QIOChannelClass methods qio_channel_rdma_readv() and
qio_channel_rdma_writev() violate their method contract when
rdma->error_state is non-zero:
1. They return whatever is in rdma->error_state then. Only -1 will be
fine. -2 will be misinterpreted as "would block". Anything less
than -2 isn't defined in the contract. A positive value would be
misinterpreted as success, but I believe that's not actually
possible.
2. They neglect to set an error then. If something up the call stack
dereferences the error when failure is returned, it will crash. If
it ignores the return value and checks the error instead, it will
miss the error.
Crap like this happens when return statements hide in macros,
especially when their uses are far away from the definition.
I elected not to investigate how callers are impacted.
Expand the two bad macro uses, so we can set an error and return -1.
The next commit will then get rid of the macro altogether.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Reviewed-by: Li Zhijian <lizhijian@fujitsu.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20230928132019.2544702-19-armbru@redhat.com>
Several error messages include numeric error codes returned by failed
functions:
* ibv_poll_cq() returns an unspecified negative value. Useless.
* rdma_accept and rdma_get_cm_event() return -1. Useless.
* qemu_rdma_poll() returns either -1 or an unspecified negative
value. Useless.
* qemu_rdma_block_for_wrid(), qemu_rdma_write_flush(),
qemu_rdma_exchange_send(), qemu_rdma_exchange_recv(),
qemu_rdma_write() return a negative value that may or may not be an
errno value. While reporting human-readable errno
information (which a number is not) can be useful, reporting an
error code that may or may not be an errno value is useless.
Drop these error codes from the error messages.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Reviewed-by: Li Zhijian <lizhijian@fujitsu.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20230928132019.2544702-18-armbru@redhat.com>
We use errno after calling Libibverbs functions that are not
documented to set errno (manual page does not mention errno), or where
the documentation is unclear ("returns [...] the value of errno on
failure"). While this could be read as "sets errno and returns it",
a glance at the source code[*] kills that hope:
static inline int ibv_post_send(struct ibv_qp *qp, struct ibv_send_wr *wr,
struct ibv_send_wr **bad_wr)
{
return qp->context->ops.post_send(qp, wr, bad_wr);
}
The callback can be
static int mana_post_send(struct ibv_qp *ibqp, struct ibv_send_wr *wr,
struct ibv_send_wr **bad)
{
/* This version of driver supports RAW QP only.
* Posting WR is done directly in the application.
*/
return EOPNOTSUPP;
}
Neither of them touches errno.
One of these errno uses is easy to fix, so do that now. Several more
will go away later in the series; add temporary FIXME commments.
Three will remain; add TODO comments. TODO, not FIXME, because the
bug might be in Libibverbs documentation.
[*] https://github.com/linux-rdma/rdma-core.git
commit 55fa316b4b18f258d8ac1ceb4aa5a7a35b094dcf
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20230928132019.2544702-17-armbru@redhat.com>
@error_reported and @received_error are flags. The latter is even
assigned bool true. Change them from int to bool.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Reviewed-by: Li Zhijian <lizhijian@fujitsu.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20230928132019.2544702-16-armbru@redhat.com>
qemu_rdma_buffer_mergeable() is semantically a predicate. It returns
int 0 or 1. Return bool instead, and fix the function name's
spelling.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Reviewed-by: Li Zhijian <lizhijian@fujitsu.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20230928132019.2544702-15-armbru@redhat.com>
qemu_rdma_search_ram_block() can't fail. Return void, and drop the
unreachable error handling.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Li Zhijian <lizhijian@fujitsu.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20230928132019.2544702-14-armbru@redhat.com>
rdma_add_block() can't fail. Return void, and drop the unreachable
error handling.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Reviewed-by: Li Zhijian <lizhijian@fujitsu.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20230928132019.2544702-13-armbru@redhat.com>
When all we do with an Error we receive into a local variable is
propagating to somewhere else, we can just as well receive it there
right away.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Reviewed-by: Li Zhijian <lizhijian@fujitsu.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20230928132019.2544702-12-armbru@redhat.com>
include/qapi/error.h demands:
* - Functions that use Error to report errors have an Error **errp
* parameter. It should be the last parameter, except for functions
* taking variable arguments.
qemu_rdma_connect() does not conform. Clean it up.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Reviewed-by: Li Zhijian <lizhijian@fujitsu.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20230928132019.2544702-11-armbru@redhat.com>
qemu_rdma_accept() returns 0 in some cases even when it didn't
complete its job due to errors. Impact is not obvious. I figure the
caller will soon fail again with a misleading error message.
Fix it to return -1 on any failure.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Reviewed-by: Li Zhijian <lizhijian@fujitsu.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20230928132019.2544702-10-armbru@redhat.com>
qemu_rdma_exchange_get_response() compares int parameter @expecting
with uint32_t head->type. Actual arguments are non-negative
enumeration constants, RDMAControlHeader uint32_t member type, or
qemu_rdma_exchange_recv() int parameter expecting. Actual arguments
for the latter are non-negative enumeration constants. Change both
parameters to uint32_t.
In qio_channel_rdma_readv(), loop control variable @i is ssize_t, and
counts from 0 up to @niov, which is size_t. Change @i to size_t.
While there, make qio_channel_rdma_readv() and
qio_channel_rdma_writev() more consistent: change the former's @done
to ssize_t, and delete the latter's useless initialization of @len.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20230928132019.2544702-8-armbru@redhat.com>
qio_channel_rdma_readv() assigns the size_t value of qemu_rdma_fill()
to an int variable before it adds it to @done / subtracts it from
@want, both size_t. Truncation when qemu_rdma_fill() copies more than
INT_MAX bytes. Seems vanishingly unlikely, but needs fixing all the
same.
Fixes: 6ddd2d76ca6f (migration: convert RDMA to use QIOChannel interface)
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20230928132019.2544702-7-armbru@redhat.com>
We use int instead of uint64_t in a few places. Change them to
uint64_t.
This cleans up a comparison of signed qemu_rdma_block_for_wrid()
parameter @wrid_requested with unsigned @wr_id. Harmless, because the
actual arguments are non-negative enumeration constants.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Reviewed-by: Li Zhijian <lizhijian@fujitsu.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20230928132019.2544702-6-armbru@redhat.com>
wrid_desc[] uses 4001 pointers to map four integer values to strings.
print_wrid() accesses wrid_desc[] out of bounds when passed a negative
argument. It returns null for values 2..1999 and 2001..3999.
qemu_rdma_poll() and qemu_rdma_block_for_wrid() print wrid_desc[wr_id]
and passes print_wrid(wr_id) to tracepoints. Could conceivably crash
trying to format a null string. I believe access out of bounds is not
possible.
Not worth cleaning up. Dumb down to show just numeric wr_id.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Reviewed-by: Li Zhijian <lizhijian@fujitsu.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20230928132019.2544702-5-armbru@redhat.com>
rdma_delete_block() always returns 0, which its only caller ignores.
Return void instead.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Reviewed-by: Li Zhijian <lizhijian@fujitsu.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20230928132019.2544702-4-armbru@redhat.com>
qemu_rdma_data_init() return type is void *. It actually returns
RDMAContext *, and all its callers assign the value to an
RDMAContext *. Unclean.
Return RDMAContext * instead.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Reviewed-by: Li Zhijian <lizhijian@fujitsu.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20230928132019.2544702-3-armbru@redhat.com>
qemu_rdma_poll()'s return type is uint64_t, even though it returns 0,
-1, or @ret, which is int. Its callers assign the return value to int
variables, then check whether it's negative. Unclean.
Return int instead.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Reviewed-by: Li Zhijian <lizhijian@fujitsu.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20230928132019.2544702-2-armbru@redhat.com>
There's a bug on dest that if a double fault triggered on dest qemu (a
network issue during postcopy-recover), we won't set PAUSED correctly
because we assumed we always came from ACTIVE.
Fix that by always overwriting the state to PAUSE.
We could also check for these two states, but maybe it's an overkill. We
did the same on the src QEMU to unconditionally switch to PAUSE anyway.
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231004220240.167175-10-peterx@redhat.com>
We are sending a migration event of MIGRATION_STATUS_SETUP at
qemu_start_incoming_migration but never actually setting the state.
This creates a window between qmp_migrate_incoming and
process_incoming_migration_co where the migration status is still
MIGRATION_STATUS_NONE. Calling query-migrate during this time will
return an empty response even though the incoming migration command
has already been issued.
Commit 7cf1fe6d68 ("migration: Add migration events on target side")
has added support to the 'events' capability to the incoming part of
migration, but chose to send the SETUP event without setting the
state. I'm assuming this was a mistake.
This introduces a change in behavior, any QMP client waiting for the
SETUP event will hang, unless it has previously enabled the 'events'
capability. Having the capability enabled is sufficient to continue to
receive the event.
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20230712190742.22294-5-farosas@suse.de>
QEMU will crash if anyone tries to set tls-authz (which is a type
StrOrNull) with 'null' value. Fix it in the easy way by converting it to
qstring just like the other two tls parameters.
Cc: qemu-stable@nongnu.org # v4.0+
Fixes: d2f1d29b95 ("migration: add support for a "tls-authz" migration parameter")
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20230905162335.235619-2-peterx@redhat.com>
Currently query-dirty-rate uses QEMU_CLOCK_REALTIME as
the source for start-time field. This translates to
clock_gettime(CLOCK_MONOTONIC), i.e. number of seconds
since host boot. This is not very useful. The only
reasonable use case of start-time I can imagine is to
check whether previously completed measurements are
too old or not. But this makes sense only if start-time
is reported as host wall-clock time.
This patch replaces source of start-time from
QEMU_CLOCK_REALTIME to QEMU_CLOCK_HOST.
Signed-off-by: Andrei Gudkov <gudkov.andrei@huawei.com>
Reviewed-by: Hyman Huang <yong.huang@smartx.com>
Message-Id: <399861531e3b24a1ecea2ba453fb2c3d129fb03a.1693905328.git.gudkov.andrei@huawei.com>
Signed-off-by: Hyman Huang <yong.huang@smartx.com>
For both save/load we actually share the logic on deciding whether a field
should exist. Merge the checks into a helper and use it for both save and
load. When doing so, add documentations and reformat the code to make it
much easier to read.
The real benefit here (besides code cleanups) is we add a trace-point for
this; this is a known spot where we can easily break migration
compatibilities between binaries, and this trace point will be critical for
us to identify such issues.
For example, this will be handy when debugging things like:
https://gitlab.com/qemu-project/qemu/-/issues/932
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20230906204722.514474-1-peterx@redhat.com>
Allow an offset option to be specified as part of the file URI, in
the form "file:filename,offset=offset", where offset accepts the common
size suffixes, or the 0x prefix, but not both. Migration data is written
to and read from the file starting at offset. If unspecified, it defaults
to 0.
This is needed by libvirt to store its own data at the head of the file.
Suggested-by: Daniel P. Berrange <berrange@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Steve Sistare <steven.sistare@oracle.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <1694182931-61390-3-git-send-email-steven.sistare@oracle.com>
Extend the migration URI to support file:<filename>. This can be used for
any migration scenario that does not require a reverse path. It can be
used as an alternative to 'exec:cat > file' in minimized containers that
do not contain /bin/sh, and it is easier to use than the fd:<fdname> URI.
It can be used in HMP commands, and as a qemu command-line parameter.
For best performance, guest ram should be shared and x-ignore-shared
should be true, so guest pages are not written to the file, in which case
the guest may remain running. If ram is not so configured, then the user
is advised to stop the guest first. Otherwise, a busy guest may re-dirty
the same page, causing it to be appended to the file multiple times,
and the file may grow unboundedly. That issue is being addressed in the
"fixed-ram" patch series.
Signed-off-by: Steve Sistare <steven.sistare@oracle.com>
Tested-by: Michael Galaxy <mgalaxy@akamai.com>
Reviewed-by: Michael Galaxy <mgalaxy@akamai.com>
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <1694182931-61390-2-git-send-email-steven.sistare@oracle.com>
Previously, we got a confusion error that complains
the RDMAControlHeader.repeat:
qemu-system-x86_64: rdma: Too many requests in this message (3638950032).Bailing.
Actually, it's caused by an unexpected RDMAControlHeader.type.
After this patch, error will become:
qemu-system-x86_64: Unknown control message QEMU FILE
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Li Zhijian <lizhijian@fujitsu.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20230926100103.201564-2-lizhijian@fujitsu.com>
A few code paths exist in the source code,where a migration is
marked as failed via MIGRATION_STATUS_FAILED, but the failure happens
outside of migration.c
In such cases, an error_report() call is made, however the current
MigrationState is never updated with the error description, and hence
clients like libvirt never know the actual reason for the failure.
This patch covers such cases outside of migration.c and updates the
error description at the appropriate places.
Acked-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Tejus GK <tejus.gk@nutanix.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231003065538.244752-3-tejus.gk@nutanix.com>
Currently, a few code paths exist in the function vmstate_save_state_v,
which ultimately leads to a migration failure. However, an update in the
current MigrationState for the error description is never done.
vmstate.c somehow doesn't seem to allow the use of migrate_set_error due
to some dependencies for unit tests. Hence, this patch introduces a new
function vmstate_save_state_with_err, which will eventually propagate
the error message to savevm.c where a migrate_set_error call can be
eventually done.
Acked-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Tejus GK <tejus.gk@nutanix.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231003065538.244752-2-tejus.gk@nutanix.com>
When we sent a page through QEMUFile hooks (RDMA) there are three
posiblities:
- We are not using RDMA. return RAM_SAVE_CONTROL_DELAYED and
control_save_page() returns false to let anything else to proceed.
- There is one error but we are using RDMA. Then we return a negative
value, control_save_page() needs to return true.
- Everything goes well and RDMA start the sent of the page
asynchronously. It returns RAM_SAVE_CONTROL_DELAYED and we need to
return 1 for ram_save_page_legacy.
Clear?
I know, I know, the interface is as bad as it gets. I think that now
it is a bit clearer, but this needs to be done some other way.
Reviewed-by: Leonardo Bras <leobras@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-Id: <20230515195709.63843-16-quintela@redhat.com>
After this change, nothing abuses QEMUFile to account for data
transferrefd during migration.
Reviewed-by: Leonardo Bras <leobras@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-Id: <20230515195709.63843-15-quintela@redhat.com>
RDMA protocol is completely asynchronous, so in qemu_rdma_save_page()
they "invent" that a byte has been transferred. And then they call
qemu_file_credit_transfer() and ram_transferred_add() with that byte.
Just remove that calls as nothing has been sent.
Reviewed-by: Leonardo Bras <leobras@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-Id: <20230515195709.63843-14-quintela@redhat.com>
Reviewed-by: Leonardo Bras <leobras@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-Id: <20230515195709.63843-13-quintela@redhat.com>
Remove the one in control_save_page().
Reviewed-by: Leonardo Bras <leobras@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-Id: <20230515195709.63843-12-quintela@redhat.com>
Just create a variable for it, the same way that multifd does. This
way it is safe to use for other thread, etc, etc.
Reviewed-by: Leonardo Bras <leobras@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-Id: <20230515195709.63843-11-quintela@redhat.com>
We only care about the amount of bytes transferred. Flushing is done
by the system somewhere else.
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20230530183941.7223-4-quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
In the function qmp_migrate(), yank_unregister_instance() gets called
twice which isn't required. Hence, refactoring it so that it gets called
during the local_error cleanup.
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Acked-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Tejus GK <tejus.gk@nutanix.com>
Message-ID: <20230621130940.178659-3-tejus.gk@nutanix.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Local variables shadowing other local variables or parameters make the
code needlessly hard to understand. Tracked down with -Wshadow=local.
Clean up: delete inner declarations when they are actually redundant,
else rename variables.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Li Zhijian <lizhijian@fujitsu.com>
Message-ID: <20230921121312.1301864-3-armbru@redhat.com>
qemu_rdma_save_page() reports polling error with error_report(), then
succeeds anyway. This is because the variable holding the polling
status *shadows* the variable the function returns. The latter
remains zero.
Broken since day one, and duplicated more recently.
Fixes: 2da776db4846 (rdma: core logic)
Fixes: b390afd8c50b (migration/rdma: Fix out of order wrid)
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Li Zhijian <lizhijian@fujitsu.com>
Message-ID: <20230921121312.1301864-2-armbru@redhat.com>
Now that the return path thread is allowed to finish during a paused
migration, we can move the cleanup of the QEMUFiles to the main
migration thread.
Reviewed-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-ID: <20230918172822.19052-9-farosas@suse.de>
Replace the return path retry logic with finishing and restarting the
thread. This fixes a race when resuming the migration that leads to a
segfault.
Currently when doing postcopy we consider that an IO error on the
return path file could be due to a network intermittency. We then keep
the thread alive but have it do cleanup of the 'from_dst_file' and
wait on the 'postcopy_pause_rp' semaphore. When the user issues a
migrate resume, a new return path is opened and the thread is allowed
to continue.
There's a race condition in the above mechanism. It is possible for
the new return path file to be setup *before* the cleanup code in the
return path thread has had a chance to run, leading to the *new* file
being closed and the pointer set to NULL. When the thread is released
after the resume, it tries to dereference 'from_dst_file' and crashes:
Thread 7 "return path" received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0x7fffd1dbf700 (LWP 9611)]
0x00005555560e4893 in qemu_file_get_error_obj (f=0x0, errp=0x0) at ../migration/qemu-file.c:154
154 return f->last_error;
(gdb) bt
#0 0x00005555560e4893 in qemu_file_get_error_obj (f=0x0, errp=0x0) at ../migration/qemu-file.c:154
#1 0x00005555560e4983 in qemu_file_get_error (f=0x0) at ../migration/qemu-file.c:206
#2 0x0000555555b9a1df in source_return_path_thread (opaque=0x555556e06000) at ../migration/migration.c:1876
#3 0x000055555602e14f in qemu_thread_start (args=0x55555782e780) at ../util/qemu-thread-posix.c:541
#4 0x00007ffff38d76ea in start_thread (arg=0x7fffd1dbf700) at pthread_create.c:477
#5 0x00007ffff35efa6f in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:95
Here's the race (important bit is open_return_path happening before
migration_release_dst_files):
migration | qmp | return path
--------------------------+-----------------------------+---------------------------------
qmp_migrate_pause()
shutdown(ms->to_dst_file)
f->last_error = -EIO
migrate_detect_error()
postcopy_pause()
set_state(PAUSED)
wait(postcopy_pause_sem)
qmp_migrate(resume)
migrate_fd_connect()
resume = state == PAUSED
open_return_path <-- TOO SOON!
set_state(RECOVER)
post(postcopy_pause_sem)
(incoming closes to_src_file)
res = qemu_file_get_error(rp)
migration_release_dst_files()
ms->rp_state.from_dst_file = NULL
post(postcopy_pause_rp_sem)
postcopy_pause_return_path_thread()
wait(postcopy_pause_rp_sem)
rp = ms->rp_state.from_dst_file
goto retry
qemu_file_get_error(rp)
SIGSEGV
-------------------------------------------------------------------------------------------
We can keep the retry logic without having the thread alive and
waiting. The only piece of data used by it is the 'from_dst_file' and
it is only allowed to proceed after a migrate resume is issued and the
semaphore released at migrate_fd_connect().
Move the retry logic to outside the thread by waiting for the thread
to finish before pausing the migration.
Reviewed-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-ID: <20230918172822.19052-8-farosas@suse.de>
We'll start calling the await_return_path_close_on_source() function
from other parts of the code, so move all of the related checks and
tracepoints into it.
Reviewed-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-ID: <20230918172822.19052-7-farosas@suse.de>
This file is owned by the return path thread which is already doing
cleanup.
Reviewed-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-ID: <20230918172822.19052-6-farosas@suse.de>
It's not safe to call qemu_file_shutdown() on the to_dst_file without
first checking for the file's presence under the lock. The cleanup of
this file happens at postcopy_pause() and migrate_fd_cleanup() which
are not necessarily running in the same thread as migrate_fd_cancel().
Reviewed-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-ID: <20230918172822.19052-5-farosas@suse.de>
We cannot call qemu_file_shutdown() on the return path file without
taking the file lock. The return path thread could be running it's
cleanup code and have just cleared the from_dst_file pointer.
Checking ms->to_dst_file for errors could also race with
migrate_fd_cleanup() which clears the to_dst_file pointer.
Protect both accesses by taking the file lock.
This was caught by inspection, it should be rare, but the next patches
will start calling this code from other places, so let's do the
correct thing.
Reviewed-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-ID: <20230918172822.19052-4-farosas@suse.de>
We don't need to set the rp_state.error right after a shutdown because
qemu_file_shutdown() always sets the QEMUFile error, so the return
path thread would have seen it and set the rp error itself.
Setting the error outside of the thread is also racy because the
thread could clear it after we set it.
Reviewed-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-ID: <20230918172822.19052-3-farosas@suse.de>
We hit intermit CI issue on failing at migration-test over the unit test
preempt/plain:
qemu-system-x86_64: Unable to read from socket: Connection reset by peer
Memory content inconsistency at 5b43000 first_byte = bd last_byte = bc current = 4f hit_edge = 1
**
ERROR:../tests/qtest/migration-test.c:300:check_guests_ram: assertion failed: (bad == 0)
(test program exited with status code -6)
Fabiano debugged into it and found that the preempt thread can quit even
without receiving all the pages, which can cause guest not receiving all
the pages and corrupt the guest memory.
To make sure preempt thread finished receiving all the pages, we can rely
on the page_requested_count being zero because preempt channel will only
receive requested page faults. Note, not all the faulted pages are required
to be sent via the preempt channel/thread; imagine the case when a
requested page is just queued into the background main channel for
migration, the src qemu will just still send it via the background channel.
Here instead of spinning over reading the count, we add a condvar so the
main thread can wait on it if that unusual case happened, without burning
the cpu for no good reason, even if the duration is short; so even if we
spin in this rare case is probably fine. It's just better to not do so.
The condvar is only used when that special case is triggered. Some memory
ordering trick is needed to guarantee it from happening (against the
preempt thread status field), so the main thread will always get a kick
when that triggers correctly.
Closes: https://gitlab.com/qemu-project/qemu/-/issues/1886
Debugged-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Fabiano Rosas <farosas@suse.de>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-ID: <20230918172822.19052-2-farosas@suse.de>
Add a new .save_prepare() handler to struct SaveVMHandlers. This handler
is called early, even before migration starts, and can be used by
devices to perform early checks.
Refactor migrate_init() to be able to return errors and call
.save_prepare() from there.
Suggested-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Avihai Horon <avihaih@nvidia.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Cédric Le Goater <clg@redhat.com>
Signed-off-by: Cédric Le Goater <clg@redhat.com>
Initialization of mig_stats, compression_counters and VFIO bytes
transferred is hard-coded in migration code path and snapshot code path.
Make the code cleaner by initializing them in migrate_init().
Suggested-by: Cédric Le Goater <clg@redhat.com>
Signed-off-by: Avihai Horon <avihaih@nvidia.com>
Reviewed-by: Cédric Le Goater <clg@redhat.com>
Signed-off-by: Cédric Le Goater <clg@redhat.com>
