Discussion:
[GIT]: Networking
(too old to reply)
David Miller
2008-04-25 10:24:32 UTC
Permalink
More fixes and improvements:

1) Various spots don't check the copy_*_user() return value correctly
and pass non-zere returns back to userspace as if it were an error.
From Tom Quetchenbach and Pavel Emelyanov.

2) tehuti driver security fix, grrr it lets any user read any chip
register... From Jeff Garzik.

3) Fix enforcement of RFC3542 rules in ipv6 socket option processing,
from Yoshifuji HIDEAKI.

4) Fix EEPROM ethtool dumping regression, from Mandeep Singh Baines.

5) Fix net/key/af_key.c compiler warning, from Brian Haley.

6) net/ipv6/Kconfig type fix from Michael Beasley

7) IPSEC algorithm key length signedness tweak from Eric Dumazet.

8) Several wired networking fixes from Jeff Garzik.

Please pull, thanks a lot!

The following changes since commit 9edb74cc6ccb3a893c3d40727b7003c3c16f=
85a0:
David S. Miller (1):
tun: Multicast handling in tun_chr_ioctl() needs proper locking=
=2E

are available in the git repository at:

master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6.git master

Andrew Morton (1):
ehea: make things static

Andy Fleming (3):
ucc_geth: Fix sneaky merge conflict regarding bus_id
phylib: Add support for board-level PHY fixups
gianfar: Fix skb allocation strategy

Auke Kok (3):
e1000e: Add interrupt moderation run-time ethtool interface
e1000e: lower ring minimum size to 64
igb: save and restore pcie/msi state to support EEH recovery

Ayaz Abdulla (2):
forcedeth: new backoff implementation
forcedeth: realtek phy crossover detection

Benjamin Herrenschmidt (1):
ibm_newemac: Increase MDIO timeouts

Brian Haley (1):
af_key: Fix af_key.c compiler warning

Bruce Allan (1):
e1000e: cleanup several stats issues

Chris Snook (2):
atlx: remove flash vendor parameter
[netdrvr] atlx: code movement: move atl1 parameter parsing

C=C3=A9dric Augonnet (1):
Removing dead code in drivers/net/wan/hdlc_fr.c

David Graham (1):
e1000e: Fix HW Error on es2lan, ARP capture issue by BMC

David S. Miller (1):
Merge branch 'upstream-davem' of master.kernel.org:/.../jgarzik/n=
etdev-2.6

Eric Dumazet (2):
tg3: sparse cleanup
xfrm: alg_key_len & alg_icv_len should be unsigned

=46rancois Romieu (4):
tehuti: check register size
korina: fix misplaced return statement
korina: misc cleanup
via-velocity: fix vlan receipt

Grant Grundler (1):
[netdrvr] typhoon: typhoon_resume - remove call to start_queue

Harvey Harrison (1):
netxen: reduce stack usage of netxen_nic_flash_print

Jeff Garzik (1):
[netdrvr] tehuti: move ioctl perm check closer to function start

Jesse Brandeburg (1):
e1000e: Increment version to 0.2.1

Josh Boyer (2):
ibm_newemac: Fix section mismatch warnings
ibm_newemac Use status property for unused/unwired EMACs

Kay Sievers (1):
net drivers: fix platform driver hotplug/coldplug

Krzysztof Halasa (1):
WAN: Fix confusing insmod error code for C101 too.

Mandeep Singh Baines (1):
ethtool: EEPROM dump no longer works for tg3 and natsemi

Matheos Worku (2):
niu: Determine the # of ports from the card's VPD data
niu: Add support for Neptune FEM/NEM cards for C10 server blades

Michael Beasley (1):
ipv6: Fix typo in net/ipv6/Kconfig

Pavel Emelyanov (1):
net: Fix wrong interpretation of some copy_to_user() results.

Roel Kluin (1):
smc911x: test after postfix decrement fails in smc911x_{reset,dro=
p_pkt}

Sreenivasa Honnur (4):
S2io: Fix memory leak during free_tx_buffers
S2io: Version update for memory leak fix during free_tx_buffers
S2io: Removed receive buffer replenishment tasklet
S2io: Removed rx_lock and put_lock

Stefan Roese (2):
ibm_newemac: Fix problem with jumbo frame support and EMAC V4.pat=
ch
ibm_newemac: Add support for 460EX/GT-type MAL rx-channel handlin=
g

Tom Quetchenbach (1):
tcp: tcp_probe buffer overflow and incorrect return value

Valentine Barshak (2):
ibm_newemac: PowerPC 440GX EMAC PHY clock workaround
ibm_newemac: PowerPC 440EP/440GR EMAC PHY clock workaround

Wendy Xiong (2):
ixgbe: save and restore pcie/msi state to support EEH recovery
e1000e: save and restore pcie/msi state to support EEH recovery

YOSHIFUJI Hideaki (1):
ipv6 RAW: Disallow IPPROTO_IPV6-level IPV6_CHECKSUM socket option=
on ICMPv6 sockets.

Documentation/networking/phy.txt | 38 +++-
drivers/net/arm/at91_ether.c | 1 +
drivers/net/arm/ep93xx_eth.c | 2 +
drivers/net/atlx/atl1.c | 138 ++++++++++++
drivers/net/atlx/atlx.c | 177 ---------------
drivers/net/ax88796.c | 1 +
drivers/net/bfin_mac.c | 7 +-
drivers/net/cpmac.c | 2 +
drivers/net/dm9000.c | 1 +
drivers/net/e1000e/82571.c | 6 -
drivers/net/e1000e/defines.h | 3 +
drivers/net/e1000e/e1000.h | 34 +++-
drivers/net/e1000e/es2lan.c | 129 ++++++++---
drivers/net/e1000e/ethtool.c | 49 ++++-
drivers/net/e1000e/hw.h | 12 +-
drivers/net/e1000e/netdev.c | 159 ++++++++-----
drivers/net/e1000e/phy.c | 73 ++++---
drivers/net/ehea/ehea_main.c | 4 +-
drivers/net/forcedeth.c | 432 ++++++++++++++++++++++++++++=
+-------
drivers/net/gianfar.c | 104 ++++++---
drivers/net/ibm_newemac/core.c | 83 ++++++-
drivers/net/ibm_newemac/core.h | 14 +-
drivers/net/ibm_newemac/mal.c | 20 ++-
drivers/net/ibm_newemac/rgmii.c | 2 +-
drivers/net/ibm_newemac/tah.c | 2 +-
drivers/net/ibm_newemac/zmii.c | 2 +-
drivers/net/igb/igb_main.c | 2 +
drivers/net/irda/ali-ircc.c | 2 +
drivers/net/irda/pxaficp_ir.c | 2 +
drivers/net/irda/sa1100_ir.c | 2 +
drivers/net/ixgbe/ixgbe_main.c | 2 +
drivers/net/jazzsonic.c | 2 +
drivers/net/korina.c | 39 ++--
drivers/net/macb.c | 2 +
drivers/net/meth.c | 2 +
drivers/net/mv643xx_eth.c | 5 +-
drivers/net/netx-eth.c | 2 +-
drivers/net/netxen/netxen_nic_hw.c | 15 +-
drivers/net/niu.c | 371 +++++++++++++++++++++++++++-=
---
drivers/net/niu.h | 12 +
drivers/net/phy/mdio_bus.c | 3 +
drivers/net/phy/phy.c | 4 +-
drivers/net/phy/phy_device.c | 129 ++++++++++-
drivers/net/s2io.c | 128 +----------
drivers/net/s2io.h | 9 -
drivers/net/sgiseeq.c | 4 +-
drivers/net/smc911x.c | 8 +-
drivers/net/smc91x.c | 2 +
drivers/net/sni_82596.c | 2 +
drivers/net/tehuti.c | 15 ++
drivers/net/tg3.c | 2 +
drivers/net/tsi108_eth.c | 2 +
drivers/net/typhoon.c | 1 -
drivers/net/ucc_geth.c | 2 +-
drivers/net/via-velocity.c | 46 +++--
drivers/net/wan/c101.c | 6 +-
drivers/net/wan/hdlc_fr.c | 4 -
include/linux/phy.h | 24 ++-
include/linux/xfrm.h | 8 +-
net/can/raw.c | 3 +-
net/core/ethtool.c | 5 +
net/dccp/probe.c | 2 +-
net/ipv4/tcp_probe.c | 9 +-
net/ipv6/Kconfig | 2 +-
net/ipv6/raw.c | 18 ++
net/key/af_key.c | 5 +-
net/tipc/socket.c | 4 +-
67 files changed, 1673 insertions(+), 729 deletions(-)
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
David Miller
2008-04-30 08:31:38 UTC
Permalink
Several cures and additions:

1) TCP congestion window fixups in the presence of TSO, from John
Heffner, based upon some recent reports from google.

2) IPVS OOPS fix from Julian Anastasov.

3) IPSEC auditing fix from Yoshifuji HIDEAKI.

4) Various network driver updates from Jeff Garzik, and in particular
a new driver for Solarstorm SFC4000 chips, and one for
some IXP4xxx ethernet ports.

5) More multicast socket option compat handling work from David
Stevens.

6) Hopefully the last of the iwlwifi build issues are cured thanks to
Adrian Bunk.

7) SFQ scheduler shutdown should use dev_timer_sync(), from Jarek
Poplawski.

8) Imporperly fragmented SKBs can get created by layers such as
sunrpc when sending to UFO capable devices. Fix from
Kostya B, and reviewed by Herbert Xu.

9) Netfilter connection tracking hash calculation bug on ARM, diagnosed
and fixed by Philip Craig.

10) Two bridge fixes from Bodo Stroesser and Volodymyr G Lukiianyk, via
Stephen Hemminger.

Please pull, thanks a lot!

The following changes since commit 358c12953b88c5a06a57c33eb27c753b2e7934d1:
Jason Riedy (1):
iwlwifi: Allow building iwl3945 without iwl4965.

are available in the git repository at:

master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6.git master

Adrian Bunk (1):
iwlwifi: move the selects to the tristate drivers

Arnaud Ebalard (1):
netfilter: {nfnetlink,ip,ip6}_queue: fix skb_over_panic when enlarging packets

Ben Hutchings (1):
New driver "sfc" for Solarstorm SFC4000 controller.

Bodo Stroesser (1):
bridge: kernel panic when unloading bridge module

Bryan Wu (2):
Blackfin EMAC Driver: code cleanup
Blackfin EMAC Driver: Initial version of ethtool support

Cornelia Huck (1):
netiucv: Fix missing driver attributes.

Daniel Walker (1):
atm: ambassador: vcc_sf semaphore to mutex

David L Stevens (2):
net: Several cleanups for the setsockopt compat support.
net: Add compat support for getsockopt (MCAST_MSFILTER)

Francois Romieu (3):
sis190: Rx path update
sis190: remove needless MII reset
sis190: account for Tx errors

Frank Blaschka (5):
qeth: layer 3 support vlan IPv6 on hiper socket
qeth: provide get ethtool settings
qeth: rework fast path
qeth: layer 3 add missing dev_open/close to ccwgroup handler
qeth: layer 2 allow ethtool to set TSO

Harvey Harrison (1):
net: eepro autoport typo

Ivan Vecera (1):
r8169: fix oops in r8169_get_mac_version

Jarek Poplawski (1):
sch_sfq: use del_timer_sync() in sfq_destroy()

Jeff Garzik (2):
Merge branch 'sis190' of git://git.kernel.org/.../romieu/netdev-2.6 into upstream
Merge branch 'upstream-jeff' of git://git.kernel.org/.../romieu/netdev-2.6 into upstream

John Heffner (2):
tcp: Allow send-limited cwnd to grow up to max_burst when gso disabled
tcp: Limit cwnd growth when deferring for GSO

Julian Anastasov (1):
ipvs: fix oops in backup for fwmark conn templates

Klaus D. Wacker (1):
lcs: CCL-sequ. numbers required for protocol 802.2 only.

Kostya B (1):
[IPv4] UFO: prevent generation of chained skb destined to UFO device

Krzysztof Halasa (1):
Driver for IXP4xx built-in Ethernet ports

Lachlan Andrew (1):
tcp: Overflow bug in Vegas

Paul Gortmaker (1):
[netdrvr] gianfar: Determine TBIPA value dynamically

Paulius Zaleckas (3):
3c505: use netstats in net_device structure
3c509: use netstats in net_device structure
3c515: use netstats in net_device structure

Pavel Emelyanov (2):
netfilter: x_tables: fix net namespace leak when reading /proc/net/xxx_tables_names
MAINTAINERS: The socketcan-core list is subscribers-only.

Peter Tiedemann (1):
qeth module size reduction.

Philip Craig (1):
netfilter: nf_conntrack: padding breaks conntrack hash on ARM

Randy Dunlap (1):
drivers/net/phy: fix kernel-doc notation

Roel Kluin (3):
r8169: fix past rtl_chip_info array size for unknown chipsets
ARM: am79c961a: platform_get_irq() may return signed unnoticed
netfilter: xt_TCPOPTSTRIP: signed tcphoff for ipv6_skip_exthdr() retval

Sam Ravnborg (1):
ipv4: annotate a few functions __init in ipconfig.c

Sreenivasa Honnur (2):
S2io: Enable multi ring support
S2io: Version update for multi ring patches

Stephen Hemminger (3):
sis190: use the allocated buffer as a status code in sis190_alloc_rx_skb
sis190: hard-code the alignment of tiny packets
sis190: use netdev_alloc_skb

Steve Glendinning (2):
Rename SMSC phy functions to be more generic
Add support for SMSC LAN8187 and LAN8700 PHYs

Timo Teras (1):
ipv4: Update MTU to all related cache entries in ip_rt_frag_needed()

Ursula Braun (3):
netiucv: get rid of in_atomic() use
ccwgroup: Unify parsing for group attribute.
qeth: read number of ports from card

Volodymyr G Lukiianyk (1):
bridge: fix error handling in br_add_if()

MAINTAINERS | 9 +-
arch/arm/mach-ixp4xx/ixp4xx_npe.c | 4 +
arch/arm/mach-ixp4xx/ixp4xx_qmgr.c | 2 +
drivers/atm/ambassador.c | 19 +-
drivers/atm/ambassador.h | 2 +-
drivers/net/3c505.c | 30 +-
drivers/net/3c505.h | 1 -
drivers/net/3c509.c | 47 +-
drivers/net/3c515.c | 64 +-
drivers/net/Kconfig | 1 +
drivers/net/Makefile | 2 +
drivers/net/arm/Kconfig | 8 +
drivers/net/arm/Makefile | 1 +
drivers/net/arm/am79c961a.c | 10 +-
drivers/net/arm/ixp4xx_eth.c | 1265 +++++++++++
drivers/net/bfin_mac.c | 296 ++--
drivers/net/bfin_mac.h | 2 +-
drivers/net/eepro.c | 2 +-
drivers/net/gianfar.c | 27 +-
drivers/net/gianfar.h | 1 -
drivers/net/gianfar_mii.c | 38 +-
drivers/net/gianfar_mii.h | 3 +
drivers/net/phy/Kconfig | 2 +-
drivers/net/phy/phy_device.c | 2 +-
drivers/net/phy/smsc.c | 83 +-
drivers/net/r8169.c | 8 +-
drivers/net/s2io.c | 337 ++--
drivers/net/s2io.h | 82 +-
drivers/net/sfc/Kconfig | 12 +
drivers/net/sfc/Makefile | 5 +
drivers/net/sfc/bitfield.h | 508 +++++
drivers/net/sfc/boards.c | 167 ++
drivers/net/sfc/boards.h | 26 +
drivers/net/sfc/efx.c | 2208 +++++++++++++++++++
drivers/net/sfc/efx.h | 67 +
drivers/net/sfc/enum.h | 50 +
drivers/net/sfc/ethtool.c | 460 ++++
drivers/net/sfc/ethtool.h | 27 +
drivers/net/sfc/falcon.c | 2722 ++++++++++++++++++++++++
drivers/net/sfc/falcon.h | 130 ++
drivers/net/sfc/falcon_hwdefs.h | 1135 ++++++++++
drivers/net/sfc/falcon_io.h | 243 +++
drivers/net/sfc/falcon_xmac.c | 585 +++++
drivers/net/sfc/gmii.h | 195 ++
drivers/net/sfc/i2c-direct.c | 381 ++++
drivers/net/sfc/i2c-direct.h | 91 +
drivers/net/sfc/mac.h | 33 +
drivers/net/sfc/mdio_10g.c | 282 +++
drivers/net/sfc/mdio_10g.h | 232 ++
drivers/net/sfc/net_driver.h | 883 ++++++++
drivers/net/sfc/phy.h | 48 +
drivers/net/sfc/rx.c | 875 ++++++++
drivers/net/sfc/rx.h | 29 +
drivers/net/sfc/sfe4001.c | 252 +++
drivers/net/sfc/spi.h | 71 +
drivers/net/sfc/tenxpress.c | 434 ++++
drivers/net/sfc/tx.c | 452 ++++
drivers/net/sfc/tx.h | 24 +
drivers/net/sfc/workarounds.h | 56 +
drivers/net/sfc/xenpack.h | 62 +
drivers/net/sfc/xfp_phy.c | 132 ++
drivers/net/sis190.c | 136 +-
drivers/net/wireless/iwlwifi/Kconfig | 12 +-
drivers/s390/cio/ccwgroup.c | 96 +-
drivers/s390/net/cu3088.c | 20 +-
drivers/s390/net/lcs.c | 3 +-
drivers/s390/net/netiucv.c | 3 +-
drivers/s390/net/qeth_core.h | 50 +-
drivers/s390/net/qeth_core_main.c | 200 +-
drivers/s390/net/qeth_l2_main.c | 30 +-
drivers/s390/net/qeth_l3.h | 3 -
drivers/s390/net/qeth_l3_main.c | 30 +-
include/asm-s390/ccwgroup.h | 7 +-
include/asm-s390/qdio.h | 1 +
include/net/compat.h | 3 +
include/net/ip_vs.h | 3 +-
include/net/netfilter/nf_conntrack_tuple.h | 10 -
include/net/route.h | 2 +-
net/bridge/br.c | 2 +-
net/bridge/br_if.c | 5 +-
net/compat.c | 90 +-
net/ipv4/icmp.c | 3 +-
net/ipv4/ip_output.c | 22 +-
net/ipv4/ip_sockglue.c | 9 +-
net/ipv4/ipconfig.c | 8 +-
net/ipv4/ipvs/ip_vs_proto.c | 2 +-
net/ipv4/ipvs/ip_vs_proto_ah.c | 1 +
net/ipv4/ipvs/ip_vs_proto_esp.c | 1 +
net/ipv4/ipvs/ip_vs_proto_tcp.c | 1 +
net/ipv4/ipvs/ip_vs_proto_udp.c | 1 +
net/ipv4/ipvs/ip_vs_sync.c | 80 +-
net/ipv4/netfilter/ip_queue.c | 5 +-
net/ipv4/netfilter/nf_conntrack_l3proto_ipv4.c | 2 +-
net/ipv4/route.c | 40 +-
net/ipv4/tcp_cong.c | 12 +-
net/ipv4/tcp_vegas.c | 10 +-
net/ipv4/tcp_veno.c | 8 +-
net/ipv6/ipv6_sockglue.c | 4 +
net/ipv6/netfilter/ip6_queue.c | 5 +-
net/netfilter/nf_conntrack_core.c | 4 +-
net/netfilter/nfnetlink_queue.c | 5 +-
net/netfilter/x_tables.c | 2 +-
net/netfilter/xt_TCPOPTSTRIP.c | 2 +-
net/sched/sch_sfq.c | 3 +-
104 files changed, 15349 insertions(+), 812 deletions(-)
create mode 100644 drivers/net/arm/ixp4xx_eth.c
create mode 100644 drivers/net/sfc/Kconfig
create mode 100644 drivers/net/sfc/Makefile
create mode 100644 drivers/net/sfc/bitfield.h
create mode 100644 drivers/net/sfc/boards.c
create mode 100644 drivers/net/sfc/boards.h
create mode 100644 drivers/net/sfc/efx.c
create mode 100644 drivers/net/sfc/efx.h
create mode 100644 drivers/net/sfc/enum.h
create mode 100644 drivers/net/sfc/ethtool.c
create mode 100644 drivers/net/sfc/ethtool.h
create mode 100644 drivers/net/sfc/falcon.c
create mode 100644 drivers/net/sfc/falcon.h
create mode 100644 drivers/net/sfc/falcon_hwdefs.h
create mode 100644 drivers/net/sfc/falcon_io.h
create mode 100644 drivers/net/sfc/falcon_xmac.c
create mode 100644 drivers/net/sfc/gmii.h
create mode 100644 drivers/net/sfc/i2c-direct.c
create mode 100644 drivers/net/sfc/i2c-direct.h
create mode 100644 drivers/net/sfc/mac.h
create mode 100644 drivers/net/sfc/mdio_10g.c
create mode 100644 drivers/net/sfc/mdio_10g.h
create mode 100644 drivers/net/sfc/net_driver.h
create mode 100644 drivers/net/sfc/phy.h
create mode 100644 drivers/net/sfc/rx.c
create mode 100644 drivers/net/sfc/rx.h
create mode 100644 drivers/net/sfc/sfe4001.c
create mode 100644 drivers/net/sfc/spi.h
create mode 100644 drivers/net/sfc/tenxpress.c
create mode 100644 drivers/net/sfc/tx.c
create mode 100644 drivers/net/sfc/tx.h
create mode 100644 drivers/net/sfc/workarounds.h
create mode 100644 drivers/net/sfc/xenpack.h
create mode 100644 drivers/net/sfc/xfp_phy.c
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Ingo Molnar
2008-04-30 20:38:39 UTC
Permalink
hi David,
Post by David Miller
net: Several cleanups for the setsockopt compat support.
net: Add compat support for getsockopt (MCAST_MSFILTER)
this broke !COMPAT builds in current -git. Fix below. NOTE: fix is just
a quick guess.

Ingo

----------------------->
Subject: net: ipv6 build fix
From: Ingo Molnar <***@elte.hu>
Date: Wed Apr 30 22:32:09 CEST 2008

x86.git testing found the following build error on -git:

ERROR: "compat_mc_getsockopt" [net/ipv6/ipv6.ko] undefined!

with this config:

http://redhat.com/~mingo/misc/config-Wed_Apr_30_22_23_37_CEST_2008.bad

this hits on !COMPAT + IPV6 modular builds. It is caused by commit
42908c69f6 ("net: Add compat support for getsockopt (MCAST_MSFILTER)").

this fix here is just a quick guess - i _suspect_ that
do_ipv6_getsockopt() will just DTRT and return -EINVAL but i
havent checked. It obviously solves the build issue.

Signed-off-by: Ingo Molnar <***@elte.hu>
---
net/ipv6/ipv6_sockglue.c | 2 ++
1 file changed, 2 insertions(+)

Index: linux/net/ipv6/ipv6_sockglue.c
===================================================================
--- linux.orig/net/ipv6/ipv6_sockglue.c
+++ linux/net/ipv6/ipv6_sockglue.c
@@ -1089,9 +1089,11 @@ int ipv6_getsockopt(struct sock *sk, int
if(level != SOL_IPV6)
return -ENOPROTOOPT;

+#ifdef CONFIG_COMPAT
if (optname == MCAST_MSFILTER)
return compat_mc_getsockopt(sk, level, optname, optval, optlen,
ipv6_getsockopt);
+#endif

err = do_ipv6_getsockopt(sk, level, optname, optval, optlen);
#ifdef CONFIG_NETFILTER
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
David Stevens
2008-04-30 21:17:41 UTC
Permalink
Ingo,
That fix is not correct. Pavel Emelyanov posted a
correct fix to netdev earlier today. Apparently, the context
matched ipv6_getsockopt() in the split version, but those
changes belong in compat_ipv6_getsockopt(), as in Pavel's
fix.
Since the original patch was correct, I'm not sure
if the fix will have the same problem or not.

+-DLS

--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Ingo Molnar
2008-04-30 21:29:22 UTC
Permalink
Post by Ingo Molnar
this fix here is just a quick guess - i _suspect_ that
do_ipv6_getsockopt() will just DTRT and return -EINVAL but i havent
checked. It obviously solves the build issue.
i see now that it's a mismerge:

http://marc.info/?l=linux-netdev&m=120956601217678&w=2

Ingo
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
David Miller
2008-04-30 21:48:35 UTC
Permalink
From: Ingo Molnar <***@elte.hu>
Date: Wed, 30 Apr 2008 22:38:39 +0200
Post by Ingo Molnar
hi David,
Post by David Miller
net: Several cleanups for the setsockopt compat support.
net: Add compat support for getsockopt (MCAST_MSFILTER)
this broke !COMPAT builds in current -git. Fix below. NOTE: fix is just
a quick guess.
I know, Pavel posted a fix to netdev already which I'll push to
Linus.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
David Miller
2008-05-03 10:00:01 UTC
Permalink
Here is what has accumulated the past few days:

1) Set of patches from Denis V. Lunev which fix the
procfs "pde->data not setup before grafting new node
under /proc" bug in several networking spots.

2) When we get a transmit timeout, that's a pretty serious error
and it's useful to have a backtrace, thus add a WARN_ON_ONCE().
From Arjan van de Ven.

3) sysfs network device renaming fix, and a ipv6 fragmentation
queue namespace fix, from Daniel Lezcano.

4) Several wireless bug fixes from John Linville and the wireless
crew.

5) Several changes from Harvy Harrison, mostly to use recently
added infrastructure such as clamp and get_unaligned_${ENDIAN}${SZ}

6) Ilpo spotted some scary cases of missing braces hidden from humans
by mis-indented code.

7) TG3 and BNX2 driver updates from Matt Carlson and Michael Chan.

8) Some bridge layer interfaces could only handle up to 256 devices,
fix from Stephen Hemminger.

9) Some de-NR_CPUS'ification in net/core/dev.c from Mike Travis.

10) usb bluetooth driver race fix from Pavel Machek

11) Bernard Pidoux found incorrect locking as well as bad sparse
locking annotations in AF_ROSE.

Please pull, thanks a lot!

The following changes since commit e4c576b911e364737b1bf4f5bfdab1c44071=
3f26:
Theodore Ts'o (1):
Update .gitignore to include include/linux/bounds.h

are available in the git repository at:

master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6.git master

Arjan van de Ven (1):
net: Add a WARN_ON_ONCE() to the transmit timeout function

Bernard Pidoux (1):
rose: Wrong list_lock argument in rose_node seqops

Bill Moss (1):
iwlwifi: fix debug messages during scanning

Daniel Lezcano (2):
netns: Fix device renaming for sysfs
netns: Fix reassembly timer to use the right namespace

David S. Miller (2):
Merge branch 'master' of master.kernel.org:/.../linville/wireless=
-2.6
Merge branch 'master' of master.kernel.org:/.../linville/wireless=
-2.6

Denis V. Lunev (10):
sunrpc: assign PDE->data before gluing PDE into /proc tree
netfilter: assign PDE->data before gluing PDE into /proc tree
net: assign PDE->data before gluing PDE into /proc tree
ipv6: assign PDE->data before gluing PDE into /proc tree
atm: assign PDE->data before gluing PDE into /proc tree
vlan: assign PDE->data before gluing PDE into /proc tree
ipv4: assign PDE->data before gluing PDE into /proc tree
netfilter: assign PDE->fops before gluing PDE into /proc tree
netfilter: assign PDE->data before gluing PDE into /proc tree
netns: assign PDE->data before gluing entry into /proc tree

Guy Cohen (1):
iwlwifi: fix current channel is not scanned

Harvey Harrison (6):
mac80211: incorrect shift direction
net: fix returning void-valued expression warnings
irda: use get_unaligned_* helpers
bluetooth: use get/put_unaligned_* helpers
net: use get/put_unaligned_* helpers
dccp: ccid2.c, ccid3.c use clamp(), clamp_t()

Holger Schurig (1):
libertas: fix use-before-check violation

Ilpo J=E4rvinen (1):
net: Add missing braces to multi-statement if()s

Ivo van Doorn (2):
rt2x00: Don't enable short preamble for 1MBs
rt2x00: Fix quality/activity led handling

Johannes Berg (3):
mac80211: assign conf.beacon_control for mesh
mac80211: don't allow invalid WDS peer addresses
mac80211: insert WDS peer after adding interface

Kirill A. Shutemov (1):
Make linux/wireless.h be able to compile

Luis Carlos Cobo (1):
mac80211: use 4-byte mesh sequence number

Matt Carlson (5):
tg3: Use constant 500KHz MI clock on adapters with a CPMU
tg3: Fix 5761 NVRAM sizes
tg3: Fix ethtool loopback test for 5761 BX devices
tg3: Add link state reporting to UMP firmware
tg3: Update version to 3.92

Michael Buesch (2):
b43: Fix dual-PHY devices
b43: Fix some TX/RX locking issues

Michael Chan (6):
bnx2: Refine remote PHY locking.
bnx2: Fix remote PHY initial link state.
bnx2: Fix register test on 5709.
bnx2: Zero out context memory for 5709.
bnx2: Update RV2P firmware for 5709.
bnx2: Update version to 1.7.5.

Mike Travis (1):
net: remove NR_CPUS arrays in net/core/dev.c

Pavel Machek (1):
hci_usb.h: fix hard-to-trigger race

Stephen Hemminger (1):
bridge: forwarding table information for >256 devices

drivers/bluetooth/hci_usb.h | 21 +-
drivers/net/8390.c | 2 +-
drivers/net/bnx2.c | 43 ++-
drivers/net/bnx2_fw2.h | 502 ++++++++++++++-----=
--------
drivers/net/tg3.c | 149 ++++++--
drivers/net/tg3.h | 15 +-
drivers/net/wireless/b43/b43.h | 4 +
drivers/net/wireless/b43/main.c | 51 ++-
drivers/net/wireless/iwlwifi/iwl-3945.h | 1 -
drivers/net/wireless/iwlwifi/iwl-4965.h | 1 -
drivers/net/wireless/iwlwifi/iwl3945-base.c | 29 +--
drivers/net/wireless/iwlwifi/iwl4965-base.c | 26 +-
drivers/net/wireless/libertas/scan.c | 3 +-
drivers/net/wireless/rt2x00/rt2400pci.c | 4 +-
drivers/net/wireless/rt2x00/rt2500pci.c | 4 +-
drivers/net/wireless/rt2x00/rt2500usb.c | 2 +-
drivers/net/wireless/rt2x00/rt2x00dev.c | 2 +
drivers/net/wireless/rt2x00/rt2x00leds.c | 15 +
drivers/net/wireless/rt2x00/rt2x00lib.h | 6 +
drivers/net/wireless/rt2x00/rt61pci.c | 2 +-
drivers/net/wireless/rt2x00/rt73usb.c | 2 +-
drivers/net/xen-netfront.c | 2 +-
fs/proc/generic.c | 17 +
fs/proc/proc_net.c | 11 -
include/linux/ieee80211.h | 2 +-
include/linux/if_bridge.h | 4 +-
include/linux/wireless.h | 7 +-
net/8021q/vlanproc.c | 7 +-
net/atm/proc.c | 6 +-
net/bluetooth/bnep/core.c | 8 +-
net/bluetooth/hci_event.c | 9 +-
net/bluetooth/hci_sock.c | 2 +-
net/bluetooth/l2cap.c | 2 +-
net/bridge/br_fdb.c | 4 +
net/bridge/br_stp_bpdu.c | 4 +-
net/core/dev.c | 19 +-
net/core/filter.c | 4 +-
net/core/neighbour.c | 5 +-
net/core/net-sysfs.c | 7 +-
net/core/net-sysfs.h | 2 +-
net/core/pktgen.c | 12 +-
net/core/sock.c | 3 +-
net/dccp/ccids/ccid2.c | 2 +-
net/dccp/ccids/ccid3.c | 4 +-
net/ipv4/cipso_ipv4.c | 18 +-
net/ipv4/netfilter/ipt_CLUSTERIP.c | 6 +-
net/ipv4/tcp_hybla.c | 6 +-
net/ipv4/tcp_input.c | 18 +-
net/ipv4/tcp_ipv4.c | 10 +-
net/ipv4/tcp_vegas.c | 6 +-
net/ipv4/tcp_veno.c | 6 +-
net/ipv4/udp.c | 7 +-
net/ipv6/proc.c | 6 +-
net/ipv6/reassembly.c | 4 +-
net/irda/iriap.c | 8 +-
net/mac80211/ieee80211_i.h | 2 +-
net/mac80211/main.c | 47 ++-
net/mac80211/mesh.c | 17 +-
net/mac80211/mesh.h | 2 +-
net/mac80211/tkip.c | 6 +-
net/netfilter/nf_conntrack_standalone.c | 6 +-
net/netfilter/x_tables.c | 13 +-
net/netfilter/xt_hashlimit.c | 12 +-
net/rose/rose_route.c | 8 +-
net/rxrpc/ar-transport.c | 3 +-
net/sched/sch_generic.c | 1 +
net/sunrpc/cache.c | 15 +-
net/sunrpc/stats.c | 8 +-
net/sysctl_net.c | 2 +-
69 files changed, 704 insertions(+), 560 deletions(-)
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
David Miller
2008-05-06 09:33:54 UTC
Permalink
More fixes, as usual:

1) When DCCP socket option feature length is invalid, return
real error code to userspace instead of "1". From Chris
Wright.

2) IRDA + !PNP driver build fixes from Ingo.

3) lapbeth leaks netdevice references, leading to module unload
etc. hangs

4) Fix NIU device probing regression.

5) Two classifier actions fixes from Jamal Hadi Salim.

6) Fix tree handling in HTB classifier, from Jarek Poplawski.

7) Kill build warnings in IPSEC when auditing is disabled.
From Marcin Slusarz.

8) Several device leak and double-free fixes from Pavel Emelyanov.
These cure the cases where the driver registered a destructor
that freed the device, but would then again free it anyways.

9) Get rid of open coded references to route metrics, use existing
helper functions.

10 IWLWIFI should be a tristate, from Adrian Bunk.

Please pull, thanks a lot!

The following changes since commit f37f2c62a28e848e06399ea2f9be1e098212625c:
Bernard Pidoux (1):
rose: Wrong list_lock argument in rose_node seqops

are available in the git repository at:

master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6.git master

Adrian Bunk (1):
iwlwifi: make IWLWIFI a tristate

Chris Wright (1):
dccp: return -EINVAL on invalid feature length

David S. Miller (3):
lapbeth: Release ->ethdev when unregistering device.
niu: Fix probing regression for maramba on-board chips.
Revert "atm: Do not free already unregistered net device."

Ingo Molnar (2):
irda: fix !PNP support in drivers/net/irda/nsc-ircc.c
irda: fix !PNP support for drivers/net/irda/smsc-ircc2.c

Jamal Hadi Salim (2):
net_cls_act: Make act_simple use of netlink policy.
net_cls_act: act_simple dont ignore realloc code

Jarek Poplawski (1):
sch_htb: remove from event queue in htb_parent_to_leaf()

Johannes Berg (1):
net: Fix useless comment reference loop.

Marcin Slusarz (1):
xfrm: convert empty xfrm_audit_* macros to functions

Pavel Emelyanov (5):
bridge: Net device leak in br_add_bridge().
bridge: Consolidate error paths in br_add_bridge().
mac80211: Do not free net device after it is unregistered.
atm: Do not free already unregistered net device.
atm: Bad locking on br2684_devs modifications.

Satoru SATOH (2):
ip: Make use of the inline function dst_metric_locked()
ip: Use inline function dst_metric() instead of direct access to dst->metric[]

drivers/net/irda/nsc-ircc.c | 6 ++++
drivers/net/irda/smsc-ircc2.c | 5 +++-
drivers/net/niu.c | 11 +++++---
drivers/net/wan/lapbether.c | 1 +
drivers/net/wireless/iwlwifi/Kconfig | 3 +-
include/net/ip.h | 2 +-
include/net/xfrm.h | 48 ++++++++++++++++++++++++++++-----
net/atm/br2684.c | 4 +-
net/bridge/br_if.c | 12 +++++---
net/core/skbuff.c | 4 ++-
net/dccp/feat.c | 2 +-
net/decnet/dn_route.c | 12 ++++----
net/ipv4/route.c | 18 ++++++------
net/ipv4/tcp_input.c | 15 +++++-----
net/ipv6/route.c | 6 ++--
net/mac80211/main.c | 7 +++-
net/sched/act_simple.c | 45 +++++++++++++++----------------
net/sched/sch_htb.c | 8 ++++-
18 files changed, 132 insertions(+), 77 deletions(-)
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
David Miller
2008-05-09 00:14:08 UTC
Permalink
Here are some networking fixes and enhancements:

1) Bonding bug fixes from Pavel Emelyanov and Jay Vosburgh.

2) Support for some newer PHY chips in E1000E, from Bruce Allan.

3) TCP FRTO SACK fix when in newreno mode from Ilpo.

4) Fix can_send() handling when dev_queue_xmit() fails, from
Oliver Hartkopp.

5) macvlan leak/oops fix from Patrick McHardy.

6) Two netfilter fixes from Patrick McHardy.

7) fs_enet leak fix from Scott Wood.

8) Namespace network device corruption fix from Pavel Emelyanov.

9) Several other device driver fixes from Jeff Garzik's tree.

Please pull, thanks a lot!

The following changes since commit 9d1045ad68fcccfaf1393cc463ab6357693e=
8d1d:
Jamal Hadi Salim (1):
net_cls_act: act_simple dont ignore realloc code

are available in the git repository at:

master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6.git master

Alan Cox (1):
cxgb3: Use CAP_SYS_RAWIO for firmware

Andy Fleming (3):
gianfar: Fix a locking bug in gianfar's sysfs code
phylib: Fix some sparse warnings
ucc_geth: Fix a bunch of sparse warnings

Anton Vorontsov (2):
uli526x: initialize the hardware prior to requesting interrupts
uli526x: fix endianness issues in the setup frame

Ben Hutchings (1):
net: Added ASSERT_RTNL() to dev_open() and dev_close().

Bruce Allan (1):
e1000e: Add support for BM PHYs on ICH9

Bruce Robson (1):
[netdrvr] eexpress: IPv6 fails - multicast problems

David S. Miller (1):
Merge branch 'upstream-davem' of master.kernel.org:/.../jgarzik/n=
etdev-2.6

Don Fry (1):
pcnet32: delete non NAPI code from driver.

Gunnar Larisch (1):
3c980-TX needs EXTRA_PREAMBLE

Ilpo J=C3=A4rvinen (1):
tcp FRTO: SACK variant is errorneously used with NewReno

J.H.M. Dassen (Ray) (1):
net/ipv4: correct RFC 1122 section reference in comment

Jay Vosburgh (2):
bonding: fix error unwind in bonding_store_bonds
bonding: fix enslavement error unwinds

Jeff Garzik (2):
fix warning in drivers/net/appletalk/cops.c
Merge branch 'for-2.6.26' of git://git.farnsworth.org/dale/linux-=
2.6-mv643xx_eth into upstream

Joakim Tjernlund (1):
ucc_geth: Don't use RX clock as TX clock.

Kok, Auke (1):
e1000e: don't return half-read eeprom on error

Lennert Buytenhek (5):
mv643xx_eth: get rid of static variables, allow multiple instance=
s
mv643xx_eth: mbus decode window support
mv643xx_eth: configurable t_clk
mv643xx_eth: shorten shared platform driver name
mv643xx_eth: inter-mv643xx SMI port sharing

Oliver Hartkopp (1):
can: Fix can_send() handling on dev_queue_xmit() failures

Patrick McHardy (3):
macvlan: Fix memleak on device removal/crash on module removal
netfilter: nf_conntrack_sip: restrict RTP expect flushing on erro=
r to last request
netfilter: Kconfig: default DCCP/SCTP conntrack support to the pr=
otocol config values

Paulius Zaleckas (1):
3c59x: use netstats in net_device structure

Pavel Emelyanov (3):
bonding: Do not call free_netdev for already registered device.
bonding: Deadlock between bonding_store_bonds and bond_destroy_sy=
sfs.
netns: Fix arbitrary net_device-s corruptions on net_ns stop.

Scott Wood (1):
fs_enet: Fix a memory leak in fs_enet_mdio_probe

arch/arm/mach-orion5x/common.c | 2 +
arch/powerpc/platforms/chrp/pegasos_eth.c | 4 +
arch/powerpc/sysdev/mv64x60_dev.c | 2 +
arch/ppc/syslib/mv64x60.c | 3 +
drivers/net/3c59x.c | 73 +++---
drivers/net/Kconfig | 14 --
drivers/net/appletalk/cops.c | 16 +-
drivers/net/bonding/bond_main.c | 24 ++-
drivers/net/bonding/bond_sysfs.c | 16 +-
drivers/net/cxgb3/cxgb3_main.c | 4 +-
drivers/net/e1000e/defines.h | 10 +
drivers/net/e1000e/e1000.h | 7 +-
drivers/net/e1000e/ethtool.c | 45 +++-
drivers/net/e1000e/hw.h | 22 ++
drivers/net/e1000e/ich8lan.c | 83 +++++++-
drivers/net/e1000e/netdev.c | 330 ++++++++++++++++++++=
+++++++-
drivers/net/e1000e/phy.c | 278 ++++++++++++++++++++=
+++
drivers/net/eexpress.c | 11 +-
drivers/net/fs_enet/mii-fec.c | 3 +-
drivers/net/gianfar.c | 5 +-
drivers/net/gianfar.h | 3 +
drivers/net/gianfar_sysfs.c | 10 +-
drivers/net/macvlan.c | 2 +-
drivers/net/mv643xx_eth.c | 160 +++++++++++---
drivers/net/pcnet32.c | 61 +-----
drivers/net/phy/phy.c | 2 +-
drivers/net/tulip/uli526x.c | 38 ++--
drivers/net/ucc_geth.c | 278 ++++++++++++--------=
---
drivers/net/ucc_geth.h | 48 +++--
drivers/net/ucc_geth_ethtool.c | 6 -
drivers/net/ucc_geth_mii.c | 4 +-
include/linux/mv643xx_eth.h | 16 ++-
include/linux/netfilter/nf_conntrack_sip.h | 1 +
include/linux/phy.h | 5 +
net/can/af_can.c | 16 +-
net/core/dev.c | 12 +-
net/ipv4/ip_input.c | 2 +-
net/ipv4/tcp_input.c | 13 +-
net/netfilter/Kconfig | 4 +
net/netfilter/nf_conntrack_sip.c | 22 +-
40 files changed, 1242 insertions(+), 413 deletions(-)
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Linus Torvalds
2008-05-09 01:51:08 UTC
Permalink
David, you need to stop sending patches like this after -rc1.

This is too big. I'm taking it, but there's something wrong in this. Your
merges after -rc1 have generally been way too large, and I need to shut
you down if this gets to be a big pattern.

It's one thing taking a totally new driver (that couldn't have worked
before and cannot be a regression) or adding new PCI ID's. It's a totally
different thing to take 800 lines of code to an old driver that people
depend on. These things need to happen during the merge window, and just
because they are drivers doesn't mean that they can just do anythign they
want.

Linus
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
David Miller
2008-05-09 04:24:52 UTC
Permalink
From: Linus Torvalds <***@linux-foundation.org>
Date: Thu, 8 May 2008 18:51:08 -0700 (PDT)
Post by Linus Torvalds
David, you need to stop sending patches like this after -rc1.
This is too big. I'm taking it, but there's something wrong in this. Your
merges after -rc1 have generally been way too large, and I need to shut
you down if this gets to be a big pattern.
It's one thing taking a totally new driver (that couldn't have worked
before and cannot be a regression) or adding new PCI ID's. It's a totally
different thing to take 800 lines of code to an old driver that people
depend on. These things need to happen during the merge window, and just
because they are drivers doesn't mean that they can just do anythign they
want.
I saw that e1000e commit from Jeff's tree and was about to push
back on him. But I decided not to.

I will next time if it's such a big deal, but frankly I thought this
one was OK. It's stupid to make people wait until 2.6.27 to be able
to use the e1000e cards with the new PHY type. It causes some hw bug
workaround to need to be added to the e1000e driver, and that's where
the bulk of that 800 line change is from.

Distributions are going to add that e1000e PHY support patch anyways,
regardless of what we do, and isn't that one of the reasons we were
talking about more aggressively merging new drivers?

Who the heck are we kidding if we don't put changes like that into the
tree when pretty much every distribution will? Yes, even after rc1.

--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Jeff Garzik
2008-05-09 14:32:14 UTC
Permalink
Post by Linus Torvalds
It's one thing taking a totally new driver (that couldn't have worked
before and cannot be a regression) or adding new PCI ID's. It's a totally
different thing to take 800 lines of code to an old driver that people
depend on. These things need to happen during the merge window, and just
because they are drivers doesn't mean that they can just do anythign they
want.
The e1000e change is just a bunch of code additions for a new phy type,
not a bunch of /changes/ really. Is that the change you're concerned with?

Jeff



--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Linus Torvalds
2008-05-09 15:05:47 UTC
Permalink
The e1000e change is just a bunch of code additions for a new phy type, not a
bunch of /changes/ really. Is that the change you're concerned with?
Yes. I saw 800 new lines to a driver I use, and I go "Ugh".

Linus
David Miller
2008-05-09 21:43:59 UTC
Permalink
From: Linus Torvalds <***@linux-foundation.org>
Date: Fri, 9 May 2008 08:05:47 -0700 (PDT)
Post by Linus Torvalds
The e1000e change is just a bunch of code additions for a new phy type, not a
bunch of /changes/ really. Is that the change you're concerned with?
Yes. I saw 800 new lines to a driver I use, and I go "Ugh".
And if it enabled support for a device you have, for which you've been
patching the driver for some time, you'd go "Yippie!" :-)))

BTW, this reminds me, did we get that driver in for the device the
EEPC machines use? I have this strange feeling like that slipped
through the cracks.

--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Jay Cliburn
2008-05-09 21:58:32 UTC
Permalink
On Fri, 09 May 2008 14:43:59 -0700 (PDT)
Post by David Miller
BTW, this reminds me, did we get that driver in for the device the
EEPC machines use? I have this strange feeling like that slipped
through the cracks.
That'd be the Atheros/Attansic L2 NIC. I don't think it's quite ready
yet.

[Adding Chris Snook to the cc list.]
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Chris Snook
2008-05-09 22:51:20 UTC
Permalink
Post by Jay Cliburn
On Fri, 09 May 2008 14:43:59 -0700 (PDT)
Post by David Miller
BTW, this reminds me, did we get that driver in for the device the
EEPC machines use? I have this strange feeling like that slipped
through the cracks.
That'd be the Atheros/Attansic L2 NIC. I don't think it's quite ready
yet.
[Adding Chris Snook to the cc list.]
I had a sudden shortage of Copious Free Time for the past few weeks.
I'm going to post it this weekend. Once it's merged with the atlx code,
it'll be much easier for others to help out with it.

-- Chris
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
David Miller
2008-05-09 10:18:09 UTC
Permalink
The SIT bug fix is, unfortunately, remotely exploitable memory leak.
As long as the SIT module is loaded and SIT packets aren't blocked by
the firewall, it's pretty easy to construct a frame that will be
leaked every single time (IPV4 header, with protocol == SIT, and
length just convering the IPV4 header).

I'm queueing that SIT fix up to -stable in parallel as I type this.

The TIPC fix makes sure the is enough header space for all possible
output devices.

Please pull, thanks a lot!

The following changes since commit e46b66bc42b6b1430b04cc5c207ecb2b2f4553dc:
Ben Hutchings (1):
net: Added ASSERT_RTNL() to dev_open() and dev_close().

are available in the git repository at:

master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6.git master

Allan Stephens (1):
tipc: Increase buffer header to support worst-case device

David S. Miller (1):
sit: Add missing kfree_skb() on pskb_may_pull() failure.

net/ipv6/sit.c | 2 +-
net/tipc/core.h | 11 +++++------
2 files changed, 6 insertions(+), 7 deletions(-)
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to ***@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
David Miller
2008-05-14 11:06:02 UTC
Permalink
1) The mac80211 layer and drivers using it had this awful
problem on TX that causes basically every packet to be
copied before we give it to the device.

Thanks to some detective work by Johannes Berg we were
able to fix up the bulk of these problems.

The solution is two fold:

i) First we adjust LL_MAX_HEADER properly for wireless
cases

ii) We let the driver indicate any extra header and tail
room it might need, beyond netdev->hard_header_len,
in the netdev struct. Then mac80211 is updated to
set this as appropriate.

The other wireless stack we have (ieee80211) and bare wireless
drivers still have some similar issues, but with this
infrastructure they should be quite easy to fix.

2) Fix syncppp back up by providing a proper mid-layer private
pointer in the netdev struct, which is what this thing
wanted from the beginning. Thus, we can remove the BROKEN
tag in Kconfig

3) Misalignment fix in irda from Graf Yang, use get_unaligned().

4) Ilpo fixed the bad TCP inteactions with some weird printers.

They only will accept data in-order, and with FRTO (which is
enabled by default) we could get into a wedged state and stop
making forward progress after loss.

I very much appreciate all of the effort he and the bug reporter
put into diagnosing this problem and coming up with a fix.

5) Uninitialized pointer fix in SCTP by Patrick McHardy

6) CIPSO calculates it's hash function imporperly, fix from
Pavel Emelyanov

7) PPP frees registered netdevices in error path, also from
Pavel Emelyanov

8) Netfilter xt_iprange is missing module aliases, fix from
Phil Oester

9) Two SCTP fixes from Wei Yongjun

10) Wireless updates and fixes from John Linville and the rest
of the wireless crew.

11) Wired driver updates from Jeff Garzik and co.

You will notice the larger bits for the SFQ driver in
here, but that driver is new and thus these changes
cannot cause regressions so I think they are appropriate
to add at this time. If you disagree just let us know. :)

Also of note is that Auke has passed the Intel ethernet
driver torch on to another crazy fool :-)

Please pull, thanks a lot!

The following changes since commit 36ca34cc3b8335eb1fe8bd9a1d0a2592980c=
3f02:
David S. Miller (1):
sit: Add missing kfree_skb() on pskb_may_pull() failure.

are available in the git repository at:

master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6.git master

Andy Fleming (1):
gianfar: Fix a bug where the pointer never moves for dma_unmap...

Auke Kok (1):
New maintainer for Intel ethernet adapters

Ben Dooks (1):
DM9000: Use delayed work to update MII PHY state

Ben Hutchings (9):
[netdrvr] sfc: Add TSO support
[netdrvr] sfc: Add phy_flash_cfg module parameter and implementat=
ion
[netdrvr] sfc: Removed bogus 'fall-thru' comments
[netdrvr] sfc: Remove garbage from comment
[netdrvr] sfc: Remove kernel-doc comments for removed members of =
struct efx_nic
[netdrvr] sfc: Fix code formatting
[netdrvr] sfc: Remove unused macro EFX_XAUI_RETRAIN_MAX
[netdrvr] sfc: Increment rx_reset when reported as driver event
[netdrvr] sfc: sfc: Add self-test support

Brice Goglin (11):
myri10ge: update firmware headers
myri10ge: fix module parameter descriptions
myri10ge: increase and fix handoff timeout
myri10ge: properly align scratch buffers
myri10ge: report FIBER in ethtool for XFP based NIC
myri10ge: add barrier in myri10ge_send_cmd
myri10ge: trivial formatting fix
myri10ge: fix potential infinite loop in enable_ecrc
myri10ge: move data structures into a single slice
myri10ge: cleanup retrieving of firmware capabilities
myri10ge: fix the number of interrupt slots

David S. Miller (5):
syncppp: Fix crashes.
Merge branch 'master' of git://git.kernel.org/.../linville/wirele=
ss-2.6
net: Set LL_MAX_HEADER properly for wireless.
mac80211: Use skb_header_cloned() on TX path.
Merge branch 'upstream-davem' of master.kernel.org:/.../jgarzik/n=
etdev-2.6

Divy Le Ray (2):
cxgb3 - fix port up/down error path
cxgb3 - fix EEH

Emmanuel Grumbach (1):
iwlwifi: Fix frequency in rx_status fill

Enrico Scholz (2):
DM9000: Add __devinit and __devexit attributes to probe and remov=
e
DM9000: Update and fix driver debugging messages

Eric Leblond (1):
netfilter: ctnetlink: dump conntrack ID in event messages

Graf Yang (1):
irda: Fix a misalign access issue. (v2)

Hannes Hering (3):
memory: Introduce exports for memory notifiers
ehea: Add dependency to Kconfig
ehea: Add DLPAR memory remove support

Helmut Schaa (1):
mac80211: fix association with some APs

Ilpo J=E4rvinen (2):
tcp FRTO: Fix fallback to conventional recovery
tcp FRTO: work-around inorder receivers

Ivo van Doorn (4):
rt2x00: Don't use pskb_expand_head()
rt2x00: Fix broken recover-on-error path
rt2x00: Clean up error handling of PCI queue DMA allocation.
mac80211: Don't encrypt beacons

Jay Cliburn (4):
atl1: add PHY power save mode
atl1: fix broken suspend and resume
atl1: add shutdown callback
atl1: bump version number

Jesse Brandeburg (1):
sky2: fix simple define thinko

Johannes Berg (4):
mac80211: fix debugfs default key oops
zd1211rw: fix potential use-after-free bug
net: Allow netdevices to specify needed head/tailroom
mac80211: assign needed_headroom/tailroom for netdevs

John W. Linville (1):
wavelan: avoid index past end of array if DEBUG_SHOW_UNUSED is de=
fined

Julia Lawall (1):
pppol2tp: Remove null pointer dereference.

Luis Carlos Cobo (2):
mac80211: fix incorrect mesh header length
mac80211: fix access to null skb

Luis R. Rodriguez (1):
prism54: fix regression with missing carrier in AP-mode

Masakazu Mokuno (1):
PS3: gelic: fix memory leak

Matheos Worku (1):
niu: Determine the # of ports from the card's VPD data

Patrick McHardy (2):
mac80211: mesh hwmp: fix kfree(skb)
sctp: Fix use of uninitialized pointer

Pavel Emelyanov (5):
Fix GFP_KERNEL allocation under read lock.
Fix not checked kmalloc() result.
Fix potential scheduling while atomic in mesh_path_add.
cipso: Relax too much careful cipso hash function.
ppp: Do not free not yet unregistered net device.

Pavel Roskin (1):
mac80211: add missing newlines in printk()

Phil Oester (1):
netfilter: xt_iprange: module aliases for xt_iprange

Rami Rosen (1):
net: Fix typo in net/core/sock.c.

Steven Rostedt (1):
fix irq flags in mac80211 code

Wei Yongjun (2):
sctp: Do not enable peer IPv6 address support on PF_INET socket
sctp: Add address type check while process paramaters of ASCONF c=
hunk

Zhu Yi (1):
iwlwifi: fix compile error when CONFIG_MAC80211_DEBUGFS is not se=
lected

MAINTAINERS | 6 +-
drivers/base/memory.c | 2 +
drivers/net/Kconfig | 2 +-
drivers/net/atlx/atl1.c | 157 ++++--
drivers/net/atlx/atl1.h | 2 +-
drivers/net/atlx/atlx.c | 2 +-
drivers/net/atlx/atlx.h | 7 +-
drivers/net/cxgb3/adapter.h | 1 +
drivers/net/cxgb3/common.h | 1 +
drivers/net/cxgb3/cxgb3_main.c | 40 +-
drivers/net/cxgb3/regs.h | 8 +
drivers/net/cxgb3/sge.c | 29 +-
drivers/net/cxgb3/t3_hw.c | 28 +
drivers/net/dm9000.c | 37 ++-
drivers/net/ehea/ehea.h | 27 +-
drivers/net/ehea/ehea_main.c | 25 +
drivers/net/ehea/ehea_qmr.c | 286 +++++++---
drivers/net/gianfar.c | 2 +
drivers/net/myri10ge/myri10ge.c | 730 +++++++++++++---=
-------
drivers/net/myri10ge/myri10ge_mcp.h | 56 ++-
drivers/net/myri10ge/myri10ge_mcp_gen_header.h | 39 +-
drivers/net/niu.c | 53 +-
drivers/net/niu.h | 9 +
drivers/net/ppp_generic.c | 1 +
drivers/net/pppol2tp.c | 13 +-
drivers/net/ps3_gelic_wireless.c | 2 +
drivers/net/sfc/Makefile | 4 +-
drivers/net/sfc/boards.h | 2 +
drivers/net/sfc/efx.c | 4 +-
drivers/net/sfc/enum.h | 49 ++
drivers/net/sfc/ethtool.c | 259 +++++++++-
drivers/net/sfc/falcon.c | 8 +-
drivers/net/sfc/falcon_hwdefs.h | 16 +-
drivers/net/sfc/falcon_xmac.c | 82 +++-
drivers/net/sfc/mdio_10g.c | 78 +++
drivers/net/sfc/mdio_10g.h | 24 +-
drivers/net/sfc/net_driver.h | 28 +-
drivers/net/sfc/rx.c | 11 +-
drivers/net/sfc/selftest.c | 717 ++++++++++++++++=
+++++++
drivers/net/sfc/selftest.h | 50 ++
drivers/net/sfc/sfe4001.c | 14 +
drivers/net/sfc/tenxpress.c | 91 +++-
drivers/net/sfc/tx.c | 664 ++++++++++++++++=
+++++
drivers/net/sfc/xfp_phy.c | 36 ++
drivers/net/sky2.h | 4 +-
drivers/net/wan/Kconfig | 4 +-
drivers/net/wan/cosa.c | 14 +-
drivers/net/wan/hdlc_ppp.c | 2 +-
drivers/net/wan/hostess_sv11.c | 12 +-
drivers/net/wan/lmc/lmc_main.c | 1 +
drivers/net/wan/sealevel.c | 1 +
drivers/net/wireless/iwlwifi/iwl-3945.c | 2 +-
drivers/net/wireless/iwlwifi/iwl-4965-rs.c | 2 +-
drivers/net/wireless/iwlwifi/iwl-4965.c | 2 +-
drivers/net/wireless/prism54/islpci_dev.c | 11 +-
drivers/net/wireless/rt2x00/rt2x00dev.c | 11 +-
drivers/net/wireless/rt2x00/rt2x00pci.c | 5 +-
drivers/net/wireless/rt2x00/rt61pci.c | 31 +-
drivers/net/wireless/wavelan.c | 4 +-
drivers/net/wireless/wavelan_cs.c | 6 +-
drivers/net/wireless/zd1211rw/zd_usb.c | 6 +-
include/linux/netdevice.h | 35 +-
include/net/irda/discovery.h | 3 -
include/net/syncppp.h | 2 +-
net/core/netpoll.c | 2 +-
net/core/sock.c | 2 +-
net/econet/af_econet.c | 2 +-
net/ipv4/arp.c | 2 +-
net/ipv4/cipso_ipv4.c | 4 +-
net/ipv4/igmp.c | 4 +-
net/ipv4/ipconfig.c | 6 +-
net/ipv4/raw.c | 10 +-
net/ipv4/tcp_input.c | 17 +-
net/ipv6/ip6_output.c | 2 +-
net/ipv6/mcast.c | 4 +-
net/ipv6/ndisc.c | 4 +-
net/ipv6/raw.c | 10 +-
net/irda/discovery.c | 8 +-
net/irda/irlmp.c | 5 +-
net/irda/irnet/irnet_irda.c | 4 +-
net/mac80211/debugfs_key.c | 15 +-
net/mac80211/iface.c | 9 +
net/mac80211/mesh.c | 2 +-
net/mac80211/mesh_hwmp.c | 2 +-
net/mac80211/mesh_pathtbl.c | 17 +-
net/mac80211/mlme.c | 64 ++-
net/mac80211/rc80211_pid_debugfs.c | 4 +-
net/mac80211/rx.c | 12 +-
net/mac80211/tx.c | 5 +-
net/mac80211/util.c | 10 +-
net/mac80211/wme.c | 3 +-
net/netfilter/nf_conntrack_netlink.c | 3 +
net/netfilter/xt_iprange.c | 2 +
net/packet/af_packet.c | 2 +-
net/sctp/sm_make_chunk.c | 16 +-
net/xfrm/xfrm_output.c | 6 +-
96 files changed, 3380 insertions(+), 738 deletions(-)
create mode 100644 drivers/net/sfc/selftest.c
create mode 100644 drivers/net/sfc/selftest.h

Continue reading on narkive:
Loading...