Merge tag 'migration-20240126-pull-request' of https://gitlab.com/peterx/qemu into staging

Migration Pull

[dropped fabiano's patch on modifying cpu model for arm migration tests for
 now]

- Fabiano's patchset to fix migration state references in BHs
- Fabiano's new 'n-1' migration test for CI
- Het's fix on making "uri" optional in QMP migrate cmd
- Markus's HMP leak fix reported by Coverity
- Paolo's cleanup on uffd to replace u64 usage
- Peter's small migration cleanup series all over the places

# -----BEGIN PGP SIGNATURE-----
#
# iIgEABYKADAWIQS5GE3CDMRX2s990ak7X8zN86vXBgUCZbcVeBIccGV0ZXJ4QHJl
# ZGhhdC5jb20ACgkQO1/MzfOr1wYHjgD9F2Fnrf4EuPNC/gF3yUvHVz1mgHqevb/g
# pw/ThcJF31wBALuWmwuUaNWm+VNtRc10YH6bY7HZW8oa1RefRN6QZn0L
# =JGTX
# -----END PGP SIGNATURE-----
# gpg: Signature made Mon 29 Jan 2024 03:03:20 GMT
# gpg:                using EDDSA key B9184DC20CC457DACF7DD1A93B5FCCCDF3ABD706
# gpg:                issuer "peterx@redhat.com"
# gpg: Good signature from "Peter Xu <xzpeter@gmail.com>" [marginal]
# gpg:                 aka "Peter Xu <peterx@redhat.com>" [marginal]
# gpg: WARNING: This key is not certified with sufficiently trusted signatures!
# gpg:          It is not certain that the signature belongs to the owner.
# Primary key fingerprint: B918 4DC2 0CC4 57DA CF7D  D1A9 3B5F CCCD F3AB D706

* tag 'migration-20240126-pull-request' of https://gitlab.com/peterx/qemu:
  Make 'uri' optional for migrate QAPI
  migration: Centralize BH creation and dispatch
  migration: Add a wrapper to qemu_bh_schedule
  migration: Reference migration state around loadvm_postcopy_handle_run_bh
  migration: Take reference to migration state around bg_migration_vm_start_bh
  migration: Fix use-after-free of migration state object
  migration/yank: Use channel features
  ci: Disable migration compatibility tests for aarch64
  ci: Add a migration compatibility test job
  analyze-migration.py: Remove trick on parsing ramblocks
  migration: Drop unnecessary check in ram's pending_exact()
  migration: Make threshold_size an uint64_t
  migration: Plug memory leak on HMP migrate error path
  userfaultfd: use 1ULL to build ioctl masks

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

1 files changed, 3 insertions, 8 deletions

diff --git a/scripts/analyze-migration.py b/scripts/analyze-migration.py
index a39dfb8766..8a254a5b6a 100755
--- a/scripts/analyze-migration.py
+++ b/scripts/analyze-migration.py
@@ -151,17 +151,12 @@ class RamSection(object):
             addr &= ~(self.TARGET_PAGE_SIZE - 1)
 
             if flags & self.RAM_SAVE_FLAG_MEM_SIZE:
-                while True:
+                total_length = addr
+                while total_length > 0:
                     namelen = self.file.read8()
-                    # We assume that no RAM chunk is big enough to ever
-                    # hit the first byte of the address, so when we see
-                    # a zero here we know it has to be an address, not the
-                    # length of the next block.
-                    if namelen == 0:
-                        self.file.file.seek(-1, 1)
-                        break
                     self.name = self.file.readstr(len = namelen)
                     len = self.file.read64()
+                    total_length -= len
                     self.sizeinfo[self.name] = '0x%016x' % len
                     if self.write_memory:
                         print(self.name)