The kernel's implementation of notifier chains is unsafe. There is no
protection against entries being added to or removed from a chain while the
chain is in use. The issues were discussed in this thread:
http://marc.theaimsgroup.com/?l=linux-kernel&m=113018709002036&w=2
We noticed that notifier chains in the kernel fall into two basic usage
classes:
"Blocking" chains are always called from a process context
and the callout routines are allowed to sleep;
"Atomic" chains can be called from an atomic context and
the callout routines are not allowed to sleep.
We decided to codify this distinction and make it part of the API. Therefore
this set of patches introduces three new, parallel APIs: one for blocking
notifiers, one for atomic notifiers, and one for "raw" notifiers (which is
really just the old API under a new name). New kinds of data structures are
used for the heads of the chains, and new routines are defined for
registration, unregistration, and calling a chain. The three APIs are
explained in include/linux/notifier.h and their implementation is in
kernel/sys.c.
With atomic and blocking chains, the implementation guarantees that the chain
links will not be corrupted and that chain callers will not get messed up by
entries being added or removed. For raw chains the implementation provides no
guarantees at all; users of this API must provide their own protections. (The
idea was that situations may come up where the assumptions of the atomic and
blocking APIs are not appropriate, so it should be possible for users to
handle these things in their own way.)
There are some limitations, which should not be too hard to live with. For
atomic/blocking chains, registration and unregistration must always be done in
a process context since the chain is protected by a mutex/rwsem. Also, a
callout routine for a non-raw chain must not try to register or unregister
entries on its own chain. (This did happen in a couple of places and the code
had to be changed to avoid it.)
Since atomic chains may be called from within an NMI handler, they cannot use
spinlocks for synchronization. Instead we use RCU. The overhead falls almost
entirely in the unregister routine, which is okay since unregistration is much
less frequent that calling a chain.
Here is the list of chains that we adjusted and their classifications. None
of them use the raw API, so for the moment it is only a placeholder.
ATOMIC CHAINS
-------------
arch/i386/kernel/traps.c: i386die_chain
arch/ia64/kernel/traps.c: ia64die_chain
arch/powerpc/kernel/traps.c: powerpc_die_chain
arch/sparc64/kernel/traps.c: sparc64die_chain
arch/x86_64/kernel/traps.c: die_chain
drivers/char/ipmi/ipmi_si_intf.c: xaction_notifier_list
kernel/panic.c: panic_notifier_list
kernel/profile.c: task_free_notifier
net/bluetooth/hci_core.c: hci_notifier
net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_chain
net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_expect_chain
net/ipv6/addrconf.c: inet6addr_chain
net/netfilter/nf_conntrack_core.c: nf_conntrack_chain
net/netfilter/nf_conntrack_core.c: nf_conntrack_expect_chain
net/netlink/af_netlink.c: netlink_chain
BLOCKING CHAINS
---------------
arch/powerpc/platforms/pseries/reconfig.c: pSeries_reconfig_chain
arch/s390/kernel/process.c: idle_chain
arch/x86_64/kernel/process.c idle_notifier
drivers/base/memory.c: memory_chain
drivers/cpufreq/cpufreq.c cpufreq_policy_notifier_list
drivers/cpufreq/cpufreq.c cpufreq_transition_notifier_list
drivers/macintosh/adb.c: adb_client_list
drivers/macintosh/via-pmu.c sleep_notifier_list
drivers/macintosh/via-pmu68k.c sleep_notifier_list
drivers/macintosh/windfarm_core.c wf_client_list
drivers/usb/core/notify.c usb_notifier_list
drivers/video/fbmem.c fb_notifier_list
kernel/cpu.c cpu_chain
kernel/module.c module_notify_list
kernel/profile.c munmap_notifier
kernel/profile.c task_exit_notifier
kernel/sys.c reboot_notifier_list
net/core/dev.c netdev_chain
net/decnet/dn_dev.c: dnaddr_chain
net/ipv4/devinet.c: inetaddr_chain
It's possible that some of these classifications are wrong. If they are,
please let us know or submit a patch to fix them. Note that any chain that
gets called very frequently should be atomic, because the rwsem read-locking
used for blocking chains is very likely to incur cache misses on SMP systems.
(However, if the chain's callout routines may sleep then the chain cannot be
atomic.)
The patch set was written by Alan Stern and Chandra Seetharaman, incorporating
material written by Keith Owens and suggestions from Paul McKenney and Andrew
Morton.
[jes@sgi.com: restructure the notifier chain initialization macros]
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Chandra Seetharaman <sekharan@us.ibm.com>
Signed-off-by: Jes Sorensen <jes@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Handle netif_carrier_{on,of} also if media is forced to 10baseT/100baseTx.
Signed-off-by: Steffen Klassert <klassert@mathematik.tu-chemnitz.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Set the polling interval for media changes to 5 seconds if link is down and
60 seconds if link is up.
Signed-off-by: Steffen Klassert <klassert@mathematik.tu-chemnitz.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Check for media changes and netif_carrier by using mii_check_media() if mii is
used.
Signed-off-by: Steffen Klassert <klassert@mathematik.tu-chemnitz.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Semaphore to mutex conversion.
The conversion was generated via scripts, and the result was validated
automatically via a script as well.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Cc: Dave Jones <davej@codemonkey.org.uk>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Jens Axboe <axboe@suse.de>
Cc: Neil Brown <neilb@cse.unsw.edu.au>
Acked-by: Alasdair G Kergon <agk@redhat.com>
Cc: Greg KH <greg@kroah.com>
Cc: Dominik Brodowski <linux@dominikbrodowski.net>
Cc: Adam Belay <ambx1@neo.rr.com>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: "David S. Miller" <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
* master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6:
[NETFILTER] x_table.c: sem2mutex
[IPV4]: Aggregate route entries with different TOS values
[TCP]: Mark tcp_*mem[] __read_mostly.
[TCP]: Set default max buffers from memory pool size
[SCTP]: Fix up sctp_rcv return value
[NET]: Take RTNL when unregistering notifier
[WIRELESS]: Fix config dependencies.
[NET]: Fill in a 32-bit hole in struct sock on 64-bit platforms.
[NET]: Ensure device name passed to SO_BINDTODEVICE is NULL terminated.
[MODULES]: Don't allow statically declared exports
[BRIDGE]: Unaligned accesses in the ethernet bridge
Remove the assumption that driver_register() returns the number of devices
bound to the driver. In fact, it returns zero for success or a negative
error value.
zorro_module_init() used the device count to automatically unregister and
unload drivers that found no devices. That might have worked at one time,
but has been broken for some time because zorro_register_driver() returned
either a negative error or a positive count (never zero). So it could only
unregister on failure, when it's not needed anyway.
This functionality could be resurrected in individual drivers by counting
devices in their .probe() methods.
Signed-off-by: Bjorn Helgaas <bjorn.helgaas@hp.com>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Roman Zippel <zippel@linux-m68k.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Remove the assumption that driver_register() returns the number of devices
bound to the driver. In fact, it returns zero for success or a negative
error value.
Signed-off-by: Bjorn Helgaas <bjorn.helgaas@hp.com>
Acked-by: Jeff Garzik <jgarzik@pobox.com>
Acked-by: Marc Zyngier <maz@misterjones.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Remove the assumption that driver_register() returns the number of devices
bound to the driver. In fact, it returns zero for success or a negative
error value.
dio_module_init() used the device count to automatically unregister and
unload drivers that found no devices. That might have worked at one time,
but has been broken for some time because dio_register_driver() returned
either a negative error or a positive count (never zero). So it could only
unregister on failure, when it's not needed anyway.
This functionality could be resurrected in individual drivers by counting
devices in their .probe() methods.
Signed-off-by: Bjorn Helgaas <bjorn.helgaas@hp.com>
Cc: Philip Blundell <philb@gnu.org>
Cc: Jochen Friedrich <jochen@scram.de>
Cc: "Antonino A. Daplas" <adaplas@pol.net>
Cc: Jeff Garzik <jeff@garzik.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
MODULE_PARM was actually breaking: recent gcc version optimize them out as
unused. It's time to replace the last users, which are generally in the
most unloved drivers anyway.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
I accidentally ended up with a config that set NET_RADIO off,
and NET_WIRELESS_RTNETLINK on, which blew up thus..
net/built-in.o: In function `do_setlink':net/core/rtnetlink.c:479: undefined reference to `wireless_rtnetlink_set'
net/built-in.o: In function `do_getlink':net/core/rtnetlink.c:521: undefined reference to `wireless_rtnetlink_get'
Signed-off-by: Dave Jones <davej@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
net: ne2k.c won't compile if pci_clone_list is const
f71e130966 which (amongst other things)
made pci_clone_list in ne2k-pci.c const causes the following compile error.
This patch reverses that portion of that changeset
drivers/net/ne2k-pci.c:123: error: pci_clone_list causes a section type
conflict
~/ gcc --version
gcc (GCC) 4.0.3 (Debian 4.0.3-1)
~/ dpkg gcc-4.0 | grep Version
Version: 4.0.3-1
Signed-Off-By: Horms <horms@verge.net.au
ne2k-pci.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
cee0890cc97247b6a9decd94f5dc0719ac8f0b1b
Signed-off-by: Jeff Garzik <jeff@garzik.org>
This patch adds support for the Ethernet controller integrated in the
Atmel AT91RM9200 SoC processor.
Changes since the previous submission (01/02/2006) are:
- Make use of the clk.h clock infrastructure.
- The multicast hash function is not crc32. [Patch by Pedro Perez]
Signed-off-by: Andrew Victor <andrew@sanpeople.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
switch to ioremap()
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
switch to ioremap()
Adrian Bunk:
The order of the hunks in the patch was slightly rearranged due to an
unrelated change in the driver.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Don't need to keep track of available buffers, it is simpler
to just compute the value (ala e1000). Don't need tes on link up
because should always have available buffers then.
Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
Don't free transmit buffers until the whole set of transmit descriptors
has been marked as done. Otherwise, we risk freeing a skb before the
whole transmit is done.
This changes the transmit completion handling from incremental to a
two pass algorithm. First pass scans and records the start of the last
done descriptor, second cleans up until that point.
Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
In the error case we call skge_rx_reuse twice. This is harmless
but unnecessary.
Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
The skge driver was using dev_alloc_skb which reserves space for the
Ethernet header. This unnecessary and it should just use alloc_skb,
also by using GFP_KERNEL during startup it won't run into problems when
a user asks for a huge ring size or mtu and potentially drains the
reserved atomic pool.
Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
The skge driver aligns the header on the initial receive buffers, but
but doesn't on followon receive buffer allocations.
Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
Transmit buffers are always freed with interrupts enabled (softirq),
so we can just call dev_kfree_skb.
Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
Unicast packets are shown as multicast, real multicast packets are missing.
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
The Coverity checker (CID: 659, 660) spotted this resource leak on
PCI probe error path. Free private data structure if pci_enable_device()
fails.
Signed-off-by: Jouni Malinen <jkmaline@cc.hut.fi>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
The Coverity checker (CID: 58) spotted this duplicated idx != 0
validation for unicast keys in prism2_ioctl_siwencodeext().
Signed-off-by: Jouni Malinen <jkmaline@cc.hut.fi>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
The Coverity checker (CID: 452, 453, 454, 455, 456) spotted this
unlikely read overrun of CIS buffer. Abort if CISTPL_CONFIG or
CISTPL_MANFID would not fit in buffer.
Signed-off-by: Jouni Malinen <jkmaline@cc.hut.fi>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
The Coverity checker (CID: 930) spotted this double free on error path
(allocation failure). Do not free these here since generic error path
will take care of this.
Signed-off-by: Eugene Teo <eugene.teo@eugeneteo.net>
Signed-off-by: Jouni Malinen <jkmaline@cc.hut.fi>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
The Coverity checker (CID: 273) spotted this inconsequent NULL checking
(unconditionally dereferencing directly after checking for NULL
isn't a good idea). Return immediately to avoid this.
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Jouni Malinen <jkmaline@cc.hut.fi>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
The Coverity checker (CID: 59) noted that the call to prism2_hw_reset()
was dead code. Move prism2_hw_reset() call to a place where it is
actually executed.
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Jouni Malinen <jkmaline@cc.hut.fi>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Observed problems when multiple processes request scans and subsequently
scan results. This causes a scan result request to hit card registers
before the scan is complete, returning an incomplete scan list and
possibly making the card very angry. Instead, cache the results of a
wireless scan and serve result requests from the cache, rather than
hitting the hardware for them.
Signed-off-by: Dan Williams <dcbw@redhat.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
The number 2312 was used all over the place to refer to the card's
default MTU. Make it a #define and use that everywhere rather than the
number.
Signed-off-by: Dan Williams <dcbw@redhat.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Show the specific device that driver messages are about.
Signed-off-by: Dan Williams <dcbw@redhat.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
When we stop allocating percpu memory for not-possible CPUs we must not touch
the percpu data for not-possible CPUs at all. The correct way of doing this
is to test cpu_possible() or to use for_each_cpu().
This patch is a kernel-wide sweep of all instances of NR_CPUS. I found very
few instances of this bug, if any. But the patch converts lots of open-coded
test to use the preferred helper macros.
Cc: Mikael Starvik <starvik@axis.com>
Cc: David Howells <dhowells@redhat.com>
Acked-by: Kyle McMartin <kyle@parisc-linux.org>
Cc: Anton Blanchard <anton@samba.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Paul Mundt <lethal@linux-sh.org>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: William Lee Irwin III <wli@holomorphy.com>
Cc: Andi Kleen <ak@muc.de>
Cc: Christian Zankel <chris@zankel.net>
Cc: Philippe Elie <phil.el@wanadoo.fr>
Cc: Nathan Scott <nathans@sgi.com>
Cc: Jens Axboe <axboe@suse.de>
Cc: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Semaphore to mutex conversion.
The conversion was generated via scripts, and the result was validated
automatically via a script as well.
Signed-off-by: Arjan van de Ven <arjan@infradead.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Cc: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This is version 20 of the Wireless Extensions. This is the
completion of the RtNetlink work I started early 2004, it enables the
full Wireless Extension API over RtNetlink.
Few comments on the patch :
o totally driver transparent, no change in drivers needed.
o iwevent were already RtNetlink based since they were created
(around 2.5.7). This adds all the regular SET and GET requests over
RtNetlink, using the exact same mechanism and data format as iwevents.
o This is a Kconfig option, as currently most people have no
need for it. Surprisingly, patch is actually small and well
encapsulated.
o Tested on SMP, attention as been paid to make it 64 bits clean.
o Code do probably too many checks and could be further
optimised, but better safe than sorry.
o RtNetlink based version of the Wireless Tools available on
my web page for people inclined to try out this stuff.
I would also like to thank Alexey Kuznetsov for his helpful
suggestions to make this patch better.
Signed-off-by: Jean Tourrilhes <jt@hpl.hp.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Add support for new chip 5755 which is very similar to 5787.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
