]> git.karo-electronics.de Git - karo-tx-linux.git/log
karo-tx-linux.git
15 years agosgi-gru: macro for scanning all gru chiplets
Jack Steiner [Thu, 2 Apr 2009 23:59:10 +0000 (16:59 -0700)]
sgi-gru: macro for scanning all gru chiplets

Add macro for scanning all active GRU chiplets.  Maximum chiplet id is
saved during GRU initialization.

Signed-off-by: Jack Steiner <steiner@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agosgi-gru: add support to the GRU driver for message queue interrupts
Jack Steiner [Thu, 2 Apr 2009 23:59:10 +0000 (16:59 -0700)]
sgi-gru: add support to the GRU driver for message queue interrupts

Add support to the GRU driver for message queue interrupts.

Signed-off-by: Jack Steiner <steiner@sgi.com>
Signed-off-by: Dean Nelson <dcn@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agosgi-gru: restructure the GRU vtop functions
Jack Steiner [Thu, 2 Apr 2009 23:59:09 +0000 (16:59 -0700)]
sgi-gru: restructure the GRU vtop functions

Restructure the GRU vtop functions in preparation for future changes.
This patch simply moves code around & does not change the algorithm.

Signed-off-by: Jack Steiner <steiner@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agosgi-gru: aSID (context management) bug fixes
Jack Steiner [Thu, 2 Apr 2009 23:59:08 +0000 (16:59 -0700)]
sgi-gru: aSID (context management) bug fixes

This patch fixes bugs related to ASID (context id) management in the GRU
driver.  These changes are all internal to the SGI GRU driver and have no
effect on the base kernel.

Signed-off-by: Jack Steiner <steiner@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agosgi-gru: add support for a user to explicitly unload a GRU context
Jack Steiner [Thu, 2 Apr 2009 23:59:07 +0000 (16:59 -0700)]
sgi-gru: add support for a user to explicitly unload a GRU context

Add support for a user to explicitly unload a GRU context.

Signed-off-by: Jack Steiner <steiner@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agosgi-gru: add statistics to the GRU context management functions
Jack Steiner [Thu, 2 Apr 2009 23:59:06 +0000 (16:59 -0700)]
sgi-gru: add statistics to the GRU context management functions

Add statistics to the GRU context management functions.

Signed-off-by: Jack Steiner <steiner@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agosgi-gru: change GRU CCH commands from inline functions to outofline functions
Jack Steiner [Thu, 2 Apr 2009 23:59:06 +0000 (16:59 -0700)]
sgi-gru: change GRU CCH commands from inline functions to outofline functions

Change the GRU instructions that manage contexts from inline functions to
out-of-line functions.  This simplifies adding statistics & error checking
to these functions.

Signed-off-by: Jack Steiner <steiner@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agosgi-gru: improvements to GRU debug messages & statistics
Jack Steiner [Thu, 2 Apr 2009 23:59:05 +0000 (16:59 -0700)]
sgi-gru: improvements to GRU debug messages & statistics

Improvements to GRU debug messages & statistics.

Signed-off-by: Jack Steiner <steiner@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agosgi-gru: misc GRU cleanup
Jack Steiner [Thu, 2 Apr 2009 23:59:04 +0000 (16:59 -0700)]
sgi-gru: misc GRU cleanup

Misc trivial GRU drivers fixes:
- fix long lines
- eliminate extra whitespace
- eliminate compiler warning
- better validation of invalidate user parameters
- bug fix for GRU TLB flush (not the cpu TLB flush)

These changes are all internal to the SGI GRU driver and have no effect
on the base kernel.

Signed-off-by: Jack Steiner <steiner@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agosgi-gru: add macros for using the UV hub to send interrupts
Jack Steiner [Thu, 2 Apr 2009 23:59:03 +0000 (16:59 -0700)]
sgi-gru: add macros for using the UV hub to send interrupts

Add macros for using the UV hub to send interrupts.  Change the IPI code
to use these macros.  These macros will also be used in additional patches
that will follow.

Signed-off-by: Jack Steiner <steiner@sgi.com>
Cc: "Luck, Tony" <tony.luck@intel.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agosgi-gru: add definitions of ia64 GRU MMRs
Jack Steiner [Thu, 2 Apr 2009 23:59:02 +0000 (16:59 -0700)]
sgi-gru: add definitions of ia64 GRU MMRs

Add definitions for IA64 GRU MMRs.

Signed-off-by: Jack Steiner <steiner@sgi.com>
Cc: "Luck, Tony" <tony.luck@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agosgi-gru: add definitions of x86_64 GRU MMRs
Jack Steiner [Thu, 2 Apr 2009 23:59:01 +0000 (16:59 -0700)]
sgi-gru: add definitions of x86_64 GRU MMRs

Add definitions for x86_64 GRU MMRs.

Signed-off-by: Jack Steiner <steiner@sgi.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agosgi-gru: exclude UV definitions on 32-bit x86
Jack Steiner [Thu, 2 Apr 2009 23:59:00 +0000 (16:59 -0700)]
sgi-gru: exclude UV definitions on 32-bit x86

Eliminate compile errors on 32-bit X86 caused by UV.

Signed-off-by: Jack Steiner <steiner@sgi.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agorelay: fix for possible loss/corruption of produced subbufs
Aravind Srinivasan [Thu, 2 Apr 2009 23:58:59 +0000 (16:58 -0700)]
relay: fix for possible loss/corruption of produced subbufs

Fix possible loss/corruption of produced subbufs in
relay_subbufs_consumed().

When buf->subbufs_produced wraps around after UINT_MAX and
buf->subbufs_consumed is still < UINT_MAX, the condition

if (buf->subbufs_consumed > buf->subbufs_produced)

will be true even for certain valid values of subbufs_consumed.  This may
lead to loss or corruption of produced subbufs.

Signed-off-by: Aravind Srinivasan <raa.aars@gmail.com>
Cc: Tom Zanussi <tzanussi@gmail.com>
Cc: Tom Zanussi <zanussi@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agokexec: vmcoreinfo_data[] can become static
Dmitri Vorobiev [Thu, 2 Apr 2009 23:58:58 +0000 (16:58 -0700)]
kexec: vmcoreinfo_data[] can become static

The vmcoreinfo_data[] array is not used outside of kernel/kexec.c, and
can therefore become static. This patch adds the relevant keyword to the
definition of the array.

Noticed by sparse.

Signed-off-by: Dmitri Vorobiev <dmitri.vorobiev@movial.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agokexec: add dmesg log symbols to /proc/vmcoreinfo lists
Neil Horman [Thu, 2 Apr 2009 23:58:57 +0000 (16:58 -0700)]
kexec: add dmesg log symbols to /proc/vmcoreinfo lists

It would be nice to be able to extract the dmesg log from a vmcore file
without needing to keep the debug symbols for the running kernel handy all
the time.  We have a facility to do this in /proc/vmcore.  This patch adds
the log_buf and log_end symbols to the vmcoreinfo area so that tools (like
makedumpfile) can easily extract the dmesg logs from a vmcore image.

[akpm@linux-foundation.org: several fixes and cleanups]
[akpm@linux-foundation.org: fix unused log_buf_kexec_setup()]
[akpm@linux-foundation.org: build fix]
Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
Cc: Simon Horman <horms@verge.net.au>
Acked-by: Vivek Goyal <vgoyal@redhat.com>
Cc: Neil Horman <nhorman@tuxdriver.com>
Cc: Simon Horman <horms@verge.net.au>
Cc: Vivek Goyal <vgoyal@redhat.com>
Cc: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agow1: coding style cleanups in w1_io.c
Daniel Mack [Thu, 2 Apr 2009 23:58:54 +0000 (16:58 -0700)]
w1: coding style cleanups in w1_io.c

This fixes a number of coding style issues I stubled over.

Acked-by: Evgeniy Polyakov <johnpol@2ka.mipt.ru>
Signed-off-by: Daniel Mack <daniel@caiaq.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoparport: netmos 9845 & 9855 1P4S fixes
Philippe De Muyter [Thu, 2 Apr 2009 23:58:53 +0000 (16:58 -0700)]
parport: netmos 9845 & 9855 1P4S fixes

netmos serial/parallel adapters come in different flavour differing only
by the number of parallel and serial ports, which are encoded in the
subdevice ID.

Last fix of Christian Pellegrin for 9855 2P2S broke support for 9855 1P4S,
and works only by side-effect for the first parallel port of a 2P2S, as
this first parallel port is found by reading the second addr entry of
(struct parport_pc_pci) cards[netmos_9855], which is not initialized, and
hence has value 0, which happens to be the BAR of the first parallel port.

netmos_9xx5_combo entry in (struct parport_pc_pci) cards[], which is used
for a 9845 1P4S must also be fixed for the parallel port support when
there are 4 serial ports because this entry currently gives 2 as BAR index
for the parallel port.  Actually, in this case, BAR 2 is the 3rd serial
port while the parallel port is at BAR 4.

I fixed 9845 1P4S and 9855 1P4S support, while preserving 9855 2P2S support,

- by creating a netmos_9855_2p entry and using it for 9855 boards with 2
  parallel ports : 9855 2P2S and 9855 2P0S boards,

- and by allowing netmos_parallel_init to change not only the number of
  parallel ports (0 or 1), but making it also change the BAR index of the
  parallel port when the serial ports are before the parallel port.

PS: the netmos_9855_2p entry in (struct pciserial_board)
pci_parport_serial_boards[] is needed because netmos_parallel_init has no
clean way to replace FL_BASE2 by FL_BASE4 in the description of the serial
ports in function of the number of parallel ports on the card.

Tested with 9845 1P4S, 9855 1P4S and 9855 2P2S boards.

Signed-off-by: Philippe De Muyter <phdm@macqel.be>
Tested-by: Christian Pellegrin <chripell@fsfe.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoedac: AMD8111 driver Kconfig & Makefile
Harry Ciao [Thu, 2 Apr 2009 23:58:51 +0000 (16:58 -0700)]
edac: AMD8111 driver Kconfig & Makefile

Introduce Kconfig and Makefile options for AMD8111 EDAC driver.

Signed-off-by: Harry Ciao <qingtao.cao@windriver.com>
Cc: Doug Thompson <norsk5@yahoo.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoedac: AMD8131 driver Kconfig & Makefile
Harry Ciao [Thu, 2 Apr 2009 23:58:51 +0000 (16:58 -0700)]
edac: AMD8131 driver Kconfig & Makefile

Introduce Kconfig and Makefile options for AMD8131 EDAC driver.

Signed-off-by: Harry Ciao <qingtao.cao@windriver.com>
Cc: Doug Thompson <norsk5@yahoo.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoedac: AMD8131 driver source file
Harry Ciao [Thu, 2 Apr 2009 23:58:50 +0000 (16:58 -0700)]
edac: AMD8131 driver source file

Introduce AMD8131 EDAC driver source file, which makes use of error
detections on the PCI-X Bridge Controllers on the AMD8131 HyperTransport
PCI-X Tunnel.

Signed-off-by: Harry Ciao <qingtao.cao@windriver.com>
Cc: Doug Thompson <norsk5@yahoo.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoedac: AMD8131 driver header file
Harry Ciao [Thu, 2 Apr 2009 23:58:49 +0000 (16:58 -0700)]
edac: AMD8131 driver header file

Introduce AMD8131 EDAC driver header file, which adds register and bits
definitions for the PCI-X Bridge Controller on the AMD8131 HyperTransport
I/O Hub.

Signed-off-by: Harry Ciao <qingtao.cao@windriver.com>
Cc: Doug Thompson <norsk5@yahoo.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agopci: Add AMD8111 PCI Bridge PCI Device ID
Harry Ciao [Thu, 2 Apr 2009 23:58:48 +0000 (16:58 -0700)]
pci: Add AMD8111 PCI Bridge PCI Device ID

Add the PCI Device ID of the PCI Bridge Controller on AMD8111 chip.

Signed-off-by: Harry Ciao <qingtao.cao@windriver.com>
Cc: Doug Thompson <norsk5@yahoo.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoedac: Add edac_pci_alloc_index()
Harry Ciao [Thu, 2 Apr 2009 23:58:47 +0000 (16:58 -0700)]
edac: Add edac_pci_alloc_index()

Add edac_pci_alloc_index(), because for MAPLE platform there may exist
several EDAC driver modules that could make use of edac_pci_ctl_info
structure at the same time.  The index allocation for these structures
should be taken care of by EDAC core.

Signed-off-by: Harry Ciao <qingtao.cao@windriver.com>
Cc: Doug Thompson <norsk5@yahoo.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoedac: AMD8111 driver source file
Harry Ciao [Thu, 2 Apr 2009 23:58:46 +0000 (16:58 -0700)]
edac: AMD8111 driver source file

Introduce AMD8111 EDAC driver source file, which makes use of error
detections on the LPC Bridge Controller and PCI Bridge Controller on the
AMD8111 HyperTransport I/O Hub.

Signed-off-by: Harry Ciao <qingtao.cao@windriver.com>
Cc: Doug Thompson <norsk5@yahoo.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoedac: AMD8111 driver header file
Harry Ciao [Thu, 2 Apr 2009 23:58:46 +0000 (16:58 -0700)]
edac: AMD8111 driver header file

Introduce AMD8111 EDAC driver header file, which adds register and bits
definitions for the LPC Bridge Controller and PCI Bridge Controller on the
AMD8111 HyperTransport I/O Hub.

Signed-off-by: Harry Ciao <qingtao.cao@windriver.com>
Cc: Doug Thompson <norsk5@yahoo.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoedac: new ppc4xx driver module
Grant Erickson [Thu, 2 Apr 2009 23:58:45 +0000 (16:58 -0700)]
edac: new ppc4xx driver module

This adds support for an EDAC memory controller adaptation driver for the
"ibm,sdram-4xx-ddr2" ECC controller realized in the AMCC PowerPC 405EX[r].

At present, this driver has been developed and tested against the
controller realization in the AMCC PPC405EX[r] on the AMCC Kilauea and
Haleakala boards (256 MiB w/o ECC memory soldered onto the board) and a
proprietary board based on those designs (128 MiB ECC memory, also
soldered onto the board).

In the future, dynamic feature detection and handling needs to be added
for the other realizations of this controller found in the 440SP, 440SPe,
460EX, 460GT and 460SX.

Eventually, this driver will likely be evolved and adapted to the above
variant realizations of this controller as well as broken apart to handle
the other known ECC-capable controllers prevalent in other PPC4xx
processors:

  - IBM SDRAM (405GP, 405CR and 405EP) "ibm,sdram-4xx"
  - IBM DDR1 (440GP, 440GX, 440EP and 440GR) "ibm,sdram-4xx-ddr"
  - Denali DDR1/DDR2 (440EPX and 440GRX) "denali,sdram-4xx-ddr2"

[akpm@linux-foundation.org: coding-style fixes]
Signed-off-by: Grant Erickson <gerickson@nuovations.com>
Signed-off-by: Doug Thompson <dougthompson@xmission.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoedac: remove EDAC's experimental status
Doug Thompson [Thu, 2 Apr 2009 23:58:43 +0000 (16:58 -0700)]
edac: remove EDAC's experimental status

After 3 years, this is a patch to remove the EXPERIMENTAL tag on EDAC.  We
now have many module drivers submitters in EDAC and believe EDAC is no
longer EXPERIMENTAL

Signed-off-by: Doug Thompson <dougthompson@xmission.com
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoedac: add more verbose debug info
Hitoshi Mitake [Thu, 2 Apr 2009 23:58:43 +0000 (16:58 -0700)]
edac: add more verbose debug info

A patch for making a debugging information more verbose for use in
development debugging.

By enabling the new option "More verbose debugging", information about
source file and line number will be added to debugging message.

This is sample output,

EDAC MC0: Giving out device to 'e7xxx_edac' 'E7205': DEV 0000:00:00.0
EDAC DEBUG: in drivers/edac/edac_pci.c, line at 48: edac_pci_alloc_ctl_info()
EDAC DEBUG: in drivers/edac/edac_pci.c, line at 334: edac_pci_add_device()
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Signed-off-by: Hitoshi Mitake <h.mitake@gmail.com>
Signed-off-by: Doug Thompson <dougthompson@xmission.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agonbd: trivial cleanups
Pavel Machek [Thu, 2 Apr 2009 23:58:42 +0000 (16:58 -0700)]
nbd: trivial cleanups

Trivial cleanups for nbd: only the return -EIO one really changes code,
and I've verified all the callers (plus 0 == success, 1 == error
convention is really ugly).

Signed-off-by: Pavel Machek <pavel@suse.cz>
Acked-by: Paul Clements <paul.clements@steeleye.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agonbd: add locking to nbd_ioctl
Pavel Machek [Thu, 2 Apr 2009 23:58:41 +0000 (16:58 -0700)]
nbd: add locking to nbd_ioctl

The code was written to rely on big kernel lock to protect it from races.
It mostly works when interface is not abused.

So this uses tx_lock to protect data structures from concurrent use
between ioctl and worker threads.

Next step will be moving from ioctl to unlocked_ioctl.

[akpm@linux-foundation.org: coding-style fixes]
[akpm@linux-foundation.org: add missing return]
Signed-off-by: Pavel Machek <pavel@suse.cz>
Acked-by: Paul Clements <paul.clements@steeleye.com>
Cc: Jens Axboe <jens.axboe@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agopids: kill signal_struct-> __pgrp/__session and friends
Oleg Nesterov [Thu, 2 Apr 2009 23:58:39 +0000 (16:58 -0700)]
pids: kill signal_struct-> __pgrp/__session and friends

We are wasting 2 words in signal_struct without any reason to implement
task_pgrp_nr() and task_session_nr().

task_session_nr() has no callers since
2e2ba22ea4fd4bb85f0fa37c521066db6775cbef, we can remove it.

task_pgrp_nr() is still (I believe wrongly) used in fs/autofsX and
fs/coda.

This patch reimplements task_pgrp_nr() via task_pgrp_nr_ns(), and kills
__pgrp/__session and the related helpers.

The change in drivers/char/tty_io.c is cosmetic, but hopefully makes sense
anyway.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Acked-by: Alan Cox <number6@the-village.bc.nu> [tty parts]
Cc: Cedric Le Goater <clg@fr.ibm.com>
Cc: Dave Hansen <haveblue@us.ibm.com>
Cc: Eric Biederman <ebiederm@xmission.com>
Cc: Pavel Emelyanov <xemul@openvz.org>
Cc: Serge Hallyn <serue@us.ibm.com>
Cc: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Cc: Roland McGrath <roland@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agopids: refactor vnr/nr_ns helpers to make them safe
Oleg Nesterov [Thu, 2 Apr 2009 23:58:38 +0000 (16:58 -0700)]
pids: refactor vnr/nr_ns helpers to make them safe

Inho, the safety rules for vnr/nr_ns helpers are horrible and buggy.

task_pid_nr_ns(task) needs rcu/tasklist depending on task == current.

As for "special" pids, vnr/nr_ns helpers always need rcu.  However, if
task != current, they are unsafe even under rcu lock, we can't trust
task->group_leader without the special checks.

And almost every helper has a callsite which needs a fix.

Also, it is a bit annoying that the implementations of, say,
task_pgrp_vnr() and task_pgrp_nr_ns() are not "symmetrical".

This patch introduces the new helper, __task_pid_nr_ns(), which is always
safe to use, and turns all other helpers into the trivial wrappers.

After this I'll send another patch which converts task_tgid_xxx() as well,
they're are a bit special.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Cc: Louis Rilling <Louis.Rilling@kerlabs.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Pavel Emelyanov <xemul@openvz.org>
Cc: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Cc: Roland McGrath <roland@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agopids: improve get_task_pid() to fix the unsafe sys_wait4()->task_pgrp()
Oleg Nesterov [Thu, 2 Apr 2009 23:58:36 +0000 (16:58 -0700)]
pids: improve get_task_pid() to fix the unsafe sys_wait4()->task_pgrp()

sys_wait4() does get_pid(task_pgrp(current)), this is not safe.  We can
add rcu lock/unlock around, but we already have get_task_pid() which can
be improved to handle the special pids in more reliable manner.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Cc: Louis Rilling <Louis.Rilling@kerlabs.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Pavel Emelyanov <xemul@openvz.org>
Cc: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Cc: Roland McGrath <roland@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agopids: document task_pgrp/task_session is not safe without tasklist/rcu
Oleg Nesterov [Thu, 2 Apr 2009 23:58:35 +0000 (16:58 -0700)]
pids: document task_pgrp/task_session is not safe without tasklist/rcu

Even if task == current, it is not safe to dereference the result of
task_pgrp/task_session.  We can race with another thread which changes the
special pid via setpgid/setsid.

Document this.  The next 2 patches give an example of the unsafe usage, we
have more bad users.

[akpm@linux-foundation.org: coding-style fixes]
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Cc: Louis Rilling <Louis.Rilling@kerlabs.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Pavel Emelyanov <xemul@openvz.org>
Cc: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Cc: Roland McGrath <roland@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agosysctl: fix suid_dumpable and lease-break-time sysctls
Matthew Wilcox [Thu, 2 Apr 2009 23:58:33 +0000 (16:58 -0700)]
sysctl: fix suid_dumpable and lease-break-time sysctls

Arne de Bruijn points out that commit
76fdbb25f963de5dc1e308325f0578a2f92b1c2d ("coredump masking: bound
suid_dumpable sysctl") mistakenly limits lease-break-time instead of
suid_dumpable.

Signed-off-by: Matthew Wilcox <matthew@wil.cx>
Reported-by: Arne de Bruijn <kernelbt@arbruijn.dds.nl>
Cc: Kawai, Hidehiro <hidehiro.kawai.ez@hitachi.com>
Cc: Alexey Dobriyan <adobriyan@gmail.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agohpet: fix the possibility of insane return value of hpet_calibrate() against SMI
Yasunori Goto [Thu, 2 Apr 2009 23:58:31 +0000 (16:58 -0700)]
hpet: fix the possibility of insane return value of hpet_calibrate() against SMI

hpet_calibrate() has a possibility of miss-calibration due to SMI.  If SMI
interrupts in the while loop of calibration, then return value will be
big.  This change calibrates until stabilizing by the return value with a
small value.

[akpm@linux-foundation.org: trivial style tweaks]
Signed-off-by: Yasunori Goto <y-goto@jp.fujitsu.com>
Acked-by: Clemens Ladisch <clemens@ladisch.de>
Acked-by: Vojtech Pavlik <vojtech@suse.cz>
Cc: Robert Picco <Robert.Picco@hp.com>
Cc: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
Cc: Ingo Molnar <mingo@elte.hu>
Acked-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agosynclink_gt: add clock options
Paul Fulghum [Thu, 2 Apr 2009 23:58:30 +0000 (16:58 -0700)]
synclink_gt: add clock options

Add support for x8 asynchronous sample rate and ability to specify base
clock frequency.

Signed-off-by: Paul Fulghum <paulkf@microgate.com>
Acked-by: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agostruct linux_binprm: drop unused fields
Kirill A. Shutemov [Thu, 2 Apr 2009 23:58:29 +0000 (16:58 -0700)]
struct linux_binprm: drop unused fields

Signed-off-by: Kirill A. Shutemov <kirill@shutemov.name>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agobin_elf_fdpic: check the return value of clear_user
Mike Frysinger [Thu, 2 Apr 2009 23:58:28 +0000 (16:58 -0700)]
bin_elf_fdpic: check the return value of clear_user

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>
Cc: David Howells <dhowells@redhat.com>
Cc: Greg Ungerer <gerg@snapgear.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoproc_sysctl: use CONFIG_PROC_SYSCTL around ipc and utsname proc_handlers
Serge E. Hallyn [Thu, 2 Apr 2009 23:58:27 +0000 (16:58 -0700)]
proc_sysctl: use CONFIG_PROC_SYSCTL around ipc and utsname proc_handlers

As pointed out by Cedric Le Goater (in response to Alexey's original
comment wrt mqns), ipc_sysctl.c and utsname_sysctl.c are using
CONFIG_PROC_FS, not CONFIG_PROC_SYSCTL, to determine whether to define
the proc_handlers.  Change that.

Signed-off-by: Serge E. Hallyn <serue@us.ibm.com>
Cc: Cedric Le Goater <clg@fr.ibm.com>
Acked-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoipc: make shm_get_stat() more robust
Tony Battersby [Thu, 2 Apr 2009 23:58:26 +0000 (16:58 -0700)]
ipc: make shm_get_stat() more robust

shm_get_stat() assumes idr_find(&shm_ids(ns).ipcs_idr) returns "struct
shmid_kernel *"; all other callers assume that it returns "struct
kern_ipc_perm *".  This works because "struct kern_ipc_perm" is currently
the first member of "struct shmid_kernel", but it would be better to use
container_of() to prevent future breakage.

Signed-off-by: Tony Battersby <tonyb@cybernetics.com>
Cc: Jiri Olsa <olsajiri@gmail.com>
Cc: Jiri Kosina <jkosina@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agocpu hotplug: remove unused cpuhotplug_mutex_lock()
Lai Jiangshan [Thu, 2 Apr 2009 23:58:25 +0000 (16:58 -0700)]
cpu hotplug: remove unused cpuhotplug_mutex_lock()

cpuhotplug_mutex_lock() is not used, remove it.

Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Acked-by: Gautham R Shenoy <ego@in.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoworkqueue: avoid recursion in run_workqueue()
Lai Jiangshan [Thu, 2 Apr 2009 23:58:24 +0000 (16:58 -0700)]
workqueue: avoid recursion in run_workqueue()

1) lockdep will complain when run_workqueue() performs recursion.

2) The recursive implementation of run_workqueue() means that
   flush_workqueue() and its documentation are inconsistent.  This may
   hide deadlocks and other bugs.

3) The recursion in run_workqueue() will poison cwq->current_work, but
   flush_work() and __cancel_work_timer(), etcetera need a reliable
   cwq->current_work.

Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Acked-by: Oleg Nesterov <oleg@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Eric Dumazet <dada1@cosmosbay.com>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoptrace_untrace: fix the SIGNAL_STOP_STOPPED check
Oleg Nesterov [Thu, 2 Apr 2009 23:58:23 +0000 (16:58 -0700)]
ptrace_untrace: fix the SIGNAL_STOP_STOPPED check

This bug is ancient too. ptrace_untrace() must not resume the task
if the group stop in progress, we should set TASK_STOPPED instead.

Unfortunately, we still have problems here:

- if the process/thread was traced, SIGNAL_STOP_STOPPED
  does not necessary means this thread group is stopped.

- ptrace breaks the bookkeeping of ->group_stop_count.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Cc: Jerome Marchand <jmarchan@redhat.com>
Cc: Roland McGrath <roland@redhat.com>
Cc: Denys Vlasenko <dvlasenk@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoptrace_detach: the wrong wakeup breaks the ERESTARTxxx logic
Oleg Nesterov [Thu, 2 Apr 2009 23:58:21 +0000 (16:58 -0700)]
ptrace_detach: the wrong wakeup breaks the ERESTARTxxx logic

Another ancient bug. Consider this trivial test-case,

int main(void)
{
int pid = fork();

if (pid) {
ptrace(PTRACE_ATTACH, pid, NULL, NULL);
wait(NULL);
ptrace(PTRACE_DETACH, pid, NULL, NULL);
} else {
pause();
printf("WE HAVE A KERNEL BUG!!!\n");
}

return 0;
}

the child must not "escape" for sys_pause(), but it can and this was seen
in practice.

This is because ptrace_detach does:

if (!child->exit_state)
wake_up_process(child);

this wakeup can happen after this child has already restarted sys_pause(),
because it gets another wakeup from ptrace_untrace().

With or without this patch, perhaps sys_pause() needs a fix.  But this
wakeup also breaks the SIGNAL_STOP_STOPPED logic in ptrace_untrace().

Remove this wakeup.  The caller saw this task in TASK_TRACED state, and
unless it was SIGKILL'ed in between __ptrace_unlink()->ptrace_untrace()
should handle this case correctly.  If it was SIGKILL'ed, we don't need to
wakup the dying tracee too.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Cc: Jerome Marchand <jmarchan@redhat.com>
Acked-by: Roland McGrath <roland@redhat.com>
Cc: Denys Vlasenko <dvlasenk@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agotracehook_notify_death: use task_detached() helper
Oleg Nesterov [Thu, 2 Apr 2009 23:58:20 +0000 (16:58 -0700)]
tracehook_notify_death: use task_detached() helper

Now that task_detached() is exported, change tracehook_notify_death() to
use this helper, nobody else checks ->exit_signal == -1 by hand.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: "Metzger, Markus T" <markus.t.metzger@intel.com>
Acked-by: Roland McGrath <roland@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoforget_original_parent: do not abuse child->ptrace_entry
Oleg Nesterov [Thu, 2 Apr 2009 23:58:19 +0000 (16:58 -0700)]
forget_original_parent: do not abuse child->ptrace_entry

By discussion with Roland.

- Use ->sibling instead of ->ptrace_entry to chain the need to be
  release_task'd childs. Nobody else can use ->sibling, this task
  is EXIT_DEAD and nobody can find it on its own list.

- rename ptrace_dead to dead_childs.

- Now that we don't have the "parallel" untrace code, change back
  reparent_thread() to return void, pass dead_childs as an argument.

Actually, I don't understand why do we notify /sbin/init when we
reparent a zombie, probably it is better to reap it unconditionally.

[akpm@linux-foundation.org: s/childs/children/]
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: "Metzger, Markus T" <markus.t.metzger@intel.com>
Cc: Roland McGrath <roland@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoforget_original_parent: split out the un-ptrace part
Oleg Nesterov [Thu, 2 Apr 2009 23:58:18 +0000 (16:58 -0700)]
forget_original_parent: split out the un-ptrace part

By discussion with Roland.

- Rename ptrace_exit() to exit_ptrace(), and change it to do all the
  necessary work with ->ptraced list by its own.

- Move this code from exit.c to ptrace.c

- Update the comment in ptrace_detach() to explain the rechecking of
  the child->ptrace.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: "Metzger, Markus T" <markus.t.metzger@intel.com>
Cc: Roland McGrath <roland@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoreparent_thread: fix a zombie leak if /sbin/init ignores SIGCHLD
Oleg Nesterov [Thu, 2 Apr 2009 23:58:17 +0000 (16:58 -0700)]
reparent_thread: fix a zombie leak if /sbin/init ignores SIGCHLD

If /sbin/init ignores SIGCHLD and we re-parent a zombie, it is leaked.
reparent_thread() does do_notify_parent() which sets ->exit_signal = -1 in
this case.  This means that nobody except us can reap it, the detached
task is not visible to do_wait().

Change reparent_thread() to return a boolean (like __pthread_detach) to
indicate that the thread is dead and must be released.  Also change
forget_original_parent() to add the child to ptrace_dead list in this
case.

The naming becomes insane, the next patch does the cleanup.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Cc: Roland McGrath <roland@redhat.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoreparent_thread: fix the "is it traced" check
Oleg Nesterov [Thu, 2 Apr 2009 23:58:16 +0000 (16:58 -0700)]
reparent_thread: fix the "is it traced" check

reparent_thread() uses ptrace_reparented() to check whether this thread is
ptraced, in that case we should not notify the new parent.

But ptrace_reparented() is not exactly correct when the reparented thread
is traced by /sbin/init, because forget_original_parent() has already
changed ->real_parent.

Currently, the only problem is the false notification.  But with the next
patch the kernel crash in this (yes, pathological) case.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Cc: Roland McGrath <roland@redhat.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoreparent_thread: don't call kill_orphaned_pgrp() if task_detached()
Oleg Nesterov [Thu, 2 Apr 2009 23:58:15 +0000 (16:58 -0700)]
reparent_thread: don't call kill_orphaned_pgrp() if task_detached()

If task_detached(p) == T, then either

  a) p is not the main thread, we will find the group leader on the
     ->children list.

or

  b) p is the group leader but its ->exit_state = EXIT_DEAD.  This
     can only happen when the last sub-thread has died, but in that case
     that thread has already called kill_orphaned_pgrp() from
     exit_notify().

In both cases kill_orphaned_pgrp() looks bogus.

Move the task_detached() check up and simplify the code, this is also
right from the "common sense" pov: we should do nothing with the detached
childs, except move them to the new parent's ->children list.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Cc: Roland McGrath <roland@redhat.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoptrace: fix possible zombie leak on PTRACE_DETACH
Oleg Nesterov [Thu, 2 Apr 2009 23:58:14 +0000 (16:58 -0700)]
ptrace: fix possible zombie leak on PTRACE_DETACH

When ptrace_detach() takes tasklist, the tracee can be SIGKILL'ed.  If it
has already passed exit_notify() we can leak a zombie, because a) ptracing
disables the auto-reaping logic, and b) ->real_parent was not notified
about the child's death.

ptrace_detach() should follow the ptrace_exit's logic, change the code
accordingly.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Cc: Jerome Marchand <jmarchan@redhat.com>
Cc: Roland McGrath <roland@redhat.com>
Tested-by: Denys Vlasenko <dvlasenk@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoptrace: reintroduce __ptrace_detach() as a callee of ptrace_exit()
Oleg Nesterov [Thu, 2 Apr 2009 23:58:13 +0000 (16:58 -0700)]
ptrace: reintroduce __ptrace_detach() as a callee of ptrace_exit()

No functional changes, preparation for the next patch.

Move the "should we release this child" logic into the separate handler,
__ptrace_detach().

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Cc: Jerome Marchand <jmarchan@redhat.com>
Cc: Roland McGrath <roland@redhat.com>
Cc: Denys Vlasenko <dvlasenk@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoptrace: simplify ptrace_exit()->ignoring_children() path
Oleg Nesterov [Thu, 2 Apr 2009 23:58:12 +0000 (16:58 -0700)]
ptrace: simplify ptrace_exit()->ignoring_children() path

ignoring_children() takes parent->sighand->siglock and checks
k_sigaction[SIGCHLD] atomically.  But this buys nothing, we can't get the
"really" wrong result even if we race with sigaction(SIGCHLD).  If we read
the "stale" sa_handler/sa_flags we can pretend it was changed right after
the check.

Remove spin_lock(->siglock), and kill "int ign" which caches the result of
ignoring_children() which becomes rather trivial.

Perhaps it makes sense to export this helper, do_notify_parent() can use
it too.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Cc: Jerome Marchand <jmarchan@redhat.com>
Cc: Roland McGrath <roland@redhat.com>
Cc: Denys Vlasenko <dvlasenk@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoptrace: kill __ptrace_detach(), fix ->exit_state check
Oleg Nesterov [Thu, 2 Apr 2009 23:58:11 +0000 (16:58 -0700)]
ptrace: kill __ptrace_detach(), fix ->exit_state check

Move the code from __ptrace_detach() to its single caller and kill this
helper.

Also, fix the ->exit_state check, we shouldn't wake up EXIT_DEAD tasks.
Actually, I think task_is_stopped_or_traced() makes more sense, but this
needs another patch.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Cc: Jerome Marchand <jmarchan@redhat.com>
Cc: Roland McGrath <roland@redhat.com>
Cc: Denys Vlasenko <dvlasenk@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agosignals: SI_USER: Masquerade si_pid when crossing pid ns boundary
Sukadev Bhattiprolu [Thu, 2 Apr 2009 23:58:09 +0000 (16:58 -0700)]
signals: SI_USER: Masquerade si_pid when crossing pid ns boundary

When sending a signal to a descendant namespace, set ->si_pid to 0 since
the sender does not have a pid in the receiver's namespace.

Note:
- If rt_sigqueueinfo() sets si_code to SI_USER when sending a
  signal across a pid namespace boundary, the value in ->si_pid
  will be cleared to 0.

Signed-off-by: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Cc: Oleg Nesterov <oleg@tv-sign.ru>
Cc: Roland McGrath <roland@redhat.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Daniel Lezcano <daniel.lezcano@free.fr>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agosignals: protect cinit from blocked fatal signals
Sukadev Bhattiprolu [Thu, 2 Apr 2009 23:58:08 +0000 (16:58 -0700)]
signals: protect cinit from blocked fatal signals

Normally SIG_DFL signals to global and container-init are dropped early.
But if a signal is blocked when it is posted, we cannot drop the signal
since the receiver may install a handler before unblocking the signal.
Once this signal is queued however, the receiver container-init has no way
of knowing if the signal was sent from an ancestor or descendant
namespace.  This patch ensures that contianer-init drops all SIG_DFL
signals in get_signal_to_deliver() except SIGKILL/SIGSTOP.

If SIGSTOP/SIGKILL originate from a descendant of container-init they are
never queued (i.e dropped in sig_ignored() in an earler patch).

If SIGSTOP/SIGKILL originate from parent namespace, the signal is queued
and container-init processes the signal.

IOW, if get_signal_to_deliver() sees a sig_kernel_only() signal for global
or container-init, the signal must have been generated internally or must
have come from an ancestor ns and we process the signal.

Further, the signal_group_exit() check was needed to cover the case of a
multi-threaded init sending SIGKILL to other threads when doing an exit()
or exec().  But since the new sig_kernel_only() check covers the SIGKILL,
the signal_group_exit() check is no longer needed and can be removed.

Finally, now that we have all pieces in place, set SIGNAL_UNKILLABLE for
container-inits.

Signed-off-by: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Cc: Oleg Nesterov <oleg@tv-sign.ru>
Cc: Roland McGrath <roland@redhat.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Daniel Lezcano <daniel.lezcano@free.fr>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agosignals: zap_pid_ns_process() should use force_sig()
Sukadev Bhattiprolu [Thu, 2 Apr 2009 23:58:06 +0000 (16:58 -0700)]
signals: zap_pid_ns_process() should use force_sig()

send_signal() assumes that signals with SEND_SIG_PRIV are generated from
within the same namespace.  So any nested container-init processes become
immune to the SIGKILL generated by kill_proc_info() in
zap_pid_ns_processes().

Use force_sig() in zap_pid_ns_processes() instead - force_sig() clears the
SIGNAL_UNKILLABLE flag ensuring the signal is processed by
container-inits.

Signed-off-by: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Cc: Oleg Nesterov <oleg@tv-sign.ru>
Cc: Roland McGrath <roland@redhat.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Daniel Lezcano <daniel.lezcano@free.fr>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agosignals: protect cinit from unblocked SIG_DFL signals
Sukadev Bhattiprolu [Thu, 2 Apr 2009 23:58:05 +0000 (16:58 -0700)]
signals: protect cinit from unblocked SIG_DFL signals

Drop early any SIG_DFL or SIG_IGN signals to container-init from within
the same container.  But queue SIGSTOP and SIGKILL to the container-init
if they are from an ancestor container.

Blocked, fatal signals (i.e when SIG_DFL is to terminate) from within the
container can still terminate the container-init.  That will be addressed
in the next patch.

Note: To be bisect-safe, SIGNAL_UNKILLABLE will be set for container-inits
    in a follow-on patch. Until then, this patch is just a preparatory
step.

Signed-off-by: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Cc: Oleg Nesterov <oleg@tv-sign.ru>
Cc: Roland McGrath <roland@redhat.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Daniel Lezcano <daniel.lezcano@free.fr>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agosignals: add from_ancestor_ns parameter to send_signal()
Sukadev Bhattiprolu [Thu, 2 Apr 2009 23:58:04 +0000 (16:58 -0700)]
signals: add from_ancestor_ns parameter to send_signal()

send_signal() (or its helper) needs to determine the pid namespace of the
sender.  But a signal sent via kill_pid_info_as_uid() comes from within
the kernel and send_signal() does not need to determine the pid namespace
of the sender.  So define a helper for send_signal() which takes an
additional parameter, 'from_ancestor_ns' and have kill_pid_info_as_uid()
use that helper directly.

The 'from_ancestor_ns' parameter will be used in a follow-on patch.

Signed-off-by: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Cc: Oleg Nesterov <oleg@tv-sign.ru>
Cc: Roland McGrath <roland@redhat.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Daniel Lezcano <daniel.lezcano@free.fr>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agosignals: protect init from unwanted signals more
Oleg Nesterov [Thu, 2 Apr 2009 23:58:02 +0000 (16:58 -0700)]
signals: protect init from unwanted signals more

(This is a modified version of the patch submitted by Oleg Nesterov
http://lkml.org/lkml/2008/11/18/249 and tries to address comments that
came up in that discussion)

init ignores the SIG_DFL signals but we queue them anyway, including
SIGKILL.  This is mostly OK, the signal will be dropped silently when
dequeued, but the pending SIGKILL has 2 bad implications:

        - it implies fatal_signal_pending(), so we confuse things
          like wait_for_completion_killable/lock_page_killable.

        - for the sub-namespace inits, the pending SIGKILL can
          mask (legacy_queue) the subsequent SIGKILL from the
          parent namespace which must kill cinit reliably.
          (preparation, cinits don't have SIGNAL_UNKILLABLE yet)

The patch can't help when init is ptraced, but ptracing of init is not
"safe" anyway.

Signed-off-by: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Acked-by: Roland McGrath <roland@redhat.com>
Signed-off-by: Oleg Nesterov <oleg@tv-sign.ru>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Daniel Lezcano <daniel.lezcano@free.fr>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agosignals: remove 'handler' parameter to tracehook functions
Oleg Nesterov [Thu, 2 Apr 2009 23:58:00 +0000 (16:58 -0700)]
signals: remove 'handler' parameter to tracehook functions

Container-init must behave like global-init to processes within the
container and hence it must be immune to unhandled fatal signals from
within the container (i.e SIG_DFL signals that terminate the process).

But the same container-init must behave like a normal process to processes
in ancestor namespaces and so if it receives the same fatal signal from a
process in ancestor namespace, the signal must be processed.

Implementing these semantics requires that send_signal() determine pid
namespace of the sender but since signals can originate from workqueues/
interrupt-handlers, determining pid namespace of sender may not always be
possible or safe.

This patchset implements the design/simplified semantics suggested by
Oleg Nesterov.  The simplified semantics for container-init are:

- container-init must never be terminated by a signal from a
  descendant process.

- container-init must never be immune to SIGKILL from an ancestor
  namespace (so a process in parent namespace must always be able
  to terminate a descendant container).

- container-init may be immune to unhandled fatal signals (like
  SIGUSR1) even if they are from ancestor namespace. SIGKILL/SIGSTOP
  are the only reliable signals to a container-init from ancestor
  namespace.

This patch:

Based on an earlier patch submitted by Oleg Nesterov and comments from
Roland McGrath (http://lkml.org/lkml/2008/11/19/258).

The handler parameter is currently unused in the tracehook functions.
Besides, the tracehook functions are called with siglock held, so the
functions can check the handler if they later need to.

Removing the parameter simiplifies changes to sig_ignored() in a follow-on
patch.

Signed-off-by: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Acked-by: Roland McGrath <roland@redhat.com>
Signed-off-by: Oleg Nesterov <oleg@tv-sign.ru>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Daniel Lezcano <daniel.lezcano@free.fr>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agodo_wait: fix waiting for the group stop with the dead leader
Oleg Nesterov [Thu, 2 Apr 2009 23:57:58 +0000 (16:57 -0700)]
do_wait: fix waiting for the group stop with the dead leader

do_wait(WSTOPPED) assumes that p->state must be == TASK_STOPPED, this is
not true if the leader is already dead.  Check SIGNAL_STOP_STOPPED instead
and use signal->group_exit_code.

Trivial test-case:

void *tfunc(void *arg)
{
pause();
return NULL;
}

int main(void)
{
pthread_t thr;
pthread_create(&thr, NULL, tfunc, NULL);
pthread_exit(NULL);
return 0;
}

It doesn't react to ^Z (and then to ^C or ^\). The task is stopped, but
bash can't see this.

The bug is very old, and it was reported multiple times. This patch was sent
more than a year ago (http://marc.info/?t=119713920000003) but it was ignored.

This change also fixes other oddities (but not all) in this area.  For
example, before this patch:

$ sleep 100
^Z
[1]+  Stopped                 sleep 100
$ strace -p `pidof sleep`
Process 11442 attached - interrupt to quit

strace hangs in do_wait(), because ->exit_code was already consumed by
bash.  After this patch, strace happily proceeds:

--- SIGTSTP (Stopped) @ 0 (0) ---
restart_syscall(<... resuming interrupted call ...>

To me, this looks much more "natural" and correct.

Another example.  Let's suppose we have the main thread M and sub-thread
T, the process is stopped, and its parent did wait(WSTOPPED).  Now we can
ptrace T but not M.  This looks at least strange to me.

Imho, do_wait() should not confuse the per-thread ptrace stops with the
per-process job control stops.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Cc: Denys Vlasenko <dvlasenk@redhat.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Jan Kratochvil <jan.kratochvil@redhat.com>
Cc: Kaz Kylheku <kkylheku@gmail.com>
Cc: Michael Kerrisk <mtk.manpages@googlemail.com>
Cc: Roland McGrath <roland@redhat.com>
Cc: Ulrich Drepper <drepper@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agocpusets: prevent PF_THREAD_BOUND tasks from attaching to non-root cpusets
David Rientjes [Thu, 2 Apr 2009 23:57:57 +0000 (16:57 -0700)]
cpusets: prevent PF_THREAD_BOUND tasks from attaching to non-root cpusets

Kthreads that have the PF_THREAD_BOUND bit set in their flags are bound to a
specific cpu.  Thus, their set of allowed cpus shall not change.

This patch prevents such threads from attaching to non-root cpusets.  They do
not have mempolicies that restrict them to a subset of system nodes and, since
their cpumask may never change, they cannot use any of the features of
cpusets.

The tasks will forever be a member of the root cpuset and will be returned
when listing the tasks attached to that cpuset.

Cc: Paul Menage <menage@google.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Dhaval Giani <dhaval@linux.vnet.ibm.com>
Signed-off-by: David Rientjes <rientjes@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agocpusets: allow cpusets to be configured/built on non-SMP systems
Paul Menage [Thu, 2 Apr 2009 23:57:55 +0000 (16:57 -0700)]
cpusets: allow cpusets to be configured/built on non-SMP systems

Allow cpusets to be configured/built on non-SMP systems

Currently it's impossible to build cpusets under UML on x86-64, since
cpusets depends on SMP and x86-64 UML doesn't support SMP.

There's code in cpusets that doesn't depend on SMP.  This patch surrounds
the minimum amount of cpusets code with #ifdef CONFIG_SMP in order to
allow cpusets to build/run on UP systems (for testing purposes under UML).

Reviewed-by: Li Zefan <lizf@cn.fujitsu.com>
Signed-off-by: Paul Menage <menage@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agocpusets: replace zone allowed functions with node allowed
David Rientjes [Thu, 2 Apr 2009 23:57:54 +0000 (16:57 -0700)]
cpusets: replace zone allowed functions with node allowed

The cpuset_zone_allowed() variants are actually only a function of the
zone's node.

Cc: Paul Menage <menage@google.com>
Acked-by: Christoph Lameter <cl@linux-foundation.org>
Cc: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: David Rientjes <rientjes@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agocpuset: remove struct cpuset_hotplug_scanner
Li Zefan [Thu, 2 Apr 2009 23:57:53 +0000 (16:57 -0700)]
cpuset: remove struct cpuset_hotplug_scanner

Use cgroup_scanner.data, instead of introducing cpuset_hotplug_scanner.

Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Paul Menage <menage@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agocpuset: avoid changing cpuset's mems when errno returned
Li Zefan [Thu, 2 Apr 2009 23:57:52 +0000 (16:57 -0700)]
cpuset: avoid changing cpuset's mems when errno returned

When writing to cpuset.mems, cpuset has to update its mems_allowed before
calling update_tasks_nodemask(), but this function might return -ENOMEM.

To avoid this rare case, we allocate the memory before changing
mems_allowed, and then pass to update_tasks_nodemask().  Similar to what
update_cpumask() does.

Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Paul Menage <menage@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agocpuset: rewrite update_tasks_nodemask()
Li Zefan [Thu, 2 Apr 2009 23:57:51 +0000 (16:57 -0700)]
cpuset: rewrite update_tasks_nodemask()

This patch uses cgroup_scan_tasks() to rebind tasks' vmas to new cpuset's
mems_allowed.

Not only simplify the code largely, but also avoid allocating an array to
hold mm pointers of all the tasks in the cpuset.  This array can be big
(size > PAGESIZE) if we have lots of tasks in that cpuset, thus has a
chance to fail the allocation when under memory stress.

Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Paul Menage <menage@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agocgroups: add 'data' field to struct cgroup_scanner
Li Zefan [Thu, 2 Apr 2009 23:57:50 +0000 (16:57 -0700)]
cgroups: add 'data' field to struct cgroup_scanner

We need to pass some data to test_task() or process_task() in some cases.
Will be used later.

Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Paul Menage <menage@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agocpuset: fix possible races in cpu/memory hotplug
Li Zefan [Thu, 2 Apr 2009 23:57:49 +0000 (16:57 -0700)]
cpuset: fix possible races in cpu/memory hotplug

Change to cpuset->cpus_allowed and cpuset->mems_allowed should be protected
by callback_mutex, otherwise the reader may read wrong cpus/mems. This is
cpuset's lock rule.

Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Cc: Paul Menage <menage@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agomemcg: cleanup cache_charge
Daisuke Nishimura [Thu, 2 Apr 2009 23:57:48 +0000 (16:57 -0700)]
memcg: cleanup cache_charge

Current mem_cgroup_cache_charge is a bit complicated especially
in the case of shmem's swap-in.

This patch cleans it up by using try_charge_swapin and commit_charge_swapin.

Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp>
Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agomemcg: remove redundant message at swapon
KAMEZAWA Hiroyuki [Thu, 2 Apr 2009 23:57:47 +0000 (16:57 -0700)]
memcg: remove redundant message at swapon

It's pointed out that swap_cgroup's message at swapon() is nonsense.
Because

  * It can be calculated very easily if all necessary information is
    written in Kconfig.

  * It's not necessary to annoying people at every swapon().

In other view, now, memory usage per swp_entry is reduced to 2bytes from
8bytes(64bit) and I think it's reasonably small.

Reported-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agocgroups: use css id in swap cgroup for saving memory v5
KAMEZAWA Hiroyuki [Thu, 2 Apr 2009 23:57:45 +0000 (16:57 -0700)]
cgroups: use css id in swap cgroup for saving memory v5

Try to use CSS ID for records in swap_cgroup.  By this, on 64bit machine,
size of swap_cgroup goes down to 2 bytes from 8bytes.

This means, when 2GB of swap is equipped, (assume the page size is 4096bytes)

From size of swap_cgroup = 2G/4k * 8 = 4Mbytes.
To   size of swap_cgroup = 2G/4k * 2 = 1Mbytes.

Reduction is large.  Of course, there are trade-offs.  This CSS ID will
add overhead to swap-in/swap-out/swap-free.

But in general,
  - swap is a resource which the user tend to avoid use.
  - If swap is never used, swap_cgroup area is not used.
  - Reading traditional manuals, size of swap should be proportional to
    size of memory. Memory size of machine is increasing now.

I think reducing size of swap_cgroup makes sense.

Note:
  - ID->CSS lookup routine has no locks, it's under RCU-Read-Side.
  - memcg can be obsolete at rmdir() but not freed while refcnt from
    swap_cgroup is available.

Changelog v4->v5:
 - reworked on to memcg-charge-swapcache-to-proper-memcg.patch
Changlog ->v4:
 - fixed not configured case.
 - deleted unnecessary comments.
 - fixed NULL pointer bug.
 - fixed message in dmesg.

[nishimura@mxp.nes.nec.co.jp: css_tryget can be called twice in !PageCgroupUsed case]
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Li Zefan <lizf@cn.fujitsu.com>
Cc: Balbir Singh <balbir@in.ibm.com>
Cc: Paul Menage <menage@google.com>
Cc: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agomemcg: charge swapcache to proper memcg
Daisuke Nishimura [Thu, 2 Apr 2009 23:57:43 +0000 (16:57 -0700)]
memcg: charge swapcache to proper memcg

memcg_test.txt says at 4.1:

This swap-in is one of the most complicated work. In do_swap_page(),
following events occur when pte is unchanged.

(1) the page (SwapCache) is looked up.
(2) lock_page()
(3) try_charge_swapin()
(4) reuse_swap_page() (may call delete_swap_cache())
(5) commit_charge_swapin()
(6) swap_free().

Considering following situation for example.

(A) The page has not been charged before (2) and reuse_swap_page()
    doesn't call delete_from_swap_cache().
(B) The page has not been charged before (2) and reuse_swap_page()
    calls delete_from_swap_cache().
(C) The page has been charged before (2) and reuse_swap_page() doesn't
    call delete_from_swap_cache().
(D) The page has been charged before (2) and reuse_swap_page() calls
    delete_from_swap_cache().

    memory.usage/memsw.usage changes to this page/swp_entry will be
 Case          (A)      (B)       (C)     (D)
         Event
       Before (2)     0/ 1     0/ 1      1/ 1    1/ 1
          ===========================================
          (3)        +1/+1    +1/+1     +1/+1   +1/+1
          (4)          -       0/ 0       -     -1/ 0
          (5)         0/-1     0/ 0     -1/-1    0/ 0
          (6)          -       0/-1       -      0/-1
          ===========================================
       Result         1/ 1     1/ 1      1/ 1    1/ 1

       In any cases, charges to this page should be 1/ 1.

In case of (D), mem_cgroup_try_get_from_swapcache() returns NULL
(because lookup_swap_cgroup() returns NULL), so "+1/+1" at (3) means
charges to the memcg("foo") to which the "current" belongs.
OTOH, "-1/0" at (4) and "0/-1" at (6) means uncharges from the memcg("baa")
to which the page has been charged.

So, if the "foo" and "baa" is different(for example because of task move),
this charge will be moved from "baa" to "foo".

I think this is an unexpected behavior.

This patch fixes this by modifying mem_cgroup_try_get_from_swapcache()
to return the memcg to which the swapcache has been charged if PCG_USED bit
is set.
IIUC, checking PCG_USED bit of swapcache is safe under page lock.

Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp>
Cc: Balbir Singh <balbir@linux.vnet.ibm.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Li Zefan <lizf@cn.fujitsu.com>
Cc: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agomemcg: remove mem_cgroup_reclaim_imbalance() remnants
KOSAKI Motohiro [Thu, 2 Apr 2009 23:57:41 +0000 (16:57 -0700)]
memcg: remove mem_cgroup_reclaim_imbalance() remnants

commit 4f98a2fee8acdb4ac84545df98cccecfd130f8db (vmscan: split LRU lists
into anon & file sets) removed mem_cgroup_reclaim_imbalance(), but there
are some leftovers in memcontrol.h.

Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Balbir Singh <balbir@linux.vnet.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agomemcg: remove mem_cgroup_calc_mapped_ratio()
KOSAKI Motohiro [Thu, 2 Apr 2009 23:57:40 +0000 (16:57 -0700)]
memcg: remove mem_cgroup_calc_mapped_ratio()

Currently, mem_cgroup_calc_mapped_ratio() is unused at all.  it can be
removed and KAMEZAWA-san suggested it.

Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Acked-by: Balbir Singh <balbir@linux.vnet.ibm.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agomemcg: show memcg information during OOM
Balbir Singh [Thu, 2 Apr 2009 23:57:39 +0000 (16:57 -0700)]
memcg: show memcg information during OOM

Add RSS and swap to OOM output from memcg

Display memcg values like failcnt, usage and limit when an OOM occurs due
to memcg.

Thanks to Johannes Weiner, Li Zefan, David Rientjes, Kamezawa Hiroyuki,
Daisuke Nishimura and KOSAKI Motohiro for review.

Sample output
-------------

Task in /a/x killed as a result of limit of /a
memory: usage 1048576kB, limit 1048576kB, failcnt 4183
memory+swap: usage 1400964kB, limit 9007199254740991kB, failcnt 0

[akpm@linux-foundation.org: compilation fix]
[akpm@linux-foundation.org: fix kerneldoc and whitespace]
[akpm@linux-foundation.org: add printk facility level]
Signed-off-by: Balbir Singh <balbir@linux.vnet.ibm.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp>
Cc: Li Zefan <lizf@cn.fujitsu.com>
Cc: Paul Menage <menage@google.com>
Cc: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agomemcg: fix OOM killer under memcg
KAMEZAWA Hiroyuki [Thu, 2 Apr 2009 23:57:38 +0000 (16:57 -0700)]
memcg: fix OOM killer under memcg

This patch tries to fix OOM Killer problems caused by hierarchy.
Now, memcg itself has OOM KILL function (in oom_kill.c) and tries to
kill a task in memcg.

But, when hierarchy is used, it's broken and correct task cannot
be killed. For example, in following cgroup

/groupA/ hierarchy=1, limit=1G,
01 nolimit
02 nolimit
All tasks' memory usage under /groupA, /groupA/01, groupA/02 is limited to
groupA's 1Gbytes but OOM Killer just kills tasks in groupA.

This patch provides makes the bad process be selected from all tasks
under hierarchy. BTW, currently, oom_jiffies is updated against groupA
in above case. oom_jiffies of tree should be updated.

To see how oom_jiffies is used, please check mem_cgroup_oom_called()
callers.

[akpm@linux-foundation.org: build fix]
[akpm@linux-foundation.org: const fix]
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Paul Menage <menage@google.com>
Cc: Li Zefan <lizf@cn.fujitsu.com>
Cc: Balbir Singh <balbir@in.ibm.com>
Cc: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp>
Cc: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agomemcg: fix shrinking memory to return -EBUSY by fixing retry algorithm
KAMEZAWA Hiroyuki [Thu, 2 Apr 2009 23:57:36 +0000 (16:57 -0700)]
memcg: fix shrinking memory to return -EBUSY by fixing retry algorithm

As pointed out, shrinking memcg's limit should return -EBUSY after
reasonable retries.  This patch tries to fix the current behavior of
shrink_usage.

Before looking into "shrink should return -EBUSY" problem, we should fix
hierarchical reclaim code.  It compares current usage and current limit,
but it only makes sense when the kernel reclaims memory because hit
limits.  This is also a problem.

What this patch does are.

  1. add new argument "shrink" to hierarchical reclaim. If "shrink==true",
     hierarchical reclaim returns immediately and the caller checks the kernel
     should shrink more or not.
     (At shrinking memory, usage is always smaller than limit. So check for
      usage < limit is useless.)

  2. For adjusting to above change, 2 changes in "shrink"'s retry path.
     2-a. retry_count depends on # of children because the kernel visits
  the children under hierarchy one by one.
     2-b. rather than checking return value of hierarchical_reclaim's progress,
  compares usage-before-shrink and usage-after-shrink.
  If usage-before-shrink <= usage-after-shrink, retry_count is
  decremented.

Reported-by: Li Zefan <lizf@cn.fujitsu.com>
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Paul Menage <menage@google.com>
Cc: Balbir Singh <balbir@in.ibm.com>
Cc: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp>
Cc: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agomemcg: hierarchical stat
KAMEZAWA Hiroyuki [Thu, 2 Apr 2009 23:57:35 +0000 (16:57 -0700)]
memcg: hierarchical stat

Clean up memory.stat file routine and show "total" hierarchical stat.

This patch does
  - renamed get_all_zonestat to be get_local_zonestat.
  - remove old mem_cgroup_stat_desc, which is only for per-cpu stat.
  - add mcs_stat to cover both of per-cpu/per-lru stat.
  - add "total" stat of hierarchy (*)
  - add a callback system to scan all memcg under a root.
== "total" is added.
[kamezawa@localhost ~]$ cat /opt/cgroup/xxx/memory.stat
cache 0
rss 0
pgpgin 0
pgpgout 0
inactive_anon 0
active_anon 0
inactive_file 0
active_file 0
unevictable 0
hierarchical_memory_limit 50331648
hierarchical_memsw_limit 9223372036854775807
total_cache 65536
total_rss 192512
total_pgpgin 218
total_pgpgout 155
total_inactive_anon 0
total_active_anon 135168
total_inactive_file 61440
total_active_file 4096
total_unevictable 0
==
(*) maybe the user can do calc hierarchical stat by his own program
   in userland but if it can be written in clean way, it's worth to be
   shown, I think.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Paul Menage <menage@google.com>
Cc: Li Zefan <lizf@cn.fujitsu.com>
Cc: Balbir Singh <balbir@in.ibm.com>
Cc: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp>
Cc: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agomemcg: use CSS ID
KAMEZAWA Hiroyuki [Thu, 2 Apr 2009 23:57:33 +0000 (16:57 -0700)]
memcg: use CSS ID

Assigning CSS ID for each memcg and use css_get_next() for scanning hierarchy.

Assume folloing tree.

group_A (ID=3)
/01 (ID=4)
   /0A (ID=7)
/02 (ID=10)
group_B (ID=5)
and task in group_A/01/0A hits limit at group_A.

reclaim will be done in following order (round-robin).
group_A(3) -> group_A/01 (4) -> group_A/01/0A (7) -> group_A/02(10)
-> group_A -> .....

Round robin by ID. The last visited cgroup is recorded and restart
from it when it start reclaim again.
(More smart algorithm can be implemented..)

No cgroup_mutex or hierarchy_mutex is required.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Paul Menage <menage@google.com>
Cc: Li Zefan <lizf@cn.fujitsu.com>
Cc: Balbir Singh <balbir@in.ibm.com>
Cc: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp>
Cc: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agodevcgroup: avoid using cgroup_lock
Li Zefan [Thu, 2 Apr 2009 23:57:32 +0000 (16:57 -0700)]
devcgroup: avoid using cgroup_lock

There is nothing special that has to be protected by cgroup_lock,
so introduce devcgroup_mtuex for it's own use.

Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Cc: Paul Menage <menage@google.com>
Acked-by: Serge Hallyn <serue@us.ibm.com>
Cc: Balbir Singh <balbir@in.ibm.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agodebug cgroup: remove unneeded cgroup_lock
Li Zefan [Thu, 2 Apr 2009 23:57:31 +0000 (16:57 -0700)]
debug cgroup: remove unneeded cgroup_lock

Since we are in cgroup write handler, so the cgrp is valid, so we don't
have to hold cgroup_mutex when calling cgroup_task_count().  One similar
example is in cgroup_tasks_open().

Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Acked-by: Paul Menage <menage@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agocgroups: don't change release_agent when remount failed
Li Zefan [Thu, 2 Apr 2009 23:57:30 +0000 (16:57 -0700)]
cgroups: don't change release_agent when remount failed

Remount can fail in either case:
  - wrong mount options is specified, or option 'noprefix' is changed.
  - a to-be-added subsys is already mounted/active.

When using remount to change 'release_agent', for the above former failure
case, remount will return errno with release_agent unchanged, but for the
latter case, remount will return EBUSY with relase_agent changed, which is
unexpected I think:

 # mount -t cgroup -o cpu xxx /cgrp1
 # mount -t cgroup -o cpuset,release_agent=agent1 yyy /cgrp2
 # cat /cgrp2/release_agent
 agent1
 # mount -t cgroup -o remount,cpuset,noprefix,release_agent=agent2 yyy /cgrp2
 mount: /cgrp2 not mounted already, or bad option
 # cat /cgrp2/release_agent
 agent1     <-- ok
 # mount -t cgroup -o remount,cpu,cpuset,release_agent=agent2 yyy /cgrp2
 mount: /cgrp2 is busy
 # cat /cgrp2/release_agent
 agent2     <-- unexpected!

Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Cc: Paul Menage <menage@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agocgroups: show correct file mode
Li Zefan [Thu, 2 Apr 2009 23:57:29 +0000 (16:57 -0700)]
cgroups: show correct file mode

We have some read-only files and write-only files, but currently they are
all set to 0644, which is counter-intuitive and cause trouble for some
cgroup tools like libcgroup.

This patch adds 'mode' to struct cftype to allow cgroup subsys to set it's
own files' file mode, and for the most cases cft->mode can be default to 0
and cgroup will figure out proper mode.

Acked-by: Paul Menage <menage@google.com>
Reviewed-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agocgroups: more documentation for remount and release_agent
Li Zefan [Thu, 2 Apr 2009 23:57:28 +0000 (16:57 -0700)]
cgroups: more documentation for remount and release_agent

This won't remove cpuacct from the mounted hierachy:
 # mount -t cgroup -o cpu,cpuacct xxx /mnt
 # mount -o remount,cpu /mnt

Because for this usage mount(8) will append the new options to the original
options.

And this will get you right:
 # mount [-t cgroup] -o remount,cpu xxx /mnt

Also document how to specify or change release_agent.

Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Reviewd-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Paul Menage <menage@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agokernel/cgroup.c: kfree(NULL) is legal
Jesper Juhl [Thu, 2 Apr 2009 23:57:27 +0000 (16:57 -0700)]
kernel/cgroup.c: kfree(NULL) is legal

Reduces object file size a bit:

Before:
$ size kernel/cgroup.o
   text    data     bss     dec     hex filename
  21593    7804    4924   34321    8611 kernel/cgroup.o
After:
$ size kernel/cgroup.o
   text    data     bss     dec     hex filename
  21537    7744    4924   34205    859d kernel/cgroup.o

Signed-off-by: Jesper Juhl <jj@chaosbits.net>
Cc: Paul Menage <menage@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agocgroup: fix frequent -EBUSY at rmdir
KAMEZAWA Hiroyuki [Thu, 2 Apr 2009 23:57:26 +0000 (16:57 -0700)]
cgroup: fix frequent -EBUSY at rmdir

In following situation, with memory subsystem,

/groupA use_hierarchy==1
/01 some tasks
/02 some tasks
/03 some tasks
/04 empty

When tasks under 01/02/03 hit limit on /groupA, hierarchical reclaim
is triggered and the kernel walks tree under groupA. In this case,
rmdir /groupA/04 fails with -EBUSY frequently because of temporal
refcnt from the kernel.

In general. cgroup can be rmdir'd if there are no children groups and
no tasks. Frequent fails of rmdir() is not useful to users.
(And the reason for -EBUSY is unknown to users.....in most cases)

This patch tries to modify above behavior, by
- retries if css_refcnt is got by someone.
- add "return value" to pre_destroy() and allows subsystem to
  say "we're really busy!"

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Paul Menage <menage@google.com>
Cc: Li Zefan <lizf@cn.fujitsu.com>
Cc: Balbir Singh <balbir@in.ibm.com>
Cc: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agocgroup: CSS ID support
KAMEZAWA Hiroyuki [Thu, 2 Apr 2009 23:57:25 +0000 (16:57 -0700)]
cgroup: CSS ID support

Patch for Per-CSS(Cgroup Subsys State) ID and private hierarchy code.

This patch attaches unique ID to each css and provides following.

 - css_lookup(subsys, id)
   returns pointer to struct cgroup_subysys_state of id.
 - css_get_next(subsys, id, rootid, depth, foundid)
   returns the next css under "root" by scanning

When cgroup_subsys->use_id is set, an id for css is maintained.

The cgroup framework only parepares
- css_id of root css for subsys
- id is automatically attached at creation of css.
- id is *not* freed automatically. Because the cgroup framework
  don't know lifetime of cgroup_subsys_state.
  free_css_id() function is provided. This must be called by subsys.

There are several reasons to develop this.
- Saving space .... For example, memcg's swap_cgroup is array of
  pointers to cgroup. But it is not necessary to be very fast.
  By replacing pointers(8bytes per ent) to ID (2byes per ent), we can
  reduce much amount of memory usage.

- Scanning without lock.
  CSS_ID provides "scan id under this ROOT" function. By this, scanning
  css under root can be written without locks.
  ex)
  do {
rcu_read_lock();
next = cgroup_get_next(subsys, id, root, &found);
/* check sanity of next here */
css_tryget();
rcu_read_unlock();
id = found + 1
 } while(...)

Characteristics:
- Each css has unique ID under subsys.
- Lifetime of ID is controlled by subsys.
- css ID contains "ID" and "Depth in hierarchy" and stack of hierarchy
- Allowed ID is 1-65535, ID 0 is UNUSED ID.

Design Choices:
- scan-by-ID v.s. scan-by-tree-walk.
  As /proc's pid scan does, scan-by-ID is robust when scanning is done
  by following kind of routine.
  scan -> rest a while(release a lock) -> conitunue from interrupted
  memcg's hierarchical reclaim does this.

- When subsys->use_id is set, # of css in the system is limited to
  65535.

[bharata@linux.vnet.ibm.com: remove rcu_read_lock() from css_get_next()]
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Acked-by: Paul Menage <menage@google.com>
Cc: Li Zefan <lizf@cn.fujitsu.com>
Cc: Balbir Singh <balbir@in.ibm.com>
Cc: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp>
Signed-off-by: Bharata B Rao <bharata@linux.vnet.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agocgroups: relax ns_can_attach checks to allow attaching to grandchild cgroups
Grzegorz Nosek [Thu, 2 Apr 2009 23:57:23 +0000 (16:57 -0700)]
cgroups: relax ns_can_attach checks to allow attaching to grandchild cgroups

The ns_proxy cgroup allows moving processes to child cgroups only one
level deep at a time.  This commit relaxes this restriction and makes it
possible to attach tasks directly to grandchild cgroups, e.g.:

($pid is in the root cgroup)
echo $pid > /cgroup/CG1/CG2/tasks

Previously this operation would fail with -EPERM and would have to be
performed as two steps:
echo $pid > /cgroup/CG1/tasks
echo $pid > /cgroup/CG1/CG2/tasks

Also, the target cgroup no longer needs to be empty to move a task there.

Signed-off-by: Grzegorz Nosek <root@localdomain.pl>
Acked-by: Serge Hallyn <serue@us.ibm.com>
Reviewed-by: Li Zefan <lizf@cn.fujitsu.com>
Cc: Paul Menage <menage@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agocgroups: fix cgroup.h comments
Paul Menage [Thu, 2 Apr 2009 23:57:22 +0000 (16:57 -0700)]
cgroups: fix cgroup.h comments

Fix the style of some multi-line comments in cgroup.h to match
Documentation/CodingStyle

Signed-off-by: Paul Menage <menage@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agodocumentation: fix unix_dgram_qlen description
Li Xiaodong [Thu, 2 Apr 2009 23:57:21 +0000 (16:57 -0700)]
documentation: fix unix_dgram_qlen description

Previous description about system parameter in /proc/sys/net/unix/ is
wrong (or missed).  Simply add a new description about unix_dgram_qlen
according to latest kernel.

Signed-off-by: Li Xiaodong <lixd@cn.fujitsu.com>
Cc: "David S. Miller" <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agodocumentation: update Documentation/filesystem/proc.txt and Documentation/sysctls
Shen Feng [Thu, 2 Apr 2009 23:57:20 +0000 (16:57 -0700)]
documentation: update Documentation/filesystem/proc.txt and Documentation/sysctls

Now /proc/sys is described in many places and much information is
redundant.  This patch updates the proc.txt and move the /proc/sys
desciption out to the files in Documentation/sysctls.

Details are:

merge
-  2.1  /proc/sys/fs - File system data
-  2.11 /proc/sys/fs/mqueue - POSIX message queues filesystem
-  2.17 /proc/sys/fs/epoll - Configuration options for the epoll interface
with Documentation/sysctls/fs.txt.

remove
-  2.2  /proc/sys/fs/binfmt_misc - Miscellaneous binary formats
since it's not better then the Documentation/binfmt_misc.txt.

merge
-  2.3  /proc/sys/kernel - general kernel parameters
with Documentation/sysctls/kernel.txt

remove
-  2.5  /proc/sys/dev - Device specific parameters
since it's obsolete the sysfs is used now.

remove
-  2.6  /proc/sys/sunrpc - Remote procedure calls
since it's not better then the Documentation/sysctls/sunrpc.txt

move
-  2.7  /proc/sys/net - Networking stuff
-  2.9  Appletalk
-  2.10 IPX
to newly created Documentation/sysctls/net.txt.

remove
-  2.8  /proc/sys/net/ipv4 - IPV4 settings
since it's not better then the Documentation/networking/ip-sysctl.txt.

add
- Chapter 3 Per-Process Parameters
to descibe /proc/<pid>/xxx parameters.

Signed-off-by: Shen Feng <shen@cn.fujitsu.com>
Cc: Randy Dunlap <randy.dunlap@oracle.com>
Cc: "David S. Miller" <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agodocumentation: ignore byproducts from latex
Henrik Austad [Thu, 2 Apr 2009 23:57:18 +0000 (16:57 -0700)]
documentation: ignore byproducts from latex

When using 'make pdfdocs', auto-generated files should be ignored

Signed-off-by: Henrik Austad <henrik@austad.us>
Cc: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agohppfs: hppfs_read_file() may return -ERROR
Roel Kluin [Thu, 2 Apr 2009 23:57:18 +0000 (16:57 -0700)]
hppfs: hppfs_read_file() may return -ERROR

hppfs_read_file() may return (ssize_t) -ENOMEM, or -EFAULT.  When stored
in size_t 'count', these errors will not be noticed, a large value will be
added to *ppos.

Signed-off-by: Roel Kluin <roel.kluin@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoext3: avoid false EIO errors
Jan Kara [Thu, 2 Apr 2009 23:57:17 +0000 (16:57 -0700)]
ext3: avoid false EIO errors

Sometimes block_write_begin() can map buffers in a page but later we
fail to copy data into those buffers (because the source page has been
paged out in the mean time).  We then end up with !uptodate mapped
buffers.  To add a bit more to the confusion, block_write_end() does
not commit any data (and thus does not any mark buffers as uptodate) if
we didn't succeed with copying all the data.

Commit f4fc66a894546bdc88a775d0e83ad20a65210bcb (ext3: convert to new
aops) missed these cases and thus we were inserting non-uptodate
buffers to transaction's list which confuses JBD code and it reports IO
errors, aborts a transaction and generally makes users afraid about
their data ;-P.

This patch fixes the problem by reorganizing ext3_..._write_end() code
to first call block_write_end() to mark buffers with valid data
uptodate and after that we file only uptodate buffers to transaction's
lists.

We also fix a problem where we could leave blocks allocated beyond i_size
(i_disksize in fact) because of failed write. We now add inode to orphan
list when write fails (to be safe in case we crash) and then truncate blocks
beyond i_size in a separate transaction.

Signed-off-by: Jan Kara <jack@suse.cz>
Reviewed-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Cc: Nick Piggin <nickpiggin@yahoo.com.au>
Cc: <linux-ext4@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoext3: return -EIO not -ESTALE on directory traversal through deleted inode
Bryan Donlan [Thu, 2 Apr 2009 23:57:15 +0000 (16:57 -0700)]
ext3: return -EIO not -ESTALE on directory traversal through deleted inode

ext3_iget() returns -ESTALE if invoked on a deleted inode, in order to
report errors to NFS properly.  However, in ext[234]_lookup(), this
-ESTALE can be propagated to userspace if the filesystem is corrupted such
that a directory entry references a deleted inode.  This leads to a
misleading error message - "Stale NFS file handle" - and confusion on the
part of the admin.

The bug can be easily reproduced by creating a new filesystem, making a
link to an unused inode using debugfs, then mounting and attempting to ls
-l said link.

This patch thus changes ext3_lookup to return -EIO if it receives -ESTALE
from ext3_iget(), as ext3 does for other filesystem metadata corruption;
and also invokes the appropriate ext*_error functions when this case is
detected.

Signed-off-by: Bryan Donlan <bdonlan@gmail.com>
Cc: <linux-ext4@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoext3: use unsigned instead of int for type of blocksize in fs/ext3/namei.c
Wei Yongjun [Thu, 2 Apr 2009 23:57:14 +0000 (16:57 -0700)]
ext3: use unsigned instead of int for type of blocksize in fs/ext3/namei.c

Use unsigned instead of int for the parameter which carries a blocksize.

[akpm@linux-foundation.org: coding-style fixes]
Signed-off-by: Wei Yongjun <yjwei@cn.fujitsu.com>
Cc: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>