aboutsummaryrefslogtreecommitdiff
path: root/migration
AgeCommit message (Collapse)Author
2023-05-11cpu: expose qemu_cpu_list_lock for lock-guard useJamie Iles
Expose qemu_cpu_list_lock globally so that we can use WITH_QEMU_LOCK_GUARD and QEMU_LOCK_GUARD to simplify a few code paths now and in future. Signed-off-by: Jamie Iles <quic_jiles@quicinc.com> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-Id: <20230427020925.51003-2-quic_jiles@quicinc.com> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2023-05-10migration: block incoming colo when capability is disabledVladimir Sementsov-Ogievskiy
We generally require same set of capabilities on source and target. Let's require x-colo capability to use COLO on target. Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru> Reviewed-by: Juan Quintela <quintela@redhat.com> Reviewed-by: Peter Xu <peterx@redhat.com> Reviewed-by: Lukas Straub <lukasstraub2@web.de> Reviewed-by: Zhang Chen <chen.zhang@intel.com> Message-Id: <20230428194928.1426370-11-vsementsov@yandex-team.ru> Signed-off-by: Juan Quintela <quintela@redhat.com>
2023-05-10migration: disallow change capabilities in COLO stateVladimir Sementsov-Ogievskiy
COLO is not listed as running state in migrate_is_running(), so, it's theoretically possible to disable colo capability in COLO state and the unexpected error in migration_iteration_finish() is reachable. Let's disallow that in qmp_migrate_set_capabilities. Than the error becomes absolutely unreachable: we can get into COLO state only with enabled capability and can't disable it while we are in COLO state. So substitute the error by simple assertion. Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru> Reviewed-by: Juan Quintela <quintela@redhat.com> Reviewed-by: Peter Xu <peterx@redhat.com> Message-Id: <20230428194928.1426370-10-vsementsov@yandex-team.ru> Signed-off-by: Juan Quintela <quintela@redhat.com>
2023-05-10migration: process_incoming_migration_co: simplify code flow around retVladimir Sementsov-Ogievskiy
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru> Reviewed-by: Juan Quintela <quintela@redhat.com> Reviewed-by: Peter Xu <peterx@redhat.com> Reviewed-by: Zhang Chen <chen.zhang@intel.com> Message-Id: <20230428194928.1426370-7-vsementsov@yandex-team.ru> Signed-off-by: Juan Quintela <quintela@redhat.com>
2023-05-10migration: drop colo_incoming_thread from MigrationIncomingStateVladimir Sementsov-Ogievskiy
have_colo_incoming_thread variable is unused. colo_incoming_thread can be local. Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru> Reviewed-by: Juan Quintela <quintela@redhat.com> Reviewed-by: Peter Xu <peterx@redhat.com> Reviewed-by: Zhang Chen <chen.zhang@intel.com> Message-Id: <20230428194928.1426370-6-vsementsov@yandex-team.ru> Signed-off-by: Juan Quintela <quintela@redhat.com>
2023-05-10build: move COLO under CONFIG_REPLICATIONVladimir Sementsov-Ogievskiy
We don't allow to use x-colo capability when replication is not configured. So, no reason to build COLO when replication is disabled, it's unusable in this case. Note also that the check in migrate_caps_check() is not the only restriction: some functions in migration/colo.c will just abort if called with not defined CONFIG_REPLICATION, for example: migration_iteration_finish() case MIGRATION_STATUS_COLO: migrate_start_colo_process() colo_process_checkpoint() abort() It could probably make sense to have possibility to enable COLO without REPLICATION, but this requires deeper audit of colo & replication code, which may be done later if needed. Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru> Acked-by: Dr. David Alan Gilbert <dave@treblig.org> Reviewed-by: Juan Quintela <quintela@redhat.com> Message-Id: <20230428194928.1426370-4-vsementsov@yandex-team.ru> Signed-off-by: Juan Quintela <quintela@redhat.com>
2023-05-10colo: make colo_checkpoint_notify static and provide simpler APIVladimir Sementsov-Ogievskiy
colo_checkpoint_notify() is mostly used in colo.c. Outside we use it once when x-checkpoint-delay migration parameter is set. So, let's simplify the external API to only that function - notify COLO that parameter was set. This make external API more robust and hides implementation details from external callers. Also this helps us to make COLO module optional in further patch (i.e. we are going to add possibility not build the COLO module). Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru> Reviewed-by: Juan Quintela <quintela@redhat.com> Reviewed-by: Peter Xu <peterx@redhat.com> Reviewed-by: Zhang Chen <chen.zhang@intel.com> Message-Id: <20230428194928.1426370-3-vsementsov@yandex-team.ru> Signed-off-by: Juan Quintela <quintela@redhat.com>
2023-05-10multifd: Add the ramblock to MultiFDRecvParamsLukas Straub
This will be used in the next commits to add colo support to multifd. Signed-off-by: Lukas Straub <lukasstraub2@web.de> Reviewed-by: Juan Quintela <quintela@redhat.com> Message-Id: <88135197411df1a71d7832962b39abf60faf0021.1683572883.git.lukasstraub2@web.de> Signed-off-by: Juan Quintela <quintela@redhat.com>
2023-05-10ram: Let colo_flush_ram_cache take the bitmap_mutexLukas Straub
This is not required, colo_flush_ram_cache does not run concurrently with the multifd threads since the cache is only flushed after everything has been received. But it makes me more comfortable. This will be used in the next commits to add colo support to multifd. Signed-off-by: Lukas Straub <lukasstraub2@web.de> Reviewed-by: Juan Quintela <quintela@redhat.com> Message-Id: <35cb23ba854151d38a31e3a5c8a1020e4283cb4a.1683572883.git.lukasstraub2@web.de> Signed-off-by: Juan Quintela <quintela@redhat.com>
2023-05-10ram: Add public helper to set colo bitmapLukas Straub
The overhead of the mutex in non-multifd mode is negligible, because in that case its just the single thread taking the mutex. This will be used in the next commits to add colo support to multifd. Signed-off-by: Lukas Straub <lukasstraub2@web.de> Reviewed-by: Juan Quintela <quintela@redhat.com> Message-Id: <22d83cb428f37929563155531bfb69fd8953cc61.1683572883.git.lukasstraub2@web.de> Signed-off-by: Juan Quintela <quintela@redhat.com>
2023-05-10migration: Attempt disk reactivation in more failure scenariosEric Blake
Commit fe904ea824 added a fail_inactivate label, which tries to reactivate disks on the source after a failure while s->state == MIGRATION_STATUS_ACTIVE, but didn't actually use the label if qemu_savevm_state_complete_precopy() failed. This failure to reactivate is also present in commit 6039dd5b1c (also covering the new s->state == MIGRATION_STATUS_DEVICE state) and 403d18ae (ensuring s->block_inactive is set more reliably). Consolidate the two labels back into one - no matter HOW migration is failed, if there is any chance we can reach vm_start() after having attempted inactivation, it is essential that we have tried to restart disks before then. This also makes the cleanup more like migrate_fd_cancel(). Suggested-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Eric Blake <eblake@redhat.com> Message-Id: <20230502205212.134680-1-eblake@redhat.com> Acked-by: Peter Xu <peterx@redhat.com> Reviewed-by: Juan Quintela <quintela@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
2023-05-08migration: Initialize and cleanup decompression in migration.cLukas Straub
This fixes compress with colo. Signed-off-by: Lukas Straub <lukasstraub2@web.de> Reviewed-by: Juan Quintela <quintela@redhat.com> Signed-off-by: Juan Quintela <quintela@redhat.com>
2023-05-08ram-compress.c: Make target independentLukas Straub
Make ram-compress.c target independent. Signed-off-by: Lukas Straub <lukasstraub2@web.de> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Juan Quintela <quintela@redhat.com> Signed-off-by: Juan Quintela <quintela@redhat.com>
2023-05-08ram compress: Assert that the file buffer matches the resultLukas Straub
Before this series, "nothing to send" was handled by the file buffer being empty. Now it is tracked via param->result. Assert that the file buffer state matches the result. Signed-off-by: Lukas Straub <lukasstraub2@web.de> Reviewed-by: Juan Quintela <quintela@redhat.com> Signed-off-by: Juan Quintela <quintela@redhat.com>
2023-05-08ram.c: Move core decompression code into its own fileLukas Straub
No functional changes intended. Signed-off-by: Lukas Straub <lukasstraub2@web.de> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Juan Quintela <quintela@redhat.com> Signed-off-by: Juan Quintela <quintela@redhat.com>
2023-05-08ram.c: Move core compression code into its own fileLukas Straub
No functional changes intended. Signed-off-by: Lukas Straub <lukasstraub2@web.de> Reviewed-by: Juan Quintela <quintela@redhat.com> Signed-off-by: Juan Quintela <quintela@redhat.com>
2023-05-08ram.c: Remove last ram.c dependency from the core compress codeLukas Straub
Make compression interfaces take send_queued_data() as an argument. Remove save_page_use_compression() from flush_compressed_data(). This removes the last ram.c dependency from the core compress code. Signed-off-by: Lukas Straub <lukasstraub2@web.de> Reviewed-by: Juan Quintela <quintela@redhat.com> Signed-off-by: Juan Quintela <quintela@redhat.com>
2023-05-08ram.c: Call update_compress_thread_counts from compress_send_queued_dataLukas Straub
This makes the core compress code more independend from ram.c. Signed-off-by: Lukas Straub <lukasstraub2@web.de> Reviewed-by: Juan Quintela <quintela@redhat.com> Signed-off-by: Juan Quintela <quintela@redhat.com>
2023-05-08ram.c: Do not call save_page_header() from compress threadsLukas Straub
save_page_header() accesses several global variables, so calling it from multiple threads is pretty ugly. Instead, call save_page_header() before writing out the compressed data from the compress buffer to the migration stream. This also makes the core compress code more independend from ram.c. Signed-off-by: Lukas Straub <lukasstraub2@web.de> Reviewed-by: Juan Quintela <quintela@redhat.com> Signed-off-by: Juan Quintela <quintela@redhat.com>
2023-05-08ram.c: Reset result after sending queued dataLukas Straub
And take the param->mutex lock for the whole section to ensure thread-safety. Now, it is explicitly clear if there is no queued data to send. Before, this was handled by param->file stream being empty and thus qemu_put_qemu_file() not sending anything. This will be used in the next commits to move save_page_header() out of compress code. Signed-off-by: Lukas Straub <lukasstraub2@web.de> Reviewed-by: Juan Quintela <quintela@redhat.com> Signed-off-by: Juan Quintela <quintela@redhat.com>
2023-05-08ram.c: Dont change param->block in the compress threadLukas Straub
Instead introduce a extra parameter to trigger the compress thread. Now, when the compress thread is done, we know what RAMBlock and offset it did compress. This will be used in the next commits to move save_page_header() out of compress code. Signed-off-by: Lukas Straub <lukasstraub2@web.de> Reviewed-by: Juan Quintela <quintela@redhat.com> Signed-off-by: Juan Quintela <quintela@redhat.com>
2023-05-08ram.c: Let the compress threads return a CompressResult enumLukas Straub
This will be used in the next commits to move save_page_header() out of compress code. Signed-off-by: Lukas Straub <lukasstraub2@web.de> Reviewed-by: Juan Quintela <quintela@redhat.com> Signed-off-by: Juan Quintela <quintela@redhat.com>
2023-05-05qemu-file: Make ram_control_save_page() use accessors for rate_limitJuan Quintela
Signed-off-by: Juan Quintela <quintela@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Reviewed-by: Peter Xu <peterx@redhat.com> Message-Id: <20230504113841.23130-9-quintela@redhat.com>
2023-05-05qemu-file: Make total_transferred an uint64_tJuan Quintela
Change all the functions that use it. It was already passed as uint64_t. Signed-off-by: Juan Quintela <quintela@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Message-Id: <20230504113841.23130-8-quintela@redhat.com>
2023-05-05qemu-file: remove shutdown memberJuan Quintela
The first thing that we do after setting the shutdown value is set the error as -EIO if there is not a previous error. So this value is redundant. Just remove it and use qemu_file_get_error() in the places that it was tested. Signed-off-by: Juan Quintela <quintela@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Reviewed-by: Peter Xu <peterx@redhat.com> Message-Id: <20230504113841.23130-7-quintela@redhat.com>
2023-05-05qemu-file: No need to check for shutdown in qemu_file_rate_limitJuan Quintela
After calling qemu_file_shutdown() we set the error as -EIO if there is no another previous error, so no need to check it here. Signed-off-by: Juan Quintela <quintela@redhat.com> Reviewed-by: Peter Xu <peterx@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Message-Id: <20230504113841.23130-6-quintela@redhat.com>
2023-05-05migration: qemu_file_total_transferred() function is monotonicJuan Quintela
So delta_bytes can only be greater or equal to zero. Never negative. Signed-off-by: Juan Quintela <quintela@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Message-Id: <20230504113841.23130-3-quintela@redhat.com>
2023-05-05migration: max_postcopy_bandwidth is a size parameterJuan Quintela
So make everything that uses it uint64_t no int64_t. Signed-off-by: Juan Quintela <quintela@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Message-Id: <20230504113841.23130-2-quintela@redhat.com>
2023-05-05migration/rdma: Check for postcopy soonerJuan Quintela
It makes no sense first try to see if there is an rdma error and then do nothing on postcopy stage. Change it so we check we are in postcopy before doing anything. Signed-off-by: Juan Quintela <quintela@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Message-Id: <20230504114443.23891-6-quintela@redhat.com>
2023-05-05migration/rdma: It makes no sense to recive that flag without RDMAJuan Quintela
This could only happen if the source sent RAM_SAVE_FLAG_HOOK (i.e. rdma) and destination don't have CONFIG_RDMA. Signed-off-by: Juan Quintela <quintela@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Message-Id: <20230504114443.23891-5-quintela@redhat.com>
2023-05-05migration/rdma: We can calculate the rioc from the QEMUFileJuan Quintela
Signed-off-by: Juan Quintela <quintela@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Message-Id: <20230504114443.23891-4-quintela@redhat.com>
2023-05-05migration/rdma: simplify ram_control_load_hook()Juan Quintela
Signed-off-by: Juan Quintela <quintela@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Message-Id: <20230504114443.23891-3-quintela@redhat.com>
2023-05-05migration: Make RAM_SAVE_FLAG_HOOK a normal case entryJuan Quintela
Fixes this commit, clearly a bad merge after a rebase or similar, it should have been its own case since that point. commit 5b0e9dd46fbda5152566a4a26fd96bc0d0452bf7 Author: Peter Lieven <pl@kamp.de> Date: Tue Jun 24 11:32:36 2014 +0200 migration: catch unknown flag combinations in ram_load Signed-off-by: Juan Quintela <quintela@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Message-Id: <20230504114443.23891-2-quintela@redhat.com>
2023-05-05migration: Rename xbzrle_enabled xbzrle_startedJuan Quintela
Otherwise it is confusing with the function xbzrle_enabled(). Suggested-by: Daniel P. Berrangé <berrange@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Reviewed-by: Peter Xu <peterx@redhat.com> Signed-off-by: Juan Quintela <quintela@redhat.com> Message-Id: <20230504115323.24407-1-quintela@redhat.com>
2023-05-05migration: Put zero_pages in alphabetical orderJuan Quintela
I forgot to move it when I rename it from duplicated_pages. Message-Id: <20230504103357.22130-3-quintela@redhat.com> Reviewed-by: David Edmondson <david.edmondson@oracle.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Signed-off-by: Juan Quintela <quintela@redhat.com>
2023-05-05migration: Document all migration_statsJuan Quintela
Message-Id: <20230504103357.22130-2-quintela@redhat.com> Reviewed-by: David Edmondson <david.edmondson@oracle.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Signed-off-by: Juan Quintela <quintela@redhat.com>
2023-05-05migration/rdma: Don't pass the QIOChannelRDMA as an opaqueJuan Quintela
We can calculate it from the QEMUFile like the caller. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Message-Id: <20230503131847.11603-6-quintela@redhat.com> Signed-off-by: Juan Quintela <quintela@redhat.com>
2023-05-05migration: Fix block_bitmap_mapping migrationJuan Quintela
It is valid that params->has_block_bitmap_mapping is true and params->block_bitmap_mapping is NULL. So we can't use the trick of having a single function. Move to two functions one for each value and the tests are fixed. Fixes: b804b35b1c8a0edfd127ac20819c234be55ac7fc migration: Create migrate_block_bitmap_mapping() function Reported-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru> Message-Id: <20230503181036.14890-1-quintela@redhat.com> Signed-off-by: Juan Quintela <quintela@redhat.com>
2023-05-03migration: Drop unused parameter for migration_tls_client_create()Juan Quintela
It is not needed since we moved the accessor for tls properties to options.c. Suggested-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru> Signed-off-by: Juan Quintela <quintela@redhat.com> Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
2023-05-03migration: Drop unused parameter for migration_tls_get_creds()Juan Quintela
It is not needed since we moved the accessor for tls properties to options.c. Suggested-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru> Signed-off-by: Juan Quintela <quintela@redhat.com> Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
2023-05-03migration/rdma: Unfold last user of acct_update_position()Juan Quintela
Signed-off-by: Juan Quintela <quintela@redhat.com> Reviewed-by: Lukas Straub <lukasstraub2@web.de>
2023-05-03migration/rdma: Split the zero page case from acct_update_positionJuan Quintela
Now that we have atomic counters, we can do it on the place that we need it, no need to do it inside ram.c. Signed-off-by: Juan Quintela <quintela@redhat.com> Reviewed-by: Lukas Straub <lukasstraub2@web.de>
2023-05-03migration: Rename RAMStats to MigrationAtomicStatsJuan Quintela
It is lousely based on MigrationStats, but that name is taken, so this is the best one that I came with. Signed-off-by: Juan Quintela <quintela@redhat.com> Reviewed-by: Lukas Straub <lukasstraub2@web.de> --- If you have any good suggestion for the name, I am all ears.
2023-05-03migration: Rename ram_counters to mig_statsJuan Quintela
migration_stats is just too long, and it is going to have more than ram counters in the near future. Signed-off-by: Juan Quintela <quintela@redhat.com> Reviewed-by: Lukas Straub <lukasstraub2@web.de>
2023-05-03migration: Move ram_stats to its own file migration-stats.[ch]Juan Quintela
There is already include/qemu/stats.h, so stats.h was a bad idea. We want this file to not depend on anything else, we will move all the migration counters/stats to this struct. Signed-off-by: Juan Quintela <quintela@redhat.com> Reviewed-by: Lukas Straub <lukasstraub2@web.de>
2023-05-03multifd: We already account for this packet on the multifd threadJuan Quintela
Signed-off-by: Juan Quintela <quintela@redhat.com> Reviewed-by: Lukas Straub <lukasstraub2@web.de>
2023-05-02migration/xbzrle: Use __attribute__((target)) for avx512Richard Henderson
Use the attribute, which is supported by clang, instead of the #pragma, which is not supported and, for some reason, also not detected by the meson probe, so we fail by -Werror. Signed-off-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Juan Quintela <quintela@redhat.com> Message-Id: <20230501210555.289806-1-richard.henderson@linaro.org>
2023-04-27migration: Make dirty_bytes_last_sync atomicJuan Quintela
As we set its value, it needs to be operated with atomics. We rename it from remaining to better reflect its meaning. Statistics always return the real reamaining bytes. This was used to store how much pages where dirty on the previous generation, so we can calculate the expected downtime as: dirty_bytes_last_sync / current_bandwith. If we use the actual remaining bytes, we would see a very small value at the end of the iteration. Signed-off-by: Juan Quintela <quintela@redhat.com> Reviewed-by: Peter Xu <peterx@redhat.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> --- I am open to use ram_bytes_remaining() in its only use and be more "optimistic" about the downtime. Don't use __nocheck() functions. Use stat64_get() now that it exists.
2023-04-27migration: Make dirty_pages_rate atomicJuan Quintela
Signed-off-by: Juan Quintela <quintela@redhat.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Peter Xu <peterx@redhat.com> --- Don't use __nocheck() variants Use stat64_get()
2023-04-27multifd: Only flush once each full round of memoryJuan Quintela
We need to add a new flag to mean to flush at that point. Notice that we still flush at the end of setup and at the end of complete stages. Signed-off-by: Juan Quintela <quintela@redhat.com> Acked-by: Peter Xu <peterx@redhat.com> --- Add missing qemu_fflush(), now it passes all tests always. In the previous version, the check that changes the default value to false got lost in some rebase. Get it back.
generated by cgit v1.2.3 (git 2.26.2) at 2024-12-20 09:12:52 +0000