1) Restore promiscuous setting after reset. 2) Don't use firmware when forcing firmware dump. 3) Reduce print level in data path statements. 4) Fix possible endian issue for rx UDP csum. 5) Make adapter drop frame errors and pass up csum errors. 6) Change version to v1.00.00.25.00.00-01. --

Signed-off-by: Ron Mercer <ron.mercer@qlogic.com> --- drivers/net/qlge/qlge_main.c | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/net/qlge/qlge_main.c b/drivers/net/qlge/qlge_main.c index e39451b..a41b6b5 100644 --- a/drivers/net/qlge/qlge_main.c +++ b/drivers/net/qlge/qlge_main.c @@ -1677,7 +1677,7 @@ static void ql_process_mac_rx_skb(struct ql_adapter *qdev, /* Unfragmented ipv4 UDP frame. */ struct iphdr *iph = (struct iphdr *) skb->data; ...

Signed-off-by: Ron Mercer <ron.mercer@qlogic.com> --- drivers/net/qlge/qlge.h | 1 + drivers/net/qlge/qlge_main.c | 7 ++++++- 2 files changed, 7 insertions(+), 1 deletions(-) diff --git a/drivers/net/qlge/qlge.h b/drivers/net/qlge/qlge.h index bfb8b32..01b0634 100644 --- a/drivers/net/qlge/qlge.h +++ b/drivers/net/qlge/qlge.h @@ -2246,6 +2246,7 @@ netdev_tx_t ql_lb_send(struct sk_buff *skb, struct net_device *ndev); void ql_check_lb_frame(struct ql_adapter *, struct sk_buff ...

Signed-off-by: Ron Mercer <ron.mercer@qlogic.com> --- drivers/net/qlge/qlge.h | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/drivers/net/qlge/qlge.h b/drivers/net/qlge/qlge.h index 27f83d6..06b2188 100644 --- a/drivers/net/qlge/qlge.h +++ b/drivers/net/qlge/qlge.h @@ -16,7 +16,7 @@ */ #define DRV_NAME "qlge" #define DRV_STRING "QLogic 10 Gigabit PCI-E Ethernet Driver " -#define DRV_VERSION "v1.00.00.23.00.00-01" +#define ...

So is there some place else that should have the clear instead of ipv4_link_failure()? Sol -----Original Message----- From: Herbert Xu [mailto:herbert@gondor.apana.org.au] Sent: Monday, July 05, 2010 5:04 AM To: Sol Kavy Cc: linux-kernel@vger.kernel.org; Greg Ren; Guojun Jin; Murat Sezgin; Sener Ilgen; David S. Miller; Stephen Hemminger; netdev@vger.kernel.org Subject: Re: Possible bug in net/ipv4/route.c? Generally this area should be cleared on entry to each stack intending on using ...

Since net_device has an instance of net_device_stats, we can remove the instance of this from the adapter structure. Based on original patch by Kulikov Vasiliy. Signed-off-by: Ralf Baechle <ralf@linux-mips.org> Signed-off-by: Kulikov Vasiliy <segooon@gmail.com> drivers/net/ioc3-eth.c | 49 ++++++++++++++++++++++++----------------------- 1 files changed, 25 insertions(+), 24 deletions(-) diff --git a/drivers/net/ioc3-eth.c b/drivers/net/ioc3-eth.c index e3b5e94..0b3f6df 100644 --- ...

Adds ability to bind to non-local IPv6 address the same way as for IPv4 Signed-off-by: Michal Humpula <michal.humpula@web4u.cz> diff --git a/Documentation/networking/ip-sysctl.txt b/Documentation/networking/ip-sysctl.txt index f350c69..1ebed5e 100644 --- a/Documentation/networking/ip-sysctl.txt +++ b/Documentation/networking/ip-sysctl.txt @@ -963,6 +963,11 @@ bindv6only - BOOLEAN Default: FALSE (as specified in RFC2553bis) +ipv6_nonlocal_bind - BOOLEAN + If set, allows processes to ...

Thanks for review. I should obviously wait a little bit longer before posting the v2. --

Hi, Well, Samba still disables splice support (hard coded), I applied your patch (adding the SPLICE_F_NONBLOCK to the splice(sock, pipe)) and I managed to write 4G file to Samba share. I did notice that the splice is done on buffers in two sizes: 1380 and 2760 (when writing to share file), I guess that if I can get samba to use bigger buffers it will reduce the splice calls and achieve better performance. I also saw that when re-writing a file splice does use the maximum buffer size (~16K) ...

Hi I have been trying to test splice on kernel 2.6.35_4 (x86) from Samba (v3.4.7) but could not copy more than ~60MB to the Samba server share. Strace shows that the splice got stuck in blocking mode on the splice call from socket to pipe. Has anyone managed to get splice from socket to fd work for large files (up to 4G file size) ? -Ofer --

I am using Samba, so from my understanding of the source code, it loops and performs splice(sock, pipe) and splice(pipe, fd). There is no flush of any sort in between. When you say drain you mean to flush all data to pipe? -Ofer -----Original Message----- From: Changli Gao [mailto:xiaosuo@gmail.com] Sent: Monday, July 05, 2010 12:59 PM To: Ofer Heifetz Cc: netdev@vger.kernel.org Subject: Re: Splice status Did you drain the pipe before calling splice(2) to move data from socket to ...

Note that if your load increases or network is faster, splice will naturally use more data per call. Dont worry. Also, you can change MIN(count, 16384) to MIN(count, 65536) now the real samba bug is known and can be fixed (by the SPLICE_F_NONBLOCK patch I sent) (I guess using 16384 instead of 65536 was a try to reduce hang --

Right, but a read_lock() would not be enough, even if all users would use it. --

Yes, I provided one patch but am working on a different one, requiring a new rtnl_link_stats64 param to ndo_get_stats64() methods and dev_get_stats() as well. dev->stats64 should not be overwritten without some synchronization, so just disallow it for the moment... --

I really didn't want to add this overhead and complication to readers when only some drivers need it. 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. --

I think you should use a similar approach here as you did in the loopback driver, i.e. update private variables in the RX and TX path and then copy/aggregate them in the implementation ndo_get_stats64 (only without the need for percpu stats). If you want to include a seqlock in the driver stats interface, you can do that but it's not going to be pretty and we're still going to need additional seqlocks for per-queue (or percpy) stats in some drivers. Ben. -- Ben Hutchings, Senior ...

Changes from v1: Use atm_dev notifier chain instead of callback function pointer in struct vcc. In drivers/usb/atm call atm_dev_signal_change(). In userspace it's helpfull to know if a network device has a carrier signal. Often it is monitored via netlink. This patchset allows a way for the struct atm_dev drivers to pass carrier on/off to the netdevice. For DSL, carrier is on when the line has reached showtime state. Currently this patchset only propagates the changes to br2684 ...

When a signal change event occurs call netif_carrier_on/off. Signed-off-by: Karl Hiramoto <karl@hiramoto.org> --- net/atm/br2684.c | 66 ++++++++++++++++++++++++++++++++++++++++++++++++++++- 1 files changed, 64 insertions(+), 2 deletions(-) diff --git a/net/atm/br2684.c b/net/atm/br2684.c index 6719af6..5afa512 100644 --- a/net/atm/br2684.c +++ b/net/atm/br2684.c @@ -139,6 +139,42 @@ static struct net_device *br2684_find_dev(const struct br2684_if_spec *s) return NULL; } +static ...

Propagate changes to upper atm layer. Signed-off-by: Karl Hiramoto <karl@hiramoto.org> --- drivers/atm/idt77105.c | 11 ++++++----- 1 files changed, 6 insertions(+), 5 deletions(-) diff --git a/drivers/atm/idt77105.c b/drivers/atm/idt77105.c index dab5cf5..bca9cb8 100644 --- a/drivers/atm/idt77105.c +++ b/drivers/atm/idt77105.c @@ -126,7 +126,7 @@ static void idt77105_restart_timer_func(unsigned long dummy) istat = GET(ISTAT); /* side effect: clears all interrupt status ...

Propagate signal changes to upper atm layer. Signed-off-by: Karl Hiramoto <karl@hiramoto.org> --- drivers/usb/atm/cxacru.c | 18 +++++++++--------- 1 files changed, 9 insertions(+), 9 deletions(-) diff --git a/drivers/usb/atm/cxacru.c b/drivers/usb/atm/cxacru.c index c89990f..101ffc9 100644 --- a/drivers/usb/atm/cxacru.c +++ b/drivers/usb/atm/cxacru.c @@ -866,50 +866,50 @@ static void cxacru_poll_status(struct work_struct *work) instance->line_status = buf[CXINF_LINE_STATUS]; ...

Propagate signal changes to upper atm layer. Signed-off-by: Karl Hiramoto <karl@hiramoto.org> --- drivers/usb/atm/ueagle-atm.c | 13 ++++++++++--- 1 files changed, 10 insertions(+), 3 deletions(-) diff --git a/drivers/usb/atm/ueagle-atm.c b/drivers/usb/atm/ueagle-atm.c index e213d3f..ebae944 100644 --- a/drivers/usb/atm/ueagle-atm.c +++ b/drivers/usb/atm/ueagle-atm.c @@ -575,6 +575,13 @@ MODULE_PARM_DESC(annex, sc->usbatm->atm_dev->type = val; \ } while (0) +#define ...

This patch adds hooks for transmit and receive time stamps. The transmit hook allows a software fallback for transmit time stamps, for MACs lacking time stamping hardware. The receive hook does not yet have any effect, but it prepares the way for hardware time stamping in PHY devices. Using the hooks will still require adding two inline function calls to each MAC driver. Signed-off-by: Richard Cochran <richard.cochran@omicron.at> --- include/linux/skbuff.h | 36 ...

The phy_mii_ioctl() function unnecessarily throws away the original ifreq. We need access to the ifreq in order to support PHYs that can perform hardware time stamping. Two maverick drivers filter the ioctl commands passed to phy_mii_ioctl(). This is unnecessary since phylib will check the command in any case. Signed-off-by: Richard Cochran <richard.cochran@omicron.at> --- drivers/net/arm/ixp4xx_eth.c | 3 ++- drivers/net/au1000_eth.c | 2 +- ...

Perhaps the inner two if statements could be combined to remove unnecessary nesting? And perhaps check for sysctl_ipv6_nonlocal_bind first, as presumably its a cheaper, though less likely to succeed check. if (!(addr_type & IPV6_ADDR_MULTICAST)) { if (!sysctl_ipv6_nonlocal_bind && !ipv6_chk_addr(net, &addr->sin6_addr, dev, 0)) { err = -EADDRNOTAVAIL; goto out_unlock; } } Actually, I wonder if all three if statements could be combined. if ...

Hello. This is not sufficient. In IPv4, even if you do non-local bind, you cannot connect/send packets from that address until the admin really assigns that address on the node. Local address is checked when you try to connect (or to send), and this is important thing to do. But in IPv6, it is not checked, and it is very bad to open this "hole". --yoshfuji --

Say, on a slightly different subject... but still having to do with setsockopt(IP_TOS)... is it intentional that system call effectively does nothing if the socket is in listen(), connect(), or bind() states? In other words, you have to issue the setsockopt() immediately after the socket() call, or it doesn't do anything. As I remember, that's slightly different semantics from BSD, which allows you to change the markings on a bound or listening socket. I've not walked through the kernel ...

At the risk of typing into Michael's keyboard, chances are, what the NIC does is follow Microsoft's specs for RSS, which as far as I can tell, discusses hashing either just on the IP addresses (v4 or v6) or the IP addresses and TCP port numbers, so unless Broadcom has an enhancement/extension, if the traffic arriving is not TCP/IP and not multiple connections, it probably does not get spread-out. rick jones --

Applied, thanks. --

Hi, sorry for the late reply, as I've been on vacation in the last week (and shut off mails intentionally :) At Mon, 28 Jun 2010 12:44:48 -0500, I like the patch very well, too. Is this correct...? Shouldn't it be a negative check? thanks, Takashi --

Yes, it should be a negative check ... I'll update the patch. I guess this still means that no-one has managed to test it on a functional system ... James --

Can you give us till Tuesday to respond? I know that there are some additional e1000e patches in my queue, which may resolve the issue, but this weekend the power is down to do some infrastructure upgrades which prevents us from doing any investigation.debugging until Tuesday. -- Cheers, Jeff --

Yes and no -- yes in the sense that if you want to use more functionality of the switch chip, you'll have to add some code that extracts that info from the Linux network interface config and turns it into commands for the switch chip, and no in the sense that I'm not sure yet what the best way to implement this would be. (Doing it all in userspace is one option.) --

Here is an updated patch, without added lock but temp storage. Thanks [PATCH net-next-2.6 V2] net: fix 64 bit counters on 32 bit arches There is a small possibility that a reader gets incorrect values on 32 bit arches. SNMP applications could catch incorrect counters when a 32bit high part is changed by another stats consumer/provider. One way to solve this is to add a rtnl_link_stats64 param to all ndo_get_stats64() methods, and also add such a parameter to dev_get_stats(). Rule ...