  [<c018bbc6>] vfs_write+0x96/0x130
  [<c0259ac0>] ? tty_write+0x0/0x1c0
  [<c018c24d>] sys_write+0x3d/0x70
  [<c0104d1e>] sysenter_past_esp+0x5f/0xa5
  =======================

Since rose_route_frame() does not use rose_node_list we can safely
remove rose_node_list_lock spin lock here and let it be free for
rose_get_neigh().

Signed-off-by: Bernard Pidoux <f6bvp@amsat.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agorose: Socket lock was not released before returning to user space
Bernard Pidoux [Sun, 20 Apr 2008 01:41:51 +0000 (18:41 -0700)]
rose: Socket lock was not released before returning to user space

================================================
[ BUG: lock held when returning to user space! ]
------------------------------------------------
xfbbd/3683 is leaving the kernel with locks still held!
1 lock held by xfbbd/3683:
  #0:  (sk_lock-AF_ROSE){--..}, at: [<c8cd1eb3>] rose_connect+0x73/0x420 [rose]

INFO: task xfbbd:3683 blocked for more than 120 seconds.
"echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
xfbbd         D 00000246     0  3683   3669
        c6965ee0 00000092 c02c5c40 00000246 c0f6b5f0 c0f6b5c0 c0f6b5f0 c0f6b5c0
        c0f6b614 c6965f18 c024b74b ffffffff c06ba070 00000000 00000000 00000001
        c6ab07c0 c012d450 c0f6b634 c0f6b634 c7b5bf10 c0d6004c c7b5bf10 c6965f40
Call Trace:
  [<c024b74b>] lock_sock_nested+0x6b/0xd0
  [<c012d450>] ? autoremove_wake_function+0x0/0x40
  [<c02488f1>] sock_fasync+0x41/0x150
  [<c0249e69>] sock_close+0x19/0x40
  [<c0175d54>] __fput+0xb4/0x170
  [<c0176018>] fput+0x18/0x20
  [<c017300e>] filp_close+0x3e/0x70
  [<c01744e9>] sys_close+0x69/0xb0
  [<c0103bda>] sysenter_past_esp+0x5f/0xa5
  =======================
INFO: lockdep is turned off.

Signed-off-by: Bernard Pidoux <f6bvp@amsat.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agohci_usb: remove code obfuscation
Pavel Machek [Sun, 20 Apr 2008 01:17:26 +0000 (18:17 -0700)]
hci_usb: remove code obfuscation

_urb_free is an alias for kfree... making code longer & harder to
read. Remove it.

Signed-off-by: Pavel Machek <pavel@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agodrivers/net/appletalk: use time_before, time_before_eq, etc
Julia Lawall [Sun, 20 Apr 2008 01:15:41 +0000 (18:15 -0700)]
drivers/net/appletalk: use time_before, time_before_eq, etc

The functions time_before, time_before_eq, time_after, and time_after_eq
are more robust for comparing jiffies against other values.

A simplified version of the semantic patch making this change is as follows:
(http://www.emn.fr/x-info/coccinelle/)

// <smpl>
@ change_compare_np @
expression E;
@@

(
- jiffies <= E
+ time_before_eq(jiffies,E)
|
- jiffies >= E
+ time_after_eq(jiffies,E)
|
- jiffies < E
+ time_before(jiffies,E)
|
- jiffies > E
+ time_after(jiffies,E)
)

@ include depends on change_compare_np @
@@

#include <linux/jiffies.h>

@ no_include depends on !include && change_compare_np @
@@

  #include <linux/...>
+ #include <linux/jiffies.h>
// </smpl>

Signed-off-by: Julia Lawall <julia@diku.dk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agodrivers/atm: use time_before, time_before_eq, etc
Julia Lawall [Sun, 20 Apr 2008 01:14:50 +0000 (18:14 -0700)]
drivers/atm: use time_before, time_before_eq, etc

The functions time_before, time_before_eq, time_after, and time_after_eq
are more robust for comparing jiffies against other values.

A simplified version of the semantic patch making this change is as follows:
(http://www.emn.fr/x-info/coccinelle/)

// <smpl>
@ change_compare_np @
expression E;
@@

(
- jiffies <= E
+ time_before_eq(jiffies,E)
|
- jiffies >= E
+ time_after_eq(jiffies,E)
|
- jiffies < E
+ time_before(jiffies,E)
|
- jiffies > E
+ time_after(jiffies,E)
)

@ include depends on change_compare_np @
@@

#include <linux/jiffies.h>

@ no_include depends on !include && change_compare_np @
@@

  #include <linux/...>
+ #include <linux/jiffies.h>
// </smpl>

Signed-off-by: Julia Lawall <julia@diku.dk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agohci_usb: do not initialize static variables to 0
Pavel Machek [Sun, 20 Apr 2008 01:13:40 +0000 (18:13 -0700)]
hci_usb: do not initialize static variables to 0

hci_usb: do not initialize static variables to 0.

Signed-off-by: Pavel Machek <pavel@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agotg3: 5701 DMA corruption fix
Matt Carlson [Sun, 20 Apr 2008 01:12:33 +0000 (18:12 -0700)]
tg3: 5701 DMA corruption fix

Herbert Xu's commit fb93134dfc2a6e6fbedc7c270a31da03fce88db9, entitled
"[TCP]: Fix size calculation in sk_stream_alloc_pskb", has triggered a
bug in the 5701 where the 5701 DMA engine will corrupt outgoing
packets.  This problem only happens when the starting address of the
packet matches a certain range of offsets and only when the 5701 is
placed downstream of a particular Intel bridge.

This patch detects the problematic bridge and if present, readjusts the
starting address of the packet data to a dword aligned boundary.

Signed-off-by: Matt Carlson <mcarlson@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>