kernel.git/fs/nfsd, branch linux-4.4.y

nfsd: fix use-after-free due to delegation race

2021-12-22T08:04:18Z

commit 548ec0805c399c65ed66c6641be467f717833ab5 upstream. A delegation break could arrive as soon as we've called vfs_setlease. A delegation break runs a callback which immediately (in nfsd4_cb_recall_prepare) adds the delegation to del_recall_lru. If we then exit nfs4_set_delegation without hashing the delegation, it will be freed as soon as the callback is done with it, without ever being removed from del_recall_lru. Symptoms show up later as use-after-free or list corruption warnings, usually in the laundromat thread. I suspect aba2072f4523 "nfsd: grant read delegations to clients holding writes" made this bug easier to hit, but I looked as far back as v3.0 and it looks to me it already had the same problem. So I'm not sure where the bug was introduced; it may have been there from the beginning. Cc: stable@vger.kernel.org Signed-off-by: J. Bruce Fields [Salvatore Bonaccorso: Backport for context changes to versions which do not have 20b7d86f29d3 ("nfsd: use boottime for lease expiry calculation")] Signed-off-by: Salvatore Bonaccorso Signed-off-by: Greg Kroah-Hartman

NFSD: Keep existing listeners on portlist error

2021-10-27T07:32:39Z

[ Upstream commit c20106944eb679fa3ab7e686fe5f6ba30fbc51e5 ] If nfsd has existing listening sockets without any processes, then an error returned from svc_create_xprt() for an additional transport will remove those existing listeners. We're seeing this in practice when userspace attempts to create rpcrdma transports without having the rpcrdma modules present before creating nfsd kernel processes. Fix this by checking for existing sockets before calling nfsd_destroy(). Signed-off-by: Benjamin Coddington Signed-off-by: Chuck Lever Signed-off-by: Sasha Levin

nfsd4: Handle the NFSv4 READDIR 'dircount' hint being zero

2021-10-17T08:04:43Z

commit f2e717d655040d632c9015f19aa4275f8b16e7f2 upstream. RFC3530 notes that the 'dircount' field may be zero, in which case the recommendation is to ignore it, and only enforce the 'maxcount' field. In RFC5661, this recommendation to ignore a zero valued field becomes a requirement. Fixes: aee377644146 ("nfsd4: fix rd_dircount enforcement") Cc: Signed-off-by: Trond Myklebust Signed-off-by: Chuck Lever Signed-off-by: Greg Kroah-Hartman

nfsd4: readdirplus shouldn't return parent of export

2021-01-23T14:36:56Z

commit 51b2ee7d006a736a9126e8111d1f24e4fd0afaa6 upstream. If you export a subdirectory of a filesystem, a READDIRPLUS on the root of that export will return the filehandle of the parent with the ".." entry. The filehandle is optional, so let's just not return the filehandle for ".." if we're at the root of an export. Note that once the client learns one filehandle outside of the export, they can trivially access the rest of the export using further lookups. However, it is also not very difficult to guess filehandles outside of the export. So exporting a subdirectory of a filesystem should considered equivalent to providing access to the entire filesystem. To avoid confusion, we recommend only exporting entire filesystems. Reported-by: Youjipeng Signed-off-by: J. Bruce Fields Cc: stable@vger.kernel.org Signed-off-by: Chuck Lever Signed-off-by: Greg Kroah-Hartman

nfsd: Fix svc_xprt refcnt leak when setup callback client failed

2020-06-30T00:07:49Z

[ Upstream commit a4abc6b12eb1f7a533c2e7484cfa555454ff0977 ] nfsd4_process_cb_update() invokes svc_xprt_get(), which increases the refcount of the "c->cn_xprt". The reference counting issue happens in one exception handling path of nfsd4_process_cb_update(). When setup callback client failed, the function forgets to decrease the refcnt increased by svc_xprt_get(), causing a refcnt leak. Fix this issue by calling svc_xprt_put() when setup callback client failed. Signed-off-by: Xiyu Yang Signed-off-by: Xin Tan Signed-off-by: J. Bruce Fields Signed-off-by: Sasha Levin

nfsd: Return EPERM, not EACCES, in some SETATTR cases

2019-12-21T09:34:45Z

[ Upstream commit 255fbca65137e25b12bced18ec9a014dc77ecda0 ] As the man(2) page for utime/utimes states, EPERM is returned when the second parameter of utime or utimes is not NULL, the caller's effective UID does not match the owner of the file, and the caller is not privileged. However, in a NFS directory mounted from knfsd, it will return EACCES (from nfsd_setattr-> fh_verify->nfsd_permission). This patch fixes that. Signed-off-by: zhengbin Signed-off-by: J. Bruce Fields Signed-off-by: Sasha Levin

nfsd: fix a warning in __cld_pipe_upcall()

2019-12-21T09:34:39Z

[ Upstream commit b493fd31c0b89d9453917e977002de58bebc3802 ] __cld_pipe_upcall() emits a "do not call blocking ops when !TASK_RUNNING" warning due to the dput() call in rpc_queue_upcall(). Fix it by using a completion instead of hand coding the wait. Signed-off-by: Scott Mayhew Signed-off-by: J. Bruce Fields Signed-off-by: Sasha Levin

nfsd: Fix overflow causing non-working mounts on 1 TB machines

2019-08-04T07:34:55Z

[ Upstream commit 3b2d4dcf71c4a91b420f835e52ddea8192300a3b ] Since commit 10a68cdf10 (nfsd: fix performance-limiting session calculation) (Linux 5.1-rc1 and 4.19.31), shares from NFS servers with 1 TB of memory cannot be mounted anymore. The mount just hangs on the client. The gist of commit 10a68cdf10 is the change below. -avail = clamp_t(int, avail, slotsize, avail/3); +avail = clamp_t(int, avail, slotsize, total_avail/3); Here are the macros. #define min_t(type, x, y) __careful_cmp((type)(x), (type)(y), <) #define clamp_t(type, val, lo, hi) min_t(type, max_t(type, val, lo), hi) `total_avail` is 8,434,659,328 on the 1 TB machine. `clamp_t()` casts the values to `int`, which for 32-bit integers can only hold values −2,147,483,648 (−2^31) through 2,147,483,647 (2^31 − 1). `avail` (in the function signature) is just 65536, so that no overflow was happening. Before the commit the assignment would result in 21845, and `num = 4`. When using `total_avail`, it is causing the assignment to be 18446744072226137429 (printed as %lu), and `num` is then 4164608182. My next guess is, that `nfsd_drc_mem_used` is then exceeded, and the server thinks there is no memory available any more for this client. Updating the arguments of `clamp_t()` and `min_t()` to `unsigned long` fixes the issue. Now, `avail = 65536` (before commit 10a68cdf10 `avail = 21845`), but `num = 4` remains the same. Fixes: c54f24e338ed (nfsd: fix performance-limiting session calculation) Cc: stable@vger.kernel.org Signed-off-by: Paul Menzel Signed-off-by: J. Bruce Fields Signed-off-by: Sasha Levin

nfsd: fix performance-limiting session calculation

2019-08-04T07:34:55Z

[ Upstream commit c54f24e338ed2a35218f117a4a1afb5f9e2b4e64 ] We're unintentionally limiting the number of slots per nfsv4.1 session to 10. Often more than 10 simultaneous RPCs are needed for the best performance. This calculation was meant to prevent any one client from using up more than a third of the limit we set for total memory use across all clients and sessions. Instead, it's limiting the client to a third of the maximum for a single session. Fix this. Reported-by: Chris Tracy Cc: stable@vger.kernel.org Fixes: de766e570413 "nfsd: give out fewer session slots as limit approaches" Signed-off-by: J. Bruce Fields Signed-off-by: Sasha Levin

nfsd: give out fewer session slots as limit approaches

2019-08-04T07:34:55Z

[ Upstream commit de766e570413bd0484af0b580299b495ada625c3 ] Instead of granting client's full requests until we hit our DRC size limit and then failing CREATE_SESSIONs (and hence mounts) completely, start granting clients smaller slot tables as we approach the limit. The factor chosen here is pretty much arbitrary. Signed-off-by: J. Bruce Fields Signed-off-by: Sasha Levin