kernel_optimize_test/drivers/char
Alan Stern e041c68341 [PATCH] Notifier chain update: API changes
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>
2006-03-27 08:44:50 -08:00
..
agp [PATCH] x86_64: Fix wrong PCI ID for ALI M1695 AGP bridge 2006-03-25 09:10:53 -08:00
drm [SPARC]: Respect vm_page_prot in io_remap_page_range(). 2006-03-22 01:15:13 -08:00
ftape drivers/char/ftape/lowlevel/fdc-io.c: Correct a comment 2006-03-26 19:18:07 +02:00
ip2 Merge with git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git 2006-02-03 23:49:49 +01:00
ipmi [PATCH] Notifier chain update: API changes 2006-03-27 08:44:50 -08:00
mwave [PATCH] mwave: missing __user in ioctl struct declaration 2005-12-15 10:04:31 -08:00
pcmcia [PATCH] s/;;/;/g 2006-03-24 07:33:24 -08:00
rio [PATCH] Yet more rio cleaning (2 of 2) 2006-03-24 07:33:29 -08:00
tpm [PATCH] tpm: sparc32 build fix 2006-03-25 08:22:55 -08:00
watchdog [PATCH] cirrus ep93xx watchdog driver 2006-03-25 08:22:49 -08:00
.gitignore gitignore: misc files 2006-01-01 22:21:50 +01:00
amiserial.c [PATCH] kill _INLINE_ 2006-03-23 07:38:16 -08:00
applicom.c [PATCH] applicom: fix error handling 2005-09-10 10:06:25 -07:00
applicom.h
cd1865.h
ChangeLog
consolemap.c [PATCH] kfree cleanup: drivers/char 2005-11-07 07:54:02 -08:00
cp437.uni
cs5535_gpio.c [PATCH] i386: GPIO driver for AMD CS5535/CS5536 2006-01-10 08:01:24 -08:00
cyclades.c [PATCH] new tty buffering locking fix 2006-02-03 08:32:09 -08:00
decserial.c
defkeymap.c_shipped
defkeymap.map
digi1.h [PATCH] Clean up the old digi support and rescue it 2005-09-07 16:57:20 -07:00
digi.h
digiFep1.h [PATCH] Clean up the old digi support and rescue it 2005-09-07 16:57:20 -07:00
digiPCI.h
ds1286.c
ds1302.c
ds1620.c
dsp56k.c [PATCH] m68k: dsp56k __user annotations 2006-01-12 09:09:04 -08:00
dtlk.c
ec3104_keyb.c
efirtc.c
epca.c BUG_ON() Conversion in drivers/char 2006-03-26 18:17:21 +02:00
epca.h [PATCH] epca iomem annotations + several missing readw() 2005-09-16 10:38:10 -07:00
epcaconfig.h
esp.c [PATCH] Fix locking error in esp 2006-02-14 10:01:39 -08:00
generic_nvram.c
generic_serial.c [PATCH] sem2mutex: serial ->port_write_mutex 2006-03-23 07:38:14 -08:00
genrtc.c
hangcheck-timer.c [PATCH] s390: hangcheck timer support 2006-02-01 08:53:24 -08:00
hpet.c [PATCH] HPET: handle multiple ACPI EXTENDED_IRQ resources 2006-02-14 16:09:34 -08:00
hvc_console.c [PATCH] Fix race condition in hvc console. 2006-02-24 11:36:40 +11:00
hvc_vio.c powerpc: don't duplicate name between vio_driver and device_driver 2005-10-24 15:42:12 +10:00
hvcs.c [PATCH] powerpc: trivial: modify comments to refer to new location of files 2006-02-10 16:53:51 +11:00
hvsi.c [PATCH] drivers/char: Use ARRAY_SIZE macro 2006-01-10 08:01:56 -08:00
hw_random.c [PATCH] Add missing ifdef for VIA RNG code 2006-03-03 21:05:58 -05:00
i8k.c [PATCH] I8K: fix /proc reporting of blank service tags 2005-11-12 11:42:32 -08:00
ip27-rtc.c
isicom.c [PATCH] char/isicom: More whitespaces and coding style 2006-01-10 08:02:01 -08:00
istallion.c [PATCH] Remove MODULE_PARM 2006-03-25 08:22:52 -08:00
ite_gpio.c
Kconfig [PATCH] isicom: select FW_LOADER 2006-03-26 08:56:58 -08:00
keyboard.c [SPARC]: introduce a SPARC Kconfig symbol 2005-12-22 23:09:54 -08:00
lcd.c Resurrect Cobalt support for 2.6. 2005-10-29 19:30:42 +01:00
lcd.h Resurrect Cobalt support for 2.6. 2005-10-29 19:30:42 +01:00
lp.c [PATCH] Driver Core: fix up all callers of class_device_create() 2005-10-28 09:52:52 -07:00
Makefile tabify drivers/char/Makefile 2006-03-24 18:10:34 +01:00
mbcs.c [IA64] mbcs_init() should give up unless running on sn2 2005-10-04 09:39:18 -07:00
mbcs.h
mem.c [PATCH] EFI, /dev/mem: simplify efi_mem_attribute_range() 2006-03-26 08:56:54 -08:00
misc.c [PATCH] Driver Core: fix up all callers of class_device_create() 2005-10-28 09:52:52 -07:00
mmtimer.c [IA64] SGI SN drivers: don't report !sn2 hardware as an error 2006-03-07 15:27:59 -08:00
moxa.c [PATCH] moxa serial: add proper capability check 2006-01-10 09:45:36 -08:00
mxser.c [PATCH] Remove MODULE_PARM 2006-03-25 08:22:52 -08:00
mxser.h
n_hdlc.c [PATCH] n_hdlc.c: remove unused declaration 2006-01-10 08:02:00 -08:00
n_r3964.c [PATCH] TTY layer buffering revamp 2006-01-10 08:01:59 -08:00
n_tty.c [PATCH] sem2mutex: tty 2006-03-23 07:38:11 -08:00
nvram.c [PATCH] drivers/char: Use ARRAY_SIZE macro 2006-01-10 08:01:56 -08:00
nwbutton.c
nwbutton.h
nwflash.c [PATCH] sem2mutex: drivers/char/ 2006-03-23 07:38:11 -08:00
ppdev.c [PATCH] parport: move PP_MAJOR from ppdev.h to major.h 2006-03-25 08:22:53 -08:00
pty.c [PATCH] TTY layer buffering revamp 2006-01-10 08:01:59 -08:00
qtronix.c qtronix.c: Handle kmalloc failure. 2005-10-29 19:30:55 +01:00
qtronixmap.c_shipped
qtronixmap.map
random.c [PATCH] random: get rid of sparse warning 2006-01-11 18:42:11 -08:00
raw.c [PATCH] sem2mutex: drivers: raw, connector, dcdbas, ppp_generic 2006-03-23 07:38:10 -08:00
riscom8_reg.h
riscom8.c [PATCH] Remove MODULE_PARM 2006-03-25 08:22:52 -08:00
riscom8.h
rocket_int.h [PATCH] drivers/char/rocket.c: cleanups 2005-06-25 16:25:04 -07:00
rocket.c [PATCH] Fix RocketPort driver 2006-02-03 08:31:59 -08:00
rocket.h
rtc.c [PATCH] Allow reading CMOS day of week register 2006-01-11 18:42:10 -08:00
s3c2410-rtc.c [PATCH] handle errors returned by platform_get_irq*() 2006-03-20 13:42:57 -08:00
scan_keyb.c
scan_keyb.h
scc.h [PATCH] m68k: static vs. extern in scc.h 2006-01-12 09:09:00 -08:00
scx200_gpio.c
selection.c [PATCH] TTY layer buffering revamp 2006-01-10 08:01:59 -08:00
ser_a2232.c [PATCH] sem2mutex: serial ->port_write_mutex 2006-03-23 07:38:14 -08:00
ser_a2232.h
ser_a2232fw.ax
ser_a2232fw.h
serial167.c [PATCH] remove unused tmp_buf_sem's 2006-01-14 10:41:42 -08:00
snsc_event.c [PATCH] snsc kmalloc2kzalloc 2006-03-23 07:38:15 -08:00
snsc.c [PATCH] snsc kmalloc2kzalloc 2006-03-23 07:38:15 -08:00
snsc.h [IA64-SGI] Handle SC env. powerdown events 2006-01-26 13:32:26 -08:00
sonypi.c [PATCH] sonypi: Enable ACPI events for Sony laptop hotkeys 2006-01-08 20:14:03 -08:00
specialix_io8.h
specialix.c [PATCH] remove unused tmp_buf_sem's 2006-01-14 10:41:42 -08:00
stallion.c [PATCH] sem2mutex: drivers/char/ 2006-03-23 07:38:11 -08:00
sx.c [PATCH] sem2mutex: serial ->port_write_mutex 2006-03-23 07:38:14 -08:00
sx.h
sxboards.h
sxwindow.h
synclink_gt.c [PATCH] s/;;/;/g 2006-03-24 07:33:24 -08:00
synclink.c [PATCH] s/;;/;/g 2006-03-24 07:33:24 -08:00
synclinkmp.c [PATCH] s/;;/;/g 2006-03-24 07:33:24 -08:00
sysrq.c [PATCH] sysrq cleanup 2006-03-25 08:22:52 -08:00
tb0219.c [PATCH] tb0219: convert to the new platform device interface 2006-03-22 07:53:56 -08:00
tipar.c [PATCH] tipar fixes 2006-02-11 21:41:13 -08:00
tlclk.c [PATCH] tlclk: fix handling of device major 2006-03-26 08:56:57 -08:00
toshiba.c [PATCH] remove ISA legacy functions: drivers/char/toshiba.c 2006-03-24 07:33:19 -08:00
tty_io.c BUG_ON() Conversion in drivers/char 2006-03-26 18:17:21 +02:00
tty_ioctl.c [PATCH] coverity: tty_ldisc_ref return null check 2005-06-28 21:20:34 -07:00
vc_screen.c [PATCH] fix ia64 compile failure with gcc4.1 2005-12-29 10:19:21 -08:00
viocons.c [PATCH] powerpc: remove bitfields from HvLpEvent 2006-01-12 20:09:29 +11:00
viotape.c [PATCH] changing CONFIG_LOCALVERSION rebuilds too much, for no good reason 2005-11-09 07:55:57 -08:00
vme_scc.c [PATCH] sem2mutex: serial ->port_write_mutex 2006-03-23 07:38:14 -08:00
vr41xx_giu.c [PATCH] vr41xx: convert to the new platform device interface 2006-03-22 07:53:56 -08:00
vr41xx_rtc.c [PATCH] vr41xx: convert to the new platform device interface 2006-03-22 07:53:56 -08:00
vt_ioctl.c [PATCH] Only disallow _setting_ of function key string 2005-11-07 07:53:39 -08:00
vt.c [PATCH] sem2mutex: tty 2006-03-23 07:38:11 -08:00