summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* Merge tag 'for_linus-3.4-rc1' of ↵Linus Torvalds2012-03-2310-29/+161
|\ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | git://git.kernel.org/pub/scm/linux/kernel/git/jwessel/kgdb Pull KGDB/KDB updates from Jason Wessel: "Fixes: - Fix KDB keyboard repeat scan codes and leaked keyboard events - Fix kernel crash with kdb_printf() for users who compile new kdb_printf()'s in early code - Return all segment registers to gdb on x86_64 Features: - KDB/KGDB hook the reboot notifier and end user can control if it stops, detaches or does nothing (updated docs as well) - Notify users who use CONFIG_DEBUG_RODATA to use hw breakpoints" * tag 'for_linus-3.4-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/jwessel/kgdb: kdb: Add message about CONFIG_DEBUG_RODATA on failure to install breakpoint kdb: Avoid using dbg_io_ops until it is initialized kgdb,debug_core: add the ability to control the reboot notifier KDB: Fix usability issues relating to the 'enter' key. kgdb,debug-core,gdbstub: Hook the reboot notifier for debugger detach kgdb: Respect that flush op is optional kgdb: x86: Return all segment registers also in 64-bit mode
| * kdb: Add message about CONFIG_DEBUG_RODATA on failure to install breakpointJason Wessel2012-03-221-0/+7
| | | | | | | | | | | | | | | | | | | | | | | | | | On x86, if CONFIG_DEBUG_RODATA is set, one cannot set breakpoints via KDB. Apparently this is a well-known problem, as at least one distribution now ships with both KDB enabled and CONFIG_DEBUG_RODATA=y for security reasons. This patch adds an printk message to the breakpoint failure case, in order to provide suggestions about how to use the debugger. Reported-by: Tim Bird <tim.bird@am.sony.com> Signed-off-by: Jason Wessel <jason.wessel@windriver.com> Acked-by: Tim Bird <tim.bird@am.sony.com>
| * kdb: Avoid using dbg_io_ops until it is initializedTim Bird2012-03-221-1/+1
| | | | | | | | | | | | | | | | | | | | | | This fixes a bug with setting a breakpoint during kdb initialization (from kdb_cmds). Any call to kdb_printf() before the initialization of the kgdboc serial console driver (which happens much later during bootup than kdb_init), results in kernel panic due to the use of dbg_io_ops before it is initialized. Signed-off-by: Tim Bird <tim.bird@am.sony.com> Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
| * kgdb,debug_core: add the ability to control the reboot notifierJason Wessel2012-03-222-0/+33
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Sometimes it is desirable to stop the kernel debugger before allowing a system to reboot either with kdb or kgdb. This patch adds the ability to turn the reboot notifier on and off or enter the debugger and stop kernel execution before rebooting. It is possible to change the setting after booting the kernel with the following: echo 1 > /sys/module/debug_core/parameters/kgdbreboot It is also possible to change this setting using kdb / kgdb to manipulate the variable directly. Using KDB: mm kgdbreboot 1 Using gdb: set kgdbreboot=1 Reported-by: Jan Kiszka <jan.kiszka@siemens.com> Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
| * KDB: Fix usability issues relating to the 'enter' key.Andrei Warkentin2012-03-223-22/+83
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This fixes the following problems: 1) Typematic-repeat of 'enter' gives warning message and leaks make/break if KDB exits. Repeats look something like 0x1c 0x1c .... 0x9c 2) Use of 'keypad enter' gives warning message and leaks the ENTER break/make code out if KDB exits. KP ENTER repeats look someting like 0xe0 0x1c 0xe0 0x1c ... 0xe0 0x9c. 3) Lag on the order of seconds between "break" and "make" when expecting the enter "break" code. Seen under virtualized environments such as VMware ESX. The existing special enter handler tries to glob the enter break code, but this fails if the other (KP) enter was used, or if there was a key repeat. It also fails if you mashed some keys along with enter, and you ended up with a non-enter make or non-enter break code coming after the enter make code. So first, we modify the handler to handle these cases. But performing these actions on every enter is annoying since now you can't hold ENTER down to scroll <more>d messages in KDB. Since this special behaviour is only necessary to handle the exiting KDB ('g' + ENTER) without leaking scancodes to the OS. This cleanup needs to get executed anytime the kdb_main loop exits. Tested on QEMU. Set a bp on atkbd.c to verify no scan code was leaked. Cc: Andrei Warkentin <andreiw@vmware.com> [jason.wessel@windriver.com: move cleanup calls to kdb_main.c] Signed-off-by: Andrei Warkentin <andrey.warkentin@gmail.com> Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
| * kgdb,debug-core,gdbstub: Hook the reboot notifier for debugger detachJason Wessel2012-03-222-0/+24
| | | | | | | | | | | | | | | | | | | | | | | | The gdbstub and kdb should get detached if the system is rebooting. Calling gdbstub_exit() will set the proper debug core state and send a message to any debugger that is connected to correctly detach. An attached debugger will receive the exit code from include/linux/reboot.h based on SYS_HALT, SYS_REBOOT, etc... Reported-by: Jan Kiszka <jan.kiszka@siemens.com> Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
| * kgdb: Respect that flush op is optionalJan Kiszka2012-03-221-1/+2
| | | | | | | | | | | | | | | | Not all kgdb I/O drivers implement a flush operation. Adjust gdbstub_exit accordingly. Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com> Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
| * kgdb: x86: Return all segment registers also in 64-bit modeJan Kiszka2012-03-222-5/+11
| | | | | | | | | | | | | | | | | | | | Even if the content is always 0, gdb expects us to return also ds, es, fs, and gs while in x86-64 mode. Do this to avoid ugly errors on "info registers". [jason.wessel@windriver.com: adjust NUMREGBYTES for two new regs] Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com> Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
* | Merge tag 'ktest-v3.4' of ↵Linus Torvalds2012-03-232-14/+56
|\ \ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-ktest Pull ktest changes from Steven Rostedt. * tag 'ktest-v3.4' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-ktest: ktest: Allow a test to override REBOOT_ON_SUCCESS ktest: Fix SWITCH_TO_GOOD to also reboot the machine ktest: Add SCP_TO_TARGET_INSTALL option ktest: Add warning when bugs are ignored ktest: Add INSTALL_MOD_STRIP=1 when installing modules
| * | ktest: Allow a test to override REBOOT_ON_SUCCESSSteven Rostedt2012-03-211-2/+10
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The option REBOOT_ON_SUCCESS is global, and will have the machine reboot the the box if all tests are successful. But a test may not want the machine to reboot, and perhaps have the kernel it loaded be used to install the next kernel. Or the last test may set up a kernel that the user may want to look at. In this case, the user could have the global option REBOOT_ON_SUCCESS be true, but if a test is defined to run at the end, that test can override the global option and keep the kernel it installed for the user to log in with. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
| * | ktest: Fix SWITCH_TO_GOOD to also reboot the machineSteven Rostedt2012-03-211-1/+0
| | | | | | | | | | | | | | | | | | | | | | | | | | | When the option SWITCH_TO_GOOD is set, it will be called when the system needs to reboot to the good server. But currently, this keeps the reboot from happening. The SWITCH_TO_GOOD is just a way to get to a new kernel, it may not mean to not reboot. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
| * | ktest: Add SCP_TO_TARGET_INSTALL optionSteven Rostedt2012-03-212-7/+33
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Currently the option used to scp both the modules to the target as well as the kernel image are the same (SCP_TO_TARGET). But some embedded boards may require them to be different. The modules may need to be put directly on the board, but the kernel image may need to go to a tftpserver. Add the option SCP_TO_TARGET_INSTALL that will allow the user to change the config so that they may have the modules and image got to different machines. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
| * | ktest: Add warning when bugs are ignoredSteven Rostedt2012-03-211-3/+12
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | When IGNORE_ERRORS is set, ktest will not fail a test if a backtrace is detected. But this can be an issue if the user added it in the config but forgot to remove it. They may be left wondering why their test did not fail, or even worse, why their bisect gave the wrong commit. Add a warning in the output if IGNORE_WARNINGS is set, and ktest detects a kernel error. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
| * | ktest: Add INSTALL_MOD_STRIP=1 when installing modulesSteven Rostedt2012-03-211-1/+1
| | | | | | | | | | | | | | | | | | | | | To keep the modules from bloating the target's filesystem strip them during the install. Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
* | | Merge tag 'upstream-3.4-rc1' of git://git.infradead.org/linux-ubifsLinus Torvalds2012-03-236-230/+234
|\ \ \ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Pull UBIFS changes from Artem Bityutskiy: - Improve error messages - Clean-up i_nlink management - Minor clean-ups * tag 'upstream-3.4-rc1' of git://git.infradead.org/linux-ubifs: UBIFS: improve error messages UBIFS: kill CUR_MAX_KEY_LEN macro UBIFS: do not use inc_link when i_nlink is zero UBIFS: make the dbg_lock spinlock static UBIFS: increase dumps loglevel UBIFS: amend recovery debugging message
| * | | UBIFS: improve error messagesArtem Bityutskiy2012-03-071-5/+14
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Ricard complaints that the following error message is odd: "UBIFS error (pid 1578): validate_sb: bad superblock, error 8" and he is right. This patch improves the error messages a bit and makes them more user-friendly. Reported-by: Ricard Wanderlof <ricard.wanderlof@axis.com> Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
| * | | UBIFS: kill CUR_MAX_KEY_LEN macroArtem Bityutskiy2012-02-291-7/+4
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | It is useless and confusing and may make people believe they may just change it, which is not true, because this will also change the on-flash format. Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
| * | | UBIFS: do not use inc_link when i_nlink is zeroArtem Bityutskiy2012-02-291-9/+9
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This patch changes the 'i_nlink' counter handling in 'ubifs_unlink()', 'ubifs_rmdir()' and 'ubifs_rename()'. In these function 'i_nlink' may become 0, and if 'ubifs_jnl_update()' failed, we would use 'inc_nlink()' to restore the previous 'i_nlink' value, which is incorrect from the VFS point of view and would cause a 'WARN_ON()' (see 'inc_nlink() implementation). This patches saves the previous 'i_nlink' value in a local variable and uses it at the error path instead of calling 'inc_nlink()'. We do this only for the inodes where 'i_nlink' may potentially become zero. This change has been requested by Al Viro <viro@ZenIV.linux.org.uk>. Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
| * | | UBIFS: make the dbg_lock spinlock staticArtem Bityutskiy2012-02-292-4/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | Remove the usage of the 'dbg_lock' spinlock from 'dbg_err()' and make it static. Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
| * | | UBIFS: increase dumps loglevelArtem Bityutskiy2012-02-291-204/+204
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Most of the time we use the dumping function to dump something in case of error. We use 'KERN_DEBUG' printk level, and the drawback is that users do not see them in the console, while they see the other error messages in the console. The result is that they send bug reports which does not contain a lot of useful information. This patch changes the printk level of the dump functions to 'KERN_ERR' to correct the situation. I documented it in the MTD web site that people have to send the 'dmesg' output when submitting bug reposts - it did not help. Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
| * | | UBIFS: amend recovery debugging messageArtem Bityutskiy2012-02-291-1/+2
| | | | | | | | | | | | | | | | | | | | | | | | Print LEB and offset as well. Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
* | | | Merge tag 'upstream-3.4-rc1' of git://git.infradead.org/linux-ubiLinus Torvalds2012-03-236-58/+49
|\ \ \ \ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Pull UBI changes from Artem Bityutskiy: - Reduce memory consumption - Fix picking unknown blocks - Fix error-path in 'ubi_scan()' - Minor clean-ups * tag 'upstream-3.4-rc1' of git://git.infradead.org/linux-ubi: UBI: rename MOVE_CANCEL_BITFLIPS to MOVE_TARGET_BITFLIPS UBI: rename peb_buf1 to peb_buf UBI: reduce memory consumption UBI: fix eraseblock picking criteria UBI: fix documentation and improve readability UBI: fix error handling in ubi_scan()
| * | | | UBI: rename MOVE_CANCEL_BITFLIPS to MOVE_TARGET_BITFLIPSArtem Bityutskiy2012-03-093-6/+6
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | While looking at a problem reported by UBI around the PEB moving area I noticed that the 'MOVE_CANCEL_BITFLIPS' is a bit inconsistent name and 'MOVE_TARGET_BITFLIPS' better - let's rename it. Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
| * | | | UBI: rename peb_buf1 to peb_bufArtem Bityutskiy2012-03-095-30/+30
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Now we have only one buffer so let's rename it to just 'peb_buf1'. Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
| * | | | UBI: reduce memory consumptionJosselin Costanzi2012-03-093-12/+4
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Remove the pre-allocated 'peb_buf2' buffer because we do not really need it. The only reason UBI has it is to check that the data were written correctly. But we do not have to have 2 buffers for this and waste RAM - we can just compare CRC checksums instead. This reduces UBI memory consumption. Artem bityutskiy: massaged the patch and commit message Signed-off-by: Josselin Costanzi <josselin.costanzi@mobile-devices.fr> Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
| * | | | UBI: fix eraseblock picking criteriaArtem Bityutskiy2012-03-091-6/+4
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The 'find_wl_entry()' function expects the maximum difference as the second argument, not the maximum absolute value. So the "unknown" eraseblock picking was incorrect, as Shmulik Ladkani spotted. This patch fixes the issue. Reported-by: Shmulik Ladkani <shmulik.ladkani@gmail.com> Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Reviewed-by: Shmulik Ladkani <shmulik.ladkani@gmail.com> Cc: stable@kernel.org
| * | | | UBI: fix documentation and improve readabilityArtem Bityutskiy2012-03-091-4/+5
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The "max" parameter of 'find_wl_entry()' was documented incorrectly and it actually means the maximum possible difference from the smallest erase counter. Rename it to "diff" instead, and amend the documentation. Reported-by: Shmulik Ladkani <shmulik.ladkani@gmail.com> Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Reviewed-by: Shmulik Ladkani <shmulik.ladkani@gmail.com>
| * | | | UBI: fix error handling in ubi_scan()Richard Weinberger2012-02-291-4/+4
| |/ / / | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Two bad things can happen in ubi_scan(): 1. If kmem_cache_create() fails we jump to out_si and call ubi_scan_destroy_si() which calls kmem_cache_destroy(). But si->scan_leb_slab is NULL. 2. If process_eb() fails we jump to out_vidh, call kmem_cache_destroy() and ubi_scan_destroy_si() which calls again kmem_cache_destroy(). Signed-off-by: Richard Weinberger <richard@nod.at> Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Cc: stable@kernel.org
* | | | Merge tag 'pstore-for-linus' of ↵Linus Torvalds2012-03-231-8/+22
|\ \ \ \ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux Pull one pstore patch from Tony Luck * tag 'pstore-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux: pstore: Introduce get_reason_str() to pstore
| * | | | pstore: Introduce get_reason_str() to pstoreSeiji Aguchi2012-03-161-8/+22
| | |/ / | |/| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Recently, there has been some changes in kmsg_dump() below and they have been applied to linus-tree. (1) kmsg_dump(KMSG_DUMP_KEXEC) was removed. http://git.kernel.org/?p=linux/kernel/git/torvalds/linux.git;a=commitdiff;h=a3dd3323058d281abd584b15ad4c5b65064d7a61 (2) The order of "enum kmsg_dump_reason" was modified. http://git.kernel.org/?p=linux/kernel/git/torvalds/linux.git;a=commitdiff;h=c22ab332902333f83766017478c1ef6607ace681 Replace the fragile reason_str array with a more robust solution that will not be broken by future re-arrangements of the enum values. Signed-off-by: Seiji Aguchi <seiji.aguchi@hds.com> Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Acked-by: Don Zickus <dzickus@redhat.com> Link: https://lkml.org/lkml/2012/3/16/417 Signed-off-by: Tony Luck <tony.luck@intel.com>
* | | | Merge branch 'for-linus' of git://oss.sgi.com/xfs/xfsLinus Torvalds2012-03-2351-2250/+1375
|\ \ \ \ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Pull XFS updates from Ben Myers: "Scalability improvements for dquots, log grant code cleanups, plus bugfixes and cleanups large and small" Fix up various trivial conflicts that were due to some of the earlier patches already having been integrated into v3.3 as bugfixes, and then there were development patches on top of those. Easily merged by just taking the newer version from the pulled branch. * 'for-linus' of git://oss.sgi.com/xfs/xfs: (45 commits) xfs: fallback to vmalloc for large buffers in xfs_getbmap xfs: fallback to vmalloc for large buffers in xfs_attrmulti_attr_get xfs: remove remaining scraps of struct xfs_iomap xfs: fix inode lookup race xfs: clean up minor sparse warnings xfs: remove the global xfs_Gqm structure xfs: remove the per-filesystem list of dquots xfs: use per-filesystem radix trees for dquot lookup xfs: per-filesystem dquot LRU lists xfs: use common code for quota statistics xfs: reimplement fdatasync support xfs: split in-core and on-disk inode log item fields xfs: make xfs_inode_item_size idempotent xfs: log timestamp updates xfs: log file size updates at I/O completion time xfs: log file size updates as part of unwritten extent conversion xfs: do not require an ioend for new EOF calculation xfs: use per-filesystem I/O completion workqueues quota: make Q_XQUOTASYNC a noop xfs: include reservations in quota reporting ...
| * | | | xfs: fallback to vmalloc for large buffers in xfs_getbmapDave Chinner2012-03-151-3/+10
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | xfs_getbmap uses for a large buffer for extents, which is kmalloc'd. This can fail after the system has been running for some time as it is a high order allocation. Add a fallback to vmalloc so that it doesn't require contiguous memory and so won't randomly fail on files with large extent lists. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
| * | | | xfs: fallback to vmalloc for large buffers in xfs_attrmulti_attr_getDave Chinner2012-03-151-4/+10
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | xfsdump uses for a large buffer for extended attributes, which has a kmalloc'd shadow buffer in the kernel. This can fail after the system has been running for some time as it is a high order allocation. Add a fallback to vmalloc so that it doesn't require contiguous memory and so won't randomly fail while xfsdump is running. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
| * | | | xfs: remove remaining scraps of struct xfs_iomapDave Chinner2012-03-152-4/+0
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
| * | | | xfs: fix inode lookup raceDave Chinner2012-03-151-6/+12
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | When we get concurrent lookups of the same inode that is not in the per-AG inode cache, there is a race condition that triggers warnings in unlock_new_inode() indicating that we are initialising an inode that isn't in a the correct state for a new inode. When we do an inode lookup via a file handle or a bulkstat, we don't serialise lookups at a higher level through the dentry cache (i.e. pathless lookup), and so we can get concurrent lookups of the same inode. The race condition is between the insertion of the inode into the cache in the case of a cache miss and a concurrently lookup: Thread 1 Thread 2 xfs_iget() xfs_iget_cache_miss() xfs_iread() lock radix tree radix_tree_insert() rcu_read_lock radix_tree_lookup lock inode flags XFS_INEW not set igrab() unlock inode flags rcu_read_unlock use uninitialised inode ..... lock inode flags set XFS_INEW unlock inode flags unlock radix tree xfs_setup_inode() inode flags = I_NEW unlock_new_inode() WARNING as inode flags != I_NEW This can lead to inode corruption, inode list corruption, etc, and is generally a bad thing to occur. Fix this by setting XFS_INEW before inserting the inode into the radix tree. This will ensure any concurrent lookup will find the new inode with XFS_INEW set and that forces the lookup to wait until the XFS_INEW flag is removed before allowing the lookup to succeed. cc: <stable@vger.kernel.org> # for 3.0.x, 3.2.x Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Ben Myers <bpm@sgi.com>
| * | | | xfs: clean up minor sparse warningsDave Chinner2012-03-143-6/+10
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Ben Myers <bpm@sgi.com>
| * | | | xfs: remove the global xfs_Gqm structureChristoph Hellwig2012-03-147-174/+46
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | If we initialize the slab caches for the quota code when XFS is loaded there is no need for a global and reference counted quota manager structure. Drop all this overhead and also fix the error handling during quota initialization. Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Ben Myers <bpm@sgi.com>
| * | | | xfs: remove the per-filesystem list of dquotsChristoph Hellwig2012-03-145-304/+226
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Instead of keeping a separate per-filesystem list of dquots we can walk the radix tree for the two places where we need to iterate all quota structures. Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Ben Myers <bpm@sgi.com>
| * | | | xfs: use per-filesystem radix trees for dquot lookupChristoph Hellwig2012-03-146-263/+66
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Replace the global hash tables for looking up in-memory dquot structures with per-filesystem radix trees to allow scaling to a large number of in-memory dquot structures. Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Ben Myers <bpm@sgi.com>
| * | | | xfs: per-filesystem dquot LRU listsChristoph Hellwig2012-03-144-78/+74
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Replace the global dquot lru lists with a per-filesystem one. Note that the shrinker isn't wire up to the per-superblock VFS shrinker infrastructure as would have problems summing up and splitting the counts for inodes and dquots. I don't think this is a major problem as the quota cache isn't as interwinded with the inode cache as the dentry cache is, because an inode that is dropped from the cache will generally release a dquot reference, but most of the time it won't be the last one. Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Ben Myers <bpm@sgi.com>
| * | | | xfs: use common code for quota statisticsChristoph Hellwig2012-03-149-187/+110
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Switch the quota code over to use the generic XFS statistics infrastructure. While the legacy /proc/fs/xfs/xqm and /proc/fs/xfs/xqmstats interfaces are preserved for now the statistics that still have a meaning with the current code are now also available from /proc/fs/xfs/stats. Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Ben Myers <bpm@sgi.com>
| * | | | xfs: reimplement fdatasync supportChristoph Hellwig2012-03-134-5/+18
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Add an in-memory only flag to say we logged timestamps only, and use it to check if fdatasync can optimize away the log force. Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
| * | | | xfs: split in-core and on-disk inode log item fieldsChristoph Hellwig2012-03-135-92/+92
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Add a new ili_fields member to the inode log item to isolate the in-memory flags from the ones that actually go to the log. This will allow tracking timestamp-only updates for fdatasync and O_DSYNC in the next patch and prepares for divorcing the on-disk log format from the in-memory log item a little further down the road. Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
| * | | | xfs: make xfs_inode_item_size idempotentChristoph Hellwig2012-03-131-132/+83
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Move all code messing with the inode log item flags into xfs_inode_item_format to make sure xfs_inode_item_size really only calculates the the number of vectors, but doesn't modify any state of the inode item. Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
| * | | | xfs: log timestamp updatesChristoph Hellwig2012-03-1313-321/+65
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Timestamps on regular files are the last metadata that XFS does not update transactionally. Now that we use the delaylog mode exclusively and made the log scode scale extremly well there is no need to bypass that code for timestamp updates. Logging all updates allows to drop a lot of code, and will allow for further performance improvements later on. Note that this patch drops optimized handling of fdatasync - it will be added back in a separate commit. Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
| * | | | xfs: log file size updates at I/O completion timeChristoph Hellwig2012-03-132-24/+111
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Do not use unlogged metadata updates and the VFS dirty bit for updating the file size after writeback. In addition to causing various problems with updates getting delayed for far too long this also drags in the unscalable VFS dirty tracking, and is one of the few remaining unlogged metadata updates. Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
| * | | | xfs: log file size updates as part of unwritten extent conversionChristoph Hellwig2012-03-052-6/+24
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | If we convert and unwritten extent past the current i_size log the size update as part of the extent manipulation transactions instead of doing an unlogged metadata update later. Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Ben Myers <bpm@sgi.com>
| * | | | xfs: do not require an ioend for new EOF calculationChristoph Hellwig2012-03-052-20/+18
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Replace xfs_ioend_new_eof with a new inline xfs_new_eof helper that doesn't require and ioend, and is available also outside of xfs_aops.c. Also make the code a bit more clear by using a normal if statement instead of a slightly misleading MIN(). Reviewed-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Ben Myers <bpm@sgi.com>
| * | | | xfs: use per-filesystem I/O completion workqueuesChristoph Hellwig2012-03-055-49/+51
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The new concurrency managed workqueues are cheap enough that we can create per-filesystem instead of global workqueues. This allows us to remove the trylock or defer scheme on the ilock, which is not helpful once we have outstanding log reservations until finishing a size update. Also allow the default concurrency on this workqueues so that I/O completions blocking on the ilock for one inode do not block process for another inode. Reviewed-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Mark Tinguely <tinguely@sgi.com> Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Ben Myers <bpm@sgi.com>
| * | | | quota: make Q_XQUOTASYNC a noopChristoph Hellwig2012-02-291-2/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Now that XFS takes quota reservations into account there is no need to flush anything before reporting quotas - in addition to beeing fully transactional all quota information is also 100% coherent with the rest of the filesystem now. Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Ben Myers <bpm@sgi.com>
generated by cgit v1.2.3 (git 2.25.1) at 2025-03-04 15:48:28 +0000