search for: command_common

The next patch wants to make queuing commands smarter; rather than duplicating the logic at each command's call site, it's better to centralize things in command_common, and to also make this helper routine visible for use in NBD_CMD_DISC. --- lib/disconnect.c | 17 +++-------------- lib/internal.h | 6 ++++++ lib/rw.c | 19 +++---------------- 3 files changed, 12 insertions(+), 30 deletions(-) diff --git a/lib/disconnect.c b/lib/disconnect.c index bc...

Our construction of struct command_in_flight was ad hoc; most parameters were set in command_common(), but extent callbacks were done after the fact, and NBD_CMD_DISC was open-coding things. Furthermore, every caller was triggering nbd_internal_run() for the cmd_issue event; doing that in a central place makes it easier for the next patch to improve that logic without duplicating the fix at each...

...of whether the state machine is currently in READY. I also polished up the tmp demo into a bit more full-fledged example file, worth including since it also let me discover a hard-to-hit race with large NBD_CMD_WRITE vs. EAGAIN failures (now fixed). Eric Blake (5): lib: Refactor state event into command_common commands: Allow for a command queue states: Split ISSUE_COMMAND.SEND_REQUEST states: Allow in-flight read while writing next command examples: Add example to demonstrate just-fixed deadlock scenario .gitignore | 1 + examples/Makefile.am | 10 ++ examp...

Since v2: - rebase to Rich's new API calls - more refactoring in patch 1 (retitled) - new patches 3 and 4 - fix data corruption in patch 6 (was 4) - more tweaks to the reproducer example (including using new API from 3) Eric Blake (7): lib: Refactor command_common() to do more common work commands: Allow for a command queue commands: Expose FIFO ordering of server completions disconnect: Allow shutdown during processing states: Split ISSUE_COMMAND.SEND_REQUEST states: Allow in-flight read while writing next command examples: Add example to demons...

...; + while (prev_cmd->next) + prev_cmd = prev_cmd->next; + prev_cmd->next = cmd; + } + else + conn->cmds_done = cmd; SET_NEXT_STATE (%.READY); return 0; diff --git a/lib/rw.c b/lib/rw.c index 9dfce97..fa7dc52 100644 --- a/lib/rw.c +++ b/lib/rw.c @@ -246,7 +246,7 @@ command_common (struct nbd_connection *conn, uint16_t flags, uint16_t type, uint64_t offset, uint64_t count, void *data) { - struct command_in_flight *cmd; + struct command_in_flight *cmd, *prev_cmd; switch (type) { /* Commands which send or receive data are limited t...

...d, the first list contains commands waiting * to be issued. The second list contains commands which have been * issued and waiting for replies. The third list contains commands diff --git a/lib/rw.c b/lib/rw.c index a7587e9..ebd4ff9 100644 --- a/lib/rw.c +++ b/lib/rw.c @@ -246,7 +246,8 @@ command_common (struct nbd_connection *conn, uint16_t flags, uint16_t type, uint64_t offset, uint64_t count, void *data) { - struct command_in_flight *cmd; + struct command_in_flight *cmd, *prev_cmd; + bool event = !conn->reached_ready; switch (type) { /* Commands...

...un } else if (nbd_unlocked_aio_is_processing (conn)) { // can't call nbd_internal_run because we're in the // wrong state, but we can enqueue the command and it // will be processed next time we get to READY } else { // this is an error } > @@ -296,9 +297,24 @@ command_common (struct nbd_connection *conn, > if (conn->structured_replies && cmd->data && type == NBD_CMD_READ) > memset (cmd->data, 0, cmd->count); > > - cmd->next = conn->cmds_to_issue; > - conn->cmds_to_issue = cmd; > - if (nbd_internal_run (...

This might not be the final solution, but it certainly seems to solve a deadlock for me that I could trigger by using 'nbdkit --filter=noparallel memory 512k' and calling nbd_aio_pread for a request larger than 256k (enough for the Linux kernel to block the server until libnbd read()s), immediately followed by nbd_aio_pwrite for a request larger than 256k (enough to block libnbd until the