-----Original Message----- From: Daniel Gryniewicz [mailto:dang@redhat.com] Sent: Tuesday, January 29, 2019 2:46 PM To: Ganesha-devel <devel(a)lists.nfs-ganesha.org&gt; Subject: [NFS-Ganesha-Devel] Proposed 2.7.2 Here's my proposed 2.7.2 patchset.  I've only included bug fixes, except for the requested export cleanup.  If you have other changes you want in, please propose them in this thread.

Here's my proposed 2.7.2 patchset.  I've only included bug fixes, except for the requested export cleanup.  If you have other changes you want in, please propose them in this thread.

I'm not sure about that one. It's a complicated backport, due to all the changes in the compound code related to async. What do people think? On Wed, Jan 30, 2019 at 11:36 AM Sriram Patil <sriramp(a)vmware.com&gt; wrote: > > How about including recent session refcnt fixes? > > 85a9ef5 Add refcount to COMPOUND4res_extended to protect slot cache > > - Sriram > > On 1/30/19, 1:16 AM, "Daniel Gryniewicz" <dang(a)redhat.com&gt; wrote: > > Here's my proposed 2.7.2 patchset. I've only included bug fixes, except > for the requested export cleanup. If you have other changes you want > in, please propose them in this thread. > > 2f6d8b7d6 (HEAD -> V2.7-stable) Remove -Wabi from maintainer mode > 949de68f8 Fix READDIR duplicate entries > 1249a2723 Fix attribute comparison in NFS4_OP_VERIFY > 20ed57717 Fix GTest build > ba756d3ce nfs4_op_open: NULL pointer bug - replace goto out3 with return > ebcde738d Don't call nfs_req_creds if we don't have export > 7b5c6c833 Handle race while adding host to IP-name cache > 9edc1fbab exports: prune off old exports after reloading config > 8ea299930 exports: copy the config tree generation to the export when > adding or updating > 914469434 exports: add generation counter to config_root and helpers to > fetch it > d0d4d0201 exports: don't allow dbus remove to leave subexports > disconnected in pseudoroot > 5d20939bf ganesha_status.py - fix missing parens > 9fa5bb085 rpm/selinux: fix %pre install and %postun uninstall of selinux > policy > 1d5e66792 Fix to set ATTR_RDATTR_ERR correctly as valid_mask in open2. > 1d8f0d242 CEPH: do a getattr after creating a dir and applying extra > attributes > b62c60b49 Read/Write - Don't leak owners/states > 7e7c01f23 SAL: fix multiple state reference leaks > 3be24ba32 Set op_ctx for lock_avail and lock_grant > ad2b9a30a FSAL_VFS, FreeBSD: upstream support for d_off > c4119a3c5 Add checks for client access in NFS user access checks. > a1b0888f3 FSAL_UP: add missing locks in async delegation recall > 0eb7faf67 Handle NLM share FREE_ALL for windows clients > 6cdc3fb22 NLM share reservation access check with owner_skip > 770db8de3 Fixed dereferencing null ctx_export on nfs_read > 789b56d35 fix the bug of import config form rados url > 01f6e20d3 Remove idmap entry only if it is present in uid_tree > 218d7f8ec MDCACHE: fix invalid assert in mdcache_avl_lookup_ck() > 0fc6e3a2a Restore op_ctx->ctx_export and fsal_export in nfs4_op_readdir > adce80154 Fix incorrect parent file handle update on rename. > 7a22a86fc Fix NFSv3 EOF handling. > 738550a12 FSAL_GLUSTER/FSAL_GPFS: Do not error out if setattr ATTR_ACL > with empty acl > 6fce461ad FSAL_GLUSTER: fix memory leak of acl_t > 1b5ba9978 [GPFS] Handle failures with NFS readdir operation > ae73a8f32 FSAL_GLUSTER: Copy user creds and lease_id after reopening fd > 3ea561a75 Acquire state's fdlock in close2 > 9383863eb Remove unexported export on DBUS unexport. > 40c7b7d54 rpc_callback : check return value for > nfs_rpc_create_chan_v40() properly > c86969ef0 FSAL_PROXY : module options > e23eaf9c9 SAL: Fix dead lock in revoke_owner_delegs() > > > Daniel > _______________________________________________ > Devel mailing list -- devel(a)lists.nfs-ganesha.org > To unsubscribe send an email to devel-leave(a)lists.nfs-ganesha.org > > _______________________________________________ Devel mailing list -- devel(a)lists.nfs-ganesha.org To unsubscribe send an email to devel-leave(a)lists.nfs-ganesha.org

On Fri, Feb 1, 2019 at 2:07 PM Matt Benjamin <mbenjami(a)redhat.com <mailto:mbenjami@redhat.com>> wrote: Defer to Frank and Kaleb on this. Frank and Kaleb? Matt On Wed, Jan 30, 2019 at 11:48 AM Daniel Gryniewicz <dang(a)redhat.com <mailto:dang@redhat.com>> wrote: > > I'm not sure about that one.  It's a complicated backport, due to all > the changes in the compound code related to async.  What do people > think? If Dan thinks it's complex.... We want to rebase our downstream to 2.7.x. While normally I like to be aggressive, I think in this case I'd like to split the difference and _not_ take the complex patch and rebase to 2.7.2 with confidence that it's reasonably solid. Then if you want we can take the complex patch into 2.7.3 in another month or two and let that soak upstream. Which would be good to get some miles on it as our next rebase after that will probably be to 2.8.x. My 2¢

and 9d1b48b Take state_lock before checking for deleg conflicts

Thanks, Soumya On 2/4/19 10:02 PM, Soumya Koduri wrote: > Please consider below fix as well - > > 4ac1eaf SAL: Handle open claim of type OPEN_DELEG_CUR_FH