destination:
../qemu/build/qemu-system-x86_64 -enable-kvm -netdev tap,id=hn0,script=/etc/qemu-ifup,downscript=/etc/qemu-ifdown -device e1000,netdev=hn0,mac=50:52:54:00:11:22 -boot c -drive if=none,file=./Fedora-rdma-server-migration.qcow2,id=drive-virtio-disk0 -device virtio-blk-pci,bus=pci.0,addr=0x4,drive=drive-virtio-disk0,id=virtio-disk0 -m 2048 -smp 2 -device piix3-usb-uhci -device usb-tablet -monitor stdio -vga qxl -spice streaming-video=filter,port=5902,disable-ticketing -incoming rdma:192.168.22.23:8888
qemu-system-x86_64: -spice streaming-video=filter,port=5902,disable-ticketing: warning: short-form boolean option 'disable-ticketing' deprecated
Please use disable-ticketing=on instead
QEMU 6.0.50 monitor - type 'help' for more information
(qemu) trace-event qemu_rdma_block_for_wrid_miss on
(qemu) dest_init RDMA Device opened: kernel name rxe_eth0 uverbs device name uverbs2, infiniband_verbs class device path /sys/class/infiniband_verbs/uverbs2, infiniband class device path /sys/class/infiniband/rxe_eth0, transport: (2) Ethernet
qemu_rdma_block_for_wrid_miss A Wanted wrid CONTROL SEND (2000) but got CONTROL RECV (4000)
source:
../qemu/build/qemu-system-x86_64 -enable-kvm -netdev tap,id=hn0,script=/etc/qemu-ifup,downscript=/etc/qemu-ifdown -device e1000,netdev=hn0,mac=50:52:54:00:11:22 -boot c -drive if=none,file=./Fedora-rdma-server.qcow2,id=drive-virtio-disk0 -device virtio-blk-pci,bus=pci.0,addr=0x4,drive=drive-virtio-disk0,id=virtio-disk0 -m 2048 -smp 2 -device piix3-usb-uhci -device usb-tablet -monitor stdio -vga qxl -spice streaming-video=filter,port=5901,disable-ticketing -S
qemu-system-x86_64: -spice streaming-video=filter,port=5901,disable-ticketing: warning: short-form boolean option 'disable-ticketing' deprecated
Please use disable-ticketing=on instead
QEMU 6.0.50 monitor - type 'help' for more information
(qemu)
(qemu) trace-event qemu_rdma_block_for_wrid_miss on
(qemu) migrate -d rdma:192.168.22.23:8888
source_resolve_host RDMA Device opened: kernel name rxe_eth0 uverbs device name uverbs2, infiniband_verbs class device path /sys/class/infiniband_verbs/uverbs2, infiniband class device path /sys/class/infiniband/rxe_eth0, transport: (2) Ethernet
(qemu) qemu_rdma_block_for_wrid_miss A Wanted wrid WRITE RDMA (1) but got CONTROL RECV (4000)
NOTE: we use soft RoCE as the rdma device.
[root@iaas-rpma images]# rdma link show rxe_eth0/1
link rxe_eth0/1 state ACTIVE physical_state LINK_UP netdev eth0
This migration could not be completed when out of order(OOO) CQ event occurs.
The send queue and receive queue shared a same completion queue, and
qemu_rdma_block_for_wrid() will drop the CQs it's not interested in. But
the dropped CQs by qemu_rdma_block_for_wrid() could be later CQs it wants.
So in this case, qemu_rdma_block_for_wrid() will block forever.
OOO cases will occur in both source side and destination side. And a
forever blocking happens on only SEND and RECV are out of order. OOO between
'WRITE RDMA' and 'RECV' doesn't matter.
below the OOO sequence:
source destination
rdma_write_one() qemu_rdma_registration_handle()
1. S1: post_recv X D1: post_recv Y
2. wait for recv CQ event X
3. D2: post_send X ---------------+
4. wait for send CQ send event X (D2) |
5. recv CQ event X reaches (D2) |
6. +-S2: post_send Y |
7. | wait for send CQ event Y |
8. | recv CQ event Y (S2) (drop it) |
9. +-send CQ event Y reaches (S2) |
10. send CQ event X reaches (D2) -----+
11. wait recv CQ event Y (dropped by (8))
Although a hardware IB works fine in my a hundred of runs, the IB specification
doesn't guaratee the CQ order in such case.
Here we introduce a independent send completion queue to distinguish
ibv_post_send completion queue from the original mixed completion queue.
It helps us to poll the specific CQE we are really interested in.
Signed-off-by: Li Zhijian <lizhijian@cn.fujitsu.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
The responder mr registering with ODP will sent RNR NAK back to
the requester in the face of the page fault.
---------
ibv_poll_cq wc.status=13 RNR retry counter exceeded!
ibv_poll_cq wrid=WRITE RDMA!
---------
ibv_advise_mr(3) helps to make pages present before the actual IO is
conducted so that the responder does page fault as little as possible.
Signed-off-by: Li Zhijian <lizhijian@cn.fujitsu.com>
Reviewed-by: Marcel Apfelbaum <marcel.apfelbaum@gmail.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
To: <quintela@redhat.com>, <dgilbert@redhat.com>, <qemu-devel@nongnu.org>
CC: Li Zhijian <lizhijian@cn.fujitsu.com>
Date: Sat, 31 Jul 2021 22:05:51 +0800 (5 weeks, 4 days, 17 hours ago)
multifd with unsupported protocol will cause a segment fault.
(gdb) bt
#0 0x0000563b4a93faf8 in socket_connect (addr=0x0, errp=0x7f7f02675410) at ../util/qemu-sockets.c:1190
#1 0x0000563b4a797a03 in qio_channel_socket_connect_sync
(ioc=0x563b4d16e8c0, addr=0x0, errp=0x7f7f02675410) at
../io/channel-socket.c:145
#2 0x0000563b4a797abf in qio_channel_socket_connect_worker (task=0x563b4cd86c30, opaque=0x0) at ../io/channel-socket.c:168
#3 0x0000563b4a792631 in qio_task_thread_worker (opaque=0x563b4cd86c30) at ../io/task.c:124
#4 0x0000563b4a91da69 in qemu_thread_start (args=0x563b4c44bb80) at ../util/qemu-thread-posix.c:541
#5 0x00007f7fe9b5b3f9 in ?? ()
#6 0x0000000000000000 in ?? ()
It's enough to check migrate_multifd_is_allowed() in multifd cleanup() and
multifd setup() though there are so many other places using migrate_use_multifd().
Signed-off-by: Li Zhijian <lizhijian@cn.fujitsu.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
backtrace:
'0x00007ffff5f44ec2 in __ibv_dereg_mr_1_1 (mr=0x7fff1007d390) at /home/lizhijian/rdma-core/libibverbs/verbs.c:478
478 void *addr = mr->addr;
(gdb) bt
#0 0x00007ffff5f44ec2 in __ibv_dereg_mr_1_1 (mr=0x7fff1007d390) at /home/lizhijian/rdma-core/libibverbs/verbs.c:478
#1 0x0000555555891fcc in rdma_delete_block (block=<optimized out>, rdma=0x7fff38176010) at ../migration/rdma.c:691
#2 qemu_rdma_cleanup (rdma=0x7fff38176010) at ../migration/rdma.c:2365
#3 0x00005555558925b0 in qio_channel_rdma_close_rcu (rcu=0x555556b8b6c0) at ../migration/rdma.c:3073
#4 0x0000555555d652a3 in call_rcu_thread (opaque=opaque@entry=0x0) at ../util/rcu.c:281
#5 0x0000555555d5edf9 in qemu_thread_start (args=0x7fffe88bb4d0) at ../util/qemu-thread-posix.c:541
#6 0x00007ffff54c73f9 in start_thread () at /lib64/libpthread.so.0
#7 0x00007ffff53f3b03 in clone () at /lib64/libc.so.6 '
Signed-off-by: Li Zhijian <lizhijian@cn.fujitsu.com>
Message-Id: <20210708144521.1959614-1-lizhijian@cn.fujitsu.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Since the prior calls are successful, in this case a errno doesn't
indicate a real error which would just make us confused.
before:
(qemu) migrate -d rdma:192.168.22.23:8888
source_resolve_host RDMA Device opened: kernel name rxe_eth0 uverbs device name uverbs2, infiniband_verbs class device path /sys/class/infiniband_verbs/uverbs2, infiniband class device path /sys/class/infiniband/rxe_eth0, transport: (2) Ethernet
rdma_get_cm_event != EVENT_ESTABLISHED after rdma_connect: No space left on device
Signed-off-by: Li Zhijian <lizhijian@cn.fujitsu.com>
Message-Id: <20210628071959.23455-1-lizhijian@cn.fujitsu.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
destination side:
$ build/qemu-system-x86_64 -enable-kvm -netdev tap,id=hn0,script=/etc/qemu-ifup,downscript=/etc/qemu-ifdown -device e1000,netdev=hn0,mac=50:52:54:00:11:22 -boot c -drive if=none,file=./Fedora-rdma-server-migration.qcow2,id=drive-virtio-disk0 -device virtio-blk-pci,bus=pci.0,addr=0x4,drive=drive-virtio-disk0,id=virtio-disk0 -m 2048 -smp 2 -device piix3-usb-uhci -device usb-tablet -monitor stdio -vga qxl -spice streaming-video=filter,port=5902,disable-ticketing -incoming rdma:192.168.1.10:8888
(qemu) migrate_set_capability postcopy-ram on
(qemu)
dest_init RDMA Device opened: kernel name rocep1s0f0 uverbs device name uverbs0, infiniband_verbs class device path /sys/class/infiniband_verbs/uverbs0, infiniband class device path /sys/class/infiniband/rocep1s0f0, transport: (2) Ethernet
Segmentation fault (core dumped)
(gdb) bt
#0 qemu_rdma_accept (rdma=0x0) at ../migration/rdma.c:3272
#1 rdma_accept_incoming_migration (opaque=0x0) at ../migration/rdma.c:3986
#2 0x0000563c9e51f02a in aio_dispatch_handler
(ctx=ctx@entry=0x563ca0606010, node=0x563ca12b2150) at ../util/aio-posix.c:329
#3 0x0000563c9e51f752 in aio_dispatch_handlers (ctx=0x563ca0606010) at ../util/aio-posix.c:372
#4 aio_dispatch (ctx=0x563ca0606010) at ../util/aio-posix.c:382
#5 0x0000563c9e4f4d9e in aio_ctx_dispatch (source=<optimized out>, callback=<optimized out>, user_data=<optimized out>) at ../util/async.c:306
#6 0x00007fe96ef3fa9f in g_main_context_dispatch () at /lib64/libglib-2.0.so.0
#7 0x0000563c9e4ffeb8 in glib_pollfds_poll () at ../util/main-loop.c:231
#8 os_host_main_loop_wait (timeout=12188789) at ../util/main-loop.c:254
#9 main_loop_wait (nonblocking=nonblocking@entry=0) at ../util/main-loop.c:530
#10 0x0000563c9e3c7211 in qemu_main_loop () at ../softmmu/runstate.c:725
#11 0x0000563c9dfd46fe in main (argc=<optimized out>, argv=<optimized out>, envp=<optimized out>) at ../softmmu/main.c:50
The rdma return path will not be created when qemu incoming is starting
since migrate_copy() is false at that moment, then a NULL return path
rdma was referenced if the user enabled postcopy later.
Signed-off-by: Li Zhijian <lizhijian@cn.fujitsu.com>
Message-Id: <20210525080552.28259-3-lizhijian@cn.fujitsu.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
A segmentation fault was triggered when i try to abort a postcopy + rdma
migration.
since rdma_ack_cm_event releases a uninitialized cm_event in these case.
like below:
2496 ret = rdma_get_cm_event(rdma->channel, &cm_event);
2497 if (ret) {
2498 perror("rdma_get_cm_event after rdma_connect");
2499 ERROR(errp, "connecting to destination!");
2500 rdma_ack_cm_event(cm_event); <<<< cause segmentation fault
2501 goto err_rdma_source_connect;
2502 }
Refer to the rdma_get_cm_event() code, cm_event will be
updated/changed only if rdma_get_cm_event() returns 0. So it's okey to
remove the ack in error patch.
Signed-off-by: Li Zhijian <lizhijian@cn.fujitsu.com>
Message-Id: <20210519064740.10828-1-lizhijian@cn.fujitsu.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Since linux commit: cf8fa920cb42 ("i386: handle an initrd in highmem (version 2)")
linux has supported initrd up to 4 GB, but the header field
ramdisk_max is still set to 2 GB to avoid "possible bootloader bugs".
When use '-kernel vmlinux -initrd initrd.cgz' to launch a VM,
the firmware(it could be linuxboot_dma.bin) helps to read initrd
contents into guest memory(below ramdisk_max) and jump to kernel.
that's similar with what bootloader does, like grub.
In addition, initrd_max is uint32_t simply because QEMU doesn't support
the 64-bit boot protocol (specifically the ext_ramdisk_image field).
Therefore here just limit initrd_max to UINT32_MAX simply as well to
allow initrd to be loaded below 4 GB.
NOTE: it's possible that linux protocol within [0x208, 0x20c]
supports up to 4 GB initrd as well.
CC: Paolo Bonzini <pbonzini@redhat.com>
CC: Richard Henderson <rth@twiddle.net>
CC: Eduardo Habkost <ehabkost@redhat.com>
CC: "Michael S. Tsirkin" <mst@redhat.com>
CC: Marcel Apfelbaum <marcel.apfelbaum@gmail.com>
Signed-off-by: Li Zhijian <lizhijian@cn.fujitsu.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Some address/memory APIs have different type between
'hwaddr/target_ulong addr' and 'int len'. It is very unsafe, especially
some APIs will be passed a non-int len by caller which might cause
overflow quietly.
Below is an potential overflow case:
dma_memory_read(uint32_t len)
-> dma_memory_rw(uint32_t len)
-> dma_memory_rw_relaxed(uint32_t len)
-> address_space_rw(int len) # len overflow
CC: Paolo Bonzini <pbonzini@redhat.com>
CC: Peter Crosthwaite <crosthwaite.peter@gmail.com>
CC: Richard Henderson <rth@twiddle.net>
CC: Peter Maydell <peter.maydell@linaro.org>
CC: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: Li Zhijian <lizhijian@cn.fujitsu.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Previously, if the size of initrd >=2G, qemu exits with error:
root@haswell-OptiPlex-9020:/home/lizj# /home/lizhijian/lkp/qemu-colo/x86_64-softmmu/qemu-system-x86_64 -kernel ./vmlinuz-4.16.0-rc4 -initrd large.cgz -nographic
qemu: error reading initrd large.cgz: No such file or directory
root@haswell-OptiPlex-9020:/home/lizj# du -sh large.cgz
2.5G large.cgz
this patch changes the caller side that use this function to calculate
size of initrd file as well.
v2: update error message and int64_t printing format
Signed-off-by: Li Zhijian <lizhijian@cn.fujitsu.com>
Message-Id: <1536833233-14121-1-git-send-email-lizhijian@cn.fujitsu.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Previously, if we attach more than one filters for a single netdev,
both ingress and egress traffic will go through net filters in same
order like:
ingress: netdev ->filter1 ->filter2 ->...filter[n] ->emulated device
egress: emulated device ->filter1 ->filter2 ->...filter[n] ->netdev.
This is against the natural feeling and will complicate filters
configuration since in some scenes, we hope filters handle the egress
traffic in a reverse order. For example, in colo-proxy (will be
implemented later), we have a redirector filter and a colo-rewriter
filter, we need the filter behave like:
ingress(->)/egress(<-): chardev<->redirector<->colo-rewriter<->emulated device
Since both buffer filter and dump do not require strict order of
filters, this patch switches to always let egress traffic walk through
net filters in reverse to simplify the possible filters configuration
in the future.
Signed-off-by: Wen Congyang <wency@cn.fujitsu.com>
Signed-off-by: Li Zhijian <lizhijian@cn.fujitsu.com>
Reviewed-by: Yang Hongyang <hongyang.yang@easystack.cn>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Prevously, if we hotplug a device(e.g. device_add e1000) during
migration is processing in source side, qemu will add a new ram
block but migration_bitmap is not extended.
In this case, migration_bitmap will overflow and lead qemu abort
unexpectedly.
Signed-off-by: Li Zhijian <lizhijian@cn.fujitsu.com>
Signed-off-by: Wen Congyang <wency@cn.fujitsu.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Li Zhijian