summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* mac80211: check size of channel switch IE when parsingJohannes Berg2012-08-203-11/+9
| | | | | | | | The channel switch IE has a fixed size, so we can discard it in parsing if it's not the right size and use the right struct pointer. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: fix CSA handling timerJohannes Berg2012-08-201-3/+2
| | | | | | | | The time until the channel switch is in TU, not in milliseconds, so use TU_TO_EXP_TIME() to correctly program the timer. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: clean up CSA handling codeJohannes Berg2012-08-201-18/+16
| | | | | | | | | | | Clean up the CSA handling code by moving some of it out of the if and using a C99 initializer for the struct passed to the driver method. While at it, also add a comment that we should wait for a beacon after switching the channel. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: remove unneeded 'bssid' variableJohannes Berg2012-08-201-4/+2
| | | | | | | There's no need to copy the BSSID just to print it, remove the unnecessary variable. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: support A-MPDU status reportingJohannes Berg2012-08-202-15/+80
| | | | | | | | Support getting A-MPDU status information from the drivers and reporting it to userspace via radiotap in the standard fields. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* wireless: add radiotap A-MPDU status fieldJohannes Berg2012-08-202-0/+13
| | | | | | | | Define the A-MPDU status field in radiotap, also update the radiotap parser for it and the MCS field that was apparently missed last time. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: add supported rates change notification in IBSSAntonio Quartulli2012-08-203-1/+11
| | | | | | | | | | | | | | | | In IBSS it is possible that the supported rates set for a station changes over time (e.g. it gets first initialised as an empty set because of no available information about rates and updated later). In this case the driver has to be notified about the change in order to update its internal table accordingly (if needed). This behaviour is needed by all those drivers that handle rc internally but leave stations management to mac80211 Reported-by: Gui Iribarren <gui@altermundi.net> Signed-off-by: Antonio Quartulli <ordex@autistici.org> [Johannes - add docs, validate IBSS mode only, fix compilation] Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: clean up mpath_move_to_queue()Thomas Pedersen2012-08-203-18/+21
| | | | | | | | | | | | Use skb_queue_walk_safe instead, and fix a few issues: - didn't free old skbs on moving - didn't react to failed skb alloc - needlessly held a local pointer to the destination frame queue - didn't check destination queue length before adding skb Signed-off-by: Thomas Pedersen <thomas@cozybit.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: use skb_queue_walk() in mesh_path_assign_nexthopThomas Pedersen2012-08-201-7/+1
| | | | | | | | Since all we really want is just to iterate over all skbs, do just that and avoid (de)queueing to a clusmy tmpq. Signed-off-by: Thomas Pedersen <thomas@cozybit.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: avoid using synchronize_rcu in ieee80211_set_probe_respEyal Shapira2012-08-204-14/+21
| | | | | | | | | | | | | | This could take a while (100ms+) and may delay sending assoc resp in AP mode with WPS or P2P GO (as setting the probe resp takes place there). We've encountered situations where the delay was big enough to cause connection problems with devices like Galaxy Nexus. Switch to using call_rcu with a free handler. [Arik - rework to use plain buffer and instead of skb] Signed-off-by: Eyal Shapira <eyal@wizery.com> Signed-off-by: Arik Nemtsov <arik@wizery.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: skb leak in mesh_plink_frame_tx()Thomas Pedersen2012-08-031-6/+11
| | | | | | | | | Although adding an IE is almost guaranteed to succeed since we already accounted for its length while allocating the skb, we should still free the skb in case of failure. Signed-off-by: Thomas Pedersen <thomas@cozybit.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: don't request ack for peering closeThomas Pedersen2012-08-031-0/+3
| | | | | | | | | It doesn't make a lot of sense to wait for an ack in response to a peering close frame since either peer in this exchange could be going down. Signed-off-by: Thomas Pedersen <thomas@cozybit.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: remove ieee80211_clean_sdata()Thomas Pedersen2012-08-031-17/+0
| | | | | | | | This function was only used by mesh, and not really needed since any interface-specific cleanup already happens in the netdev handlers. Signed-off-by: Thomas Pedersen <thomas@cozybit.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: improve cleanup when leaving meshThomas Pedersen2012-08-032-9/+5
| | | | | | | | | | | | | | | It is not necessary to stop the mesh beacon in the mac80211 ndo_stop handler, since cfg80211 has already left the mesh on NETDEV_GOING_DOWN notification. Also some improvements to ieee80211_stop_mesh(): - flush mpath entries. - flush sta entries per-sdata so we don't remove entries belonging to other vifs on the same hw. Signed-off-by: Thomas Pedersen <thomas@cozybit.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: reset station MLME flags upon new associationJohannes Berg2012-08-011-4/+7
| | | | | | | | | | When associating anew, the old station MLME flags should be cleared. The only exception is the 40 MHz disable flag as it might have been set while the channel was set in a previous authentication attempt so it needs to be kept intact. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: use eth_broadcast_addrJohannes Berg2012-07-314-6/+6
| | | | | | Instead of memset(). Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: enable WDS carrier only after adding stationJohannes Berg2012-07-311-0/+3
| | | | | | | | Enable the carrier on WDS type interfaces only after having added the station entry for the WDS peer so outgoing frames will find it. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: manage carrier state in meshJohannes Berg2012-07-312-4/+12
| | | | | | | | Instead of assuming the carrier is on all the time in mesh manage it with joining and leaving the mesh. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: use correct channel in TXJohannes Berg2012-07-312-12/+7
| | | | | | | | | Since we only need the band, remove the channel pointer from struct ieee80211_tx_data and also assign it properly, depending on context, to the correct operating or current channel. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: fix current vs. operating channel in preq/beaconJohannes Berg2012-07-316-26/+44
| | | | | | | | | | | | | When sending probe requests, e.g. during software scanning, these will go out on the *current* channel, so their IEs need to be built from the current channel. At other times, e.g. for beacons or probe request templates, the IEs will be used on the *operating* channel and using the current channel instead might result in errors. Add the appropriate parameters to respect the difference. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: use oper_channel in utils and configJohannes Berg2012-07-313-6/+6
| | | | | | | | | | | Using hw.conf.channel is wrong as it could be the temporary channel if any function like the beacon get function is called while scanning or during other temporary out-of-channel activities. Use oper_channel instead. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: use oper_channel in managed mlmeJohannes Berg2012-07-311-6/+6
| | | | | | | | | | | Using hw.conf.channel is wrong as it could be the temporary channel if any function like the beacon get function is called while scanning or during other temporary out-of-channel activities. Use oper_channel instead. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: use oper_channel in ibssJohannes Berg2012-07-311-4/+4
| | | | | | | | | | | Using hw.conf.channel is wrong as it could be the temporary channel if any function like the beacon get function is called while scanning or during other temporary out-of-channel activities. Use oper_channel instead. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: use oper_channel in meshJohannes Berg2012-07-311-3/+4
| | | | | | | | | | | Using hw.conf.channel is wrong as it could be the temporary channel if any function like the beacon get function is called while scanning or during other temporary out-of-channel activities. Use oper_channel instead. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: set channel only once during auth/assocJohannes Berg2012-07-311-26/+43
| | | | | | | | | | | There's no need to set up the channel during auth and again during assoc, just do it once. Currently this doesn't result in any changes since calling hw_config() with an unchanged channel will return early, but with the channel context work this has an impact on channel context assignment. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: rename sta to new_staJohannes Berg2012-07-311-8/+8
| | | | | | | | In ieee80211_prep_connection(), the station (if not NULL) is the new station (representing the AP) that needs to be added. Rename the variable to "new_sta" to clarify this. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: supress HT/VHT disable if not supportedJohannes Berg2012-07-311-4/+6
| | | | | | | | If HT/VHT isn't supported by us we shouldn't print a message that we disabled it, do that only if the AP didn't support WMM and we therefore disable it. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: move TX station pointer and restructure TXThomas Huehn2012-07-3153-132/+232
| | | | | | | | | | | | | | Remove the control.sta pointer from ieee80211_tx_info to free up sufficient space in the TX skb control buffer for the upcoming Transmit Power Control (TPC). Instead, the pointer is now on the stack in a new control struct that is passed as a function parameter to the drivers' tx method. Signed-off-by: Thomas Huehn <thomas@net.t-labs.tu-berlin.de> Signed-off-by: Alina Friedrichsen <x-alina@gmx.net> Signed-off-by: Felix Fietkau <nbd@openwrt.org> [reworded commit message] Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: add PS flag to bss_confEliad Peller2012-07-315-3/+27
| | | | | | | | | | | | | | | | | | | | | | | | | | | Currently, ps mode is indicated per device (rather than per interface), which doesn't make a lot of sense. Moreover, there are subtle bugs caused by the inability to indicate ps change along with other changes (e.g. when the AP deauth us, we'd like to indicate CHANGED_PS | CHANGED_ASSOC, as changing PS before notifying about disassociation will result in null-packets being sent (if IEEE80211_HW_SUPPORTS_DYNAMIC_PS) while the sta is already disconnected.) Keep the current per-device notifications, and add parallel per-vif notifications. In order to keep it simple, the per-device ps and the per-vif ps are orthogonal - the per-vif ps configuration is determined only by the user configuration (enable/disable) and the connection state, and is not affected by other vifs state and (temporary) dynamic_ps/offchannel operations (unlike per-device ps). Signed-off-by: Eliad Peller <eliad@wizery.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* wlcore: op_tx: pass sta explicitly when inferring frame hlidArik Nemtsov2012-07-313-13/+8
| | | | | | | | avoid using the skb CB for getting the appropriate sta. Signed-off-by: Arik Nemtsov <arik@wizery.com> Acked-by: Luciano Coelho <coelho@ti.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* wlcore: don't get the hlid from a queued skbArik Nemtsov2012-07-311-20/+26
| | | | | | | | | | | | There was a bug hiding here since the hlid was sometimes inferred from the sta, which might be invalid at this point. Instead, propagate the hlid from the skb-queue where we got the skb in the first place. Signed-off-by: Arik Nemtsov <arik@wizery.com> Acked-by: Luciano Coelho <coelho@ti.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* cfg80211: unify IE searchVladimir Kondratiev2012-07-311-15/+3
| | | | | | | | Remove ah-hoc IE search code found in the ieee80211_bss_get_ie() and use cfg80211_find_ie() instead. Signed-off-by: Vladimir Kondratiev <qca_vkondrat@qca.qualcomm.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: don't call mgd_prepare_tx when associatedEmmanuel Grumbach2012-07-311-4/+5
| | | | | | | | | | | | This doesn't make any sense since we are expected to be on the medium or at least to Tx only when we are on the right channel and the AP/GO can hear us. Move the call to mgd_prepare_tx() for deauth to be only done in case we're sending a deauth while not associated. Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: don't react to beacon loss if HW monitoringJohannes Berg2012-07-311-0/+3
| | | | | | | | | | | | | | | | | | | If the HW is monitoring connection loss (as advertised by IEEE80211_HW_CONNECTION_MONITOR) but not filtering beacons (IEEE80211_VIF_BEACON_FILTER) then mac80211 will still start the beacon loss timer and if a few beacons are lost, e.g. due to scanning, drop the connection. If the hardware doesn't advertise connection monitoring, then it won't drop the connection right away but probe the AP, which is intended, but due to the logic in the timer when connection monitoring is done it assumes the connection was actually lost. Fix this problem by not starting the timer when the HW does connection monitoring. Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: VHT (11ac) associationMahesh Palivela2012-07-312-2/+39
| | | | | | | | | Insert VHT IEs into association frames to allow mac80211 to connect as a VHT client. Signed-off-by: Mahesh Palivela <maheshp@posedge.com> [clarify commit message] Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* mac80211: don't re-init rate control when receiving mesh beaconChun-Yeow Yeoh2012-07-311-0/+5
| | | | | | | | | | | Rate control is re-initialized whenever a beacon from a mesh peer received, breaking the algorithms and resulting in low performance. Return early from mesh_peer_init if we already established a link with this peer to avoid this. Signed-off-by: Chun-Yeow Yeoh <yeohchunyeow@gmail.com> [clarify commit message] Signed-off-by: Johannes Berg <johannes.berg@intel.com>
* Merge tag 'ext4_for_linus' of ↵Linus Torvalds2012-07-2714-196/+563
|\ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4 Pull ext4 updates from Ted Ts'o: "The usual collection of bug fixes and optimizations. Perhaps of greatest note is a speed up for parallel, non-allocating DIO writes, since we no longer take the i_mutex lock in that case. For bug fixes, we fix an incorrect overhead calculation which caused slightly incorrect results for df(1) and statfs(2). We also fixed bugs in the metadata checksum feature." * tag 'ext4_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4: (23 commits) ext4: undo ext4_calc_metadata_amount if we fail to claim space ext4: don't let i_reserved_meta_blocks go negative ext4: fix hole punch failure when depth is greater than 0 ext4: remove unnecessary argument from __ext4_handle_dirty_metadata() ext4: weed out ext4_write_super ext4: remove unnecessary superblock dirtying ext4: convert last user of ext4_mark_super_dirty() to ext4_handle_dirty_super() ext4: remove useless marking of superblock dirty ext4: fix ext4 mismerge back in January ext4: remove dynamic array size in ext4_chksum() ext4: remove unused variable in ext4_update_super() ext4: make quota as first class supported feature ext4: don't take the i_mutex lock when doing DIO overwrites ext4: add a new nolock flag in ext4_map_blocks ext4: split ext4_file_write into buffered IO and direct IO ext4: remove an unused statement in ext4_mb_get_buddy_page_lock() ext4: fix out-of-date comments in extents.c ext4: use s_csum_seed instead of i_csum_seed for xattr block ext4: use proper csum calculation in ext4_rename ext4: fix overhead calculation used by ext4_statfs() ...
| * ext4: undo ext4_calc_metadata_amount if we fail to claim spaceTheodore Ts'o2012-07-231-11/+21
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The function ext4_calc_metadata_amount() has side effects, although it's not obvious from its function name. So if we fail to claim space, regardless of whether we retry to claim the space again, or return an error, we need to undo these side effects. Otherwise we can end up incorrectly calculating the number of metadata blocks needed for the operation, which was responsible for an xfstests failure for test #271 when using an ext2 file system with delalloc enabled. Reported-by: Brian Foster <bfoster@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu> Cc: stable@vger.kernel.org
| * ext4: don't let i_reserved_meta_blocks go negativeBrian Foster2012-07-221-0/+9
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | If we hit a condition where we have allocated metadata blocks that were not appropriately reserved, we risk underflow of ei->i_reserved_meta_blocks. In turn, this can throw sbi->s_dirtyclusters_counter significantly out of whack and undermine the nondelalloc fallback logic in ext4_nonda_switch(). Warn if this occurs and set i_allocated_meta_blocks to avoid this problem. This condition is reproduced by xfstests 270 against ext2 with delalloc enabled: Mar 28 08:58:02 localhost kernel: [ 171.526344] EXT4-fs (loop1): delayed block allocation failed for inode 14 at logical offset 64486 with max blocks 64 with error -28 Mar 28 08:58:02 localhost kernel: [ 171.526346] EXT4-fs (loop1): This should not happen!! Data will be lost 270 ultimately fails with an inconsistent filesystem and requires an fsck to repair. The cause of the error is an underflow in ext4_da_update_reserve_space() due to an unreserved meta block allocation. Signed-off-by: Brian Foster <bfoster@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu> Cc: stable@vger.kernel.org
| * ext4: fix hole punch failure when depth is greater than 0Ashish Sangwan2012-07-221-17/+29
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Whether to continue removing extents or not is decided by the return value of function ext4_ext_more_to_rm() which checks 2 conditions: a) if there are no more indexes to process. b) if the number of entries are decreased in the header of "depth -1". In case of hole punch, if the last block to be removed is not part of the last extent index than this index will not be deleted, hence the number of valid entries in the extent header of "depth - 1" will remain as it is and ext4_ext_more_to_rm will return 0 although the required blocks are not yet removed. This patch fixes the above mentioned problem as instead of removing the extents from the end of file, it starts removing the blocks from the particular extent from which removing blocks is actually required and continue backward until done. Signed-off-by: Ashish Sangwan <ashish.sangwan2@gmail.com> Signed-off-by: Namjae Jeon <linkinjeon@gmail.com> Reviewed-by: Lukas Czerner <lczerner@redhat.com> Cc: stable@vger.kernel.org
| * ext4: remove unnecessary argument from __ext4_handle_dirty_metadata()Artem Bityutskiy2012-07-225-11/+7
| | | | | | | | | | | | | | | | | | The '__ext4_handle_dirty_metadata()' does not need the 'now' argument anymore and we can kill it. Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu> Reviewed-by: Jan Kara <jack@suse.cz>
| * ext4: weed out ext4_write_superArtem Bityutskiy2012-07-221-10/+0
| | | | | | | | | | | | | | | | | | | | We do not depend on VFS's '->write_super()' anymore and do not need the 's_dirt' flag anymore, so weed out 'ext4_write_super()' and 's_dirt'. Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu> Reviewed-by: Jan Kara <jack@suse.cz>
| * ext4: remove unnecessary superblock dirtyingArtem Bityutskiy2012-07-222-4/+3
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This patch changes the 'ext4_handle_dirty_super()' function which submits the superblock for I/O in the following cases: 1. When creating the first large file on a file system without EXT4_FEATURE_RO_COMPAT_LARGE_FILE feature. 2. When re-sizing the file-system. 3. When creating an xattr on a file-system without the EXT4_FEATURE_COMPAT_EXT_ATTR feature. If the file-system has journal enabled, the superblock is written via the journal. We do not modify this path. If the file-system has no journal, this function, falls back to just marking the superblock as dirty using the 's_dirt' superblock flag. This means that it delays the actual superblock I/O submission by 5 seconds (default setting). Namely, the 'sync_supers()' kernel thread will call 'ext4_write_super()' later and will actually submit the superblock for I/O. And this is the behavior this patch modifies: we stop using 's_dirt' and just mark the superblock buffer as dirty right away. Indeed, all 3 cases above are extremely rare and it does not add any value to delay the I/O submission for them. Note: 'ext4_handle_dirty_super()' executes '__ext4_handle_dirty_super()' with 'now = 0'. This patch basically makes the 'now' argument unneeded and it will be deleted in one of the next patches. This patch also removes 's_dirt' condition on the unmount path because we never set it anymore, so we should not test it. Tested using xfstests for both journalled and non-journalled ext4. Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu> Reviewed-by: Jan Kara <jack@suse.cz>
| * ext4: convert last user of ext4_mark_super_dirty() to ext4_handle_dirty_super()Jan Kara2012-07-222-10/+13
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The last user of ext4_mark_super_dirty() in ext4_file_open() is so rare it can well be modifying the superblock properly by journalling the change. Change it and get rid of ext4_mark_super_dirty() as it's not needed anymore. Artem: small amendments. Artem: tested using xfstests for both journalled and non-journalled ext4. Signed-off-by: Jan Kara <jack@suse.cz> Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu> Tested-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
| * ext4: remove useless marking of superblock dirtyJan Kara2012-07-222-4/+0
| | | | | | | | | | | | | | | | | | | | | | | | | | | | Commit a0375156 properly notes that superblock doesn't need to be marked as dirty when only number of free inodes / blocks / number of directories changes since that is recomputed on each mount anyway. However that comment leaves some unnecessary markings as dirty in place. Remove these. Artem: tested using xfstests for both journalled and non-journalled ext4. Signed-off-by: Jan Kara <jack@suse.cz> Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu> Tested-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
| * ext4: fix ext4 mismerge back in JanuaryAl Viro2012-07-221-1/+0
| | | | | | | | | | | | | | | | | | Duplicate caused, AFAICS, by mismerge in ff9cb1c4eead5e4c292e75cd3170a82d66944101> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu> Cc: stable@vger.kernel.org
| * ext4: remove dynamic array size in ext4_chksum()Theodore Ts'o2012-07-221-1/+3
| | | | | | | | | | | | | | | | | | | | The ext4_checksum() inline function was using a dynamic array size, which is not legal C. (It is a gcc extension). Remove it. Cc: "Darrick J. Wong" <djwong@us.ibm.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
| * ext4: remove unused variable in ext4_update_super()Theodore Ts'o2012-07-221-1/+1
| | | | | | | | Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
| * ext4: make quota as first class supported featureAditya Kali2012-07-223-10/+150
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This patch adds support for quotas as a first class feature in ext4; which is to say, the quota files are stored in hidden inodes as file system metadata, instead of as separate files visible in the file system directory hierarchy. It is based on the proposal at: https://ext4.wiki.kernel.org/index.php/Design_For_1st_Class_Quota_in_Ext4 This patch introduces a new feature - EXT4_FEATURE_RO_COMPAT_QUOTA which, when turned on, enables quota accounting at mount time iteself. Also, the quota inodes are stored in two additional superblock fields. Some changes introduced by this patch that should be pointed out are: 1) Two new ext4-superblock fields - s_usr_quota_inum and s_grp_quota_inum for storing the quota inodes in use. 2) Default quota inodes are: inode#3 for tracking userquota and inode#4 for tracking group quota. The superblock fields can be set to use other inodes as well. 3) If the QUOTA feature and corresponding quota inodes are set in superblock, the quota usage tracking is turned on at mount time. On 'quotaon' ioctl, the quota limits enforcement is turned on. 'quotaoff' ioctl turns off only the limits enforcement in this case. 4) When QUOTA feature is in use, the quota mount options 'quota', 'usrquota', 'grpquota' are ignored by the kernel. 5) mke2fs or tune2fs can be used to set the QUOTA feature and initialize quota inodes. The default reserved inodes will not be visible to user as regular files. 6) The quota-tools will need to be modified to support hidden quota files on ext4. E2fsprogs will also include support for creating and fixing quota files. 7) Support is only for the new V2 quota file format. Tested-by: Jan Kara <jack@suse.cz> Reviewed-by: Jan Kara <jack@suse.cz> Reviewed-by: Johann Lombardi <johann@whamcloud.com> Signed-off-by: Aditya Kali <adityakali@google.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
| * ext4: don't take the i_mutex lock when doing DIO overwritesZheng Liu2012-07-222-4/+71
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Aligned and overwrite direct I/O can be parallelized. In ext4_file_dio_write, we first check whether these conditions are satisfied or not. If so, we take i_data_sem and release i_mutex lock directly. Meanwhile iocb->private is set to indicate that this is a dio overwrite, and it will be handled in ext4_ext_direct_IO. [ Added fix from Dan Carpenter to fix locking bug on the error path. ] CC: Tao Ma <tm@tao.ma> CC: Eric Sandeen <sandeen@redhat.com> CC: Robin Dong <hao.bigrat@gmail.com> Signed-off-by: Zheng Liu <wenqing.lz@taobao.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu> Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
OpenPOWER on IntegriCloud