The call "skb_copy_from_linear_data(skb, inl + 1, spc)" triggers a FORTIFY
memcpy() warning on ppc64 platform:
In function ‘fortify_memcpy_chk’,
inlined from ‘skb_copy_from_linear_data’ at ./include/linux/skbuff.h:4029:2,
inlined from ‘build_inline_wqe’ at drivers/net/ethernet/mellanox/mlx4/en_tx.c:722:4,
inlined from ‘mlx4_en_xmit’ at drivers/net/ethernet/mellanox/mlx4/en_tx.c:1066:3:
./include/linux/fortify-string.h:513:25: error: call to ‘__write_overflow_field’ declared with
attribute warning: detected write beyond size of field (1st parameter); maybe use struct_group()?
[-Werror=attribute-warning]
513 | __write_overflow_field(p_size_field, size);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Same behaviour on x86 you can get if you use "__always_inline" instead of
"inline" for skb_copy_from_linear_data() in skbuff.h
The call here copies data into inlined tx destricptor, which has 104
bytes (MAX_INLINE) space for data payload. In this case "spc" is known
in compile-time but the destination is used with hidden knowledge
(real structure of destination is different from that the compiler
can see). That cause the fortify warning because compiler can check
bounds, but the real bounds are different. "spc" can't be bigger than
64 bytes (MLX4_INLINE_ALIGN), so the data can always fit into inlined
tx descriptor. The fact that "inl" points into inlined tx descriptor is
determined earlier in mlx4_en_xmit().
Avoid confusing the compiler with "inl + 1" constructions to get to past
the inl header by introducing a flexible array "data" to the struct so
that the compiler can see that we are not dealing with an array of inl
structs, but rather, arbitrary data following the structure. There are
no changes to the structure layout reported by pahole, and the resulting
machine code is actually smaller.
Reported-by: Josef Oskera <joskera@redhat.com>
Link: https://lore.kernel.org/lkml/20230217094541.2362873-1-joskera@redhat.com
Fixes: f68f2ff915 ("fortify: Detect struct member overflows in memcpy() at compile-time")
Cc: Yishai Hadas <yishaih@nvidia.com>
Signed-off-by: Kees Cook <keescook@chromium.org>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Link: https://lore.kernel.org/r/20230218183842.never.954-kees@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Currently, the cifs I/O paths hand lists of pages from the VM interface
routines at the top all the way through the intervening layers to the
socket interface at the bottom.
This is a problem, however, for interfacing with netfslib which passes an
iterator through to the ->issue_read() method (and will pass an iterator
through to the ->issue_write() method in future). Netfslib takes over
bounce buffering for direct I/O, async I/O and encrypted content, so cifs
doesn't need to do that. Netfslib also converts IOVEC-type iterators into
BVEC-type iterators if necessary.
Further, cifs needs foliating - and folios may come in a variety of sizes,
so a page list pointing to an array of heterogeneous pages may cause
problems in places such as where crypto is done.
Change the cifs I/O paths to hand iov_iter iterators all the way through
instead.
Notes:
(1) Some old routines are #if'd out to be removed in a follow up patch so
as to avoid confusing diff, thereby making the diff output easier to
follow. I've removed functions that don't overlap with anything
added.
(2) struct smb_rqst loses rq_pages, rq_offset, rq_npages, rq_pagesz and
rq_tailsz which describe the pages forming the buffer; instead there's
an rq_iter describing the source buffer and an rq_buffer which is used
to hold the buffer for encryption.
(3) struct cifs_readdata and cifs_writedata are similarly modified to
smb_rqst. The ->read_into_pages() and ->copy_into_pages() are then
replaced with passing the iterator directly to the socket.
The iterators are stored in these structs so that they are persistent
and don't get deallocated when the function returns (unlike if they
were stack variables).
(4) Buffered writeback is overhauled, borrowing the code from the afs
filesystem to gather up contiguous runs of folios. The XARRAY-type
iterator is then used to refer directly to the pagecache and can be
passed to the socket to transmit data directly from there.
This includes:
cifs_extend_writeback()
cifs_write_back_from_locked_folio()
cifs_writepages_region()
cifs_writepages()
(5) Pages are converted to folios.
(6) Direct I/O uses netfs_extract_user_iter() to create a BVEC-type
iterator from an IOBUF/UBUF-type source iterator.
(7) smb2_get_aead_req() uses netfs_extract_iter_to_sg() to extract page
fragments from the iterator into the scatterlists that the crypto
layer prefers.
(8) smb2_init_transform_rq() attached pages to smb_rqst::rq_buffer, an
xarray, to use as a bounce buffer for encryption. An XARRAY-type
iterator can then be used to pass the bounce buffer to lower layers.
Signed-off-by: David Howells <dhowells@redhat.com>
cc: Steve French <sfrench@samba.org>
cc: Shyam Prasad N <nspmangalore@gmail.com>
cc: Rohith Surabattula <rohiths.msft@gmail.com>
cc: Paulo Alcantara <pc@cjr.nz>
cc: Jeff Layton <jlayton@kernel.org>
cc: linux-cifs@vger.kernel.org
Link: https://lore.kernel.org/r/164311907995.2806745.400147335497304099.stgit@warthog.procyon.org.uk/ # rfc
Link: https://lore.kernel.org/r/164928620163.457102.11602306234438271112.stgit@warthog.procyon.org.uk/ # v1
Link: https://lore.kernel.org/r/165211420279.3154751.15923591172438186144.stgit@warthog.procyon.org.uk/ # v1
Link: https://lore.kernel.org/r/165348880385.2106726.3220789453472800240.stgit@warthog.procyon.org.uk/ # v1
Link: https://lore.kernel.org/r/165364827111.3334034.934805882842932881.stgit@warthog.procyon.org.uk/ # v3
Link: https://lore.kernel.org/r/166126396180.708021.271013668175370826.stgit@warthog.procyon.org.uk/ # v1
Link: https://lore.kernel.org/r/166697259595.61150.5982032408321852414.stgit@warthog.procyon.org.uk/ # rfc
Link: https://lore.kernel.org/r/166732031756.3186319.12528413619888902872.stgit@warthog.procyon.org.uk/ # rfc
Signed-off-by: Steve French <stfrench@microsoft.com>
Daniel Borkmann says:
====================
pull-request: bpf-next 2023-02-17
We've added 64 non-merge commits during the last 7 day(s) which contain
a total of 158 files changed, 4190 insertions(+), 988 deletions(-).
The main changes are:
1) Add a rbtree data structure following the "next-gen data structure"
precedent set by recently-added linked-list, that is, by using
kfunc + kptr instead of adding a new BPF map type, from Dave Marchevsky.
2) Add a new benchmark for hashmap lookups to BPF selftests,
from Anton Protopopov.
3) Fix bpf_fib_lookup to only return valid neighbors and add an option
to skip the neigh table lookup, from Martin KaFai Lau.
4) Add cgroup.memory=nobpf kernel parameter option to disable BPF memory
accouting for container environments, from Yafang Shao.
5) Batch of ice multi-buffer and driver performance fixes,
from Alexander Lobakin.
6) Fix a bug in determining whether global subprog's argument is
PTR_TO_CTX, which is based on type names which breaks kprobe progs,
from Andrii Nakryiko.
7) Prep work for future -mcpu=v4 LLVM option which includes usage of
BPF_ST insn. Thus improve BPF_ST-related value tracking in verifier,
from Eduard Zingerman.
8) More prep work for later building selftests with Memory Sanitizer
in order to detect usages of undefined memory, from Ilya Leoshkevich.
9) Fix xsk sockets to check IFF_UP earlier to avoid a NULL pointer
dereference via sendmsg(), from Maciej Fijalkowski.
10) Implement BPF trampoline for RV64 JIT compiler, from Pu Lehui.
11) Fix BPF memory allocator in combination with BPF hashtab where it could
corrupt special fields e.g. used in bpf_spin_lock, from Hou Tao.
12) Fix LoongArch BPF JIT to always use 4 instructions for function
address so that instruction sequences don't change between passes,
from Hengqi Chen.
* tag 'for-netdev' of https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next: (64 commits)
selftests/bpf: Add bpf_fib_lookup test
bpf: Add BPF_FIB_LOOKUP_SKIP_NEIGH for bpf_fib_lookup
riscv, bpf: Add bpf trampoline support for RV64
riscv, bpf: Add bpf_arch_text_poke support for RV64
riscv, bpf: Factor out emit_call for kernel and bpf context
riscv: Extend patch_text for multiple instructions
Revert "bpf, test_run: fix &xdp_frame misplacement for LIVE_FRAMES"
selftests/bpf: Add global subprog context passing tests
selftests/bpf: Convert test_global_funcs test to test_loader framework
bpf: Fix global subprog context argument resolution logic
LoongArch, bpf: Use 4 instructions for function address in JIT
bpf: bpf_fib_lookup should not return neigh in NUD_FAILED state
bpf: Disable bh in bpf_test_run for xdp and tc prog
xsk: check IFF_UP earlier in Tx path
Fix typos in selftest/bpf files
selftests/bpf: Use bpf_{btf,link,map,prog}_get_info_by_fd()
samples/bpf: Use bpf_{btf,link,map,prog}_get_info_by_fd()
bpftool: Use bpf_{btf,link,map,prog}_get_info_by_fd()
libbpf: Use bpf_{btf,link,map,prog}_get_info_by_fd()
libbpf: Introduce bpf_{btf,link,map,prog}_get_info_by_fd()
...
====================
Link: https://lore.kernel.org/r/20230217221737.31122-1-daniel@iogearbox.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
This patch implements features provisioning for mlx5_vdpa.
1) Validate the provisioned features are a subset of the parent
features.
2) Clearing features that are not wanted by userspace.
For example:
# vdpa mgmtdev show
pci/0000:41:04.2:
supported_classes net
max_supported_vqs 65
dev_features CSUM GUEST_CSUM MTU MAC HOST_TSO4 HOST_TSO6 STATUS CTRL_VQ CTRL_VLAN MQ CTRL_MAC_ADDR VERSION_1 ACCESS_PLATFORM
1) Provision vDPA device with all features derived from the parent
# vdpa dev add name vdpa1 mgmtdev pci/0000:41:04.2
# vdpa dev config show
vdpa1: mac e4:11:c6:d3:45:f0 link up link_announce false max_vq_pairs 1 mtu 1500
negotiated_features CSUM GUEST_CSUM MTU HOST_TSO4 HOST_TSO6 STATUS CTRL_VQ CTRL_VLAN MQ CTRL_MAC_ADDR VERSION_1 ACCESS_PLATFORM
2) Provision vDPA device with a subset of parent features
# vdpa dev add name vdpa1 mgmtdev pci/0000:41:04.2 device_features 0x300020000
# vdpa dev config show
vdpa1:
negotiated_features CTRL_VQ VERSION_1 ACCESS_PLATFORM
Signed-off-by: Si-Wei Liu <si-wei.liu@oracle.com>
Reviewed-by: Eli Cohen <elic@nvidia.com>
Message-Id: <1675725124-7375-7-git-send-email-si-wei.liu@oracle.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Today when device features are explicitly provisioned, the features
user supplied may contain device class specific features that are
not supported by the parent management device. On the other hand,
when parent management device supports more than one class, the
device features to provision may be ambiguous if none of the class
specific attributes is provided at the same time. Validate these
cases and prompt appropriate user errors accordingly.
Signed-off-by: Si-Wei Liu <si-wei.liu@oracle.com>
Message-Id: <1675725124-7375-5-git-send-email-si-wei.liu@oracle.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
With device feature provisioning, there's a chance for misconfiguration
that the vdpa feature attribute supplied in 'vdpa dev add' command doesn't
get selected on the device_features to be provisioned. For instance, when
a @mac attribute is specified, the corresponding feature bit _F_MAC in
device_features should be set for consistency. If there's conflict on
provisioned features against the attribute, it should be treated as an
error to fail the ambiguous command. Noted the opposite is not
necessarily true, for e.g. it's okay to have _F_MAC set in device_features
without providing a corresponding @mac attribute, in which case the vdpa
vendor driver could load certain default value for attribute that is not
explicitly specified.
Generalize this check in vdpa core so that there's no duplicate code in
each vendor driver.
Signed-off-by: Si-Wei Liu <si-wei.liu@oracle.com>
Reviewed-by: Eli Cohen <elic@nvidia.com>
Message-Id: <1675725124-7375-4-git-send-email-si-wei.liu@oracle.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
In below example, before the fix, mtu attribute is supported
by the parent mgmtdev, but the error message showing "All
provided are not supported" is just misleading.
$ vdpa mgmtdev show
vdpasim_net:
supported_classes net
max_supported_vqs 3
dev_features MTU MAC CTRL_VQ CTRL_MAC_ADDR ANY_LAYOUT VERSION_1 ACCESS_PLATFORM
$ vdpa dev add mgmtdev vdpasim_net name vdpasim0 mtu 5000 max_vqp 2
Error: vdpa: All provided attributes are not supported.
kernel answers: Operation not supported
After fix, the relevant error message will be like:
$ vdpa dev add mgmtdev vdpasim_net name vdpasim0 mtu 5000 max_vqp 2
Error: vdpa: Some provided attributes are not supported: 0x1000.
kernel answers: Operation not supported
Fixes: d8ca2fa5be ("vdpa: Enable user to set mac and mtu of vdpa device")
Signed-off-by: Si-Wei Liu <si-wei.liu@oracle.com>
Reviewed-by: Parav Pandit <parav@nvidia.com>
Reviewed-by: Eli Cohen <elic@nvidia.com>
Message-Id: <1675725124-7375-2-git-send-email-si-wei.liu@oracle.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Clearing the mr struct erases the lock owner and causes warnings to be
emitted. It is not required to clear the mr so remove the memset call.
Fixes: 94abbccdf2 ("vdpa/mlx5: Add shared memory registration code")
Signed-off-by: Eli Cohen <elic@nvidia.com>
Message-Id: <20230206121956.1149356-1-elic@nvidia.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
When creating a memory key, the key value should be assigned to the
passed pointer and not or'ed to.
No functional issue was observed due to this bug.
Fixes: 29064bfdab ("vdpa/mlx5: Add support library for mlx5 VDPA implementation")
Signed-off-by: Eli Cohen <elic@nvidia.com>
Message-Id: <20230205072906.1108194-1-elic@nvidia.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Add a print explaining why vhost_vdpa_alloc_domain failed if the device
is not IOMMU cache coherent capable.
Without this print, we have no hint why the operation failed.
For example:
$ virsh start <domain>
error: Failed to start domain <domain>
error: Unable to open '/dev/vhost-vdpa-<idx>' for vdpa device:
Unknown error 524
Suggested-by: Eugenio Perez Martin <eperezma@redhat.com>
Signed-off-by: Alvaro Karsz <alvaro.karsz@solid-run.com>
Message-Id: <20230202084212.1328530-1-alvaro.karsz@solid-run.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Eugenio Pérez Martin <eperezma@redhat.com>
There is no check about the return value of kmalloc in
virtscsi_rescan_hotunplug. Add the check to avoid use
of null pointer 'inq_result' in case of the failure
of kmalloc.
Signed-off-by: Zheng Wang <zyytlz.wz@163.com>
Message-Id: <20230202064124.22277-1-zyytlz.wz@163.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Add VIRTIO_F_RING_RESET, which indicates that the driver can reset a
queue individually.
VIRTIO_F_RING_RESET feature is added to virtio-spec 1.2. The relevant
information is in
oasis-tcs/virtio-spec#124oasis-tcs/virtio-spec#139
The implementation only adds the feature bit in supported features. It
does not require any other changes because we reuse the existing vhost
protocol.
The virtqueue reset process can be concluded as two parts:
1. The driver can reset a virtqueue. When it is triggered, we use the
set_backend to disable the virtqueue.
2. After the virtqueue is disabled, the driver may optionally re-enable
it. The process is basically similar to when the device is started,
except that the restart process does not need to set features and set
mem table since they do not change. QEMU will send messages containing
size, base, addr, kickfd and callfd of the virtqueue in order.
Specifically, the host kernel will receive these messages in order:
a. VHOST_SET_VRING_NUM
b. VHOST_SET_VRING_BASE
c. VHOST_SET_VRING_ADDR
d. VHOST_SET_VRING_KICK
e. VHOST_SET_VRING_CALL
f. VHOST_NET_SET_BACKEND
Finally, after we use set_backend to attach the virtqueue, the virtqueue
will be enabled and start to work.
Signed-off-by: Kangjie Xu <kangjie.xu@linux.alibaba.com>
Signed-off-by: Xuan Zhuo <xuanzhuo@linux.alibaba.com>
Message-Id: <20220825085610.80315-1-kangjie.xu@linux.alibaba.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Follow the advice of the Documentation/filesystems/sysfs.rst
and show() should only use sysfs_emit() or sysfs_emit_at()
when formatting the value to be returned to user space.
Signed-off-by: Bo Liu <liubo03@inspur.com>
Message-Id: <20230129091145.2837-1-liubo03@inspur.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
This patch implements per virtqueue dma device for mlx5_vdpa. This is
needed for virtio_vdpa to work for CVQ which is backed by vringh but
not DMA. We simply advertise the vDPA device itself as the DMA device
for CVQ then DMA API can simply use PA so the identical mapping for
CVQ can still be used. Otherwise the identical (1:1) mapping won't
work when platform IOMMU is enabled since the IOVA is allocated on
demand which is not necessarily the PA.
This fixes the following crash when mlx5 vDPA device is bound to
virtio-vdpa with platform IOMMU enabled but not in passthrough mode:
BUG: unable to handle page fault for address: ff2fb3063deb1002
#PF: supervisor read access in kernel mode
#PF: error_code(0x0000) - not-present page
PGD 1393001067 P4D 1393002067 PUD 0
Oops: 0000 [#1] PREEMPT SMP NOPTI
CPU: 55 PID: 8923 Comm: kworker/u112:3 Kdump: loaded Not tainted 6.1.0+ #7
Hardware name: Dell Inc. PowerEdge R750/0PJ80M, BIOS 1.5.4 12/17/2021
Workqueue: mlx5_vdpa_wq mlx5_cvq_kick_handler [mlx5_vdpa]
RIP: 0010:vringh_getdesc_iotlb+0x93/0x1d0 [vringh]
Code: 14 25 40 ef 01 00 83 82 c0 0a 00 00 01 48 2b 05 93 5a 1b ea 8b 4c 24 14 48 c1 f8 06 48 c1 e0 0c 48 03 05 90 5a 1b ea 48 01 c8 <0f> b7 00 83 aa c0 0a 00 00 01 65 ff 0d bc e4 41 3f 0f 84 05 01 00
RSP: 0018:ff46821ba664fdf8 EFLAGS: 00010282
RAX: ff2fb3063deb1002 RBX: 0000000000000a20 RCX: 0000000000000002
RDX: ff2fb318d2f94380 RSI: 0000000000000002 RDI: 0000000000000001
RBP: ff2fb3065e832410 R08: ff46821ba664fe00 R09: 0000000000000001
R10: 0000000000000000 R11: 000000000000000d R12: ff2fb3065e832488
R13: ff2fb3065e8324a8 R14: ff2fb3065e8324c8 R15: ff2fb3065e8324a8
FS: 0000000000000000(0000) GS:ff2fb3257fac0000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: ff2fb3063deb1002 CR3: 0000001392010006 CR4: 0000000000771ee0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
PKRU: 55555554
Call Trace:
<TASK>
mlx5_cvq_kick_handler+0x89/0x2b0 [mlx5_vdpa]
process_one_work+0x1e2/0x3b0
? rescuer_thread+0x390/0x390
worker_thread+0x50/0x3a0
? rescuer_thread+0x390/0x390
kthread+0xd6/0x100
? kthread_complete_and_exit+0x20/0x20
ret_from_fork+0x1f/0x30
</TASK>
Reviewed-by: Eli Cohen <elic@nvidia.com>
Tested-by: Eli Cohen <elic@nvidia.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20230119061525.75068-6-jasowang@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Setting DMA mask for vDPA device in case that there are virtqueue that
is not backed by DMA so the vDPA device could be advertised as the DMA
device that is used by DMA API for software emulated virtqueues.
Reviewed-by: Eli Cohen <elic@nvidia.com>
Tested-by: Eli Cohen <elic@nvidia.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20230119061525.75068-5-jasowang@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
This patch adds the support of per vq dma device for virito-vDPA. vDPA
parents then are allowed to use different DMA devices. This is useful
for the parents that have software or emulated virtqueues.
Reviewed-by: Eli Cohen <elic@nvidia.com>
Tested-by: Eli Cohen <elic@nvidia.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20230119061525.75068-4-jasowang@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
This patch introduces a per virtqueue dma device. This will be used
for virtio devices whose virtqueue are backed by different underlayer
devices.
One example is the vDPA that where the control virtqueue could be
implemented through software mediation.
Some of the work are actually done before since the helper like
vring_dma_device(). This work left are:
- Let vring_dma_device() return the per virtqueue dma device instead
of the vdev's parent.
- Allow passing a dma_device when creating the virtqueue through a new
helper, old vring creation helper will keep using vdev's parent.
Reviewed-by: Eli Cohen <elic@nvidia.com>
Tested-by: Eli Cohen <elic@nvidia.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20230119061525.75068-2-jasowang@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
We used to (ab)use the DMA ops for setting up identical mappings in
the IOTLB. This patch tries to get rid of the those unnecessary DMA
ops by maintaining a simple identical/passthrough mappings by
default. When bound to virtio_vdpa driver, DMA API will simply use PA
as the IOVA and we will be all fine. When the vDPA bus tries to setup
customized mapping (e.g when bound to vhost-vDPA), the
identical/passthrough mapping will be removed.
Signed-off-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20221223060021.28011-1-jasowang@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Christoph Hellwig <hch@lst.de>
This patch adds completion batching to the IRQ path. It reuses batch
completion code of virtblk_poll(). It collects requests to io_comp_batch
and processes them all at once. It can boost up the performance by 2%.
To validate the performance improvement and stabilty, I did fio test with
4 vCPU VM and 12 vCPU VM respectively. Both VMs have 8GB ram and the same
number of HW queues as vCPU.
The fio cammad is as follows and I ran the fio 5 times and got IOPS average.
(io_uring, randread, direct=1, bs=512, iodepth=64 numjobs=2,4)
Test result shows about 2% improvement.
4 vcpu VM | numjobs=2 | numjobs=4
-----------------------------------------------------------
fio without patch | 367.2K IOPS | 397.6K IOPS
-----------------------------------------------------------
fio with patch | 372.8K IOPS | 407.7K IOPS
12 vcpu VM | numjobs=2 | numjobs=4
-----------------------------------------------------------
fio without patch | 363.6K IOPS | 374.8K IOPS
-----------------------------------------------------------
fio with patch | 373.8K IOPS | 385.3K IOPS
Signed-off-by: Suwan Kim <suwan.kim027@gmail.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Message-Id: <20221221145456.281218-3-suwan.kim027@gmail.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Driver should set req->state to MQ_RQ_COMPLETE after it finishes to process
req. But virtio-blk doesn't set MQ_RQ_COMPLETE after virtblk_poll() handles
req and req->state still remains MQ_RQ_IN_FLIGHT. Fortunately so far there
is no issue about it because blk_mq_end_request_batch() sets req->state to
MQ_RQ_IDLE.
In this patch, virblk_poll() calls blk_mq_complete_request_remote() to set
req->state to MQ_RQ_COMPLETE before it adds req to a batch completion list.
So it properly sets req->state after polling I/O is finished.
Fixes: 4e04005256 ("virtio-blk: support polling I/O")
Signed-off-by: Suwan Kim <suwan.kim027@gmail.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Message-Id: <20221221145456.281218-2-suwan.kim027@gmail.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
A sentence that checks for later spec version is meant for developers
hacking the documentation source. Make it comment block (hidden from
actual output).
Signed-off-by: Bagas Sanjaya <bagasdotme@gmail.com>
Message-Id: <20221220095828.27588-4-bagasdotme@gmail.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
This new ioctl adds support for resuming the device from userspace.
This is required when trying to restore the device in a functioning
state after it's been suspended. It is already possible to reset a
suspended device, but that means the device must be reconfigured and
all the IOMMU/IOTLB mappings must be recreated. This new operation
allows the device to be resumed without going through a full reset.
This is particularly useful when trying to perform offline migration of
a virtual machine (also known as snapshot/restore) as it allows the VMM
to resume the virtual machine back to a running state after the snapshot
is performed.
Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
Message-Id: <73b75fb87d25cff59768b4955a81fe7ffe5b4770.1672742878.git.sebastien.boeuf@intel.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Add a new operation to allow a vDPA device to be resumed after it has
been suspended. Trying to resume a device that wasn't suspended will
result in a no-op.
This operation is optional. If it's not implemented, the associated
backend feature bit will not be exposed. And if the feature bit is not
exposed, invoking this operation will return an error.
Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Sebastien Boeuf <sebastien.boeuf@intel.com>
Message-Id: <6e05c4b31b47f3e29cb2bd7ebd56c81f84b8f48a.1672742878.git.sebastien.boeuf@intel.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
This patch fixes a FLR bug on the SNET DPU rev 1 by setting the
PCI_DEV_FLAGS_NO_FLR_RESET flag.
As there is a quirk to avoid FLR (quirk_no_flr), I added a new quirk
to check the rev ID before calling to quirk_no_flr.
Without this patch, a SNET DPU rev 1 may hang when FLR is applied.
Signed-off-by: Alvaro Karsz <alvaro.karsz@solid-run.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Message-Id: <20230110165638.123745-3-alvaro.karsz@solid-run.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
