| Message ID | 1497544575-18628-4-git-send-email-a.perevalov@samsung.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Thu, Jun 15, 2017 at 07:36:15PM +0300, Alexey Perevalov wrote: [...] > +void init_receivedmap(void) > +{ > + RAMBlock *rb; > + > + RAMBLOCK_FOREACH(rb) { > + unsigned long pages; > + pages = rb->max_length >> TARGET_PAGE_BITS; Nit: I would assert(!rb->receivedmap) before new it in case for leak. > + rb->receivedmap = bitmap_new(pages); > + } > +} > + > +static unsigned long int get_received_bit_offset(void *host_addr, RAMBlock *rb) > +{ > + uint64_t host_addr_offset = (uint64_t)(uintptr_t)(host_addr > + - (void *)rb->host); > + return host_addr_offset >> TARGET_PAGE_BITS; > +} Can we rename this function to ramblock_get_pfn() and export it in ramblock API (exec/ram_addr.h)? static inline uint64_t ramblock_get_pfn(RAMBlock *rb, void *host) { return (host - rb->host) >> TARGET_PAGE_BITS; } > + > +int test_receivedmap_by_addr(void *host_addr, RAMBlock *rb) > +{ > + return test_bit(get_received_bit_offset(host_addr, rb), rb->receivedmap); > +} > + > +void set_receivedmap_by_addr(void *host_addr, RAMBlock *rb) > +{ > + set_bit_atomic(get_received_bit_offset(host_addr, rb), rb->receivedmap); > +} > + > /* > * An outstanding page request, on the source, having been received > * and queued > @@ -2324,8 +2352,13 @@ static int ram_load_setup(QEMUFile *f, void *opaque) > > static int ram_load_cleanup(void *opaque) > { > + RAMBlock *rb; > xbzrle_load_cleanup(); > compress_threads_load_cleanup(); > + > + RAMBLOCK_FOREACH(rb) { Nit: I'd prefer: if (rb->receivedmap) {...} > + g_free(rb->receivedmap); and: rb->receivedmap = NULL later. > + } > return 0; > } > > diff --git a/migration/ram.h b/migration/ram.h > index c081fde..7048ff9 100644 > --- a/migration/ram.h > +++ b/migration/ram.h > @@ -52,4 +52,9 @@ int ram_discard_range(const char *block_name, uint64_t start, size_t length); > int ram_postcopy_incoming_init(MigrationIncomingState *mis); > > void ram_handle_compressed(void *host, uint8_t ch, uint64_t size); > + > +void init_receivedmap(void); > +int test_receivedmap_by_addr(void *host_addr, RAMBlock *rb); > +void set_receivedmap_by_addr(void *host_addr, RAMBlock *rb); Nit: I would name these names as: ramblock_recv_map_*(). > + > #endif > diff --git a/migration/savevm.c b/migration/savevm.c > index 31158da..668d3bb 100644 > --- a/migration/savevm.c > +++ b/migration/savevm.c > @@ -1372,6 +1372,7 @@ static int loadvm_postcopy_handle_advise(MigrationIncomingState *mis) > return -1; > } > > + init_receivedmap(); > remote_pagesize_summary = qemu_get_be64(mis->from_src_file); > local_pagesize_summary = ram_pagesize_summary(); > > -- > 1.9.1 > Again, did you trap precopy? Thanks,
On 06/16/2017 12:06 PM, Peter Xu wrote: > On Thu, Jun 15, 2017 at 07:36:15PM +0300, Alexey Perevalov wrote: > > [...] > >> +void init_receivedmap(void) >> +{ >> + RAMBlock *rb; >> + >> + RAMBLOCK_FOREACH(rb) { >> + unsigned long pages; >> + pages = rb->max_length >> TARGET_PAGE_BITS; > Nit: I would assert(!rb->receivedmap) before new it in case for leak. ok, > >> + rb->receivedmap = bitmap_new(pages); >> + } >> +} >> + >> +static unsigned long int get_received_bit_offset(void *host_addr, RAMBlock *rb) >> +{ >> + uint64_t host_addr_offset = (uint64_t)(uintptr_t)(host_addr >> + - (void *)rb->host); >> + return host_addr_offset >> TARGET_PAGE_BITS; >> +} > Can we rename this function to ramblock_get_pfn() and export it in > ramblock API (exec/ram_addr.h)? if you need it, I'll place export into ram_addr.h but name, PFN could confuse, because in case of hugepage address this function will not return real PFN, due to TARGET_PAGE_BITS usage. > > static inline uint64_t ramblock_get_pfn(RAMBlock *rb, void *host) > { > return (host - rb->host) >> TARGET_PAGE_BITS; > } > >> + >> +int test_receivedmap_by_addr(void *host_addr, RAMBlock *rb) >> +{ >> + return test_bit(get_received_bit_offset(host_addr, rb), rb->receivedmap); >> +} >> + >> +void set_receivedmap_by_addr(void *host_addr, RAMBlock *rb) >> +{ >> + set_bit_atomic(get_received_bit_offset(host_addr, rb), rb->receivedmap); >> +} >> + >> /* >> * An outstanding page request, on the source, having been received >> * and queued >> @@ -2324,8 +2352,13 @@ static int ram_load_setup(QEMUFile *f, void *opaque) >> >> static int ram_load_cleanup(void *opaque) >> { >> + RAMBlock *rb; >> xbzrle_load_cleanup(); >> compress_threads_load_cleanup(); >> + >> + RAMBLOCK_FOREACH(rb) { > Nit: I'd prefer: if (rb->receivedmap) {...} g_free already contains it > >> + g_free(rb->receivedmap); > and: rb->receivedmap = NULL later. why not ) > >> + } >> return 0; >> } >> >> diff --git a/migration/ram.h b/migration/ram.h >> index c081fde..7048ff9 100644 >> --- a/migration/ram.h >> +++ b/migration/ram.h >> @@ -52,4 +52,9 @@ int ram_discard_range(const char *block_name, uint64_t start, size_t length); >> int ram_postcopy_incoming_init(MigrationIncomingState *mis); >> >> void ram_handle_compressed(void *host, uint8_t ch, uint64_t size); >> + >> +void init_receivedmap(void); >> +int test_receivedmap_by_addr(void *host_addr, RAMBlock *rb); >> +void set_receivedmap_by_addr(void *host_addr, RAMBlock *rb); > Nit: I would name these names as: ramblock_recv_map_*(). > >> + >> #endif >> diff --git a/migration/savevm.c b/migration/savevm.c >> index 31158da..668d3bb 100644 >> --- a/migration/savevm.c >> +++ b/migration/savevm.c >> @@ -1372,6 +1372,7 @@ static int loadvm_postcopy_handle_advise(MigrationIncomingState *mis) >> return -1; >> } >> >> + init_receivedmap(); >> remote_pagesize_summary = qemu_get_be64(mis->from_src_file); >> local_pagesize_summary = ram_pagesize_summary(); >> >> -- >> 1.9.1 >> > Again, did you trap precopy? sorry, I really forgot, now it's not copied bitmap, it's received bitmap. So, I know about case RAM_SAVE_FLAG_PAGE, but it seems another cases exist (compressed/xbzrle) and there is no single point where coping is doing. > > Thanks, >
* Alexey Perevalov (a.perevalov@samsung.com) wrote: > This patch adds ability to track down already received > pages, it's necessary for calculation vCPU block time in > postcopy migration feature, maybe for restore after > postcopy migration failure. > Also it's necessary to solve shared memory issue in > postcopy livemigration. Information about received pages > will be transferred to the software virtual bridge > (e.g. OVS-VSWITCHD), to avoid fallocate (unmap) for > already received pages. fallocate syscall is required for > remmaped shared memory, due to remmaping itself blocks > ioctl(UFFDIO_COPY, ioctl in this case will end with EEXIT > error (struct page is exists after remmap). > > Bitmap is placed into RAMBlock as another postcopy/precopy > related bitmaps. > > Signed-off-by: Alexey Perevalov <a.perevalov@samsung.com> > --- > include/exec/ram_addr.h | 2 ++ > migration/postcopy-ram.c | 11 ++++++++--- > migration/ram.c | 33 +++++++++++++++++++++++++++++++++ > migration/ram.h | 5 +++++ > migration/savevm.c | 1 + > 5 files changed, 49 insertions(+), 3 deletions(-) > > diff --git a/include/exec/ram_addr.h b/include/exec/ram_addr.h > index 140efa8..c2c1dfe 100644 > --- a/include/exec/ram_addr.h > +++ b/include/exec/ram_addr.h > @@ -47,6 +47,8 @@ struct RAMBlock { > * of the postcopy phase > */ > unsigned long *unsentmap; > + /* bitmap of already received pages in postcopy */ > + unsigned long *receivedmap; > }; > > static inline bool offset_in_ramblock(RAMBlock *b, ram_addr_t offset) > diff --git a/migration/postcopy-ram.c b/migration/postcopy-ram.c > index 38a429a..5ac3ed2 100644 > --- a/migration/postcopy-ram.c > +++ b/migration/postcopy-ram.c > @@ -562,8 +562,13 @@ int postcopy_ram_enable_notify(MigrationIncomingState *mis) > } > > static int qemu_ufd_copy_ioctl(int userfault_fd, void *host_addr, > - void *from_addr, uint64_t pagesize) > + void *from_addr, uint64_t pagesize, RAMBlock *rb) > { > + /* received page isn't feature of blocktime calculation, > + * it's more general entity, so keep it here, > + * but gup betwean two following operation could be high, > + * and in this case blocktime for such small interval will be lost */ > + set_receivedmap_by_addr(host_addr, rb); > if (from_addr) { > struct uffdio_copy copy_struct; > copy_struct.dst = (uint64_t)(uintptr_t)host_addr; > @@ -595,7 +600,7 @@ int postcopy_place_page(MigrationIncomingState *mis, void *host, void *from, > * which would be slightly cheaper, but we'd have to be careful > * of the order of updating our page state. > */ > - if (qemu_ufd_copy_ioctl(mis->userfault_fd, host, from, pagesize)) { > + if (qemu_ufd_copy_ioctl(mis->userfault_fd, host, from, pagesize, rb)) { In your first patch of the series you changed postcopy_place_page to take a RAMBlock* *rather* than a pagesize; perhaps it would be better to do the same here? Other than that looks pretty good. Dave > int e = errno; > error_report("%s: %s copy host: %p from: %p (size: %zd)", > __func__, strerror(e), host, from, pagesize); > @@ -619,7 +624,7 @@ int postcopy_place_page_zero(MigrationIncomingState *mis, void *host, > > pagesize = qemu_ram_pagesize(rb); > if (pagesize == getpagesize()) { > - if (qemu_ufd_copy_ioctl(mis->userfault_fd, host, 0, getpagesize())) { > + if (qemu_ufd_copy_ioctl(mis->userfault_fd, host, 0, getpagesize(), rb)) { > int e = errno; > error_report("%s: %s zero host: %p", > __func__, strerror(e), host); > diff --git a/migration/ram.c b/migration/ram.c > index f50479d..f5ea3c2 100644 > --- a/migration/ram.c > +++ b/migration/ram.c > @@ -151,6 +151,34 @@ out: > return ret; > } > > +void init_receivedmap(void) > +{ > + RAMBlock *rb; > + > + RAMBLOCK_FOREACH(rb) { > + unsigned long pages; > + pages = rb->max_length >> TARGET_PAGE_BITS; > + rb->receivedmap = bitmap_new(pages); > + } > +} > + > +static unsigned long int get_received_bit_offset(void *host_addr, RAMBlock *rb) > +{ > + uint64_t host_addr_offset = (uint64_t)(uintptr_t)(host_addr > + - (void *)rb->host); > + return host_addr_offset >> TARGET_PAGE_BITS; > +} > + > +int test_receivedmap_by_addr(void *host_addr, RAMBlock *rb) > +{ > + return test_bit(get_received_bit_offset(host_addr, rb), rb->receivedmap); > +} > + > +void set_receivedmap_by_addr(void *host_addr, RAMBlock *rb) > +{ > + set_bit_atomic(get_received_bit_offset(host_addr, rb), rb->receivedmap); > +} > + > /* > * An outstanding page request, on the source, having been received > * and queued > @@ -2324,8 +2352,13 @@ static int ram_load_setup(QEMUFile *f, void *opaque) > > static int ram_load_cleanup(void *opaque) > { > + RAMBlock *rb; > xbzrle_load_cleanup(); > compress_threads_load_cleanup(); > + > + RAMBLOCK_FOREACH(rb) { > + g_free(rb->receivedmap); > + } > return 0; > } > > diff --git a/migration/ram.h b/migration/ram.h > index c081fde..7048ff9 100644 > --- a/migration/ram.h > +++ b/migration/ram.h > @@ -52,4 +52,9 @@ int ram_discard_range(const char *block_name, uint64_t start, size_t length); > int ram_postcopy_incoming_init(MigrationIncomingState *mis); > > void ram_handle_compressed(void *host, uint8_t ch, uint64_t size); > + > +void init_receivedmap(void); > +int test_receivedmap_by_addr(void *host_addr, RAMBlock *rb); > +void set_receivedmap_by_addr(void *host_addr, RAMBlock *rb); > + > #endif > diff --git a/migration/savevm.c b/migration/savevm.c > index 31158da..668d3bb 100644 > --- a/migration/savevm.c > +++ b/migration/savevm.c > @@ -1372,6 +1372,7 @@ static int loadvm_postcopy_handle_advise(MigrationIncomingState *mis) > return -1; > } > > + init_receivedmap(); > remote_pagesize_summary = qemu_get_be64(mis->from_src_file); > local_pagesize_summary = ram_pagesize_summary(); > > -- > 1.9.1 > -- Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK
* Alexey Perevalov (a.perevalov@samsung.com) wrote: > On 06/16/2017 12:06 PM, Peter Xu wrote: > > On Thu, Jun 15, 2017 at 07:36:15PM +0300, Alexey Perevalov wrote: > > > > [...] > > > > > +void init_receivedmap(void) > > > +{ > > > + RAMBlock *rb; > > > + > > > + RAMBLOCK_FOREACH(rb) { > > > + unsigned long pages; > > > + pages = rb->max_length >> TARGET_PAGE_BITS; > > Nit: I would assert(!rb->receivedmap) before new it in case for leak. > ok, > > > > > + rb->receivedmap = bitmap_new(pages); > > > + } > > > +} > > > + > > > +static unsigned long int get_received_bit_offset(void *host_addr, RAMBlock *rb) > > > +{ > > > + uint64_t host_addr_offset = (uint64_t)(uintptr_t)(host_addr > > > + - (void *)rb->host); > > > + return host_addr_offset >> TARGET_PAGE_BITS; > > > +} > > Can we rename this function to ramblock_get_pfn() and export it in > > ramblock API (exec/ram_addr.h)? > if you need it, I'll place export into ram_addr.h > but name, PFN could confuse, because in case of hugepage > address this function will not return real PFN, due to > > TARGET_PAGE_BITS usage. > > > > > static inline uint64_t ramblock_get_pfn(RAMBlock *rb, void *host) > > { > > return (host - rb->host) >> TARGET_PAGE_BITS; > > } > > > > > + > > > +int test_receivedmap_by_addr(void *host_addr, RAMBlock *rb) > > > +{ > > > + return test_bit(get_received_bit_offset(host_addr, rb), rb->receivedmap); > > > +} > > > + > > > +void set_receivedmap_by_addr(void *host_addr, RAMBlock *rb) > > > +{ > > > + set_bit_atomic(get_received_bit_offset(host_addr, rb), rb->receivedmap); > > > +} > > > + > > > /* > > > * An outstanding page request, on the source, having been received > > > * and queued > > > @@ -2324,8 +2352,13 @@ static int ram_load_setup(QEMUFile *f, void *opaque) > > > static int ram_load_cleanup(void *opaque) > > > { > > > + RAMBlock *rb; > > > xbzrle_load_cleanup(); > > > compress_threads_load_cleanup(); > > > + > > > + RAMBLOCK_FOREACH(rb) { > > Nit: I'd prefer: if (rb->receivedmap) {...} > g_free already contains it > > > > > + g_free(rb->receivedmap); > > and: rb->receivedmap = NULL later. > why not ) > > > > > + } > > > return 0; > > > } > > > diff --git a/migration/ram.h b/migration/ram.h > > > index c081fde..7048ff9 100644 > > > --- a/migration/ram.h > > > +++ b/migration/ram.h > > > @@ -52,4 +52,9 @@ int ram_discard_range(const char *block_name, uint64_t start, size_t length); > > > int ram_postcopy_incoming_init(MigrationIncomingState *mis); > > > void ram_handle_compressed(void *host, uint8_t ch, uint64_t size); > > > + > > > +void init_receivedmap(void); > > > +int test_receivedmap_by_addr(void *host_addr, RAMBlock *rb); > > > +void set_receivedmap_by_addr(void *host_addr, RAMBlock *rb); > > Nit: I would name these names as: ramblock_recv_map_*(). > > > > > + > > > #endif > > > diff --git a/migration/savevm.c b/migration/savevm.c > > > index 31158da..668d3bb 100644 > > > --- a/migration/savevm.c > > > +++ b/migration/savevm.c > > > @@ -1372,6 +1372,7 @@ static int loadvm_postcopy_handle_advise(MigrationIncomingState *mis) > > > return -1; > > > } > > > + init_receivedmap(); > > > remote_pagesize_summary = qemu_get_be64(mis->from_src_file); > > > local_pagesize_summary = ram_pagesize_summary(); > > > -- > > > 1.9.1 > > > > > Again, did you trap precopy? > sorry, I really forgot, now it's not copied bitmap, it's > received bitmap. So, I know about > case RAM_SAVE_FLAG_PAGE, but it seems another cases exist > (compressed/xbzrle) > and there is no single point where coping is doing. Yes we probably should trap precopy as well; I'm not sure yet whether it'll trigger during the network cases; but certainly in the recovery-from-postcopy-failure case you want to know which pages have been received. You'd also want to clear flags in this bitmap during the discard phase. Dave > > > > > > Thanks, > > > > -- > Best regards, > Alexey Perevalov -- Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK
On 06/16/2017 08:14 PM, Dr. David Alan Gilbert wrote: > * Alexey Perevalov (a.perevalov@samsung.com) wrote: >> This patch adds ability to track down already received >> pages, it's necessary for calculation vCPU block time in >> postcopy migration feature, maybe for restore after >> postcopy migration failure. >> Also it's necessary to solve shared memory issue in >> postcopy livemigration. Information about received pages >> will be transferred to the software virtual bridge >> (e.g. OVS-VSWITCHD), to avoid fallocate (unmap) for >> already received pages. fallocate syscall is required for >> remmaped shared memory, due to remmaping itself blocks >> ioctl(UFFDIO_COPY, ioctl in this case will end with EEXIT >> error (struct page is exists after remmap). >> >> Bitmap is placed into RAMBlock as another postcopy/precopy >> related bitmaps. >> >> Signed-off-by: Alexey Perevalov <a.perevalov@samsung.com> >> --- >> include/exec/ram_addr.h | 2 ++ >> migration/postcopy-ram.c | 11 ++++++++--- >> migration/ram.c | 33 +++++++++++++++++++++++++++++++++ >> migration/ram.h | 5 +++++ >> migration/savevm.c | 1 + >> 5 files changed, 49 insertions(+), 3 deletions(-) >> >> diff --git a/include/exec/ram_addr.h b/include/exec/ram_addr.h >> index 140efa8..c2c1dfe 100644 >> --- a/include/exec/ram_addr.h >> +++ b/include/exec/ram_addr.h >> @@ -47,6 +47,8 @@ struct RAMBlock { >> * of the postcopy phase >> */ >> unsigned long *unsentmap; >> + /* bitmap of already received pages in postcopy */ >> + unsigned long *receivedmap; >> }; >> >> static inline bool offset_in_ramblock(RAMBlock *b, ram_addr_t offset) >> diff --git a/migration/postcopy-ram.c b/migration/postcopy-ram.c >> index 38a429a..5ac3ed2 100644 >> --- a/migration/postcopy-ram.c >> +++ b/migration/postcopy-ram.c >> @@ -562,8 +562,13 @@ int postcopy_ram_enable_notify(MigrationIncomingState *mis) >> } >> >> static int qemu_ufd_copy_ioctl(int userfault_fd, void *host_addr, >> - void *from_addr, uint64_t pagesize) >> + void *from_addr, uint64_t pagesize, RAMBlock *rb) >> { >> + /* received page isn't feature of blocktime calculation, >> + * it's more general entity, so keep it here, >> + * but gup betwean two following operation could be high, >> + * and in this case blocktime for such small interval will be lost */ >> + set_receivedmap_by_addr(host_addr, rb); >> if (from_addr) { >> struct uffdio_copy copy_struct; >> copy_struct.dst = (uint64_t)(uintptr_t)host_addr; >> @@ -595,7 +600,7 @@ int postcopy_place_page(MigrationIncomingState *mis, void *host, void *from, >> * which would be slightly cheaper, but we'd have to be careful >> * of the order of updating our page state. >> */ >> - if (qemu_ufd_copy_ioctl(mis->userfault_fd, host, from, pagesize)) { >> + if (qemu_ufd_copy_ioctl(mis->userfault_fd, host, from, pagesize, rb)) { > In your first patch of the series you changed postcopy_place_page to > take a RAMBlock* *rather* than a pagesize; perhaps it would be better > to do the same here? that's true only for postcopy_place_page where pagesize is getting from ramblock, but not true for postcopy_place_page_zero, where initially was len = getpagesize() > > Other than that looks pretty good. > > Dave > >> int e = errno; >> error_report("%s: %s copy host: %p from: %p (size: %zd)", >> __func__, strerror(e), host, from, pagesize); >> @@ -619,7 +624,7 @@ int postcopy_place_page_zero(MigrationIncomingState *mis, void *host, >> >> pagesize = qemu_ram_pagesize(rb); >> if (pagesize == getpagesize()) { >> - if (qemu_ufd_copy_ioctl(mis->userfault_fd, host, 0, getpagesize())) { >> + if (qemu_ufd_copy_ioctl(mis->userfault_fd, host, 0, getpagesize(), rb)) { >> int e = errno; >> error_report("%s: %s zero host: %p", >> __func__, strerror(e), host); >> diff --git a/migration/ram.c b/migration/ram.c >> index f50479d..f5ea3c2 100644 >> --- a/migration/ram.c >> +++ b/migration/ram.c >> @@ -151,6 +151,34 @@ out: >> return ret; >> } >> >> +void init_receivedmap(void) >> +{ >> + RAMBlock *rb; >> + >> + RAMBLOCK_FOREACH(rb) { >> + unsigned long pages; >> + pages = rb->max_length >> TARGET_PAGE_BITS; >> + rb->receivedmap = bitmap_new(pages); >> + } >> +} >> + >> +static unsigned long int get_received_bit_offset(void *host_addr, RAMBlock *rb) >> +{ >> + uint64_t host_addr_offset = (uint64_t)(uintptr_t)(host_addr >> + - (void *)rb->host); >> + return host_addr_offset >> TARGET_PAGE_BITS; >> +} >> + >> +int test_receivedmap_by_addr(void *host_addr, RAMBlock *rb) >> +{ >> + return test_bit(get_received_bit_offset(host_addr, rb), rb->receivedmap); >> +} >> + >> +void set_receivedmap_by_addr(void *host_addr, RAMBlock *rb) >> +{ >> + set_bit_atomic(get_received_bit_offset(host_addr, rb), rb->receivedmap); >> +} >> + >> /* >> * An outstanding page request, on the source, having been received >> * and queued >> @@ -2324,8 +2352,13 @@ static int ram_load_setup(QEMUFile *f, void *opaque) >> >> static int ram_load_cleanup(void *opaque) >> { >> + RAMBlock *rb; >> xbzrle_load_cleanup(); >> compress_threads_load_cleanup(); >> + >> + RAMBLOCK_FOREACH(rb) { >> + g_free(rb->receivedmap); >> + } >> return 0; >> } >> >> diff --git a/migration/ram.h b/migration/ram.h >> index c081fde..7048ff9 100644 >> --- a/migration/ram.h >> +++ b/migration/ram.h >> @@ -52,4 +52,9 @@ int ram_discard_range(const char *block_name, uint64_t start, size_t length); >> int ram_postcopy_incoming_init(MigrationIncomingState *mis); >> >> void ram_handle_compressed(void *host, uint8_t ch, uint64_t size); >> + >> +void init_receivedmap(void); >> +int test_receivedmap_by_addr(void *host_addr, RAMBlock *rb); >> +void set_receivedmap_by_addr(void *host_addr, RAMBlock *rb); >> + >> #endif >> diff --git a/migration/savevm.c b/migration/savevm.c >> index 31158da..668d3bb 100644 >> --- a/migration/savevm.c >> +++ b/migration/savevm.c >> @@ -1372,6 +1372,7 @@ static int loadvm_postcopy_handle_advise(MigrationIncomingState *mis) >> return -1; >> } >> >> + init_receivedmap(); >> remote_pagesize_summary = qemu_get_be64(mis->from_src_file); >> local_pagesize_summary = ram_pagesize_summary(); >> >> -- >> 1.9.1 >> > -- > Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK > > >
* Alexey Perevalov (a.perevalov@samsung.com) wrote: > On 06/16/2017 08:14 PM, Dr. David Alan Gilbert wrote: > > * Alexey Perevalov (a.perevalov@samsung.com) wrote: > > > This patch adds ability to track down already received > > > pages, it's necessary for calculation vCPU block time in > > > postcopy migration feature, maybe for restore after > > > postcopy migration failure. > > > Also it's necessary to solve shared memory issue in > > > postcopy livemigration. Information about received pages > > > will be transferred to the software virtual bridge > > > (e.g. OVS-VSWITCHD), to avoid fallocate (unmap) for > > > already received pages. fallocate syscall is required for > > > remmaped shared memory, due to remmaping itself blocks > > > ioctl(UFFDIO_COPY, ioctl in this case will end with EEXIT > > > error (struct page is exists after remmap). > > > > > > Bitmap is placed into RAMBlock as another postcopy/precopy > > > related bitmaps. > > > > > > Signed-off-by: Alexey Perevalov <a.perevalov@samsung.com> > > > --- > > > include/exec/ram_addr.h | 2 ++ > > > migration/postcopy-ram.c | 11 ++++++++--- > > > migration/ram.c | 33 +++++++++++++++++++++++++++++++++ > > > migration/ram.h | 5 +++++ > > > migration/savevm.c | 1 + > > > 5 files changed, 49 insertions(+), 3 deletions(-) > > > > > > diff --git a/include/exec/ram_addr.h b/include/exec/ram_addr.h > > > index 140efa8..c2c1dfe 100644 > > > --- a/include/exec/ram_addr.h > > > +++ b/include/exec/ram_addr.h > > > @@ -47,6 +47,8 @@ struct RAMBlock { > > > * of the postcopy phase > > > */ > > > unsigned long *unsentmap; > > > + /* bitmap of already received pages in postcopy */ > > > + unsigned long *receivedmap; > > > }; > > > static inline bool offset_in_ramblock(RAMBlock *b, ram_addr_t offset) > > > diff --git a/migration/postcopy-ram.c b/migration/postcopy-ram.c > > > index 38a429a..5ac3ed2 100644 > > > --- a/migration/postcopy-ram.c > > > +++ b/migration/postcopy-ram.c > > > @@ -562,8 +562,13 @@ int postcopy_ram_enable_notify(MigrationIncomingState *mis) > > > } > > > static int qemu_ufd_copy_ioctl(int userfault_fd, void *host_addr, > > > - void *from_addr, uint64_t pagesize) > > > + void *from_addr, uint64_t pagesize, RAMBlock *rb) > > > { > > > + /* received page isn't feature of blocktime calculation, > > > + * it's more general entity, so keep it here, > > > + * but gup betwean two following operation could be high, > > > + * and in this case blocktime for such small interval will be lost */ > > > + set_receivedmap_by_addr(host_addr, rb); > > > if (from_addr) { > > > struct uffdio_copy copy_struct; > > > copy_struct.dst = (uint64_t)(uintptr_t)host_addr; > > > @@ -595,7 +600,7 @@ int postcopy_place_page(MigrationIncomingState *mis, void *host, void *from, > > > * which would be slightly cheaper, but we'd have to be careful > > > * of the order of updating our page state. > > > */ > > > - if (qemu_ufd_copy_ioctl(mis->userfault_fd, host, from, pagesize)) { > > > + if (qemu_ufd_copy_ioctl(mis->userfault_fd, host, from, pagesize, rb)) { > > In your first patch of the series you changed postcopy_place_page to > > take a RAMBlock* *rather* than a pagesize; perhaps it would be better > > to do the same here? > that's true only for postcopy_place_page where pagesize is getting from > ramblock, > but not true for postcopy_place_page_zero, where initially was len = > getpagesize() Ah, yes: Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com> > > > > Other than that looks pretty good. > > > > Dave > > > > > int e = errno; > > > error_report("%s: %s copy host: %p from: %p (size: %zd)", > > > __func__, strerror(e), host, from, pagesize); > > > @@ -619,7 +624,7 @@ int postcopy_place_page_zero(MigrationIncomingState *mis, void *host, > > > pagesize = qemu_ram_pagesize(rb); > > > if (pagesize == getpagesize()) { > > > - if (qemu_ufd_copy_ioctl(mis->userfault_fd, host, 0, getpagesize())) { > > > + if (qemu_ufd_copy_ioctl(mis->userfault_fd, host, 0, getpagesize(), rb)) { > > > int e = errno; > > > error_report("%s: %s zero host: %p", > > > __func__, strerror(e), host); > > > diff --git a/migration/ram.c b/migration/ram.c > > > index f50479d..f5ea3c2 100644 > > > --- a/migration/ram.c > > > +++ b/migration/ram.c > > > @@ -151,6 +151,34 @@ out: > > > return ret; > > > } > > > +void init_receivedmap(void) > > > +{ > > > + RAMBlock *rb; > > > + > > > + RAMBLOCK_FOREACH(rb) { > > > + unsigned long pages; > > > + pages = rb->max_length >> TARGET_PAGE_BITS; > > > + rb->receivedmap = bitmap_new(pages); > > > + } > > > +} > > > + > > > +static unsigned long int get_received_bit_offset(void *host_addr, RAMBlock *rb) > > > +{ > > > + uint64_t host_addr_offset = (uint64_t)(uintptr_t)(host_addr > > > + - (void *)rb->host); > > > + return host_addr_offset >> TARGET_PAGE_BITS; > > > +} > > > + > > > +int test_receivedmap_by_addr(void *host_addr, RAMBlock *rb) > > > +{ > > > + return test_bit(get_received_bit_offset(host_addr, rb), rb->receivedmap); > > > +} > > > + > > > +void set_receivedmap_by_addr(void *host_addr, RAMBlock *rb) > > > +{ > > > + set_bit_atomic(get_received_bit_offset(host_addr, rb), rb->receivedmap); > > > +} > > > + > > > /* > > > * An outstanding page request, on the source, having been received > > > * and queued > > > @@ -2324,8 +2352,13 @@ static int ram_load_setup(QEMUFile *f, void *opaque) > > > static int ram_load_cleanup(void *opaque) > > > { > > > + RAMBlock *rb; > > > xbzrle_load_cleanup(); > > > compress_threads_load_cleanup(); > > > + > > > + RAMBLOCK_FOREACH(rb) { > > > + g_free(rb->receivedmap); > > > + } > > > return 0; > > > } > > > diff --git a/migration/ram.h b/migration/ram.h > > > index c081fde..7048ff9 100644 > > > --- a/migration/ram.h > > > +++ b/migration/ram.h > > > @@ -52,4 +52,9 @@ int ram_discard_range(const char *block_name, uint64_t start, size_t length); > > > int ram_postcopy_incoming_init(MigrationIncomingState *mis); > > > void ram_handle_compressed(void *host, uint8_t ch, uint64_t size); > > > + > > > +void init_receivedmap(void); > > > +int test_receivedmap_by_addr(void *host_addr, RAMBlock *rb); > > > +void set_receivedmap_by_addr(void *host_addr, RAMBlock *rb); > > > + > > > #endif > > > diff --git a/migration/savevm.c b/migration/savevm.c > > > index 31158da..668d3bb 100644 > > > --- a/migration/savevm.c > > > +++ b/migration/savevm.c > > > @@ -1372,6 +1372,7 @@ static int loadvm_postcopy_handle_advise(MigrationIncomingState *mis) > > > return -1; > > > } > > > + init_receivedmap(); > > > remote_pagesize_summary = qemu_get_be64(mis->from_src_file); > > > local_pagesize_summary = ram_pagesize_summary(); > > > -- > > > 1.9.1 > > > > > -- > > Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK > > > > > > > > -- > Best regards, > Alexey Perevalov -- Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK
On 06/16/2017 09:46 PM, Dr. David Alan Gilbert wrote: > * Alexey Perevalov (a.perevalov@samsung.com) wrote: >> On 06/16/2017 08:14 PM, Dr. David Alan Gilbert wrote: >>> * Alexey Perevalov (a.perevalov@samsung.com) wrote: >>>> This patch adds ability to track down already received >>>> pages, it's necessary for calculation vCPU block time in >>>> postcopy migration feature, maybe for restore after >>>> postcopy migration failure. >>>> Also it's necessary to solve shared memory issue in >>>> postcopy livemigration. Information about received pages >>>> will be transferred to the software virtual bridge >>>> (e.g. OVS-VSWITCHD), to avoid fallocate (unmap) for >>>> already received pages. fallocate syscall is required for >>>> remmaped shared memory, due to remmaping itself blocks >>>> ioctl(UFFDIO_COPY, ioctl in this case will end with EEXIT >>>> error (struct page is exists after remmap). >>>> >>>> Bitmap is placed into RAMBlock as another postcopy/precopy >>>> related bitmaps. >>>> >>>> Signed-off-by: Alexey Perevalov <a.perevalov@samsung.com> >>>> --- >>>> include/exec/ram_addr.h | 2 ++ >>>> migration/postcopy-ram.c | 11 ++++++++--- >>>> migration/ram.c | 33 +++++++++++++++++++++++++++++++++ >>>> migration/ram.h | 5 +++++ >>>> migration/savevm.c | 1 + >>>> 5 files changed, 49 insertions(+), 3 deletions(-) >>>> >>>> diff --git a/include/exec/ram_addr.h b/include/exec/ram_addr.h >>>> index 140efa8..c2c1dfe 100644 >>>> --- a/include/exec/ram_addr.h >>>> +++ b/include/exec/ram_addr.h >>>> @@ -47,6 +47,8 @@ struct RAMBlock { >>>> * of the postcopy phase >>>> */ >>>> unsigned long *unsentmap; >>>> + /* bitmap of already received pages in postcopy */ >>>> + unsigned long *receivedmap; >>>> }; >>>> static inline bool offset_in_ramblock(RAMBlock *b, ram_addr_t offset) >>>> diff --git a/migration/postcopy-ram.c b/migration/postcopy-ram.c >>>> index 38a429a..5ac3ed2 100644 >>>> --- a/migration/postcopy-ram.c >>>> +++ b/migration/postcopy-ram.c >>>> @@ -562,8 +562,13 @@ int postcopy_ram_enable_notify(MigrationIncomingState *mis) >>>> } >>>> static int qemu_ufd_copy_ioctl(int userfault_fd, void *host_addr, >>>> - void *from_addr, uint64_t pagesize) >>>> + void *from_addr, uint64_t pagesize, RAMBlock *rb) >>>> { >>>> + /* received page isn't feature of blocktime calculation, >>>> + * it's more general entity, so keep it here, >>>> + * but gup betwean two following operation could be high, >>>> + * and in this case blocktime for such small interval will be lost */ >>>> + set_receivedmap_by_addr(host_addr, rb); >>>> if (from_addr) { >>>> struct uffdio_copy copy_struct; >>>> copy_struct.dst = (uint64_t)(uintptr_t)host_addr; >>>> @@ -595,7 +600,7 @@ int postcopy_place_page(MigrationIncomingState *mis, void *host, void *from, >>>> * which would be slightly cheaper, but we'd have to be careful >>>> * of the order of updating our page state. >>>> */ >>>> - if (qemu_ufd_copy_ioctl(mis->userfault_fd, host, from, pagesize)) { >>>> + if (qemu_ufd_copy_ioctl(mis->userfault_fd, host, from, pagesize, rb)) { >>> In your first patch of the series you changed postcopy_place_page to >>> take a RAMBlock* *rather* than a pagesize; perhaps it would be better >>> to do the same here? >> that's true only for postcopy_place_page where pagesize is getting from >> ramblock, >> but not true for postcopy_place_page_zero, where initially was len = >> getpagesize() > Ah, yes: > > Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com> don't rush ;) I resent patch and added precopy/discard cases there. I'll be in vacation during next week, but I'll try to check mailing list. > >>> Other than that looks pretty good. >>> >>> Dave >>> >>>> int e = errno; >>>> error_report("%s: %s copy host: %p from: %p (size: %zd)", >>>> __func__, strerror(e), host, from, pagesize); >>>> @@ -619,7 +624,7 @@ int postcopy_place_page_zero(MigrationIncomingState *mis, void *host, >>>> pagesize = qemu_ram_pagesize(rb); >>>> if (pagesize == getpagesize()) { >>>> - if (qemu_ufd_copy_ioctl(mis->userfault_fd, host, 0, getpagesize())) { >>>> + if (qemu_ufd_copy_ioctl(mis->userfault_fd, host, 0, getpagesize(), rb)) { >>>> int e = errno; >>>> error_report("%s: %s zero host: %p", >>>> __func__, strerror(e), host); >>>> diff --git a/migration/ram.c b/migration/ram.c >>>> index f50479d..f5ea3c2 100644 >>>> --- a/migration/ram.c >>>> +++ b/migration/ram.c >>>> @@ -151,6 +151,34 @@ out: >>>> return ret; >>>> } >>>> +void init_receivedmap(void) >>>> +{ >>>> + RAMBlock *rb; >>>> + >>>> + RAMBLOCK_FOREACH(rb) { >>>> + unsigned long pages; >>>> + pages = rb->max_length >> TARGET_PAGE_BITS; >>>> + rb->receivedmap = bitmap_new(pages); >>>> + } >>>> +} >>>> + >>>> +static unsigned long int get_received_bit_offset(void *host_addr, RAMBlock *rb) >>>> +{ >>>> + uint64_t host_addr_offset = (uint64_t)(uintptr_t)(host_addr >>>> + - (void *)rb->host); >>>> + return host_addr_offset >> TARGET_PAGE_BITS; >>>> +} >>>> + >>>> +int test_receivedmap_by_addr(void *host_addr, RAMBlock *rb) >>>> +{ >>>> + return test_bit(get_received_bit_offset(host_addr, rb), rb->receivedmap); >>>> +} >>>> + >>>> +void set_receivedmap_by_addr(void *host_addr, RAMBlock *rb) >>>> +{ >>>> + set_bit_atomic(get_received_bit_offset(host_addr, rb), rb->receivedmap); >>>> +} >>>> + >>>> /* >>>> * An outstanding page request, on the source, having been received >>>> * and queued >>>> @@ -2324,8 +2352,13 @@ static int ram_load_setup(QEMUFile *f, void *opaque) >>>> static int ram_load_cleanup(void *opaque) >>>> { >>>> + RAMBlock *rb; >>>> xbzrle_load_cleanup(); >>>> compress_threads_load_cleanup(); >>>> + >>>> + RAMBLOCK_FOREACH(rb) { >>>> + g_free(rb->receivedmap); >>>> + } >>>> return 0; >>>> } >>>> diff --git a/migration/ram.h b/migration/ram.h >>>> index c081fde..7048ff9 100644 >>>> --- a/migration/ram.h >>>> +++ b/migration/ram.h >>>> @@ -52,4 +52,9 @@ int ram_discard_range(const char *block_name, uint64_t start, size_t length); >>>> int ram_postcopy_incoming_init(MigrationIncomingState *mis); >>>> void ram_handle_compressed(void *host, uint8_t ch, uint64_t size); >>>> + >>>> +void init_receivedmap(void); >>>> +int test_receivedmap_by_addr(void *host_addr, RAMBlock *rb); >>>> +void set_receivedmap_by_addr(void *host_addr, RAMBlock *rb); >>>> + >>>> #endif >>>> diff --git a/migration/savevm.c b/migration/savevm.c >>>> index 31158da..668d3bb 100644 >>>> --- a/migration/savevm.c >>>> +++ b/migration/savevm.c >>>> @@ -1372,6 +1372,7 @@ static int loadvm_postcopy_handle_advise(MigrationIncomingState *mis) >>>> return -1; >>>> } >>>> + init_receivedmap(); >>>> remote_pagesize_summary = qemu_get_be64(mis->from_src_file); >>>> local_pagesize_summary = ram_pagesize_summary(); >>>> -- >>>> 1.9.1 >>>> >>> -- >>> Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK >>> >>> >>> >> -- >> Best regards, >> Alexey Perevalov > -- > Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK > > >
diff --git a/include/exec/ram_addr.h b/include/exec/ram_addr.h index 140efa8..c2c1dfe 100644 --- a/include/exec/ram_addr.h +++ b/include/exec/ram_addr.h @@ -47,6 +47,8 @@ struct RAMBlock { * of the postcopy phase */ unsigned long *unsentmap; + /* bitmap of already received pages in postcopy */ + unsigned long *receivedmap; }; static inline bool offset_in_ramblock(RAMBlock *b, ram_addr_t offset) diff --git a/migration/postcopy-ram.c b/migration/postcopy-ram.c index 38a429a..5ac3ed2 100644 --- a/migration/postcopy-ram.c +++ b/migration/postcopy-ram.c @@ -562,8 +562,13 @@ int postcopy_ram_enable_notify(MigrationIncomingState *mis) } static int qemu_ufd_copy_ioctl(int userfault_fd, void *host_addr, - void *from_addr, uint64_t pagesize) + void *from_addr, uint64_t pagesize, RAMBlock *rb) { + /* received page isn't feature of blocktime calculation, + * it's more general entity, so keep it here, + * but gup betwean two following operation could be high, + * and in this case blocktime for such small interval will be lost */ + set_receivedmap_by_addr(host_addr, rb); if (from_addr) { struct uffdio_copy copy_struct; copy_struct.dst = (uint64_t)(uintptr_t)host_addr; @@ -595,7 +600,7 @@ int postcopy_place_page(MigrationIncomingState *mis, void *host, void *from, * which would be slightly cheaper, but we'd have to be careful * of the order of updating our page state. */ - if (qemu_ufd_copy_ioctl(mis->userfault_fd, host, from, pagesize)) { + if (qemu_ufd_copy_ioctl(mis->userfault_fd, host, from, pagesize, rb)) { int e = errno; error_report("%s: %s copy host: %p from: %p (size: %zd)", __func__, strerror(e), host, from, pagesize); @@ -619,7 +624,7 @@ int postcopy_place_page_zero(MigrationIncomingState *mis, void *host, pagesize = qemu_ram_pagesize(rb); if (pagesize == getpagesize()) { - if (qemu_ufd_copy_ioctl(mis->userfault_fd, host, 0, getpagesize())) { + if (qemu_ufd_copy_ioctl(mis->userfault_fd, host, 0, getpagesize(), rb)) { int e = errno; error_report("%s: %s zero host: %p", __func__, strerror(e), host); diff --git a/migration/ram.c b/migration/ram.c index f50479d..f5ea3c2 100644 --- a/migration/ram.c +++ b/migration/ram.c @@ -151,6 +151,34 @@ out: return ret; } +void init_receivedmap(void) +{ + RAMBlock *rb; + + RAMBLOCK_FOREACH(rb) { + unsigned long pages; + pages = rb->max_length >> TARGET_PAGE_BITS; + rb->receivedmap = bitmap_new(pages); + } +} + +static unsigned long int get_received_bit_offset(void *host_addr, RAMBlock *rb) +{ + uint64_t host_addr_offset = (uint64_t)(uintptr_t)(host_addr + - (void *)rb->host); + return host_addr_offset >> TARGET_PAGE_BITS; +} + +int test_receivedmap_by_addr(void *host_addr, RAMBlock *rb) +{ + return test_bit(get_received_bit_offset(host_addr, rb), rb->receivedmap); +} + +void set_receivedmap_by_addr(void *host_addr, RAMBlock *rb) +{ + set_bit_atomic(get_received_bit_offset(host_addr, rb), rb->receivedmap); +} + /* * An outstanding page request, on the source, having been received * and queued @@ -2324,8 +2352,13 @@ static int ram_load_setup(QEMUFile *f, void *opaque) static int ram_load_cleanup(void *opaque) { + RAMBlock *rb; xbzrle_load_cleanup(); compress_threads_load_cleanup(); + + RAMBLOCK_FOREACH(rb) { + g_free(rb->receivedmap); + } return 0; } diff --git a/migration/ram.h b/migration/ram.h index c081fde..7048ff9 100644 --- a/migration/ram.h +++ b/migration/ram.h @@ -52,4 +52,9 @@ int ram_discard_range(const char *block_name, uint64_t start, size_t length); int ram_postcopy_incoming_init(MigrationIncomingState *mis); void ram_handle_compressed(void *host, uint8_t ch, uint64_t size); + +void init_receivedmap(void); +int test_receivedmap_by_addr(void *host_addr, RAMBlock *rb); +void set_receivedmap_by_addr(void *host_addr, RAMBlock *rb); + #endif diff --git a/migration/savevm.c b/migration/savevm.c index 31158da..668d3bb 100644 --- a/migration/savevm.c +++ b/migration/savevm.c @@ -1372,6 +1372,7 @@ static int loadvm_postcopy_handle_advise(MigrationIncomingState *mis) return -1; } + init_receivedmap(); remote_pagesize_summary = qemu_get_be64(mis->from_src_file); local_pagesize_summary = ram_pagesize_summary();
This patch adds ability to track down already received pages, it's necessary for calculation vCPU block time in postcopy migration feature, maybe for restore after postcopy migration failure. Also it's necessary to solve shared memory issue in postcopy livemigration. Information about received pages will be transferred to the software virtual bridge (e.g. OVS-VSWITCHD), to avoid fallocate (unmap) for already received pages. fallocate syscall is required for remmaped shared memory, due to remmaping itself blocks ioctl(UFFDIO_COPY, ioctl in this case will end with EEXIT error (struct page is exists after remmap). Bitmap is placed into RAMBlock as another postcopy/precopy related bitmaps. Signed-off-by: Alexey Perevalov <a.perevalov@samsung.com> --- include/exec/ram_addr.h | 2 ++ migration/postcopy-ram.c | 11 ++++++++--- migration/ram.c | 33 +++++++++++++++++++++++++++++++++ migration/ram.h | 5 +++++ migration/savevm.c | 1 + 5 files changed, 49 insertions(+), 3 deletions(-)