aboutsummaryrefslogtreecommitdiffstats
path: root/fs/cifs/smb2pdu.c (follow)
AgeCommit message (Collapse)AuthorFilesLines
2022-10-19cifs: fix memory leaks in session setupPaulo Alcantara1-8/+7
We were only zeroing out the ntlmssp blob but forgot to free the allocated buffer in the end of SMB2_sess_auth_rawntlmssp_negotiate() and SMB2_sess_auth_rawntlmssp_authenticate() functions. This fixes below kmemleak reports: unreferenced object 0xffff88800ddcfc60 (size 96): comm "mount.cifs", pid 758, jiffies 4294696066 (age 42.967s) hex dump (first 32 bytes): 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ backtrace: [<00000000d0beeb29>] __kmalloc+0x39/0xa0 [<00000000e3834047>] build_ntlmssp_smb3_negotiate_blob+0x2c/0x110 [cifs] [<00000000e85f5ab2>] SMB2_sess_auth_rawntlmssp_negotiate+0xd3/0x230 [cifs] [<0000000080fdb897>] SMB2_sess_setup+0x16c/0x2a0 [cifs] [<000000009af320a8>] cifs_setup_session+0x13b/0x370 [cifs] [<00000000f15d5982>] cifs_get_smb_ses+0x643/0xb90 [cifs] [<00000000fe15eb90>] mount_get_conns+0x63/0x3e0 [cifs] [<00000000768aba03>] mount_get_dfs_conns+0x16/0xa0 [cifs] [<00000000cf1cf146>] cifs_mount+0x1c2/0x9a0 [cifs] [<000000000d66b51e>] cifs_smb3_do_mount+0x10e/0x710 [cifs] [<0000000077a996c5>] smb3_get_tree+0xf4/0x200 [cifs] [<0000000094dbd041>] vfs_get_tree+0x23/0xc0 [<000000003a8561de>] path_mount+0x2d3/0xb50 [<00000000ed5c86d6>] __x64_sys_mount+0x102/0x140 [<00000000142142f3>] do_syscall_64+0x3b/0x90 [<00000000e2b89731>] entry_SYSCALL_64_after_hwframe+0x63/0xcd unreferenced object 0xffff88801437f000 (size 512): comm "mount.cifs", pid 758, jiffies 4294696067 (age 42.970s) hex dump (first 32 bytes): 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ backtrace: [<00000000d0beeb29>] __kmalloc+0x39/0xa0 [<00000000004f53d2>] build_ntlmssp_auth_blob+0x4f/0x340 [cifs] [<000000005f333084>] SMB2_sess_auth_rawntlmssp_authenticate+0xd4/0x250 [cifs] [<0000000080fdb897>] SMB2_sess_setup+0x16c/0x2a0 [cifs] [<000000009af320a8>] cifs_setup_session+0x13b/0x370 [cifs] [<00000000f15d5982>] cifs_get_smb_ses+0x643/0xb90 [cifs] [<00000000fe15eb90>] mount_get_conns+0x63/0x3e0 [cifs] [<00000000768aba03>] mount_get_dfs_conns+0x16/0xa0 [cifs] [<00000000cf1cf146>] cifs_mount+0x1c2/0x9a0 [cifs] [<000000000d66b51e>] cifs_smb3_do_mount+0x10e/0x710 [cifs] [<0000000077a996c5>] smb3_get_tree+0xf4/0x200 [cifs] [<0000000094dbd041>] vfs_get_tree+0x23/0xc0 [<000000003a8561de>] path_mount+0x2d3/0xb50 [<00000000ed5c86d6>] __x64_sys_mount+0x102/0x140 [<00000000142142f3>] do_syscall_64+0x3b/0x90 [<00000000e2b89731>] entry_SYSCALL_64_after_hwframe+0x63/0xcd Fixes: a4e430c8c8ba ("cifs: replace kfree() with kfree_sensitive() for sensitive data") Signed-off-by: Paulo Alcantara (SUSE) <pc@cjr.nz> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-10-18cifs: Fix memory leak when build ntlmssp negotiate blob failedZhang Xiaoxu1-1/+1
There is a memory leak when mount cifs: unreferenced object 0xffff888166059600 (size 448): comm "mount.cifs", pid 51391, jiffies 4295596373 (age 330.596s) hex dump (first 32 bytes): fe 53 4d 42 40 00 00 00 00 00 00 00 01 00 82 00 .SMB@........... 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ backtrace: [<0000000060609a61>] mempool_alloc+0xe1/0x260 [<00000000adfa6c63>] cifs_small_buf_get+0x24/0x60 [<00000000ebb404c7>] __smb2_plain_req_init+0x32/0x460 [<00000000bcf875b4>] SMB2_sess_alloc_buffer+0xa4/0x3f0 [<00000000753a2987>] SMB2_sess_auth_rawntlmssp_negotiate+0xf5/0x480 [<00000000f0c1f4f9>] SMB2_sess_setup+0x253/0x410 [<00000000a8b83303>] cifs_setup_session+0x18f/0x4c0 [<00000000854bd16d>] cifs_get_smb_ses+0xae7/0x13c0 [<000000006cbc43d9>] mount_get_conns+0x7a/0x730 [<000000005922d816>] cifs_mount+0x103/0xd10 [<00000000e33def3b>] cifs_smb3_do_mount+0x1dd/0xc90 [<0000000078034979>] smb3_get_tree+0x1d5/0x300 [<000000004371f980>] vfs_get_tree+0x41/0xf0 [<00000000b670d8a7>] path_mount+0x9b3/0xdd0 [<000000005e839a7d>] __x64_sys_mount+0x190/0x1d0 [<000000009404c3b9>] do_syscall_64+0x35/0x80 When build ntlmssp negotiate blob failed, the session setup request should be freed. Fixes: 49bd49f983b5 ("cifs: send workstation name during ntlmssp session setup") Reviewed-by: Paulo Alcantara (SUSE) <pc@cjr.nz> Reviewed-by: Shyam Prasad N <sprasad@microsoft.com> Signed-off-by: Zhang Xiaoxu <zhangxiaoxu5@huawei.com> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-10-15smb3: improve SMB3 change notification supportSteve French1-3/+27
Change notification is a commonly supported feature by most servers, but the current ioctl to request notification when a directory is changed does not return the information about what changed (even though it is returned by the server in the SMB3 change notify response), it simply returns when there is a change. This ioctl improves upon CIFS_IOC_NOTIFY by returning the notify information structure which includes the name of the file(s) that changed and why. See MS-SMB2 2.2.35 for details on the individual filter flags and the file_notify_information structure returned. To use this simply pass in the following (with enough space to fit at least one file_notify_information structure) struct __attribute__((__packed__)) smb3_notify { uint32_t completion_filter; bool watch_tree; uint32_t data_len; uint8_t data[]; } __packed; using CIFS_IOC_NOTIFY_INFO 0xc009cf0b or equivalently _IOWR(CIFS_IOCTL_MAGIC, 11, struct smb3_notify_info) The ioctl will block until the server detects a change to that directory or its subdirectories (if watch_tree is set). Acked-by: Paulo Alcantara (SUSE) <pc@cjr.nz> Acked-by: Ronnie Sahlberg <lsahlber@redhat.com> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-10-15smb3: must initialize two ACL struct fields to zeroSteve French1-1/+2
Coverity spotted that we were not initalizing Stbz1 and Stbz2 to zero in create_sd_buf. Addresses-Coverity: 1513848 ("Uninitialized scalar variable") Cc: <stable@vger.kernel.org> Reviewed-by: Paulo Alcantara (SUSE) <pc@cjr.nz> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-10-13cifs: use ALIGN() and round_up() macrosEnzo Matsumiya1-22/+16
Improve code readability by using existing macros: Replace hardcoded alignment computations (e.g. (len + 7) & ~0x7) by ALIGN()/IS_ALIGNED() macros. Also replace (DIV_ROUND_UP(len, 8) * 8) with ALIGN(len, 8), which, if not optimized by the compiler, has the overhead of a multiplication and a division. Do the same for roundup() by replacing it by round_up() (division-less version, but requires the multiple to be a power of 2, which is always the case for us). And remove some unnecessary checks where !IS_ALIGNED() would fit, but calling round_up() directly is fine as it's a no-op if the value is already aligned. Signed-off-by: Enzo Matsumiya <ematsumiya@suse.de> Reviewed-by: Ronnie Sahlberg <lsahlber@redhat.com> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-10-13cifs: prevent copying past input buffer boundariesPaulo Alcantara1-2/+2
Prevent copying past @data buffer in smb2_validate_and_copy_iov() as the output buffer in @iov might be potentially bigger and thus copying more bytes than requested in @minbufsize. Signed-off-by: Paulo Alcantara (SUSE) <pc@cjr.nz> Reviewed-by: Ronnie Sahlberg <lsahlber@redhat.com> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-10-07cifs: replace kfree() with kfree_sensitive() for sensitive dataEnzo Matsumiya1-5/+14
Replace kfree with kfree_sensitive, or prepend memzero_explicit() in other cases, when freeing sensitive material that could still be left in memory. Signed-off-by: Enzo Matsumiya <ematsumiya@suse.de> Reported-by: kernel test robot <oliver.sang@intel.com> Link: https://lore.kernel.org/r/202209201529.ec633796-oliver.sang@intel.com Reviewed-by: Paulo Alcantara (SUSE) <pc@cjr.nz> Reviewed-by: Ronnie Sahlberg <lsahlber@redhat.com> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-10-05cifs: remove initialization valueMuhammad Usama Anjum1-1/+1
Don't initialize the rc as its value is being overwritten before its use. Reviewed-by: Paulo Alcantara (SUSE) <pc@cjr.nz> Signed-off-by: Muhammad Usama Anjum <usama.anjum@collabora.com> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-10-05smb3: add dynamic trace points for tree disconnectSteve French1-6/+10
Needed this for debugging a failing xfstest. Also change camel case for "treeName" to "tree_name" in tcon struct. Example trace output (from "trace-cmd record -e smb3_tdis*"): umount-9718 [006] ..... 5909.780244: smb3_tdis_enter: xid=206 sid=0xcf38894e tid=0x3d0b8cf8 path=\\localhost\test umount-9718 [007] ..... 5909.780878: smb3_tdis_done: xid=206 sid=0xcf38894e tid=0x3d0b8cf8 Reviewed-by: Paulo Alcantara (SUSE) <pc@cjr.nz> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-10-05cifs: Fix the error length of VALIDATE_NEGOTIATE_INFO messageZhang Xiaoxu1-2/+2
Commit d5c7076b772a ("smb3: add smb3.1.1 to default dialect list") extend the dialects from 3 to 4, but forget to decrease the extended length when specific the dialect, then the message length is larger than expected. This maybe leak some info through network because not initialize the message body. After apply this patch, the VALIDATE_NEGOTIATE_INFO message length is reduced from 28 bytes to 26 bytes. Fixes: d5c7076b772a ("smb3: add smb3.1.1 to default dialect list") Signed-off-by: Zhang Xiaoxu <zhangxiaoxu5@huawei.com> Cc: <stable@vger.kernel.org> Acked-by: Paulo Alcantara (SUSE) <pc@cjr.nz> Reviewed-by: Tom Talpey <tom@talpey.com> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-08-30cifs: fix small mempool leak in SMB2_negotiate()Enzo Matsumiya1-5/+7
In some cases of failure (dialect mismatches) in SMB2_negotiate(), after the request is sent, the checks would return -EIO when they should be rather setting rc = -EIO and jumping to neg_exit to free the response buffer from mempool. Signed-off-by: Enzo Matsumiya <ematsumiya@suse.de> Cc: stable@vger.kernel.org Reviewed-by: Ronnie Sahlberg <lsahlber@redhat.com> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-08-24cifs: skip extra NULL byte in filenamesPaulo Alcantara1-10/+6
Since commit: cifs: alloc_path_with_tree_prefix: do not append sep. if the path is empty alloc_path_with_tree_prefix() function was no longer including the trailing separator when @path is empty, although @out_len was still assuming a path separator thus adding an extra byte to the final filename. This has caused mount issues in some Synology servers due to the extra NULL byte in filenames when sending SMB2_CREATE requests with SMB2_FLAGS_DFS_OPERATIONS set. Fix this by checking if @path is not empty and then add extra byte for separator. Also, do not include any trailing NULL bytes in filename as MS-SMB2 requires it to be 8-byte aligned and not NULL terminated. Cc: stable@vger.kernel.org Fixes: 7eacba3b00a3 ("cifs: alloc_path_with_tree_prefix: do not append sep. if the path is empty") Signed-off-by: Paulo Alcantara (SUSE) <pc@cjr.nz> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-08-19cifs: move from strlcpy with unused retval to strscpyWolfram Sang1-1/+1
Follow the advice of the below link and prefer 'strscpy' in this subsystem. Conversion is 1:1 because the return value is not used. Generated by a coccinelle script. Link: https://lore.kernel.org/r/CAHk-=wgfRnXz0W3D37d01q3JFkr_i_uTL=V6A6G1oUZcprmknw@mail.gmail.com/ Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-08-17cifs: remove useless parameter 'is_fsctl' from SMB2_ioctl()Enzo Matsumiya1-11/+9
SMB2_ioctl() is always called with is_fsctl = true, so doesn't make any sense to have it at all. Thus, always set SMB2_0_IOCTL_IS_FSCTL flag on the request. Also, as per MS-SMB2 3.3.5.15 "Receiving an SMB2 IOCTL Request", servers must fail the request if the request flags is zero anyway. Signed-off-by: Enzo Matsumiya <ematsumiya@suse.de> Reviewed-by: Tom Talpey <tom@talpey.com> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-08-11cifs: Do not use tcon->cfid directly, use the cfid we get from open_cached_dirRonnie Sahlberg1-1/+1
They are the same right now but tcon-> will later point to a different type of struct containing a list of cfids. Signed-off-by: Ronnie Sahlberg <lsahlber@redhat.com> Reviewed-by: Paulo Alcantara (SUSE) <pc@cjr.nz> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-08-11cifs: Move cached-dir functions into a separate fileRonnie Sahlberg1-1/+2
Also rename crfid to cfid to have consistent naming for this variable. This commit does not change any logic. Signed-off-by: Ronnie Sahlberg <lsahlber@redhat.com> Reviewed-by: Paulo Alcantara (SUSE) <pc@cjr.nz> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-08-05cifs: remove useless DeleteMidQEntry()Enzo Matsumiya1-3/+3
DeleteMidQEntry() was just a proxy for cifs_mid_q_entry_release(). - remove DeleteMidQEntry() - rename cifs_mid_q_entry_release() to release_mid() - rename kref_put() callback _cifs_mid_q_entry_release to __release_mid - rename AllocMidQEntry() to alloc_mid() - rename cifs_delete_mid() to delete_mid() Update callers to use new names. Signed-off-by: Enzo Matsumiya <ematsumiya@suse.de> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-08-01cifs: avoid use of global locks for high contention dataShyam Prasad N1-12/+12
During analysis of multichannel perf, it was seen that the global locks cifs_tcp_ses_lock and GlobalMid_Lock, which were shared between various data structures were causing a lot of contention points. With this change, we're breaking down the use of these locks by introducing new locks at more granular levels. i.e. server->srv_lock, ses->ses_lock and tcon->tc_lock to protect the unprotected fields of server, session and tcon structs; and server->mid_lock to protect mid related lists and entries at server level. Signed-off-by: Shyam Prasad N <sprasad@microsoft.com> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-08-01cifs: remove unnecessary type castingsYu Zhe1-1/+1
remove unnecessary void* type castings. Signed-off-by: Yu Zhe <yuzhe@nfschina.com> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-07-13smb3: workaround negprot bug in some Samba serversSteve French1-6/+7
Mount can now fail to older Samba servers due to a server bug handling padding at the end of the last negotiate context (negotiate contexts typically are rounded up to 8 bytes by adding padding if needed). This server bug can be avoided by switching the order of negotiate contexts, placing a negotiate context at the end that does not require padding (prior to the recent netname context fix this was the case on the client). Fixes: 73130a7b1ac9 ("smb3: fix empty netname context on secondary channels") Reported-by: Julian Sikorski <belegdol@gmail.com> Tested-by: Julian Sikorski <belegdol+github@gmail.com> Reviewed-by: Shyam Prasad N <sprasad@microsoft.com> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-06-22smb3: use netname when available on secondary channelsShyam Prasad N1-2/+9
Some servers do not allow null netname contexts, which would cause multichannel to revert to single channel when mounting to some servers (e.g. Azure xSMB). The previous patch fixed that by avoiding incorrectly sending the netname context when there would be a null hostname sent in the netname context, while this patch fixes the null hostname for the secondary channel by using the hostname of the primary channel for the secondary channel. Fixes: 4c14d7043fede ("cifs: populate empty hostnames for extra channels") Signed-off-by: Shyam Prasad N <sprasad@microsoft.com> Reviewed-by: Paulo Alcantara (SUSE) <pc@cjr.nz> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-06-20smb3: fix empty netname context on secondary channelsSteve French1-6/+8
Some servers do not allow null netname contexts, which would cause multichannel to revert to single channel when mounting to some servers (e.g. Azure xSMB). Fixes: 4c14d7043fede ("cifs: populate empty hostnames for extra channels") Reviewed-by: Shyam Prasad N <sprasad@microsoft.com> Reviewed-by: Paulo Alcantara (SUSE) <pc@cjr.nz> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-06-16smb3: add trace point for SMB2_set_eofSteve French1-0/+2
In order to debug problems with file size being reported incorrectly temporarily (in this case xfstest generic/584 intermittent failure) we need to add trace point for the non-compounded code path where we set the file size (SMB2_set_eof). The new trace point is: "smb3_set_eof" Here is sample output from the tracepoint: TASK-PID CPU# ||||| TIMESTAMP FUNCTION | | | ||||| | | xfs_io-75403 [002] ..... 95219.189835: smb3_set_eof: xid=221 sid=0xeef1cbd2 tid=0x27079ee6 fid=0x52edb58c offset=0x100000 aio-dio-append--75418 [010] ..... 95219.242402: smb3_set_eof: xid=226 sid=0xeef1cbd2 tid=0x27079ee6 fid=0xae89852d offset=0x0 Reviewed-by: Paulo Alcantara (SUSE) <pc@cjr.nz> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-06-06cifs: return errors during session setup during reconnectsShyam Prasad N1-0/+3
During reconnects, we check the return value from cifs_negotiate_protocol, and have handlers for both success and failures. But if that passes, and cifs_setup_session returns any errors other than -EACCES, we do not handle that. This fix adds a handler for that, so that we don't go ahead and try a tree_connect on a failed session. Signed-off-by: Shyam Prasad N <sprasad@microsoft.com> Reviewed-by: Enzo Matsumiya <ematsumiya@suse.de> Cc: stable@vger.kernel.org Signed-off-by: Steve French <stfrench@microsoft.com>
2022-06-01cifs: fix potential deadlock in direct reclaimVincent Whitchurch1-3/+3
The srv_mutex is used during writeback so cifs should ensure that allocations done when that mutex is held are done with GFP_NOFS, to avoid having direct reclaim ending up waiting for the same mutex and causing a deadlock. This is detected by lockdep with the splat below: ====================================================== WARNING: possible circular locking dependency detected 5.18.0 #70 Not tainted ------------------------------------------------------ kswapd0/49 is trying to acquire lock: ffff8880195782e0 (&tcp_ses->srv_mutex){+.+.}-{3:3}, at: compound_send_recv but task is already holding lock: ffffffffa98e66c0 (fs_reclaim){+.+.}-{0:0}, at: balance_pgdat which lock already depends on the new lock. the existing dependency chain (in reverse order) is: -> #1 (fs_reclaim){+.+.}-{0:0}: fs_reclaim_acquire kmem_cache_alloc_trace __request_module crypto_alg_mod_lookup crypto_alloc_tfm_node crypto_alloc_shash cifs_alloc_hash smb311_crypto_shash_allocate smb311_update_preauth_hash compound_send_recv cifs_send_recv SMB2_negotiate smb2_negotiate cifs_negotiate_protocol cifs_get_smb_ses cifs_mount cifs_smb3_do_mount smb3_get_tree vfs_get_tree path_mount __x64_sys_mount do_syscall_64 entry_SYSCALL_64_after_hwframe -> #0 (&tcp_ses->srv_mutex){+.+.}-{3:3}: __lock_acquire lock_acquire __mutex_lock mutex_lock_nested compound_send_recv cifs_send_recv SMB2_write smb2_sync_write cifs_write cifs_writepage_locked cifs_writepage shrink_page_list shrink_lruvec shrink_node balance_pgdat kswapd kthread ret_from_fork other info that might help us debug this: Possible unsafe locking scenario: CPU0 CPU1 ---- ---- lock(fs_reclaim); lock(&tcp_ses->srv_mutex); lock(fs_reclaim); lock(&tcp_ses->srv_mutex); *** DEADLOCK *** 1 lock held by kswapd0/49: #0: ffffffffa98e66c0 (fs_reclaim){+.+.}-{0:0}, at: balance_pgdat stack backtrace: CPU: 2 PID: 49 Comm: kswapd0 Not tainted 5.18.0 #70 Call Trace: <TASK> dump_stack_lvl dump_stack print_circular_bug.cold check_noncircular __lock_acquire lock_acquire __mutex_lock mutex_lock_nested compound_send_recv cifs_send_recv SMB2_write smb2_sync_write cifs_write cifs_writepage_locked cifs_writepage shrink_page_list shrink_lruvec shrink_node balance_pgdat kswapd kthread ret_from_fork </TASK> Fix this by using the memalloc_nofs_save/restore APIs around the places where the srv_mutex is held. Do this in a wrapper function for the lock/unlock of the srv_mutex, and rename the srv_mutex to avoid missing call sites in the conversion. Note that there is another lockdep warning involving internal crypto locks, which was masked by this problem and is visible after this fix, see the discussion in this thread: https://lore.kernel.org/all/20220523123755.GA13668@axis.com/ Link: https://lore.kernel.org/r/CANT5p=rqcYfYMVHirqvdnnca4Mo+JQSw5Qu12v=kPfpk5yhhmg@mail.gmail.com/ Reported-by: Shyam Prasad N <nspmangalore@gmail.com> Suggested-by: Lars Persson <larper@axis.com> Reviewed-by: Ronnie Sahlberg <lsahlber@redhat.com> Reviewed-by: Enzo Matsumiya <ematsumiya@suse.de> Signed-off-by: Vincent Whitchurch <vincent.whitchurch@axis.com> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-05-24cifs: use new enum for ses_statusShyam Prasad N1-1/+1
ses->status today shares statusEnum with server->tcpStatus. This has been confusing, and tcon->status has deviated to use a new enum. Follow suit and use new enum for ses_status as well. Signed-off-by: Shyam Prasad N <sprasad@microsoft.com> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-05-24cifs: do not use tcpStatus after negotiate completesShyam Prasad N1-1/+2
Recent changes to multichannel to allow channel reconnects to work in parallel and independent of each other did so by making use of tcpStatus for the connection, and status for the session. However, this did not take into account the multiuser scenario, where same connection is used by multiple connections. However, tcpStatus should be tracked only till the end of negotiate exchange, and not used for session setup. This change fixes this. Signed-off-by: Shyam Prasad N <sprasad@microsoft.com> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-03-28smb3: cleanup and clarify status of tree connectionsSteve French1-2/+2
Currently the way the tid (tree connection) status is tracked is confusing. The same enum is used for structs cifs_tcon and cifs_ses and TCP_Server_info, but each of these three has different states that they transition among. The current code also unnecessarily uses camelCase. Convert from use of statusEnum to a new tid_status_enum for tree connections. The valid states for a tid are: TID_NEW = 0, TID_GOOD, TID_EXITING, TID_NEED_RECON, TID_NEED_TCON, TID_IN_TCON, TID_NEED_FILES_INVALIDATE, /* unused, considering removing in future */ TID_IN_FILES_INVALIDATE It also removes CifsNeedTcon, CifsInTcon, CifsNeedFilesInvalidate and CifsInFilesInvalidate from the statusEnum used for session and TCP_Server_Info since they are not relevant for those. A follow on patch will fix the places where we use the tcon->need_reconnect flag to be more consistent with the tid->status. Also fixes a bug that was: Reported-by: kernel test robot <lkp@intel.com> Reviewed-by: Shyam Prasad N <sprasad@microsoft.com> Reviewed-by: Ronnie Sahlberg <lsahlber@redhat.com> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-03-23cifs: fix incorrect use of list iterator after the loopXiaomeng Tong1-2/+4
The bug is here: if (!tcon) { resched = true; list_del_init(&ses->rlist); cifs_put_smb_ses(ses); Because the list_for_each_entry() never exits early (without any break/goto/return inside the loop), the iterator 'ses' after the loop will always be an pointer to a invalid struct containing the HEAD (&pserver->smb_ses_list). As a result, the uses of 'ses' above will lead to a invalid memory access. The original intention should have been to walk each entry 'ses' in '&tmp_ses_list', delete '&ses->rlist' and put 'ses'. So fix it with a list_for_each_entry_safe(). Cc: stable@vger.kernel.org # 5.17 Fixes: 3663c9045f51a ("cifs: check reconnects for channels of active tcons too") Signed-off-by: Xiaomeng Tong <xiam0nd.tong@gmail.com> Reviewed-by: Shyam Prasad N <sprasad@microsoft.com> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-03-23cifs: fix bad fids sent over wirePaulo Alcantara1-35/+28
The client used to partially convert the fids to le64, while storing or sending them by using host endianness. This broke the client on big-endian machines. Instead of converting them to le64, store them as opaque integers and then avoid byteswapping when sending them over wire. Signed-off-by: Paulo Alcantara (SUSE) <pc@cjr.nz> Reviewed-by: Namjae Jeon <linkinjeon@kernel.org> Reviewed-by: Tom Talpey <tom@talpey.com> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-01-19smb3: send NTLMSSP version informationSteve French1-1/+1
For improved debugging it can be helpful to send version information as other clients do during NTLMSSP negotiation. See protocol document MS-NLMP section 2.2.1.1 Set the major and minor versions based on the kernel version, and the BuildNumber based on the internal cifs.ko module version number, and following the recommendation in the protocol documentation (MS-NLMP section 2.2.10) we set the NTLMRevisionCurrent field to 15. Reviewed-by: Shyam Prasad N <sprasad@microsoft.com> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-01-19cifs: make status checks in version independent callersShyam Prasad N1-14/+0
The status of tcp session, smb session and tcon have the same flow, irrespective of the SMB version used. Hence these status checks and updates should happen in the version independent callers of these commands. Signed-off-by: Shyam Prasad N <sprasad@microsoft.com> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-01-19cifs: remove unused variable ses_selectedMuhammad Usama Anjum1-3/+3
ses_selected is being declared and set at several places. It is not being used. Remove it. Signed-off-by: Muhammad Usama Anjum <usama.anjum@collabora.com> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-01-19cifs: protect all accesses to chan_* with chan_lockShyam Prasad N1-1/+3
A spin lock called chan_lock was introduced recently. But not all accesses were protected. Doing that with this change. To make sure that a channel is not freed when in use, we need to introduce a ref count. But today, we don't ever free channels. Signed-off-by: Shyam Prasad N <sprasad@microsoft.com> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-01-19cifs: fix the connection state transitions with multichannelShyam Prasad N1-1/+0
Recent changes to multichannel required some adjustments in the way connection states transitioned during/after reconnect. Also some minor fixes: 1. A pending switch of GlobalMid_Lock to cifs_tcp_ses_lock 2. Relocations of the code that logs reconnect 3. Changed some code in allocate_mid to suit the new scheme Signed-off-by: Shyam Prasad N <sprasad@microsoft.com> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-01-19cifs: check reconnects for channels of active tcons tooShyam Prasad N1-16/+66
With the new multichannel logic, when a channel needs reconnection, the tree connect and other channels can still be active. This fix will handle cases of checking for channel reconnect, when the tcon does not need reconnect. Signed-off-by: Shyam Prasad N <sprasad@microsoft.com> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-01-17cifs: alloc_path_with_tree_prefix: do not append sep. if the path is emptyEugene Korenevsky1-2/+7
alloc_path_with_tree_prefix() concatenates tree prefix and the path. Windows CIFS client does not add separator after the tree prefix if the path is empty. Let's do the same. This fixes mounting DFS namespaces with names containing non-ASCII symbols. BugLink: https://bugzilla.kernel.org/show_bug.cgi?id=215440 Signed-off-by: Eugene Korenevsky <ekorenevsky@astralinux.ru> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-01-07cifs: avoid race during socket reconnect between send and recvShyam Prasad N1-2/+1
When a TCP connection gets reestablished by the sender in cifs_reconnect, There is a chance for race condition with demultiplex thread waiting in cifs_readv_from_socket on the old socket. It will now return -ECONNRESET. This condition is handled by comparing socket pointer before and after sock_recvmsg. If the socket pointer has changed, we should not call cifs_reconnect again, but instead retry with new socket. Also fixed another bug in my prev mchan commits. We should always reestablish session (even if binding) on a channel that needs reconnection. Signed-off-by: Shyam Prasad N <sprasad@microsoft.com> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-01-07cifs: maintain a state machine for tcp/smb/tcon sessionsShyam Prasad N1-11/+5
If functions like cifs_negotiate_protocol, cifs_setup_session, cifs_tree_connect are called in parallel on different channels, each of these will be execute the requests. This maybe unnecessary in some cases, and only the first caller may need to do the work. This is achieved by having more states for the tcp/smb/tcon session status fields. And tracking the state of reconnection based on the state machine. For example: for tcp connections: CifsNew/CifsNeedReconnect -> CifsNeedNegotiate -> CifsInNegotiate -> CifsNeedSessSetup -> CifsInSessSetup -> CifsGood for smb sessions: CifsNew/CifsNeedReconnect -> CifsGood for tcon: CifsNew/CifsNeedReconnect -> CifsInFilesInvalidate -> CifsNeedTcon -> CifsInTcon -> CifsGood If any channel reconnect sees that it's in the middle of transition to CifsGood, then they can skip the function. Signed-off-by: Shyam Prasad N <sprasad@microsoft.com> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-01-07cifs: take cifs_tcp_ses_lock for status checksShyam Prasad N1-3/+18
While checking/updating status for tcp ses, smb ses or tcon, we take GlobalMid_Lock. This doesn't make any sense. Replaced it with cifs_tcp_ses_lock. Ideally, we should take a spin lock per struct. But since tcp ses, smb ses and tcon objects won't add up to a lot, I think there should not be too much contention. Also, in few other places, these are checked without locking. Added locking for these. Signed-off-by: Shyam Prasad N <sprasad@microsoft.com> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-01-02cifs: use the chans_need_reconnect bitmap for reconnect statusShyam Prasad N1-50/+64
We use the concept of "binding" when one of the secondary channel is in the process of connecting/reconnecting to the server. Till this binding process completes, and the channel is bound to an existing session, we redirect traffic from other established channels on the binding channel, effectively blocking all traffic till individual channels get reconnected. With my last set of commits, we can get rid of this binding serialization. We now have a bitmap of connection states for each channel. We will use this bitmap instead for tracking channel status. Having a bitmap also now enables us to keep the session alive, as long as even a single channel underneath is alive. Unfortunately, this also meant that we need to supply the tcp connection info for the channel during all negotiate and session setup functions. These changes have resulted in a slightly bigger code churn. However, I expect perf and robustness improvements in the mchan scenario after this change. Signed-off-by: Shyam Prasad N <sprasad@microsoft.com> Signed-off-by: Steve French <stfrench@microsoft.com>
2022-01-02cifs: track individual channel status using chans_need_reconnectShyam Prasad N1-9/+50
We needed a way to identify the channels under the smb session which are in reconnect, so that the traffic to other channels can continue. So I replaced the bool need_reconnect with a bitmask identifying all the channels that need reconnection (named chans_need_reconnect). When a channel needs reconnection, the bit corresponding to the index of the server in ses->chans is used to set this bitmask. Checking if no channels or all the channels need reconnect then becomes very easy. Also wrote some helper macros for checking and setting the bits. Signed-off-by: Shyam Prasad N <sprasad@microsoft.com> Signed-off-by: Steve French <stfrench@microsoft.com>
2021-11-23smb2: clarify rc initialization in smb2_reconnectSteve French1-1/+1
It is clearer to initialize rc at the beginning of the function. Reported-by: kernel test robot <lkp@intel.com> Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Reviewed-by: Paulo Alcantara (SUSE) <pc@cjr.nz> Signed-off-by: Steve French <stfrench@microsoft.com>
2021-11-12cifs: connect individual channel servers to primary channel serverShyam Prasad N1-1/+1
Today, we don't have any way to get the smb session for any of the secondary channels. Introducing a pointer to the primary server from server struct of any secondary channel. The value will be NULL for the server of the primary channel. This will enable us to get the smb session for any channel. This will be needed for some of the changes that I'm planning to make soon. Signed-off-by: Shyam Prasad N <sprasad@microsoft.com> Signed-off-by: Steve French <stfrench@microsoft.com>
2021-11-12smb3: add additional null check in SMB311_posix_mkdirSteve French1-0/+11
Although unlikely for it to be possible for rsp to be null here, the check is safer to add, and quiets a Coverity warning. Addresses-Coverity: 1437501 ("Explicit Null dereference") Reviewed-by: Paulo Alcantara (SUSE) <pc@cjr.nz> Signed-off-by: Steve French <stfrench@microsoft.com>
2021-11-12smb3: add additional null check in SMB2_tconSteve French1-1/+1
Although unlikely to be possible for rsp to be null here, the check is safer to add, and quiets a Coverity warning. Addresses-Coverity: 1420428 ("Explicit null dereferenced") Reviewed-by: Paulo Alcantara (SUSE) <pc@cjr.nz> Signed-off-by: Steve French <stfrench@microsoft.com>
2021-11-12smb3: add additional null check in SMB2_openSteve French1-1/+3
Although unlikely to be possible for rsp to be null here, the check is safer to add, and quiets a Coverity warning. Addresses-Coverity: 1418458 ("Explicit null dereferenced") Reviewed-by: Paulo Alcantara (SUSE) <pc@cjr.nz> Signed-off-by: Steve French <stfrench@microsoft.com>
2021-11-11smb3: add additional null check in SMB2_ioctlSteve French1-0/+10
Although unlikely for it to be possible for rsp to be null here, the check is safer to add, and quiets a Coverity warning. Addresses-Coverity: 1443909 ("Explicit Null dereference") Reviewed-by: Paulo Alcantara (SUSE) <pc@cjr.nz> Signed-off-by: Steve French <stfrench@microsoft.com>
2021-11-10cifs: support nested dfs links over reconnectPaulo Alcantara1-1/+5
Mounting a dfs link that has nested links was already supported at mount(2), so make it work over reconnect as well. Make the following case work: * mount //root/dfs/link /mnt -o ... - final share: /server/share * in server settings - change target folder of /root/dfs/link3 to /server/share2 - change target folder of /root/dfs/link2 to /root/dfs/link3 - change target folder of /root/dfs/link to /root/dfs/link2 * mount -o remount,... /mnt - refresh all dfs referrals - mark current connection for failover - cifs_reconnect() reconnects to root server - tree_connect() * checks that /root/dfs/link2 is a link, then chase it * checks that root/dfs/link3 is a link, then chase it * finally tree connect to /server/share2 If the mounted share is no longer accessible and a reconnect had been triggered, the client will retry it from both last referral path (/root/dfs/link3) and original referral path (/root/dfs/link). Any new referral paths found while chasing dfs links over reconnect, it will be updated to TCP_Server_Info::leaf_fullpath, accordingly. Signed-off-by: Paulo Alcantara (SUSE) <pc@cjr.nz> Signed-off-by: Steve French <stfrench@microsoft.com>
2021-11-08cifs: send workstation name during ntlmssp session setupShyam Prasad N1-11/+6
During the ntlmssp session setup (authenticate phases) send the client workstation info. This can make debugging easier on servers. Signed-off-by: Shyam Prasad N <sprasad@microsoft.com> Reviewed-by: Paulo Alcantara (SUSE) <pc@cjr.nz> Reviewed-by: Enzo Matsumiya <ematsumiya@suse.de> Signed-off-by: Steve French <stfrench@microsoft.com>
Copyright © 1996 – 2023 Jason A. Donenfeld. All Rights Reverse Engineered.