When a non-wimax interface is looked up by the stack, a bad pointer is returned when the looked-up interface is not found in the list (of registered WiMAX interfaces). This causes an oops in the caller when trying to use the pointer. Fix by properly setting the pointer to NULL if we don't exit from the list_for_each() with a found entry. Signed-off-by: Inaky Perez-Gonzalez <inaky@linux.intel.com> --- net/wimax/id-table.c | 9 +++++---- 1 files changed, 5 insertions(+), 4 ...

Fix this sparse warnings: drivers/isdn/hardware/mISDN/hfcpci.c:59:19: warning: symbol 'hfc_tl' was not declared. Should it be static? drivers/isdn/hardware/mISDN/hfcpci.c:60:9: warning: symbol 'hfc_jiffies' was not declared. Should it be static? drivers/isdn/hardware/mISDN/hfcsusb.h:201:6: warning: symbol 'conf_str' was not declared. Should it be static? drivers/isdn/hardware/mISDN/hfcsusb.h:319:12: warning: symbol 'HFC_TE_LAYER1_STATES' was not declared. Should it be static? ...

Subject: [PATCH 39/41]bnx2x: Comments and prints Signed-off-by: Eilon Greenstein <eilong@broadcom.com> --- drivers/net/bnx2x_fw_defs.h | 9 ++-- drivers/net/bnx2x_hsi.h | 2 +- drivers/net/bnx2x_init.h | 94 +++++++++++++++++++++--------------------- drivers/net/bnx2x_link.c | 5 +- drivers/net/bnx2x_main.c | 43 ++++++++++---------- 5 files changed, 78 insertions(+), 75 deletions(-) diff --git a/drivers/net/bnx2x_fw_defs.h b/drivers/net/bnx2x_fw_defs.h index ...

Subject: [PATCH 16/41]bnx2x: Supporting BCM8481 PHY Signed-off-by: Yaniv Rosner <yanivr@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> --- drivers/net/bnx2x_link.c | 95 +++++++++++++++++++++++++++++++++++++++++++++- drivers/net/bnx2x_main.c | 17 ++++++++ 2 files changed, 110 insertions(+), 2 deletions(-) diff --git a/drivers/net/bnx2x_link.c b/drivers/net/bnx2x_link.c index 55f50c7..bed11b9 100644 --- a/drivers/net/bnx2x_link.c +++ b/drivers/net/bnx2x_link.c @@ ...

Subject: [PATCH 36/41]bnx2x: Driver dump Re-arrange the driver debug print to support the Tx multi-queue and add support for debug prints at the end of the ring (when start<end) Signed-off-by: Eilon Greenstein <eilong@broadcom.com> --- drivers/net/bnx2x_main.c | 97 ++++++++++++++++++++++++++++------------------ 1 files changed, 59 insertions(+), 38 deletions(-) diff --git a/drivers/net/bnx2x_main.c b/drivers/net/bnx2x_main.c index 26451f6..78f628a 100644 --- ...

Sorry about that, I got lazy at this part. Since bnx2x_tx_int is also writing, the smb_rmb should be replaced with Hmmm... It seems so... I will have to run it on few different systems in the lab to make sure - if I won't run into any problems I will send a new patch on top of this one with the removal of this alleged redundant smb_mb. --

Subject: [PATCH 27/41]bnx2x: smp_mb and not just smp_rmb Signed-off-by: Eilon Greenstein <eilong@broadcom.com> --- drivers/net/bnx2x_main.c | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/net/bnx2x_main.c b/drivers/net/bnx2x_main.c index 6bd9219..2e95799 100644 --- a/drivers/net/bnx2x_main.c +++ b/drivers/net/bnx2x_main.c @@ -7173,7 +7173,7 @@ static int bnx2x_nic_unload(struct bnx2x *bp, int unload_mode) struct bnx2x_fastpath *fp = &bp->fp[i]; ...

Subject: [PATCH 11/41]bnx2x: Removing the board type There are too many different board types and this field is not scalable. Removing it and making decisions according to other fields Signed-off-by: Eilon Greenstein <eilong@broadcom.com> --- drivers/net/bnx2x.h | 1 - drivers/net/bnx2x_hsi.h | 34 +++++++--------------------------- drivers/net/bnx2x_main.c | 20 +++++++------------- 3 files changed, 14 insertions(+), 41 deletions(-) diff --git a/drivers/net/bnx2x.h ...

Subject: [PATCH 14/41]bnx2x: Supporting BCM8726 PHY Also adding the ability to recognize the optic module and disable it if it is not authorized for safety reasons - since this feature might upset some users which are willing to take the risk, it is optional and can be disabled by setting an nvram bit (or a trivial driver patch to set this bit). This dual port PHY requires special handling if the ports are swapped. Signed-off-by: Yaniv Rosner <yanivr@broadcom.com> Signed-off-by: Eilon ...

Subject: [PATCH 21/41]bnx2x: GMII not SGMII Fixing the function name Signed-off-by: Eilon Greenstein <eilong@broadcom.com> --- drivers/net/bnx2x_link.c | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/net/bnx2x_link.c b/drivers/net/bnx2x_link.c index 526072b..b61a7a2 100644 --- a/drivers/net/bnx2x_link.c +++ b/drivers/net/bnx2x_link.c @@ -1753,7 +1753,7 @@ static u8 bnx2x_link_settings_status(struct link_params *params, return rc; } -static void ...

Subject: [PATCH 10/41]bnx2x: Flow control enhancement Setting better HW thresholds and enabling FW capabilities for better enforcement. Also set the HW to more efficiently use the internal buffers if this is a single port design Signed-off-by: Eilon Greenstein <eilong@broadcom.com> --- drivers/net/bnx2x.h | 2 +- drivers/net/bnx2x_fw_defs.h | 5 ++ drivers/net/bnx2x_hsi.h | 35 ++++++++++++++++ drivers/net/bnx2x_main.c | 96 ...

Subject: [PATCH 38/41]bnx2x: Re-arrange module parameters Signed-off-by: Eilon Greenstein <eilong@broadcom.com> --- drivers/net/bnx2x_main.c | 14 +++++++------- 1 files changed, 7 insertions(+), 7 deletions(-) diff --git a/drivers/net/bnx2x_main.c b/drivers/net/bnx2x_main.c index 3f84027..0d6015f 100644 --- a/drivers/net/bnx2x_main.c +++ b/drivers/net/bnx2x_main.c @@ -74,26 +74,26 @@ static int multi_mode = 1; module_param(multi_mode, int, 0); static int disable_tpa; -static int ...

Subject: [PATCH 34/41]bnx2x: Sparse endianity annotation Resolving the majority of the issues, but there are still some left for future patches. Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> --- drivers/net/bnx2x.h | 22 +++--- drivers/net/bnx2x_hsi.h | 178 +++++++++++++++++++++++----------------------- drivers/net/bnx2x_main.c | 63 ++++++++-------- 3 files changed, 132 insertions(+), 131 deletions(-) diff ...

Subject: [PATCH 32/41]bnx2x: Better struct naming This is the device info from the chip shared memory - rename to avoid confusion Signed-off-by: Eilon Greenstein <eilong@broadcom.com> --- drivers/net/bnx2x_hsi.h | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/net/bnx2x_hsi.h b/drivers/net/bnx2x_hsi.h index 8452605..88b2aa3 100644 --- a/drivers/net/bnx2x_hsi.h +++ b/drivers/net/bnx2x_hsi.h @@ -488,7 +488,7 @@ struct port_feat_cfg { /* port 0: 0x454 ...

Subject: [PATCH 30/41]bnx2x: Optimize chip MPS configuration Signed-off-by: Eilon Greenstein <eilong@broadcom.com> --- drivers/net/bnx2x_init.h | 21 +++++++++++---------- 1 files changed, 11 insertions(+), 10 deletions(-) diff --git a/drivers/net/bnx2x_init.h b/drivers/net/bnx2x_init.h index 021438f..6fcd1dc 100644 --- a/drivers/net/bnx2x_init.h +++ b/drivers/net/bnx2x_init.h @@ -655,17 +655,18 @@ static void bnx2x_init_pxp(struct bnx2x *bp) REG_WR(bp, PXP2_REG_WR_USDMDP_TH, (0x18 << ...

Subject: [PATCH 28/41]bnx2x: Using nvram default WoL config The driver default WoL setting is set to the nvram default value Signed-off-by: Eilon Greenstein <eilong@broadcom.com> --- drivers/net/bnx2x_main.c | 6 ++++++ 1 files changed, 6 insertions(+), 0 deletions(-) diff --git a/drivers/net/bnx2x_main.c b/drivers/net/bnx2x_main.c index 2e95799..a14fe75 100644 --- a/drivers/net/bnx2x_main.c +++ b/drivers/net/bnx2x_main.c @@ -8027,6 +8027,12 @@ static void __devinit ...

Subject: [PATCH 24/41]bnx2x: Manual setting for MRRS Allows better debug capabilities if the user wants to force an MRRS value Signed-off-by: Eilon Greenstein <eilong@broadcom.com> --- drivers/net/bnx2x.h | 1 + drivers/net/bnx2x_init.h | 7 ++++++- drivers/net/bnx2x_main.c | 6 ++++++ 3 files changed, 13 insertions(+), 1 deletions(-) diff --git a/drivers/net/bnx2x.h b/drivers/net/bnx2x.h index 9834a86..891a58e 100644 --- a/drivers/net/bnx2x.h +++ b/drivers/net/bnx2x.h @@ ...

Subject: [PATCH 23/41]bnx2x: Using registers name Signed-off-by: Eilon Greenstein <eilong@broadcom.com> --- drivers/net/bnx2x.h | 9 ++++-- drivers/net/bnx2x_link.c | 71 +++++++++++++++++++++++---------------------- drivers/net/bnx2x_reg.h | 7 ++++ 3 files changed, 49 insertions(+), 38 deletions(-) diff --git a/drivers/net/bnx2x.h b/drivers/net/bnx2x.h index 12d2d0b..9834a86 100644 --- a/drivers/net/bnx2x.h +++ b/drivers/net/bnx2x.h @@ -446,10 +446,13 @@ struct ...

Subject: [PATCH 17/41]bnx2x: Saving PHY FW version Some PHYs (like the BCM8726) FW version cannot be read after activating the PHY, so we need to save this information Signed-off-by: Yaniv Rosner <yanivr@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> --- drivers/net/bnx2x_hsi.h | 2 + drivers/net/bnx2x_link.c | 180 +++++++++++++++++++++++----------------------- 2 files changed, 92 insertions(+), 90 deletions(-) diff --git a/drivers/net/bnx2x_hsi.h ...

Subject: [PATCH 05/41]bnx2x: Disabling interrupts after iSCSI-boot Before initializing the chip after iSCSI boot, the interrupts of the function that was used to boot must be disabled. That means that the driver needs to set the chip as if it is the iSCSI PCI function - this bug is exposed only with MSI Signed-off-by: Eilon Greenstein <eilong@broadcom.com> --- drivers/net/bnx2x_main.c | 61 ++++++++++++++++++++++++++++++++++++++++++++- drivers/net/bnx2x_reg.h | 13 ++++++++++ 2 files ...

Subject: [PATCH 12/41]bnx2x: Removing redundant device parameters Signed-off-by: Eilon Greenstein <eilong@broadcom.com> --- drivers/net/bnx2x.h | 5 ----- drivers/net/bnx2x_main.c | 19 +++++++++---------- 2 files changed, 9 insertions(+), 15 deletions(-) diff --git a/drivers/net/bnx2x.h b/drivers/net/bnx2x.h index a18ace0..2397a89 100644 --- a/drivers/net/bnx2x.h +++ b/drivers/net/bnx2x.h @@ -525,8 +525,6 @@ struct bnx2x_common { u32 hw_config; ...

Subject: [PATCH 06/41]bnx2x: Removing microcode assertion check This check is unreliable since latest MC can issue warnings on rare occasions which are not fatal errors Signed-off-by: Eilon Greenstein <eilong@broadcom.com> --- drivers/net/bnx2x.h | 2 +- drivers/net/bnx2x_main.c | 6 +----- 2 files changed, 2 insertions(+), 6 deletions(-) diff --git a/drivers/net/bnx2x.h b/drivers/net/bnx2x.h index 89a72c8..cfbeabb 100644 --- a/drivers/net/bnx2x.h +++ b/drivers/net/bnx2x.h @@ ...

On Thu, 2009-02-12 at 20:36 +0200, Eilon Greenstein wrote: This appears to be a capability offset, not a capability id! You might [...] Similarly for this. Ben. -- Ben Hutchings, Senior Software Engineer, Solarflare Communications Not speaking for my employer; that's the marketing department's job. They asked us to note that Solarflare product names are trademarked. --

On Thu, 2009-02-12 at 10:52 -0800, Ben Hutchings wrote: Hi Ben! It's always a pleasure to see how fast you respond. I really appreciate You are right - those are the offset for this device. I find it easier to debug when I have the offsets at hand without the need to parse the capabilities linked list. Since the offsets will not change for this generation of chips, I don't see the harm in providing the explicit offsets. I assume that the driver will require major changes for a new generation ...

Subject: [PATCH 03/41]bnx2x: MSI support Enabling MSI on top of MSI-X and INTA. Also changing the module parameter to allow choosing INTA or MSI even when MSI-X is available. The default status block should not be reversed for endianity. Since MSI can issue re-configuration, the interrupt disable function now requires mmiowb Signed-off-by: Eilon Greenstein <eilong@broadcom.com> --- drivers/net/bnx2x.h | 5 +- drivers/net/bnx2x_hsi.h | 6 -- drivers/net/bnx2x_main.c | 155 ...

It does not since linux/types.h header is not included, but better use Hmm... $ grep xtables_strtoui -r /tmp/iptables-1.4.2 Well, having & clearly shows it is a pointer and not a value, especially when it is not clear from the name what it should be. -- Evgeniy Polyakov --

Hi. They can be accessed via connector configuration, but it is stopped above (and it waits for currently active users to run away); and netfilter path, which should be prevented after match was also unregistered above. -- Evgeniy Polyakov --

I advocate using this to get more accurate censuses of machines accessing given web servers, given the tendency of browsers to lie about themselves in order to avoid being shut out of certain web sites. One question about the module-unload sequence below. Given a reasonable answer to that question, I am OK with this from an RCU viewpoint. For the above to be safe: o Any remaining RCU callbacks cannot reference the list (and your callbacks do in fact meet this constraint). o Any ...

The last one was. Your answer was that you prevented any concurrent access, but I frankly don't understand the netfilter/connector code well enough to do more than simply believe you. --

Good to know. Please add this explanation as a comment If GCC had a switch to compile with I16 or I64, I could test, but it does not. 'long', as used in the netfilter includes, already bit people in pre-2.6.19, and nowadays we have that compat crap thing in place.. Just don't take any chance, and go the safe net/netfilter/Kconfig: (e.g.) config NETFILTER_XT_MATCH_COMMENT tristate '"comment" match support' depends on NETFILTER_ADVANCED The underlying implementation ...

Both TX and RX hardware time stamping are implemented. Due to hardware limitations it is not possible to verify reliably which packet was time stamped when multiple were pending for sending; this could be solved by only allowing one packet marked for hardware time stamping into the queue (not implemented yet). RX time stamping relies on the flag in the packet descriptor which marks packets that were time stamped. In "all packet" mode this flag is not set. TODO: also support that mode (even ...

Signed-off-by: Patrick Ohly <patrick.ohly@intel.com> --- fs/compat_ioctl.c | 6 ++++++ net/core/dev.c | 2 ++ 2 files changed, 8 insertions(+), 0 deletions(-) diff --git a/fs/compat_ioctl.c b/fs/compat_ioctl.c index c03c10d..ab7585a 100644 --- a/fs/compat_ioctl.c +++ b/fs/compat_ioctl.c @@ -522,6 +522,11 @@ static int dev_ifsioc(unsigned int fd, unsigned int cmd, unsigned long arg) if (err) return -EFAULT; break; + case SIOCSHWTSTAMP: + if (copy_from_user(&ifr, ...

The additional per-packet information (16 bytes for time stamps, 1 byte for flags) is stored for all packets in the skb_shared_info struct. This implementation detail is hidden from users of that information via skb_* accessor functions. A separate struct resp. union is used for the additional information so that it can be stored/copied easily outside of skb_shared_info. Compared to previous implementations (reusing the tstamp field depending on the context, optional additional structures) ...

The overlap with the old SO_TIMESTAMP[NS] options is handled so that time stamping in software (net_enable_timestamp()) is enabled when SO_TIMESTAMP[NS] and/or SO_TIMESTAMPING_RX_SOFTWARE is set. It's disabled if all of these are off. Signed-off-by: Patrick Ohly <patrick.ohly@intel.com> --- include/net/sock.h | 44 +++++++++++++++++++++++++-- net/compat.c | 19 +++++++---- net/core/sock.c | 81 ++++++++++++++++++++++++++++++++++++++++++------- net/socket.c | 84 ...

Instructions for time stamping outgoing packets are take from the socket layer and later copied into the new skb. --- Documentation/networking/timestamping.txt | 2 ++ include/net/ip.h | 1 + net/can/raw.c | 3 +++ net/ipv4/icmp.c | 2 ++ net/ipv4/ip_output.c | 6 ++++++ net/ipv4/raw.c | 1 + net/ipv4/udp.c | 4 ++++ 7 ...

Signed-off-by: John Ronciak <john.ronciak@intel.com> --- drivers/net/igb/igb_main.c | 31 +++++++++++++++++++++++++++++++ 1 files changed, 31 insertions(+), 0 deletions(-) diff --git a/drivers/net/igb/igb_main.c b/drivers/net/igb/igb_main.c index 8b2ba42..90090bb 100644 --- a/drivers/net/igb/igb_main.c +++ b/drivers/net/igb/igb_main.c @@ -34,6 +34,7 @@ #include <linux/ipv6.h> #include <net/checksum.h> #include <net/ip6_checksum.h> +#include <linux/net_tstamp.h> #include ...

Adds the register definitions and code to read the time register. Signed-off-by: John Ronciak <john.ronciak@intel.com> --- drivers/net/igb/e1000_regs.h | 28 +++++++++++ drivers/net/igb/igb.h | 4 ++ drivers/net/igb/igb_main.c | 111 ++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 143 insertions(+), 0 deletions(-) diff --git a/drivers/net/igb/e1000_regs.h b/drivers/net/igb/e1000_regs.h index 5038b73..64d95cd 100644 --- a/drivers/net/igb/e1000_regs.h +++ ...

The overlap with the old SO_TIMESTAMP[NS] options is handled so that time stamping in software (net_enable_timestamp()) is enabled when SO_TIMESTAMP[NS] and/or SO_TIMESTAMPING_RX_SOFTWARE is set. It's disabled if all of these are off. --- include/net/sock.h | 44 +++++++++++++++++++++++++-- net/compat.c | 19 +++++++---- net/core/sock.c | 81 ++++++++++++++++++++++++++++++++++++++++++------- net/socket.c | 84 +++++++++++++++++++++++++++++++++++++++------------ 4 files ...

Mapping from a struct timecounter to a time returned by functions like ktime_get_real() is implemented. This is sufficient to use this code in a network device driver which wants to support hardware time stamping and transformation of hardware time stamps to system time. The interface could have been made more versatile by not depending on a time counter, but this wasn't done to avoid writing glue code elsewhere. The method implemented here is the one used and analyzed under the ...

User space can request hardware and/or software time stamping. Reporting of the result(s) via a new control message is enabled separately for each field in the message because some of the fields may require additional computation and thus cause overhead. User space can tell the different kinds of time stamps apart and choose what suits its needs. When a TX timestamp operation is requested, the TX skb will be cloned and the clone will be time stamped (in hardware or software) and added to the ...

[once more, with signed-off - sorry for the noise] This revision of the patch series transports hardware time stamps from the hardware into user space via additional fields in struct skb_shared_info. It's based on net-next-2.6 as of this morning. This is the solution that emerged from the discussion of the other approaches suggested before (reuse tstamp, extend skbuff, optional structs): it has the advantage of not touching skbuff and also has the simplest implementation. The clocksource ...

Hi all, I have a scenario with CLIP IPoA(RFC1577) atm link over ADSL on the WAN, 801.1q VLANs on the LAN, and NAT/MASQUERADE that does not work. Network config: Nat_host <--> router <---> server a ping from the Nat_host reaches the server on the WAN fine, and the ping comes back to the router, but the ping response never reaches the Nat_Host. Using TRACE rules it seems the ICMP ping response gets lost inside the router. I see the same behavior with TCP and ...

Your email ID has been awarded £950,000,00 GBP in the Compaq PROMO,Reply us with your info: (jerry.smith11@btinternet.com) Name:............. Country:.......... Sex:.............. Addres:........... Regards Mrs Sarah Wood --

Hi David Please pull git://git.kernel.org/pub/scm/linux/kernel/git/inaky/linux-firmware-wimax.git This adds v1.4 of the i2400m WiMAX firmware, which matches the user space code release available in http://linuxwimax.org/Download. The driver code slated for 2.6.29 works with both the v1.3 and v1.4 versions. Thank you, -- Inaky --

From: Jie Yang <jie.yang@atheros.com> Supporting AR8131, and AR8132. Signed-off-by: Jie Yang <jie.yang@atheros.com> --- Updated on David Miller's comments. diff --git a/drivers/net/Kconfig b/drivers/net/Kconfig index 62bc022..cdb8c46 100644 --- a/drivers/net/Kconfig +++ b/drivers/net/Kconfig @@ -2350,6 +2350,17 @@ config ATL1E To compile this driver as a module, choose M here. The module will be called atl1e. +config ATL1C + tristate "Atheros L1C Gigabit Ethernet support ...

commit 796b5d184b4df1aae55894bf476959da83e25324 Author: Eric Leblond <eric@inl.fr> Date: Thu Feb 12 08:00:35 2009 +0100 netfilter: fix tuple inversion for Node information request Upstream commit: a51f42f3c The patch fixes a typo in the inverse mapping of Node Information request. Following draft-ietf-ipngwg-icmp-name-lookups-09, "Querier" sends a type 139 (ICMPV6_NI_QUERY) packet to "Responder" which answer with a type 140 (ICMPV6_NI_REPLY) packet. ...

Base versions handle constant folding now. For headers exposed to userspace, we must only expose the __ prefixed versions. Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com> --- include/linux/if_pppox.h | 20 ++++---- include/linux/if_tunnel.h | 16 +++--- include/linux/ncp_no.h | 26 +++++----- include/linux/netdevice.h | 4 +- include/linux/netfilter_bridge.h | 4 +- include/linux/pim.h | 4 +- include/linux/sctp.h ...

On 12-02-2009 01:41, Jan Engelhardt wrote: Better be careful! History, like Greek myths, or even quite recent, could be surprising... Jarek P. --

...But love over gold looks like safe bet! (love over beef too ;-) Jarek P. --

Well, I have no way of knowing this, because that tree is not in linux-next. Please get it added? --

I tried adding a rule to log these unaccounted-for packets. Nothing It does sound like the result of a bug. Do you have any pointers to patches or locations to check in the source? Alan Stern --

0xdeadbeef is a really bad choice of an arbitrary value. It is already used in multiple places. Only a few of which are currently listed in linux/poison.h and if I happened to see that number in a debug trace having to sort through all of the possible sources looks like a major pain. I don't really care what we call it but somehow 0xdeadbeef strikes me as wrong and something that will make debugging harder. Eric --

OK, here follows the patch with the arbitrary initial parameter renamed. ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ The current jhash.h implements the lookup2() hash function by Bob Jenkins. However, lookup2() is outdated as Bob wrote a new hash function called lookup3(). The new hash function - mixes better than lookup2(): it passes the check that every input bit changes every output bit 50% of the time, while lookup2() failed it. - performs better: ...

Hi Joszef, My concern was that it's also bigger (and we inline it). Performance is pretty much a wash since we so rarely hash more than a few words. Any stats on code size changes? Rusty. --

Thank Patrick, patch updated with rol32(): Signed-off-by: Jozsef Kadlecsik <kadlec@blackhole.kfki.hu> diff --git a/include/linux/jhash.h b/include/linux/jhash.h index 2a2f99f..e48ecc2 100644 --- a/include/linux/jhash.h +++ b/include/linux/jhash.h @@ -1,82 +1,97 @@ #ifndef _LINUX_JHASH_H #define _LINUX_JHASH_H +#include <linux/bitops.h> + /* jhash.h: Jenkins hash support. * - * Copyright (C) 1996 Bob Jenkins (bob_jenkins@burtleburtle.net) + * Copyright (C) 2006. Bob Jenkins ...

Great, thanks a lot Greg! -- Evgeniy Polyakov --

AFAIR, no. for_each_netdev is protected by rtnl_lock. --

Current net-next-2.6 (v2.6.29-rc2-1715-g367681f). I just reverified it with kvm-userspace (kvm-83-389-ga1efe3d). The problem goes away if I Seems a little beyond the first packet, this time my VM made it to starting sshd before causing this fault in the host (so it had at least DHCP'd an address): [ 208.823990] BUG: unable to handle kernel paging request at 0000000000007860 [ 208.826836] IP: [<ffffffff804481be>] skb_copy_datagram_from_iovec+0x1e/0x260 [ 208.827918] PGD 827032067 PUD ...

They can resync. Not sure where you're getting EAGAIN from, its usually used to deliver ENOBUFS to sockets when the kernel failed to even allocate the first skb thats cloned to the sockets. Check out rtmsg_ifa for an example. --

Indeed. I don't see either the point of stopping other sockets from receiving the message because one clone failed, but that's a different issue I think. BTW, the netlink_set_err() function (I found one call in rtnetlink.c) also attracted my attention since it sets the EAGAIN error to all listeners when nlmsg_multicast() fails, which happens if the echoing is set and unicast fails. This made me think, what would it be the action taken by the multicast userspace listener if it hits EAGAIN? ...

Patrick, I like it, I'm fine with this approach as soon as it let me add the "reliable" ctnetlink state-change reporting. I can add the following on top of the patch that David already applied: --- a/net/netlink/af_netlink.c +++ b/net/netlink/af_netlink.c [...] @@ -999,6 +1000,7 @@ static inline int do_one_broadcast(struct sock *sk, p->skb2 = NULL; } else if ((val = netlink_broadcast_deliver(sk, p->skb2)) < 0) { netlink_overrun(sk); + p->delivery_failure = 1; ...

Applied to the 'topic/slub/perf' branch. Thanks! Pekka --

's. You reverted the page allocator passthrough patch before this right? Otherwise kmalloc-8192 should not exist and allocation calls for 8192 bytes would be converted inline to request of an order 1 page from the page allocator.

Hi Christoph, On Thu, Feb 12, 2009 at 5:25 PM, Christoph Lameter Yup, I assume that's the case here. --

Great. Thanks. Applied to my dma-api/debug branch. Joerg --