| Message ID | 1435747190-18017-1-git-send-email-quintela@redhat.com |
|---|---|
| State | New |
| Headers | show |
On 1 July 2015 at 11:39, Juan Quintela <quintela@redhat.com> wrote: > Hi > > This series includes: > - rdma fixes by Dave > - rdma memory fix by gonglei > - vmdescription for old machine types (dave) > - fix footers for power (dave) > - migration bitmap extensions (Li) > just fixed the compilation issues for linux-users > - migration events (me) > - optional secttions (me) > - global configuration (me) > > > Please, Apply. > > > The following changes since commit d2966f804d70a244f5dde395fc5d22a50ed3e74e: > > Merge remote-tracking branch 'remotes/vivier/tags/pull-m68k-20150629' into staging (2015-06-29 17:03:20 +0100) > > are available in the git repository at: > > git://github.com/juanquintela/qemu.git tags/migration/20150701 > > for you to fetch changes up to a4fe58b0ea0d78f92461607f4f90be3384fa30e5: > > migration: Add migration events on target side (2015-07-01 12:35:05 +0200) > > ---------------------------------------------------------------- > migration/next for 20150701 On OSX at least every QEMU executable aborts immediately with "qemu: qemu_mutex_lock: Invalid argument". Here's a backtrace: #0 0x00007fff92c29286 in __pthread_kill () #1 0x00007fff8a4f342f in pthread_kill () #2 0x00007fff9240eb53 in abort () #3 0x00000001002b0915 in error_exit (err=<value temporarily unavailable, due to optimizations>, msg=<value temporarily unavailable, due to optimizations>) at /Users/pm215/src/qemu/util/qemu-thread-posix.c:48 #4 0x00000001002b095d in qemu_mutex_lock (mutex=<value temporarily unavailable, due to optimizations>) at /Users/pm215/src/qemu/util/qemu-thread-posix.c:75 #5 0x0000000100050f20 in migration_bitmap_extend (old=0, new=32768) at /Users/pm215/src/qemu/migration/ram.c:1068 #6 0x0000000100002a8d in ram_block_add [inlined] () at /Users/pm215/src/qemu/exec.c:1407 #7 0x0000000100002a8d in qemu_ram_alloc_internal (size=<value temporarily unavailable, due to optimizations>, max_size=<value temporarily unavailable, due to optimizations>, resized=<value temporarily unavailable, due to optimizations>, host=<value temporarily unavailable, due to optimizations>, resizeable=false, mr=0x100e97320, errp=<value temporarily unavailable, due to optimizations>) at /Users/pm215/src/qemu/exec.c:1537 #8 0x0000000100002e71 in qemu_ram_alloc (size=1823, mr=0x100e97320, errp=<value temporarily unavailable, due to optimizations>) at /Users/pm215/src/qemu/exec.c:1554 #9 0x000000010004353f in memory_region_init_ram (mr=0x100e97320, owner=<value temporarily unavailable, due to optimizations>, name=<value temporarily unavailable, due to optimizations>, size=134217728, errp=0x0) at /Users/pm215/src/qemu/memory.c:1216 #10 0x000000010003f10a in allocate_system_memory_nonnuma [inlined] () at /Users/pm215/src/qemu/numa.c:339 #11 0x000000010003f10a in memory_region_allocate_system_memory (mr=0x100e97320, owner=0x6, name=0x0, ram_size=140734799798744) at /Users/pm215/src/qemu/numa.c:352 #12 0x00000001000781cb in pc_memory_init (machine=0x102198630, system_memory=0x102199690, below_4g_mem_size=134217728, above_4g_mem_size=0, rom_memory=0x100e96de0, ram_memory=0x7fff5fbfe4b8, guest_info=<value temporarily unavailable, due to optimizations>) at /Users/pm215/src/qemu/hw/i386/pc.c:1254 #13 0x000000010007aae6 in pc_init1 (machine=0x102198630) at /Users/pm215/src/qemu/hw/i386/pc_piix.c:182 #14 0x00000001000e3cf9 in realtime_init [inlined] () at /Users/pm215/src/qemu/vl.c:4503 #15 0x00000001000e3cf9 in qemu_main (argc=<value temporarily unavailable, due to optimizations>, argv=<value temporarily unavailable, due to optimizations>, envp=0x0) at /Users/pm215/src/qemu/vl.c:4505 #16 0x00000001002117be in -[QemuCocoaAppController startEmulationWithArgc:argv:] (self=<value temporarily unavailable, due to optimizations>, _cmd=<value temporarily unavailable, due to optimizations>, argc=1823, argv=0x102199690) at /Users/pm215/src/qemu/ui/cocoa.m:941 A breakpoint on ram_save_setup() is never hit, so it looks like the problem is the mutex is being used before it is initialized. thanks -- PMM
On 07/02/2015 05:31 PM, Peter Maydell wrote: > On 1 July 2015 at 11:39, Juan Quintela <quintela@redhat.com> wrote: >> Hi >> >> This series includes: >> - rdma fixes by Dave >> - rdma memory fix by gonglei >> - vmdescription for old machine types (dave) >> - fix footers for power (dave) >> - migration bitmap extensions (Li) >> just fixed the compilation issues for linux-users >> - migration events (me) >> - optional secttions (me) >> - global configuration (me) >> >> >> Please, Apply. >> >> >> The following changes since commit d2966f804d70a244f5dde395fc5d22a50ed3e74e: >> >> Merge remote-tracking branch 'remotes/vivier/tags/pull-m68k-20150629' into staging (2015-06-29 17:03:20 +0100) >> >> are available in the git repository at: >> >> git://github.com/juanquintela/qemu.git tags/migration/20150701 >> >> for you to fetch changes up to a4fe58b0ea0d78f92461607f4f90be3384fa30e5: >> >> migration: Add migration events on target side (2015-07-01 12:35:05 +0200) >> >> ---------------------------------------------------------------- >> migration/next for 20150701 > > On OSX at least every QEMU executable aborts immediately > with "qemu: qemu_mutex_lock: Invalid argument". Here's a backtrace: > > #0 0x00007fff92c29286 in __pthread_kill () > #1 0x00007fff8a4f342f in pthread_kill () > #2 0x00007fff9240eb53 in abort () > #3 0x00000001002b0915 in error_exit (err=<value temporarily > unavailable, due to optimizations>, msg=<value temporarily > unavailable, due to optimizations>) at > /Users/pm215/src/qemu/util/qemu-thread-posix.c:48 > #4 0x00000001002b095d in qemu_mutex_lock (mutex=<value temporarily > unavailable, due to optimizations>) at > /Users/pm215/src/qemu/util/qemu-thread-posix.c:75 > #5 0x0000000100050f20 in migration_bitmap_extend (old=0, new=32768) > at /Users/pm215/src/qemu/migration/ram.c:1068 > #6 0x0000000100002a8d in ram_block_add [inlined] () at > /Users/pm215/src/qemu/exec.c:1407 > #7 0x0000000100002a8d in qemu_ram_alloc_internal (size=<value > temporarily unavailable, due to optimizations>, max_size=<value > temporarily unavailable, due to optimizations>, resized=<value > temporarily unavailable, due to optimizations>, host=<value > temporarily unavailable, due to optimizations>, resizeable=false, > mr=0x100e97320, errp=<value temporarily unavailable, due to > optimizations>) at /Users/pm215/src/qemu/exec.c:1537 > #8 0x0000000100002e71 in qemu_ram_alloc (size=1823, mr=0x100e97320, > errp=<value temporarily unavailable, due to optimizations>) at > /Users/pm215/src/qemu/exec.c:1554 > #9 0x000000010004353f in memory_region_init_ram (mr=0x100e97320, > owner=<value temporarily unavailable, due to optimizations>, > name=<value temporarily unavailable, due to optimizations>, > size=134217728, errp=0x0) at /Users/pm215/src/qemu/memory.c:1216 > #10 0x000000010003f10a in allocate_system_memory_nonnuma [inlined] () > at /Users/pm215/src/qemu/numa.c:339 > #11 0x000000010003f10a in memory_region_allocate_system_memory > (mr=0x100e97320, owner=0x6, name=0x0, ram_size=140734799798744) at > /Users/pm215/src/qemu/numa.c:352 > #12 0x00000001000781cb in pc_memory_init (machine=0x102198630, > system_memory=0x102199690, below_4g_mem_size=134217728, > above_4g_mem_size=0, rom_memory=0x100e96de0, > ram_memory=0x7fff5fbfe4b8, guest_info=<value temporarily unavailable, > due to optimizations>) at /Users/pm215/src/qemu/hw/i386/pc.c:1254 > #13 0x000000010007aae6 in pc_init1 (machine=0x102198630) at > /Users/pm215/src/qemu/hw/i386/pc_piix.c:182 > #14 0x00000001000e3cf9 in realtime_init [inlined] () at > /Users/pm215/src/qemu/vl.c:4503 > #15 0x00000001000e3cf9 in qemu_main (argc=<value temporarily > unavailable, due to optimizations>, argv=<value temporarily > unavailable, due to optimizations>, envp=0x0) at > /Users/pm215/src/qemu/vl.c:4505 > #16 0x00000001002117be in -[QemuCocoaAppController > startEmulationWithArgc:argv:] (self=<value temporarily unavailable, > due to optimizations>, _cmd=<value temporarily unavailable, due to > optimizations>, argc=1823, argv=0x102199690) at > /Users/pm215/src/qemu/ui/cocoa.m:941 > > A breakpoint on ram_save_setup() is never hit, so it looks > like the problem is the mutex is being used before it is > initialized. Yes. I don't know why our test doesn't trigger this problem. Will fix it soon. Thanks Wen Congyang > > thanks > -- PMM > > . >
Peter Maydell <peter.maydell@linaro.org> wrote: > On 1 July 2015 at 11:39, Juan Quintela <quintela@redhat.com> wrote: >> Hi >> >> This series includes: >> - rdma fixes by Dave >> - rdma memory fix by gonglei >> - vmdescription for old machine types (dave) >> - fix footers for power (dave) >> - migration bitmap extensions (Li) >> just fixed the compilation issues for linux-users >> - migration events (me) >> - optional secttions (me) >> - global configuration (me) >> >> >> Please, Apply. >> >> >> The following changes since commit d2966f804d70a244f5dde395fc5d22a50ed3e74e: >> >> Merge remote-tracking branch >> 'remotes/vivier/tags/pull-m68k-20150629' into staging (2015-06-29 >> 17:03:20 +0100) >> >> are available in the git repository at: >> >> git://github.com/juanquintela/qemu.git tags/migration/20150701 >> >> for you to fetch changes up to a4fe58b0ea0d78f92461607f4f90be3384fa30e5: >> >> migration: Add migration events on target side (2015-07-01 12:35:05 +0200) >> >> ---------------------------------------------------------------- >> migration/next for 20150701 > > On OSX at least every QEMU executable aborts immediately > with "qemu: qemu_mutex_lock: Invalid argument". Here's a backtrace: > > #0 0x00007fff92c29286 in __pthread_kill () > #1 0x00007fff8a4f342f in pthread_kill () > #2 0x00007fff9240eb53 in abort () > #3 0x00000001002b0915 in error_exit (err=<value temporarily > unavailable, due to optimizations>, msg=<value temporarily > unavailable, due to optimizations>) at > /Users/pm215/src/qemu/util/qemu-thread-posix.c:48 > #4 0x00000001002b095d in qemu_mutex_lock (mutex=<value temporarily > unavailable, due to optimizations>) at > /Users/pm215/src/qemu/util/qemu-thread-posix.c:75 > #5 0x0000000100050f20 in migration_bitmap_extend (old=0, new=32768) > at /Users/pm215/src/qemu/migration/ram.c:1068 > #6 0x0000000100002a8d in ram_block_add [inlined] () at > /Users/pm215/src/qemu/exec.c:1407 > #7 0x0000000100002a8d in qemu_ram_alloc_internal (size=<value > temporarily unavailable, due to optimizations>, max_size=<value > temporarily unavailable, due to optimizations>, resized=<value > temporarily unavailable, due to optimizations>, host=<value > temporarily unavailable, due to optimizations>, resizeable=false, > mr=0x100e97320, errp=<value temporarily unavailable, due to > optimizations>) at /Users/pm215/src/qemu/exec.c:1537 > #8 0x0000000100002e71 in qemu_ram_alloc (size=1823, mr=0x100e97320, > errp=<value temporarily unavailable, due to optimizations>) at > /Users/pm215/src/qemu/exec.c:1554 > #9 0x000000010004353f in memory_region_init_ram (mr=0x100e97320, > owner=<value temporarily unavailable, due to optimizations>, > name=<value temporarily unavailable, due to optimizations>, > size=134217728, errp=0x0) at /Users/pm215/src/qemu/memory.c:1216 > #10 0x000000010003f10a in allocate_system_memory_nonnuma [inlined] () > at /Users/pm215/src/qemu/numa.c:339 > #11 0x000000010003f10a in memory_region_allocate_system_memory > (mr=0x100e97320, owner=0x6, name=0x0, ram_size=140734799798744) at > /Users/pm215/src/qemu/numa.c:352 > #12 0x00000001000781cb in pc_memory_init (machine=0x102198630, > system_memory=0x102199690, below_4g_mem_size=134217728, > above_4g_mem_size=0, rom_memory=0x100e96de0, > ram_memory=0x7fff5fbfe4b8, guest_info=<value temporarily unavailable, > due to optimizations>) at /Users/pm215/src/qemu/hw/i386/pc.c:1254 > #13 0x000000010007aae6 in pc_init1 (machine=0x102198630) at > /Users/pm215/src/qemu/hw/i386/pc_piix.c:182 > #14 0x00000001000e3cf9 in realtime_init [inlined] () at > /Users/pm215/src/qemu/vl.c:4503 > #15 0x00000001000e3cf9 in qemu_main (argc=<value temporarily > unavailable, due to optimizations>, argv=<value temporarily > unavailable, due to optimizations>, envp=0x0) at > /Users/pm215/src/qemu/vl.c:4505 > #16 0x00000001002117be in -[QemuCocoaAppController > startEmulationWithArgc:argv:] (self=<value temporarily unavailable, > due to optimizations>, _cmd=<value temporarily unavailable, due to > optimizations>, argc=1823, argv=0x102199690) at > /Users/pm215/src/qemu/ui/cocoa.m:941 > > A breakpoint on ram_save_setup() is never hit, so it looks > like the problem is the mutex is being used before it is > initialized. > > thanks > -- PMM Why, o why, o why it didn't fail on linux, why. I think I know where the problem is. Thanks for testing.
On 2 July 2015 at 13:17, Juan Quintela <quintela@redhat.com> wrote: > Why, o why, o why it didn't fail on linux, why. I think I know where > the problem is. I think OSX defaults to being pickier about its mutexes, whereas Linux defaults to "make lock/unlock as fast as possible and assume the program is correct". -- PMM
Peter Maydell <peter.maydell@linaro.org> writes: > On 2 July 2015 at 13:17, Juan Quintela <quintela@redhat.com> wrote: >> Why, o why, o why it didn't fail on linux, why. I think I know where >> the problem is. > > I think OSX defaults to being pickier about its mutexes, > whereas Linux defaults to "make lock/unlock as fast as > possible and assume the program is correct". From pthread_mutexattr_settype(3p): PTHREAD_MUTEX_ERRORCHECK This type of mutex provides error checking. A thread attempting to relock this mutex without first unlocking it shall return with an error. A thread attempting to unlock a mutex which another thread has locked shall return with an error. A thread attempting to unlock an unlocked mutex shall return with an error. Not sure it detects this particular problem. Should we use it with --enable-debug or something?
On 2 July 2015 at 13:55, Markus Armbruster <armbru@redhat.com> wrote: > From pthread_mutexattr_settype(3p): > > PTHREAD_MUTEX_ERRORCHECK > > This type of mutex provides error checking. A thread > attempting to relock this mutex without first unlocking it > shall return with an error. A thread attempting to unlock > a mutex which another thread has locked shall return with > an error. A thread attempting to unlock an unlocked mutex > shall return with an error. > > Not sure it detects this particular problem. > > Should we use it with --enable-debug or something? It breaks when fork is involved, unfortunately: https://lists.gnu.org/archive/html/qemu-devel/2015-06/msg06016.html -- PMM
Hi This series includes: - rdma fixes by Dave - rdma memory fix by gonglei - vmdescription for old machine types (dave) - fix footers for power (dave) - migration bitmap extensions (Li) just fixed the compilation issues for linux-users - migration events (me) - optional secttions (me) - global configuration (me) Please, Apply. The following changes since commit d2966f804d70a244f5dde395fc5d22a50ed3e74e: Merge remote-tracking branch 'remotes/vivier/tags/pull-m68k-20150629' into staging (2015-06-29 17:03:20 +0100) are available in the git repository at: git://github.com/juanquintela/qemu.git tags/migration/20150701 for you to fetch changes up to a4fe58b0ea0d78f92461607f4f90be3384fa30e5: migration: Add migration events on target side (2015-07-01 12:35:05 +0200) ---------------------------------------------------------------- migration/next for 20150701 ---------------------------------------------------------------- Dr. David Alan Gilbert (11): Only try and read a VMDescription if it should be there rdma typos Store block name in local blocks structure Translate offsets to destination address space Rework ram_control_load_hook to hook during block load Allow rdma_delete_block to work without the hash Rework ram block hash Sort destination RAMBlocks to be the same as the source Sanity check RDMA remote data Fail more cleanly in mismatched RAM cases Fix older machine type compatibility on power with section footers Gonglei (1): rdma: fix memory leak Juan Quintela (12): runstate: Add runstate store runstate: migration allows more transitions now migration: create new section to store global state global_state: Make section optional vmstate: Create optional sections migration: Add configuration section migration: Use cmpxchg correctly migration: ensure we start in NONE state migration: Use always helper to set state migration: No need to call trace_migrate_set_state() migration: create migration event migration: Add migration events on target side Li Zhijian (2): migration: protect migration_bitmap migration: extend migration_bitmap docs/qmp/qmp-events.txt | 14 ++ exec.c | 5 + hw/i386/pc_piix.c | 2 + hw/i386/pc_q35.c | 2 + hw/ppc/spapr.c | 3 + include/exec/exec-all.h | 3 + include/migration/migration.h | 6 +- include/migration/qemu-file.h | 14 +- include/migration/vmstate.h | 2 + include/sysemu/sysemu.h | 1 + migration/migration.c | 158 ++++++++++++++++++++--- migration/qemu-file.c | 16 ++- migration/ram.c | 44 ++++++- migration/rdma.c | 289 ++++++++++++++++++++++++++++++------------ migration/savevm.c | 104 +++++++++++++-- migration/vmstate.c | 11 ++ qapi/event.json | 12 ++ trace-events | 16 ++- vl.c | 21 ++- 19 files changed, 597 insertions(+), 126 deletions(-)