- 21 Nov, 2022 18 commits
-
-
Alexander Aring authored
This patch is rework of lowcomms handling, the main goal was here to handle recvmsg() and sendpage() to run parallel. Parallel in two senses: 1. per connection and 2. that recvmsg()/sendpage() doesn't block each other. Currently recvmsg()/sendpage() cannot run parallel because two workqueues "dlm_recv" and "dlm_send" are ordered workqueues. That means only one work item can be executed. The amount of queue items will be increased about the amount of nodes being inside the cluster. The current two workqueues for sending and receiving can also block each other if the same connection is executed at the same time in dlm_recv and dlm_send workqueue because a per connection mutex for the socket handling. To make it more parallel we introduce one "dlm_io" workqueue which is not an ordered workqueue, the amount of workers are not limited. Due per connection flags SEND/RECV pending we schedule workers ordered per connection and per send and receive task. To get rid of the mutex blocking same workers to do socket handling we switched to a semaphore which handles socket operations as read lock and sock releases as write operations, to prevent sock_release() being called while the socket is being used. There might be more optimization removing the semaphore and replacing it with other synchronization mechanism, however due other circumstances e.g. othercon behaviour it seems complicated to doing this change. I added comments to remove the othercon handling and moving to a different synchronization mechanism as this is done. We need to do that to the next dlm major version upgrade because it is not backwards compatible with the current connect mechanism. The processing of dlm messages need to be still handled by a ordered workqueue. An dlm_process ordered workqueue was introduced which gets filled by the receive worker. This is probably the next bottleneck of DLM but the application can't currently parse dlm messages parallel. A comment was introduced to lift the workqueue context of dlm processing in a non-sleepable softirq to get messages processing done fast. Signed-off-by:
Alexander Aring <aahringo@redhat.com> Signed-off-by:
David Teigland <teigland@redhat.com>
-
Alexander Aring authored
This patch removes a init of an error value to -EINVAL which is not necessary. Signed-off-by:
-
Alexander Aring authored
This patch changes the handling of calling the original sk_error_report() by not putting it on the stack and calling it later. If the listen_sock.sk_error_report() is NULL in this moment it indicates a bug in our implementation. Signed-off-by:
-
Alexander Aring authored
This patch removes null checks on private data for sockets. If we have a null dereference there we having a bug in our implementation that such callback occurs in this state. Signed-off-by:
-
Alexander Aring authored
This patch merges the dlm_node_addrs lookup list to the connection structure. It is a per node mapping to some configuration setup by configfs. We don't need two lookup structures. The connection hash has now a lifetime like the dlm_node_addrs entries. Means we add only new entries when configure cluster and not while new connections are coming in, remove connection when a node got fenced and cleanup all connection when the dlm exits. It should work the same and even will show more issues because we don't try to somehow keep those two data structures in sync with the current cluster configuration. Signed-off-by:
-
Alexander Aring authored
This patch removes to allocate the dlm_local_addr[] pointers on the heap. Instead we directly store the type of "struct sockaddr_storage". This removes function deinit_local() because it was freeing memory only. Signed-off-by:
-
Alexander Aring authored
This patch removes save_listen_callbacks() and add_listen_sock() as they are only used once in lowcomms functionality. For shutdown lowcomms it's not necessary to whole flush the workqueues to synchronize with restoring the old sk_data_ready() callback. Only the listen con receive work need to be cancelled. For each individual node shutdown we should be sure that last ack was been transmitted which is done by flushing per connection swork worker. Signed-off-by:
-
Alexander Aring authored
Since commit 489d8e55 ("fs: dlm: add reliable connection if reconnect") we have functionality like TCP offers for half-closed sockets on dlm application protocol layer. This feature is required because the cluster manager events about leaving resource memberships can be locally already occurred but other cluster nodes having a pending leaving membership over the cluster manager protocol happening. In this time the local dlm node already shutdown it's connection and don't transmit anymore any new dlm messages, but however it still needs to be able to accept dlm messages because the pending leave membership request of the cluster manager protocol which the dlm kernel implementation has no control about it. We have this functionality on the application for two reasons, the main reason is that SCTP does not support such functionality on socket layer. But we can do it inside application layer. Another small issue is that this feature is broken in the TCP world because some NAT devices does not implement such functionality correctly. This is the same reason why the reliable connection session layer in DLM exists. We give up on middle devices in the networking which sends e.g. TCP resets out. In DLM we cannot have any message dropping and we ensure it over a session layer that it can't happen. Back to the half-closed grace shutdown handling. It's not necessary anymore to do it on socket layer (which is only support for TCP sockets) because we do it on application layer. This patch removes this handling, if there are still issues then we have a problem on the application layer for such handling. Signed-off-by:
-
Alexander Aring authored
This patch will switch from dlm_allow_conn to check if dlm lowcomms is running or not to if we actually have a listen socket set or not. The list socket will be set and unset in lowcomms start and shutdown functionality. To synchronize with data_ready() callback we will set the socket callback to NULL while socket lock is held. Signed-off-by:
-
Alexander Aring authored
Instead of check on list_empty() we can do the same with list_first_entry_or_null() and return NULL if the returned value is NULL. Signed-off-by:
-
Alexander Aring authored
This patch removed a twice INIT_WORK() functionality. We already doing this inside of dlm_lowcomms_init() functionality which is called only once dlm is loaded. Signed-off-by:
-
Alexander Aring authored
This patch introduces leftovers of init, start, stop and exit functionality. The dlm application layer should always call the midcomms layer which getting aware of such event and redirect it to the lowcomms layer. Some functionality which is currently handled inside the start functionality of midcomms and lowcomms should be handled in the init functionality as it only need to be initialized once when dlm is loaded. Signed-off-by:
-
Alexander Aring authored
In DLM when we send a dlm message it is easy to add the lock resource name, but additional lookup is required when to trace the receive message side. The idea here is to move the lookup work to the user by using a lookup to find the right send message with recv message. As note DLM can't drop any message which is guaranteed by a special session layer. For doing the lookup a 3 tupel is required as an unique identification which is dst nodeid, src nodeid and sequence number. This patch adds the destination nodeid to the dlm message trace points. The source nodeid is given by the h_nodeid field inside the header. Signed-off-by:
-
Alexander Aring authored
This patch renames seq to h_seq as it is named in the dlm header structure. Signed-off-by:
-
Alexander Aring authored
This patch renames DLM_IFL_NEED_SCHED to DLM_IFL_CB_PENDING because CB_PENDING is a proper name to describe this flag. This flag is set when callback enqueue will return DLM_ENQUEUE_CALLBACK_NEED_SCHED because the callback worker need to be queued. The flag tells that callbacks are currently pending to be called and will be unset if the callback work for the specific lkb is done. The term need schedule is part of this time but a proper name is to say that there are some callbacks pending to being called. Signed-off-by:
-
Alexander Aring authored
This patch changes the ast hotpath functionality in very unlikely cases that we do WARN_ON_ONCE() instead of WARN_ON() to not spamming the console output if we run into states that it would occur over and over again. Signed-off-by:
-
Alexander Aring authored
This patch will drop the lkb reference in an very unlikely case which should in practice not happened. However if it happens we cleanup the reference just in case. Signed-off-by:
-
Alexander Aring authored
This patch avoid the false-positive checker warning about writing 112 bytes into a 88 bytes field "e->request", see: [ 54.891560] dlm: csmb1: dlm_recover_directory 23 out 2 messages [ 54.990542] ------------[ cut here ]------------ [ 54.991012] memcpy: detected field-spanning write (size 112) of single field "&e->request" at fs/dlm/requestqueue.c:47 (size 88) [ 54.992150] WARNING: CPU: 0 PID: 297 at fs/dlm/requestqueue.c:47 dlm_add_requestqueue+0x177/0x180 [ 54.993002] CPU: 0 PID: 297 Comm: kworker/u4:3 Not tainted 6.1.0-rc5-00008-ge01d50cb #248 [ 54.993878] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.0-1.fc36 04/01/2014 [ 54.994718] Workqueue: dlm_recv process_recv_sockets [ 54.995230] RIP: 0010:dlm_add_requestqueue+0x177/0x180 [ 54.995731] Code: e7 01 0f 85 3b ff ff ff b9 58 00 00 00 48 c7 c2 c0 41 74 82 4c 89 ee 48 c7 c7 20 42 74 82 c6 05 8b 8d 30 02 01 e8 51 07 be 00 <0f> 0b e9 12 ff ff ff 66 90 0f 1f 44 00 00 41 57 48 8d 87 10 08 00 [ 54.997483] RSP: 0018:ffffc90000b1fbe8 EFLAGS: 00010282 [ 54.997990] RAX: 0000000000000000 RBX: ffff888024fc3d00 RCX: 0000000000000000 [ 54.998667] RDX: 0000000000000001 RSI: ffffffff81155014 RDI: fffff52000163f73 [ 54.999342] RBP: ffff88800dbac000 R08: 0000000000000001 R09: ffffc90000b1fa5f [ 54.999997] R10: fffff52000163f4b R11: 203a7970636d656d R12: ffff88800cfb0018 [ 55.000673] R13: 0000000000000070 R14: ffff888024fc3d18 R15: 0000000000000000 [ 55.001344] FS: 0000000000000000(0000) GS:ffff88806d600000(0000) knlGS:0000000000000000 [ 55.002078] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 55.002603] CR2: 00007f35d4f0b9a0 CR3: 0000000025495002 CR4: 0000000000770ef0 [ 55.003258] PKRU: 55555554 [ 55.003514] Call Trace: [ 55.003756] <TASK> [ 55.003953] dlm_receive_buffer+0x1c0/0x200 [ 55.004348] dlm_process_incoming_buffer+0x46d/0x780 [ 55.004786] ? kernel_recvmsg+0x8b/0xc0 [ 55.005150] receive_from_sock.isra.0+0x168/0x420 [ 55.005582] ? process_listen_recv_socket+0x10/0x10 [ 55.006018] ? finish_task_switch.isra.0+0xe0/0x400 [ 55.006469] ? __switch_to+0x2fe/0x6a0 [ 55.006808] ? read_word_at_a_time+0xe/0x20 [ 55.007197] ? strscpy+0x146/0x190 [ 55.007505] process_one_work+0x3d0/0x6b0 [ 55.007863] worker_thread+0x8d/0x620 [ 55.008209] ? __kthread_parkme+0xd8/0xf0 [ 55.008565] ? process_one_work+0x6b0/0x6b0 [ 55.008937] kthread+0x171/0x1a0 [ 55.009251] ? kthread_exit+0x60/0x60 [ 55.009582] ret_from_fork+0x1f/0x30 [ 55.009903] </TASK> [ 55.010120] ---[ end trace 0000000000000000 ]--- [ 55.025783] dlm: csmb1: dlm_recover 5 generation 3 done: 201 ms [ 55.026466] gfs2: fsid=smbcluster:csmb1.0: recover generation 3 done It seems the checker is unable to detect the additional length bytes which was allocated additionally for the flexible array in struct dlm_message. To solve it we split the memcpy() into copy for the 88 bytes struct and another memcpy() for the flexible array m_extra field. Signed-off-by:
-
- 08 Nov, 2022 18 commits
-
-
Alexander Aring authored
To not get the console spammed about WARN_ON() of invalid states in the dlm midcomms hot path handling we switch to WARN_ON_ONCE() to get it only once that there might be an issue with the midcomms state handling. Signed-off-by:
-
Alexander Aring authored
This patch will fix a small issue when printing out that dlm_midcomms_start() failed to start and it was printing out that the dlm subcomponent lowcomms was failed but lowcomms is behind the midcomms layer. Signed-off-by:
-
Alexander Aring authored
This patch will catch a possible dlm_add_member() and delivers it to the dlm recovery handling. Signed-off-by:
-
Alexander Aring authored
This patch drops additionally the sock_mutex when there is a sending message burst. Since we have acknowledge handling we free sending buffers only when we receive an ack back, but if we are stuck in send_to_sock() looping because dlm sends a lot of messages and we never leave the loop the sending buffer fill up very quickly. We can't receive during this iteration because the sock_mutex is held. This patch will unlock the sock_mutex so it should be possible to receive messages when a burst of sending messages happens. This will allow to free up memory because acks which are already received can be processed. Signed-off-by:
-
Alexander Aring authored
This patch removes the ls_remove_wait waitqueue handling. The current handling tries to wait before a lookup is send out for a identically resource name which is going to be removed. Hereby the remove message should be send out before the new lookup message. The reason is that after a lookup request and response will actually use the specific remote rsb. A followed remove message would delete the rsb on the remote side but it's still being used. To reach a similar behaviour we simple send the remove message out while the rsb lookup lock is held and the rsb is removed from the toss list. Other find_rsb() calls would never have the change to get a rsb back to live while a remove message will be send out (without holding the lock). This behaviour requires a non-sleepable context which should be provided now and might be the reason why it was not implemented so in the first place. Signed-off-by:
-
Alexander Aring authored
This patch allows to give the use control about the allocation context based on a per message basis. Currently all messages forced to be created under GFP_NOFS context. Signed-off-by:
-
Alexander Aring authored
This patch will introducde a queue implementation for callbacks by using the Linux lists. The current callback queue handling is implemented by a static limit of 6 entries, see DLM_CALLBACKS_SIZE. The sequence number inside the callback structure was used to see if the entries inside the static entry is valid or not. We don't need any sequence numbers anymore with a dynamic datastructure with grows and shrinks during runtime to offer such functionality. We assume that every callback will be delivered to the DLM user if once queued. Therefore the callback flag DLM_CB_SKIP was dropped and the check for skipping bast was moved before worker handling and not skip while the callback worker executes. This will reduce unnecessary queues of the callback worker. All last callback saves are pointers now and don't need to copied over. There is a reference counter for callback structures which will care about to free the callback structures at the right time if they are not referenced anymore. Signed-off-by:
-
Alexander Aring authored
This patch moves the debugging information of the last cast and bast time when calling the last and bast function call. Signed-off-by:
-
Alexander Aring authored
There is no need to use a mutex in those hot path sections. We change it to spin lock to serve callbacks more faster by not allowing schedule. The locked sections will not be locked for a long time. Signed-off-by:
-
Alexander Aring authored
This patch converts the ls_cb_mutex mutex to a spinlock, there is no sleepable context when this lock is held. Signed-off-by:
-
Alexander Aring authored
Instead of using list_entry() this patch moves to using the list_first_entry() macro. Signed-off-by:
-
Alexander Aring authored
We should allow dlm_add_cb() to call queue_work() only after the recovery queued pending for delayed lkbs. This patch will move the switch LSFL_CB_DELAY after the delayed lkb work was processed. Signed-off-by:
-
Alexander Aring authored
This patch adds tracepoints for send and recv cases of dlm messages and dlm rcom messages. In case of send and dlm message we add the dlm rsb resource name this dlm messages belongs to. This has the advantage to follow dlm messages on a per lock basis. In case of recv message the resource name can be extracted by follow the send message sequence number. The dlm message DLM_MSG_PURGE doesn't belong to a lock request and will not set the resource name in a dlm_message trace. The same for all rcom messages. There is additional handling required for this debugging functionality which is tried to be small as possible. Also the midcomms layer gets aware of lock resource names, for now this is required to make a connection between sequence number and lock resource names. It is for debugging purpose only. Signed-off-by:
-
Alexander Aring authored
To allow more than just dereferencing the inner header we directly point to the inner dlm packet which allows us to dereference the header, rcom or message structure. Signed-off-by:
-
Alexander Aring authored
This patch removes the send repeat remove handling. This handling is there to repeatingly DLM_MSG_REMOVE messages in cases the dlm stack thinks it was not received at the first time. In cases of message drops this functionality is necessary, but since the DLM midcomms layer guarantees there are no messages drops between cluster nodes this feature became not strict necessary anymore. Due message delays/processing it could be that two send_repeat_remove() are sent out while the other should be still on it's way. We remove the repeat remove handling because we are sure that the message cannot be dropped due communication errors. Signed-off-by:
-
Alexander Aring authored
This patch fixes a race if we get two times an socket data ready event while the listen connection worker is queued. Currently it will be served only once but we need to do it (in this case twice) until we hit -EAGAIN which tells us there is no pending accept going on. This patch wraps an do while loop until we receive a return value which is different than 0 as it was done before commit d11ccd45 ("fs: dlm: listen socket out of connection hash"). Cc: stable@vger.kernel.org Fixes: d11ccd45 ("fs: dlm: listen socket out of connection hash") Signed-off-by:
-
Alexander Aring authored
This patch fixes a double sock_release() call when the listen() is called for the dlm lowcomms listen socket. The caller of dlm_listen_for_all should never care about releasing the socket if dlm_listen_for_all() fails, it's done now only once if listen() fails. Cc: stable@vger.kernel.org Fixes: 2dc6b115 ("fs: dlm: introduce generic listen") Signed-off-by:
-
Paulo Miguel Almeida authored
One-element arrays are deprecated. So, replace one-element array with fixed size array member in struct dlm_ls, and refactor the rest of the code, accordingly. Link: https://github.com/KSPP/linux/issues/79 Link: https://github.com/KSPP/linux/issues/228 Link: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101836 Link: https://lore.kernel.org/lkml/Y0W5jkiXUkpNl4ap@mail.google.com/Signed-off-by:
Paulo Miguel Almeida <paulo.miguel.almeida.rodenas@gmail.com> Reviewed-by:
Gustavo A. R. Silva <gustavoars@kernel.org> Reviewed-by:
Kees Cook <keescook@chromium.org> Signed-off-by:
-
- 06 Nov, 2022 4 commits
-
-
Linus Torvalds authored
-
git://git.kernel.org/pub/scm/linux/kernel/git/cxl/cxlLinus Torvalds authored
Pull cxl fixes from Dan Williams: "Several fixes for CXL region creation crashes, leaks and failures. This is mainly fallout from the original implementation of dynamic CXL region creation (instantiate new physical memory pools) that arrived in v6.0-rc1. Given the theme of "failures in the presence of pass-through decoders" this also includes new regression test infrastructure for that case. Summary: - Fix region creation crash with pass-through decoders - Fix region creation crash when no decoder allocation fails - Fix region creation crash when scanning regions to enforce the increasing physical address order constraint that CXL mandates - Fix a memory leak for cxl_pmem_region objects, track 1:N instead of 1:1 memory-device-to-region associations. - Fix a memory leak for cxl_region objects when regions with active targets are deleted - Fix assignment of NUMA nodes to CXL regions by CFMWS (CXL Window) emulated proximity domains. - Fix region creation failure for switch attached devices downstream of a single-port host-bridge - Fix false positive memory leak of cxl_region objects by recycling recently used region ids rather than freeing them - Add regression test infrastructure for a pass-through decoder configuration - Fix some mailbox payload handling corner cases" * tag 'cxl-fixes-for-6.1-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/cxl/cxl: cxl/region: Recycle region ids cxl/region: Fix 'distance' calculation with passthrough ports tools/testing/cxl: Add a single-port host-bridge regression config tools/testing/cxl: Fix some error exits cxl/pmem: Fix cxl_pmem_region and cxl_memdev leak cxl/region: Fix cxl_region leak, cleanup targets at region delete cxl/region: Fix region HPA ordering validation cxl/pmem: Use size_add() against integer overflow cxl/region: Fix decoder allocation crash ACPI: NUMA: Add CXL CFMWS 'nodes' to the possible nodes set cxl/pmem: Fix failure to account for 8 byte header for writes to the device LSA. cxl/region: Fix null pointer dereference due to pass through decoder commit cxl/mbox: Add a check on input payload size -
Linus Torvalds authored
Merge tag 'hwmon-for-v6.1-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/linux-staging Pull hwmon fixes from Guenter Roeck: "Fix two regressions: - Commit 54cc3dbf ("hwmon: (pmbus) Add regulator supply into macro") resulted in regulator undercount when disabling regulators. Revert it. - The thermal subsystem rework caused the scmi driver to no longer register with the thermal subsystem because index values no longer match. To fix the problem, the scmi driver now directly registers with the thermal subsystem, no longer through the hwmon core" * tag 'hwmon-for-v6.1-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/linux-staging: Revert "hwmon: (pmbus) Add regulator supply into macro" hwmon: (scmi) Register explicitly with Thermal Framework
-
git://git.kernel.org/pub/scm/linux/kernel/git/tip/tipLinus Torvalds authored
Pull perf fixes from Borislav Petkov: - Add Cooper Lake's stepping to the PEBS guest/host events isolation fixed microcode revisions checking quirk - Update Icelake and Sapphire Rapids events constraints - Use the standard energy unit for Sapphire Rapids in RAPL - Fix the hw_breakpoint test to fail more graciously on !SMP configs * tag 'perf_urgent_for_v6.1_rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: perf/x86/intel: Add Cooper Lake stepping to isolation_ucodes[] perf/x86/intel: Fix pebs event constraints for SPR perf/x86/intel: Fix pebs event constraints for ICL perf/x86/rapl: Use standard Energy Unit for SPR Dram RAPL domain perf/hw_breakpoint: test: Skip the test if dependencies unmet
-
