August 2024 - Devel - NFS Ganesha List Archives

Announce Push of V6-dev.20
by Frank Filz 21 Aug '24

21 Aug '24

Branch next NOTE: If you jump straight to this merge, the previous merge had an ntirpc pullup, please update your submodule. NOTE: This is just about the final merge before I tag V6.0. Tomorrow I hope to merge patches that will change the formatting from the kernel (verified with checkpatch.pl) to the style preferred by clang. This will make it easier for folks to integrate with their development environment. Look for the instructions on how to make sure you have the right commit hooks installed so your code is validated against this. Note that the Continuous Integration test platform needs to change to reflect the new check, that may take some time, so ignore any checkpatch warnings and errors raised in Gerrithub for patches submitted. Tag:V6-dev.20 Merge Highlights * Make lttng easier to use Signed-off-by: Frank S. Filz <ffilzlnx(a)mindspring.com> Contents: d5599e564 Frank S. Filz V6-dev.20 63202a37f Shahar Hochma Add libntirpc_tracepoints.so to Ganesha lttng package 6001bb2b8 Shahar Hochma Change existing tracepoints to use generated tracepoints b8ec0efcb Shahar Hochma Integrate with trace generator 9d3db6cf9 Shahar Hochma Allow loading of lttng traces on Ganesha startup

1 0

Announce Push of V6-dev.19
by Frank Filz 21 Aug '24

21 Aug '24

Branch next Tag:V6-dev.19 NOTE: This merge includes and ntirpc pullup, please update your submodule Merge Highlights * pullup ntripc 6.0.1 * RDMA fixes (includes ntirpc pullup) * podman script fixes * Limit the number of files we delegate * connection_manager: Add connection_manager__connection_init * nfs_rpc_dispatcher_thread: Move connection manager start from connection establishment. * sal_functions.h: Increase size of DISPLAY_CLIENTID_SIZE to 37 * STATE: fix a state_t refcount leak and a couple minor cleanups * SAL/nfs4_recovery.c: fix the issue where adding export fails when the gethostname function call fails * nfs_main.c: add prctl(PR_SET_IO_FLUSHER,1) call to avoid kernel deadlocks * Fix fsal not supporting fso_allocate_own_read_buffer Signed-off-by: Frank S. Filz <ffilzlnx(a)mindspring.com> Contents: ced8e8578 Frank S. Filz V6-dev.19 240e6c2b6 Frank S. Filz pull up ntirpc 6.0.1 b9c94834b Gaurav Gangalwar Fix fsal not supporting fso_allocate_own_read_buffer 05d7cbcdb Gaurav Gangalwar Fix XDR_GETPOS for RDMA f15c95ed4 Gaurav Gangalwar RDMA fixes to serve reads not aligned on BYTES_PER_XDR_UNIT affce098c Jason Woodward nfs_main.c: add prctl(PR_SET_IO_FLUSHER,1) call to avoid kernel deadlocks 141c75e36 izxl007 SAL/nfs4_recovery.c: fix the issue where adding export fails when the gethostname function call fails fd7dc2254 Frank S. Filz STATE: fix a state_t refcount leak and a couple minor cleanups 7e5f36ac4 Miki Ivry sal_functions.h: Increase size of DISPLAY_CLIENTID_SIZE to 37 65944cbd3 Miki Ivry nfs_rpc_dispatcher_thread: Move connection manager start from connection establishment. 792ee9146 Miki Ivry connection_manager: Add connection_manager__connection_init 5bebefe02 Deepak Arumugam Sankara Subramanian Limit the number of files we delegate 39f4b7c01 Martin Schwenke scripts: Allow command to be specified for a specified distro 922a04217 Martin Schwenke scripts: Drop support for CentOS 7 and Ubuntu 18.04 15ac4ac2f Martin Schwenke scripts: Try harder to create a matching user in the container f5c4d20bc Martin Schwenke scripts: Add podman SUSE support dff0429a3 Martin Schwenke scripts: Use login shell for interactive podman shell 701afc99e Martin Schwenke scripts: Hack around broken CentOS 8 Stream container repositories 88d2910ea Martin Schwenke scripts: Group all the distros that come from docker.io/library 162e462d1 Martin Schwenke scripts: Add delete-all operation for ganesha-container 3926ac977 Martin Schwenke scripts: Add some newer Linux distribution versions 4ce884b7b Martin Schwenke scripts: Add missing Debian dependency for container builds

1 0

[M] Change in ...nfs-ganesha[next]: FSAL_CEPH: Add callbacks for inode & dentry invalidation
by Sachin Punadikar (GerritHub) 21 Aug '24

21 Aug '24

Sachin Punadikar has uploaded this change for review. ( https://review.gerrithub.io/c/ffilz/nfs-ganesha/+/1199819?usp=email ) Change subject: FSAL_CEPH: Add callbacks for inode & dentry invalidation ...................................................................... FSAL_CEPH: Add callbacks for inode & dentry invalidation Ganesha need to register callabcks to handle events raised by ceph client like dentry and inode invalidation. On receiving inode invalidation from ceph, ganesha should also invalidate its cache, and on receiving debtry invalidation, ganesha should close the related fsal handle and invalidate the cache. Change-Id: I56c5f79f82d2e5cfeaee0b938549097092a96df4 Signed-off-by: Sachin Punadikar <sachin.punadikar(a)ibm.com> --- M src/FSAL/FSAL_CEPH/main.c 1 file changed, 58 insertions(+), 0 deletions(-) git pull ssh://review.gerrithub.io:29418/ffilz/nfs-ganesha refs/changes/19/1199819/1 -- To view, visit https://review.gerrithub.io/c/ffilz/nfs-ganesha/+/1199819?usp=email To unsubscribe, or for help writing mail filters, visit https://review.gerrithub.io/settings?usp=email Gerrit-MessageType: newchange Gerrit-Project: ffilz/nfs-ganesha Gerrit-Branch: next Gerrit-Change-Id: I56c5f79f82d2e5cfeaee0b938549097092a96df4 Gerrit-Change-Number: 1199819 Gerrit-PatchSet: 1 Gerrit-Owner: Sachin Punadikar <sachin.punadikar(a)ibm.com>

1 0

[S] Change in ...nfs-ganesha[next]: Fix fsal not supporting fso_allocate_own_read_buffer
by Gaurav (GerritHub) 20 Aug '24

20 Aug '24

Gaurav has uploaded this change for review. ( https://review.gerrithub.io/c/ffilz/nfs-ganesha/+/1199736?usp=email ) Change subject: Fix fsal not supporting fso_allocate_own_read_buffer ...................................................................... Fix fsal not supporting fso_allocate_own_read_buffer We use io_data to set release method and release_data. But it could result in use after free as pointed by valgrind. So keep copy of io_data instead of referencing original one. ==144581== Invalid read of size 8 ==144581== at 0x4E8A3D7: fsal_iov_release (fsal_helper.c:1900) ==144581== by 0x4EF705D: xdr_io_data_uio_release (rpc_tools.c:500) ==144581== by 0x530EA08: xdr_ioq_uv_release (xdr_ioq.c:568) ==144581== by 0x53100FE: xdr_ioq_release (xdr_ioq.c:1210) ==144581== by 0x53101CA: xdr_ioq_destroy (xdr_ioq.c:1228) ==144581== by 0x5310261: xdr_ioq_destroy_internal (xdr_ioq.c:1245) ==144581== by 0x531250C: svc_ioq_write (svc_ioq.c:404) ==144581== by 0x5312618: svc_ioq_write_now (svc_ioq.c:441) ==144581== by 0x5309C17: svc_vc_reply (svc_vc.c:1272) ==144581== by 0x52FD857: svc_sendreply (svc.c:509) ==144581== by 0x4EACEDC: complete_request (nfs_worker_thread.c:784) ==144581== by 0x4EAF5DE: nfs_rpc_process_request (nfs_worker_thread.c:1579) ==144581== Address 0x10d1f0d8 is 200 bytes inside a block of size 208 free'd ==144581== at 0x4C3B4CB: free (vg_replace_malloc.c:985) ==144581== by 0x4F93D70: gsh_free (abstract_mem.h:257) ==144581== by 0x4F959DE: nfs4_op_read (nfs4_op_read.c:892) ==144581== by 0x4F77BF4: process_one_op (nfs4_Compound.c:912) ==144581== by 0x4F78F8B: nfs4_Compound (nfs4_Compound.c:1394) ==144581== by 0x4EAF5A3: nfs_rpc_process_request (nfs_worker_thread.c:1558) ==144581== by 0x4EAFB36: nfs_rpc_valid_NFS (nfs_worker_thread.c:1778) ==144581== by 0x530999A: svc_vc_decode (svc_vc.c:1209) ==144581== by 0x53046B5: svc_request (svc_rqst.c:1236) ==144581== by 0x53098A3: svc_vc_recv (svc_vc.c:1182) ==144581== by 0x5304635: svc_rqst_xprt_task_recv (svc_rqst.c:1217) ==144581== by 0x53051CC: svc_rqst_epoll_loop (svc_rqst.c:1598) ==144581== Block was alloc'd at ==144581== at 0x4C3F963: calloc (vg_replace_malloc.c:1595) ==144581== by 0x4F9572F: nfs4_read (nfs4_op_read.c:783) ==144581== by 0x4F95986: nfs4_op_read (nfs4_op_read.c:875) ==144581== by 0x4F77BF4: process_one_op (nfs4_Compound.c:912) ==144581== by 0x4F78F8B: nfs4_Compound (nfs4_Compound.c:1394) ==144581== by 0x4EAF5A3: nfs_rpc_process_request (nfs_worker_thread.c:1558) ==144581== by 0x4EAFB36: nfs_rpc_valid_NFS (nfs_worker_thread.c:1778) ==144581== by 0x530999A: svc_vc_decode (svc_vc.c:1209) ==144581== by 0x53046B5: svc_request (svc_rqst.c:1236) ==144581== by 0x53098A3: svc_vc_recv (svc_vc.c:1182) ==144581== by 0x5304635: svc_rqst_xprt_task_recv (svc_rqst.c:1217) ==144581== by 0x53051CC: svc_rqst_epoll_loop (svc_rqst.c:1598) ==144581== ==144581== Invalid read of size 8 ==144581== at 0x4E8A3EA: fsal_iov_release (fsal_helper.c:1901) ==144581== by 0x4EF705D: xdr_io_data_uio_release (rpc_tools.c:500) ==144581== by 0x530EA08: xdr_ioq_uv_release (xdr_ioq.c:568) ==144581== by 0x53100FE: xdr_ioq_release (xdr_ioq.c:1210) ==144581== by 0x53101CA: xdr_ioq_destroy (xdr_ioq.c:1228) ==144581== by 0x5310261: xdr_ioq_destroy_internal (xdr_ioq.c:1245) ==144581== by 0x531250C: svc_ioq_write (svc_ioq.c:404) ==144581== by 0x5312618: svc_ioq_write_now (svc_ioq.c:441) ==144581== by 0x5309C17: svc_vc_reply (svc_vc.c:1272) ==144581== by 0x52FD857: svc_sendreply (svc.c:509) ==144581== by 0x4EACEDC: complete_request (nfs_worker_thread.c:784) ==144581== by 0x4EAF5DE: nfs_rpc_process_request (nfs_worker_thread.c:1579) ==144581== Address 0x10d1f0d8 is 200 bytes inside a block of size 208 free'd ==144581== at 0x4C3B4CB: free (vg_replace_malloc.c:985) ==144581== by 0x4F93D70: gsh_free (abstract_mem.h:257) ==144581== by 0x4F959DE: nfs4_op_read (nfs4_op_read.c:892) ==144581== by 0x4F77BF4: process_one_op (nfs4_Compound.c:912) ==144581== by 0x4F78F8B: nfs4_Compound (nfs4_Compound.c:1394) ==144581== by 0x4EAF5A3: nfs_rpc_process_request (nfs_worker_thread.c:1558) ==144581== by 0x4EAFB36: nfs_rpc_valid_NFS (nfs_worker_thread.c:1778) ==144581== by 0x530999A: svc_vc_decode (svc_vc.c:1209) ==144581== by 0x53046B5: svc_request (svc_rqst.c:1236) ==144581== by 0x53098A3: svc_vc_recv (svc_vc.c:1182) ==144581== by 0x5304635: svc_rqst_xprt_task_recv (svc_rqst.c:1217) ==144581== by 0x53051CC: svc_rqst_epoll_loop (svc_rqst.c:1598) ==144581== Block was alloc'd at ==144581== at 0x4C3F963: calloc (vg_replace_malloc.c:1595) ==144581== by 0x4F9572F: nfs4_read (nfs4_op_read.c:783) ==144581== by 0x4F95986: nfs4_op_read (nfs4_op_read.c:875) ==144581== by 0x4F77BF4: process_one_op (nfs4_Compound.c:912) ==144581== by 0x4F78F8B: nfs4_Compound (nfs4_Compound.c:1394) ==144581== by 0x4EAF5A3: nfs_rpc_process_request (nfs_worker_thread.c:1558) ==144581== by 0x4EAFB36: nfs_rpc_valid_NFS (nfs_worker_thread.c:1778) ==144581== by 0x530999A: svc_vc_decode (svc_vc.c:1209) ==144581== by 0x53046B5: svc_request (svc_rqst.c:1236) ==144581== by 0x53098A3: svc_vc_recv (svc_vc.c:1182) ==144581== by 0x5304635: svc_rqst_xprt_task_recv (svc_rqst.c:1217) ==144581== by 0x53051CC: svc_rqst_epoll_loop (svc_rqst.c:1598) ==144581== Change-Id: I8a9396006e4a3ce46a88eb6f93db6cd6c40180e0 Signed-off-by: Gaurav Gangalwar <gaurav.gangalwar(a)gmail.com> --- M src/FSAL/fsal_helper.c M src/RPCAL/rpc_tools.c 2 files changed, 13 insertions(+), 7 deletions(-) git pull ssh://review.gerrithub.io:29418/ffilz/nfs-ganesha refs/changes/36/1199736/1 -- To view, visit https://review.gerrithub.io/c/ffilz/nfs-ganesha/+/1199736?usp=email To unsubscribe, or for help writing mail filters, visit https://review.gerrithub.io/settings?usp=email Gerrit-MessageType: newchange Gerrit-Project: ffilz/nfs-ganesha Gerrit-Branch: next Gerrit-Change-Id: I8a9396006e4a3ce46a88eb6f93db6cd6c40180e0 Gerrit-Change-Number: 1199736 Gerrit-PatchSet: 1 Gerrit-Owner: Gaurav <gaurav.gangalwar(a)gmail.com>

1 0

[XS] Change in ...nfs-ganesha[next]: FSAL_CEPH: Invalidate mdcache instead of release
by Sachin Punadikar (GerritHub) 20 Aug '24

20 Aug '24

Sachin Punadikar has uploaded this change for review. ( https://review.gerrithub.io/c/ffilz/nfs-ganesha/+/1199716?usp=email ) Change subject: FSAL_CEPH: Invalidate mdcache instead of release ...................................................................... FSAL_CEPH: Invalidate mdcache instead of release The "try_release" upcall releases the cached entity only when its refcount is 1. This is not expected action when ceph client cache is releasing the inode. When ceph client cache releases inode, MDCACHE should release refs to the cached entity. So calling "invalidate". Change-Id: I8b76a7079e2ec7b317b71f9f53a7521e0ab5231c Signed-off-by: Sachin Punadikar <sachin.punadikar(a)ibm.com> --- M src/FSAL/FSAL_CEPH/main.c 1 file changed, 3 insertions(+), 2 deletions(-) git pull ssh://review.gerrithub.io:29418/ffilz/nfs-ganesha refs/changes/16/1199716/1 -- To view, visit https://review.gerrithub.io/c/ffilz/nfs-ganesha/+/1199716?usp=email To unsubscribe, or for help writing mail filters, visit https://review.gerrithub.io/settings?usp=email Gerrit-MessageType: newchange Gerrit-Project: ffilz/nfs-ganesha Gerrit-Branch: next Gerrit-Change-Id: I8b76a7079e2ec7b317b71f9f53a7521e0ab5231c Gerrit-Change-Number: 1199716 Gerrit-PatchSet: 1 Gerrit-Owner: Sachin Punadikar <sachin.punadikar(a)ibm.com>

1 0

[S] Change in ...nfs-ganesha[next]: nfs_main.c: add prctl(PR_SET_IO_FLUSHER,1) call to avoid kernel deadl...
by Jason Woodward (GerritHub) 19 Aug '24

19 Aug '24

Jason Woodward has uploaded this change for review. ( https://review.gerrithub.io/c/ffilz/nfs-ganesha/+/1199685?usp=email ) Change subject: nfs_main.c: add prctl(PR_SET_IO_FLUSHER,1) call to avoid kernel deadlocks ...................................................................... nfs_main.c: add prctl(PR_SET_IO_FLUSHER,1) call to avoid kernel deadlocks See https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… Setting PR_SET_IO_FLUSHER prevents recursive allocation for user space components that are in the I/O path. Otherwise a deadlock can occur when device I/O needs to allocate memory but the allocation would require device I/O to complete. Change-Id: Ia5328ba133ba3fc4950db22b0ec2aab718143b6c Signed-off-by: Jason Woodward <jawoodwa(a)microsoft.com> --- M src/MainNFSD/nfs_main.c 1 file changed, 18 insertions(+), 0 deletions(-) git pull ssh://review.gerrithub.io:29418/ffilz/nfs-ganesha refs/changes/85/1199685/1 -- To view, visit https://review.gerrithub.io/c/ffilz/nfs-ganesha/+/1199685?usp=email To unsubscribe, or for help writing mail filters, visit https://review.gerrithub.io/settings?usp=email Gerrit-MessageType: newchange Gerrit-Project: ffilz/nfs-ganesha Gerrit-Branch: next Gerrit-Change-Id: Ia5328ba133ba3fc4950db22b0ec2aab718143b6c Gerrit-Change-Number: 1199685 Gerrit-PatchSet: 1 Gerrit-Owner: Jason Woodward <jawoodwa(a)microsoft.com>

1 0

[XS] Change in ...nfs-ganesha[next]: Fix XDR_GETPOS for RDMA
by Gaurav (GerritHub) 19 Aug '24

19 Aug '24

Gaurav has uploaded this change for review. ( https://review.gerrithub.io/c/ffilz/nfs-ganesha/+/1199665?usp=email ) Change subject: Fix XDR_GETPOS for RDMA ...................................................................... Fix XDR_GETPOS for RDMA Use XDR_GETSTARTDATAPOS and XDR_GETENDDATAPOS for NFS writes. For RDMA data offsets are not inline, since we use seperate buffers for rdma_reads. If there buf1 and buf2 buf1 will have all compound ops including WRITE. buf2 will have data read using rdma_read for WRITE. Need tirpc pull https://github.com/nfs-ganesha/ntirpc/pull/314 Change-Id: If33cd2c636325e913315a7fd9b6e5ff37b2add66 Signed-off-by: Gaurav Gangalwar <gaurav.gangalwar(a)nutanix.com> --- M src/RPCAL/rpc_tools.c 1 file changed, 5 insertions(+), 1 deletion(-) git pull ssh://review.gerrithub.io:29418/ffilz/nfs-ganesha refs/changes/65/1199665/1 -- To view, visit https://review.gerrithub.io/c/ffilz/nfs-ganesha/+/1199665?usp=email To unsubscribe, or for help writing mail filters, visit https://review.gerrithub.io/settings?usp=email Gerrit-MessageType: newchange Gerrit-Project: ffilz/nfs-ganesha Gerrit-Branch: next Gerrit-Change-Id: If33cd2c636325e913315a7fd9b6e5ff37b2add66 Gerrit-Change-Number: 1199665 Gerrit-PatchSet: 1 Gerrit-Owner: Gaurav <gaurav.gangalwar(a)gmail.com>

1 0

[S] Change in ...nfs-ganesha[next]: RDMA fixes to serve reads not aligned on BYTES_PER_XDR_UNIT
by Gaurav (GerritHub) 19 Aug '24

19 Aug '24

Gaurav has uploaded this change for review. ( https://review.gerrithub.io/c/ffilz/nfs-ganesha/+/1199664?usp=email ) Change subject: RDMA fixes to serve reads not aligned on BYTES_PER_XDR_UNIT ...................................................................... RDMA fixes to serve reads not aligned on BYTES_PER_XDR_UNIT For RDMA reads we can't use allocated buffer for padding bytes. We should use registered buffer only for sending padding bytes also. For RDMA buffer size is always aligned to BYTES_PER_XDR_UNIT, so just check if iov has space left and use it for padding. Need tirpc pull https://github.com/nfs-ganesha/ntirpc/pull/314 Change-Id: Id6a931bccd53b4cc48bf5ed3dfae356023909544 Signed-off-by: Gaurav Gangalwar <gaurav.gangalwar(a)nutanix.com> --- M src/FSAL/fsal_helper.c M src/Protocols/NFS/nfs3_readdir.c M src/Protocols/NFS/nfs3_readdirplus.c M src/Protocols/NFS/nfs4_op_read.c M src/Protocols/NFS/nfs4_op_readdir.c M src/RPCAL/rpc_tools.c M src/include/fsal_api.h M src/include/gsh_rpc.h 8 files changed, 26 insertions(+), 8 deletions(-) git pull ssh://review.gerrithub.io:29418/ffilz/nfs-ganesha refs/changes/64/1199664/1 -- To view, visit https://review.gerrithub.io/c/ffilz/nfs-ganesha/+/1199664?usp=email To unsubscribe, or for help writing mail filters, visit https://review.gerrithub.io/settings?usp=email Gerrit-MessageType: newchange Gerrit-Project: ffilz/nfs-ganesha Gerrit-Branch: next Gerrit-Change-Id: Id6a931bccd53b4cc48bf5ed3dfae356023909544 Gerrit-Change-Number: 1199664 Gerrit-PatchSet: 1 Gerrit-Owner: Gaurav <gaurav.gangalwar(a)gmail.com>

1 0

cesonik999
by cesonik999＠albarulo.com 18 Aug '24

18 Aug '24

The cod and bamboo leaves are grilled by using Josper Grill. The fish fillets are sincerely defined, making it an exceedingly delicate dish. The fish body is soft and clean, and every diner will praise the deliciousness after completing the meal! https://www.bestsitesolutions.com

1 0

thepopularbooks
by uzv1puh0＠maillazy.com 15 Aug '24

15 Aug '24

https://www.thepopularbooks.com One day, a monkey drove his car near a lake. Sudden, he saw a tiger under a tree. The tiger saw him, too. The tiger ran very fast at the monkey. The monkey was very afraid and he and his car fell into the lake. The monkey couldn't swim and he shouted. A rabbit came and he jumped into the lake. The abbit swam to the monkey but did not help him. Both of them shouted forhelp. Luckily, a giraffe came along. He was very tall. He helped the rabbit and the monkey.

1 0