- 30 May, 2005 11 commits
-
-
Herbert Xu authored
Steven Hand <Steven.Hand@cl.cam.ac.uk> wrote: > > Reconstructed forward trace: > > net/ipv4/udp.c:1334 spin_lock_irq() > net/ipv4/udp.c:1336 udp_checksum_complete() > net/core/skbuff.c:1069 skb_shinfo(skb)->nr_frags > 1 > net/core/skbuff.c:1086 kunmap_skb_frag() > net/core/skbuff.h:1087 local_bh_enable() > kernel/softirq.c:0140 WARN_ON(irqs_disabled()); The receive queue lock is never taken in IRQs (and should never be) so we can simply substitute bh for irq. Signed-off-by:
Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Harald Welte authored
When we have ip_queue being used from LOCAL_IN, then we end up with a situation where the verdicts coming back from userspace traverse the TCP input path from syscall context. While this seems to work most of the time, there's an ugly deadlock: syscall context is interrupted by the timer interrupt. When the timer interrupt leaves, the timer softirq get's scheduled and calls tcp_delack_timer() and alike. They themselves do bh_lock_sock(sk), which is already held from somewhere else -> boom. I've now tested the suggested solution by Patrick McHardy and Herbert Xu to simply use local_bh_{en,dis}able(). Signed-off-by:Harald Welte <laforge@netfilter.org> Signed-off-by:
-
-
David S. Miller authored
-
David S. Miller authored
Signed-off-by: -
David S. Miller authored
We are intentionally ignoring the copy_to_user() value, make it clear to the compiler too. Noted by Jeff Garzik. Signed-off-by: -
Jon Mason authored
Signed-off-by:
Jon Mason <jdmason@us.ibm.com> Signed-off-by:
-
Pravin B. Shelar authored
It cannot work properly, so just ignore it in drr and rr multipath algorithms just like the random multipath algorithm does. Suggested by Herbert Xu. Signed-off by: Pravin B. Shelar <pravins@calsoftinc.com> Signed-off-by: -
Michael Ellerman authored
This patch adds is_multicast_ether_addr() to go along with is_valid_ether_addr() and friends. It then changes is_valid_ether_addr() to use the new macro, and fixes up the comment on that function to move implementation details out of the API doco. Signed-off-by:
Michael Ellerman <michael@ellerman.id.au> Signed-off-by:
-
Harald Welte authored
Add an option to make secondary IP addresses get promoted when primary IP addresses are removed from the device. It defaults to off to preserve existing behavior. Signed-off-by:
Harald Welte <laforge@gnumonks.org> Signed-off-by:
-
-
- 29 May, 2005 18 commits
-
-
Michael Chan authored
Add tg3_nvram_lock() and tg3_nvram_unlock() calls around tg3_halt_cpu(). It is possible that the bootcode may be loading code from nvram during this call and stopping the cpu without getting the lock may cause uncompleted nvram data to be left in the nvram data register. Subsequent calls to read/write nvram data will fail. Signed-off-by:
Michael Chan <mchan@broadcom.com> Signed-off-by:
-
Michael Chan authored
This test uses the previously added tg3_test_interrupt() to perform the test. Signed-off-by:
-
Michael Chan authored
The test will loopback one packet in MAC loopback mode and verify the packet data. Signed-off-by:
-
Michael Chan authored
Signed-off-by:
-
Michael Chan authored
Signed-off-by:
-
Michael Chan authored
Add a reset kind parameter to tg3_halt() so that the RESET_KIND_SUSPEND parameter can be passed to tg3_halt() before doing offline tests. All other calls to tg3_halt() will use the RESET_KIND_SHUTDOWN parameter. Signed-off-by:
-
Michael Chan authored
Signed-off-by:
-
Michael Chan authored
Signed-off-by:
-
Michael Chan authored
Signed-off-by:
-
Stephen Hemminger authored
This improves the bridge local receive path by avoiding going through another softirq. The bridge receive path is already being called from a netif_receive_skb() there is no point in going through another receiveq round trip. Recursion is limited because bridge can never be a port of a bridge so handle_bridge() always returns. Signed-off-by: -
Stephen Hemminger authored
Avoid poisoning of the bridge forwarding table by frames that have been dropped by filtering. This prevents spoofed source addresses on hostile side of bridge from causing packet leakage, a small but possible security risk. Signed-off-by:
Stephen Hemminger <shemminger@osdl.org> Signed-off-by:
-
Stephen Hemminger authored
Make features of the bridge pseudo-device be a subset of the underlying devices. Motivated by Xen and others who use bridging to do failover. Signed-off-by: Catalin BOIE <catab at umrella.ro> Signed-off-by:
-
Stephen Hemminger authored
The features field in netdevice is really a bitmask, and bitmask's should be unsigned. Signed-off-by:
-
Stephen Hemminger authored
Resend of earlier patch (no changes) from Catalin used to provide device feature change notification. Signed-off-by: Catalin BOIE <catab at umbrella.ro> Acked-by:
-
Jeff Garzik authored
-
-
-
Vojtech Pavlik authored
Signed-off-by:
Vojtech Pavlik <vojtech@suse.cz> Signed-off-by:
Dmitry Torokhov <dtor@mail.ru>
-
- 28 May, 2005 11 commits
-
-
Kumar Gala authored
Trying to initialize the i8259 PIC will not work if CONFIG_PCI is not enabled. The kernel hangs if the initialization is tried. Signed-off-by:
Kumar Gala <kumar.gala@freescale.com> Signed-off-by:
Andrew Morton <akpm@osdl.org> Signed-off-by:
Linus Torvalds <torvalds@osdl.org>
-
Kumar Gala authored
This change allows mpc83xx_restart to issue a software reset. Signed-off-by:
Randy Vinson <rvinson@mvista.com> Signed-off-by:
-
Kumar Gala authored
The call to io_block_mapping was creating an invalid BAT entry because the value of BCSR_SIZE (32K) is too small to be used in a BAT (128K min). This change removes the io_block_mapping call since these registers can easily be mapped using ioremap at the point of use. Signed-off-by:
-
Kumar Gala authored
The error checking for emulation of load string instructions was overly generous and would cause certain valid forms of the instructions to be treated as illegal. We drop the range checking since the architecture allows this to be boundedly undefined. Tests on CPUs that support these instructions appear not do cause illegal instruction traps on range errors and just allow the execution to occur. Thanks to Kim Phillips for debugging this and figuring out what real HW was doing. Signed-off-by:
-
Kumar Gala authored
The patch that introduced support for the VIA chipset broke building if CONFIG_PCI is disabled. Signed-off-by:
-
Vitaly Bordug authored
This patch adds on-chip PCI bridge support for the PQ2 family. The incomplete existent code is updated with interrupt handling stuff and board-specific bits for 8272ADS and PQ2FADS; the related files were renamed (from m8260_pci to m82xx_pci) to be of more generic fashion. This is tested with 8266ADS and 8272ADS, should work on PQ2FADS as well. Signed-off-by:
Vitaly Bordug <vbordug@ru.mvista.com> Signed-off-by:
-
Kumar Gala authored
Add support for the VIA IDE controller that exists on the MPC8555 CDS system. Updated the config for the system to enable support by default. Signed-off-by:
Scott Hall <shall@mvista.com> Signed-off-by:
-
Kumar Gala authored
Some debug registers needed to be initialized early on to allow proper support for KGDB. Additionally, we need to setup the ppc.md_early_serial_map function pointer on boards that have serial support for KGDB. Signed-off-by:
-
Kumar Gala authored
The existing make rule when building a uImage would check to see if the image file existed to report 'is ready' or 'not made'. However make appeared to compute the file list before the rule was executed. Signed-off-by:
Chris Clark <cpclark@xmission.com> Signed-off-by:
Sam Ravnborg <sam@ravnborg.org> Signed-off-by:
-
Paolo 'Blaisorblade' Giarrusso authored
This var is currently useless, as it's apparent from reading the code. Until 2.6.11 it was used in some code related to jail mode, in the same proc.: if(jail){ while(!reading) sched_yield(); } jail mode has been dropped, together with that use, so let's finish dropping this. Also, remove some other useless definitions I met. Acked-by:Jeff Dike <jdike@addtoit.com> Signed-off-by:
Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it> Signed-off-by:
-
Paolo 'Blaisorblade' Giarrusso authored
Fix it a bit (after some cross checking with "man arch_prctl"). There were: *) typos FS/GS and back *) FS in place of FS_BASE (and the same for GS) *) the procedure used put_user on &addr, where addr was the parameter (i.e. changed its param with put_user, completely useless) rather than interpreting addr as a pointer, as requested in this case (see the man page). Signed-off-by:
-
