This converts some of the visually simpler cases that have been split
over multiple lines. I only did the ones that are easy to verify the
resulting diff by having just that final GFP_KERNEL argument on the next
line.
Somebody should probably do a proper coccinelle script for this, but for
me the trivial script actually resulted in an assertion failure in the
middle of the script. I probably had made it a bit _too_ trivial.
So after fighting that far a while I decided to just do some of the
syntactically simpler cases with variations of the previous 'sed'
scripts.
The more syntactically complex multi-line cases would mostly really want
whitespace cleanup anyway.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This is the exact same thing as the 'alloc_obj()' version, only much
smaller because there are a lot fewer users of the *alloc_flex()
interface.
As with alloc_obj() version, this was done entirely with mindless brute
force, using the same script, except using 'flex' in the pattern rather
than 'objs*'.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This was done entirely with mindless brute force, using
git grep -l '\<k[vmz]*alloc_objs*(.*, GFP_KERNEL)' |
xargs sed -i 's/\(alloc_objs*(.*\), GFP_KERNEL)/\1)/'
to convert the new alloc_obj() users that had a simple GFP_KERNEL
argument to just drop that argument.
Note that due to the extreme simplicity of the scripting, any slightly
more complex cases spread over multiple lines would not be triggered:
they definitely exist, but this covers the vast bulk of the cases, and
the resulting diff is also then easier to check automatically.
For the same reason the 'flex' versions will be done as a separate
conversion.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This is the result of running the Coccinelle script from
scripts/coccinelle/api/kmalloc_objs.cocci. The script is designed to
avoid scalar types (which need careful case-by-case checking), and
instead replace kmalloc-family calls that allocate struct or union
object instances:
Single allocations: kmalloc(sizeof(TYPE), ...)
are replaced with: kmalloc_obj(TYPE, ...)
Array allocations: kmalloc_array(COUNT, sizeof(TYPE), ...)
are replaced with: kmalloc_objs(TYPE, COUNT, ...)
Flex array allocations: kmalloc(struct_size(PTR, FAM, COUNT), ...)
are replaced with: kmalloc_flex(*PTR, FAM, COUNT, ...)
(where TYPE may also be *VAR)
The resulting allocations no longer return "void *", instead returning
"TYPE *".
Signed-off-by: Kees Cook <kees@kernel.org>
We (Paolo and I) noticed that in the sending path touching an extra
cacheline due to cq_cached_prod_lock will impact the performance. After
moving the lock from struct xsk_buff_pool to struct xsk_queue, the
performance is increased by ~5% which can be observed by xdpsock.
An alternative approach [1] can be using atomic_try_cmpxchg() to have the
same effect. But unfortunately I don't have evident performance numbers to
prove the atomic approach is better than the current patch. The advantage
is to save the contention time among multiple xsks sharing the same pool
while the disadvantage is losing good maintenance. The full discussion can
be found at the following link.
[1]: https://lore.kernel.org/all/20251128134601.54678-1-kerneljasonxing@gmail.com/
Suggested-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Jason Xing <kernelxing@tencent.com>
Link: https://patch.msgid.link/20260104012125.44003-3-kerneljasonxing@gmail.com
Acked-by: Stanislav Fomichev <sdf@fomichev.me>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
In the shared umem mode with different queues or devices, either
uninitialized cq or fq is not allowed which was previously done in
xp_assign_dev_shared(). The patch advances the check at the beginning
so that 1) we can avoid a few memory allocation and stuff if cq or fq
is NULL, 2) it can be regarded as preparation for the next patch in
the series.
Signed-off-by: Jason Xing <kernelxing@tencent.com>
Link: https://patch.msgid.link/20260104012125.44003-2-kerneljasonxing@gmail.com
Acked-by: Stanislav Fomichev <sdf@fomichev.me>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
- Split cq_lock into two smaller locks: cq_prod_lock and
cq_cached_prod_lock
- Avoid disabling/enabling interrupts in the hot xmit path
In either xsk_cq_cancel_locked() or xsk_cq_reserve_locked() function,
the race condition is only between multiple xsks sharing the same
pool. They are all in the process context rather than interrupt context,
so now the small lock named cq_cached_prod_lock can be used without
handling interrupts.
While cq_cached_prod_lock ensures the exclusive modification of
@cached_prod, cq_prod_lock in xsk_cq_submit_addr_locked() only cares
about @producer and corresponding @desc. Both of them don't necessarily
be consistent with @cached_prod protected by cq_cached_prod_lock.
That's the reason why the previous big lock can be split into two
smaller ones. Please note that SPSC rule is all about the global state
of producer and consumer that can affect both layers instead of local
or cached ones.
Frequently disabling and enabling interrupt are very time consuming
in some cases, especially in a per-descriptor granularity, which now
can be avoided after this optimization, even when the pool is shared by
multiple xsks.
With this patch, the performance number[1] could go from 1,872,565 pps
to 1,961,009 pps. It's a minor rise of around 5%.
[1]: taskset -c 1 ./xdpsock -i enp2s0f1 -q 0 -t -S -s 64
Signed-off-by: Jason Xing <kernelxing@tencent.com>
Acked-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Link: https://patch.msgid.link/20251030000646.18859-3-kerneljasonxing@gmail.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
The commit ac98d8aab6 ("xsk: wire upp Tx zero-copy functions")
originally introducing this lock put the deletion process in the
sk_destruct which can run in irq context obviously, so the
xxx_irqsave()/xxx_irqrestore() pair was used. But later another
commit 541d7fdd76 ("xsk: proper AF_XDP socket teardown ordering")
moved the deletion into xsk_release() that only happens in process
context. It means that since this commit, it doesn't necessarily
need that pair.
Now, there are two places that use this xsk_tx_list_lock and only
run in the process context. So avoid manipulating the irq then.
Signed-off-by: Jason Xing <kernelxing@tencent.com>
Acked-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Link: https://patch.msgid.link/20251030000646.18859-2-kerneljasonxing@gmail.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Move rx_lock from xsk_socket to xsk_buff_pool.
Fix synchronization for shared umem mode in
generic RX path where multiple sockets share
single xsk_buff_pool.
RX queue is exclusive to xsk_socket, while FILL
queue can be shared between multiple sockets.
This could result in race condition where two
CPU cores access RX path of two different sockets
sharing the same umem.
Protect both queues by acquiring spinlock in shared
xsk_buff_pool.
Lock contention may be minimized in the future by some
per-thread FQ buffering.
It's safe and necessary to move spin_lock_bh(rx_lock)
after xsk_rcv_check():
* xs->pool and spinlock_init is synchronized by
xsk_bind() -> xsk_is_bound() memory barriers.
* xsk_rcv_check() may return true at the moment
of xsk_release() or xsk_unbind_dev(),
however this will not cause any data races or
race conditions. xsk_unbind_dev() removes xdp
socket from all maps and waits for completion
of all outstanding rx operations. Packets in
RX path will either complete safely or drop.
Signed-off-by: Eryk Kubanski <e.kubanski@partner.samsung.com>
Fixes: bf0bdd1343 ("xdp: fix race on generic receive path")
Acked-by: Magnus Karlsson <magnus.karlsson@intel.com>
Link: https://patch.msgid.link/20250416101908.10919-1-e.kubanski@partner.samsung.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Read accesses go via xsk_get_pool_from_qid(), the call coming
from the core and gve look safe (other "ops locked" drivers
don't support XSK).
Write accesses go via xsk_reg_pool_at_qid() and xsk_clear_pool_at_qid().
Former is already under the ops lock, latter is not (both coming from
the workqueue via xp_clear_dev() and NETDEV_UNREGISTER via xsk_notifier()).
Acked-by: Stanislav Fomichev <sdf@fomichev.me>
Signed-off-by: Stanislav Fomichev <sdf@fomichev.me>
Link: https://patch.msgid.link/20250408195956.412733-3-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Cross-merge networking fixes after downstream PR (net-6.14-rc8).
Conflict:
tools/testing/selftests/net/Makefile
03544faad7 ("selftest: net: add proc_net_pktgen")
3ed61b8938 ("selftests: net: test for lwtunnel dst ref loops")
tools/testing/selftests/net/config:
85cb3711ac ("selftests: net: Add test cases for link and peer netns")
3ed61b8938 ("selftests: net: test for lwtunnel dst ref loops")
Adjacent commits:
tools/testing/selftests/net/Makefile
c935af429e ("selftests: net: add support for testing SO_RCVMARK and SO_RCVPRIORITY")
355d940f4d ("Revert "selftests: Add IPv6 link-local address generation tests for GRE devices."")
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Move the more esoteric helpers for netdev instance lock to
a dedicated header. This avoids growing netdevice.h to infinity
and makes rebuilding the kernel much faster (after touching
the header with the helpers).
The main netdev_lock() / netdev_unlock() functions are used
in static inlines in netdevice.h and will probably be used
most commonly, so keep them in netdevice.h.
Acked-by: Stanislav Fomichev <sdf@fomichev.me>
Link: https://patch.msgid.link/20250307183006.2312761-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Currently, when your driver supports XSk Tx metadata and you want to
send an XSk frame, you need to do the following:
* call external xsk_buff_raw_get_dma();
* call inline xsk_buff_get_metadata(), which calls external
xsk_buff_raw_get_data() and then do some inline checks.
This effectively means that the following piece:
addr = pool->unaligned ? xp_unaligned_add_offset_to_addr(addr) : addr;
is done twice per frame, plus you have 2 external calls per frame, plus
this:
meta = pool->addrs + addr - pool->tx_metadata_len;
if (unlikely(!xsk_buff_valid_tx_metadata(meta)))
is always inlined, even if there's no meta or it's invalid.
Add xsk_buff_raw_get_ctx() (xp_raw_get_ctx() to be precise) to do that
in one go. It returns a small structure with 2 fields: DMA address,
filled unconditionally, and metadata pointer, non-NULL only if it's
present and valid. The address correction is performed only once and
you also have only 1 external call per XSk frame, which does all the
calculations and checks outside of your hotpath. You only need to
check `if (ctx.meta)` for the metadata presence.
To not copy any existing code, derive address correction and getting
virtual and DMA address into small helpers. bloat-o-meter reports no
object code changes for the existing functionality.
Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>
Link: https://patch.msgid.link/20250206182630.3914318-5-aleksander.lobakin@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
When the umem is shared, the DMA mapping is also shared between the xsk
pools, therefore it should stay valid as long as at least 1 user remains.
However, the pool also keeps the copies of DMA-related information that are
initialized in the same way in xp_init_dma_info(), but cleared by
xp_dma_unmap() only for the last remaining pool, this causes the problems
below.
The first one is that the commit adbf5a4234 ("ice: remove af_xdp_zc_qps
bitmap") relies on pool->dev to determine the presence of a ZC pool on a
given queue, avoiding internal bookkeeping. This works perfectly fine if
the UMEM is not shared, but reliably fails otherwise as stated in the
linked report.
The second one is pool->dma_pages which is dynamically allocated and
only freed in xp_dma_unmap(), this leads to a small memory leak. kmemleak
does not catch it, but by printing the allocation results after terminating
the userspace program it is possible to see that all addresses except the
one belonging to the last detached pool are still accessible through the
kmemleak dump functionality.
Always clear the DMA mapping information from the pool and free
pool->dma_pages when unmapping the pool, so that the only difference
between results of the last remaining user's call and the ones before would
be the destruction of the DMA mapping.
Fixes: adbf5a4234 ("ice: remove af_xdp_zc_qps bitmap")
Fixes: 921b68692a ("xsk: Enable sharing of dma mappings")
Reported-by: Alasdair McWilliam <alasdair.mcwilliam@outlook.com>
Closes: https://lore.kernel.org/PA4P194MB10056F208AF221D043F57A3D86512@PA4P194MB1005.EURP194.PROD.OUTLOOK.COM
Acked-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Signed-off-by: Larysa Zaremba <larysa.zaremba@intel.com>
Link: https://lore.kernel.org/r/20241122112912.89881-1-larysa.zaremba@intel.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Continue the process of dieting xdp_buff_xsk by removing orig_addr
member. It can be calculated from xdp->data_hard_start where it was
previously used, so it is not anything that has to be carried around in
struct used widely in hot path.
This has been used for initializing xdp_buff_xsk::frame_dma during pool
setup and as a shortcut in xp_get_handle() to retrieve address provided
to xsk Rx queue.
Signed-off-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Magnus Karlsson <magnus.karlsson@intel.com>
Link: https://lore.kernel.org/bpf/20241007122458.282590-4-maciej.fijalkowski@intel.com
In cases when synchronizing DMA operations is necessary,
xsk_buff_alloc_batch() returns a single buffer instead of the requested
count. This puts the pressure on drivers that use batch API as they have
to check for this corner case on their side and take care of allocations
by themselves, which feels counter productive. Let us improve the core
by looping over xp_alloc() @max times when slow path needs to be taken.
Another issue with current interface, as spotted and fixed by Dries, was
that when driver called xsk_buff_alloc_batch() with @max == 0, for slow
path case it still allocated and returned a single buffer, which should
not happen. By introducing the logic from first paragraph we kill two
birds with one stone and address this problem as well.
Fixes: 47e4075df3 ("xsk: Batched buffer allocation for the pool")
Reported-and-tested-by: Dries De Winter <ddewinter@synamedia.com>
Co-developed-by: Dries De Winter <ddewinter@synamedia.com>
Signed-off-by: Dries De Winter <ddewinter@synamedia.com>
Signed-off-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Acked-by: Magnus Karlsson <magnus.karlsson@intel.com>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Link: https://patch.msgid.link/20240911191019.296480-1-maciej.fijalkowski@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Daniel Borkmann says:
====================
pull-request: bpf-next 2024-09-11
We've added 12 non-merge commits during the last 16 day(s) which contain
a total of 20 files changed, 228 insertions(+), 30 deletions(-).
There's a minor merge conflict in drivers/net/netkit.c:
00d066a4d4 ("netdev_features: convert NETIF_F_LLTX to dev->lltx")
d966087948 ("netkit: Disable netpoll support")
The main changes are:
1) Enable bpf_dynptr_from_skb for tp_btf such that this can be used
to easily parse skbs in BPF programs attached to tracepoints,
from Philo Lu.
2) Add a cond_resched() point in BPF's sock_hash_free() as there have
been several syzbot soft lockup reports recently, from Eric Dumazet.
3) Fix xsk_buff_can_alloc() to account for queue_empty_descs which
got noticed when zero copy ice driver started to use it,
from Maciej Fijalkowski.
4) Move the xdp:xdp_cpumap_kthread tracepoint before cpumap pushes skbs
up via netif_receive_skb_list() to better measure latencies,
from Daniel Xu.
5) Follow-up to disable netpoll support from netkit, from Daniel Borkmann.
6) Improve xsk selftests to not assume a fixed MAX_SKB_FRAGS of 17 but
instead gather the actual value via /proc/sys/net/core/max_skb_frags,
also from Maciej Fijalkowski.
* tag 'for-netdev' of https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next:
sock_map: Add a cond_resched() in sock_hash_free()
selftests/bpf: Expand skb dynptr selftests for tp_btf
bpf: Allow bpf_dynptr_from_skb() for tp_btf
tcp: Use skb__nullable in trace_tcp_send_reset
selftests/bpf: Add test for __nullable suffix in tp_btf
bpf: Support __nullable argument suffix for tp_btf
bpf, cpumap: Move xdp:xdp_cpumap_kthread tracepoint before rcv
selftests/xsk: Read current MAX_SKB_FRAGS from sysctl knob
xsk: Bump xsk_queue::queue_empty_descs in xp_can_alloc()
tcp_bpf: Remove an unused parameter for bpf_tcp_ingress()
bpf, sockmap: Correct spelling skmsg.c
netkit: Disable netpoll support
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
====================
Link: https://patch.msgid.link/20240911211525.13834-1-daniel@iogearbox.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Add a netdev_dmabuf_binding struct which represents the
dma-buf-to-netdevice binding. The netlink API will bind the dma-buf to
rx queues on the netdevice. On the binding, the dma_buf_attach
& dma_buf_map_attachment will occur. The entries in the sg_table from
mapping will be inserted into a genpool to make it ready
for allocation.
The chunks in the genpool are owned by a dmabuf_chunk_owner struct which
holds the dma-buf offset of the base of the chunk and the dma_addr of
the chunk. Both are needed to use allocations that come from this chunk.
We create a new type that represents an allocation from the genpool:
net_iov. We setup the net_iov allocation size in the
genpool to PAGE_SIZE for simplicity: to match the PAGE_SIZE normally
allocated by the page pool and given to the drivers.
The user can unbind the dmabuf from the netdevice by closing the netlink
socket that established the binding. We do this so that the binding is
automatically unbound even if the userspace process crashes.
The binding and unbinding leaves an indicator in struct netdev_rx_queue
that the given queue is bound, and the binding is actuated by resetting
the rx queue using the queue API.
The netdev_dmabuf_binding struct is refcounted, and releases its
resources only when all the refs are released.
Signed-off-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Kaiyuan Zhang <kaiyuanz@google.com>
Signed-off-by: Mina Almasry <almasrymina@google.com>
Reviewed-by: Pavel Begunkov <asml.silence@gmail.com> # excluding netlink
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Link: https://patch.msgid.link/20240910171458.219195-4-almasrymina@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
We have STAT_FILL_EMPTY test case in xskxceiver that tries to process
traffic with fill queue being empty which currently fails for zero copy
ice driver after it started to use xsk_buff_can_alloc() API. That is
because xsk_queue::queue_empty_descs is currently only increased from
alloc APIs and right now if driver sees that xsk_buff_pool will be
unable to provide the requested count of buffers, it bails out early,
skipping calls to xsk_buff_alloc{_batch}().
Mentioned statistic should be handled in xsk_buff_can_alloc() from the
very beginning, so let's add this logic now. Do it by open coding
xskq_cons_has_entries() and bumping queue_empty_descs in the middle when
fill queue currently has no entries.
Signed-off-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Magnus Karlsson <magnus.karlsson@intel.com>
Link: https://lore.kernel.org/bpf/20240904162808.249160-1-maciej.fijalkowski@intel.com
XSk infra's been using its own DMA sync shortcut to try avoiding
redundant function calls. Now that there is a generic one, remove
the custom implementation and rely on the generic helpers.
xsk_buff_dma_sync_for_cpu() doesn't need the second argument anymore,
remove it.
Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
XDP multi-buffer support introduced XDP_FLAGS_HAS_FRAGS flag that is
used by drivers to notify data path whether xdp_buff contains fragments
or not. Data path looks up mentioned flag on first buffer that occupies
the linear part of xdp_buff, so drivers only modify it there. This is
sufficient for SKB and XDP_DRV modes as usually xdp_buff is allocated on
stack or it resides within struct representing driver's queue and
fragments are carried via skb_frag_t structs. IOW, we are dealing with
only one xdp_buff.
ZC mode though relies on list of xdp_buff structs that is carried via
xsk_buff_pool::xskb_list, so ZC data path has to make sure that
fragments do *not* have XDP_FLAGS_HAS_FRAGS set. Otherwise,
xsk_buff_free() could misbehave if it would be executed against xdp_buff
that carries a frag with XDP_FLAGS_HAS_FRAGS flag set. Such scenario can
take place when within supplied XDP program bpf_xdp_adjust_tail() is
used with negative offset that would in turn release the tail fragment
from multi-buffer frame.
Calling xsk_buff_free() on tail fragment with XDP_FLAGS_HAS_FRAGS would
result in releasing all the nodes from xskb_list that were produced by
driver before XDP program execution, which is not what is intended -
only tail fragment should be deleted from xskb_list and then it should
be put onto xsk_buff_pool::free_list. Such multi-buffer frame will never
make it up to user space, so from AF_XDP application POV there would be
no traffic running, however due to free_list getting constantly new
nodes, driver will be able to feed HW Rx queue with recycled buffers.
Bottom line is that instead of traffic being redirected to user space,
it would be continuously dropped.
To fix this, let us clear the mentioned flag on xsk_buff_pool side
during xdp_buff initialization, which is what should have been done
right from the start of XSK multi-buffer support.
Fixes: 1bbc04de60 ("ice: xsk: add RX multi-buffer support")
Fixes: 1c9ba9c146 ("i40e: xsk: add RX multi-buffer support")
Fixes: 24ea50127e ("xsk: support mbuf on ZC RX")
Signed-off-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Link: https://lore.kernel.org/r/20240124191602.566724-3-maciej.fijalkowski@intel.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
For zerocopy mode, tx_desc->addr can point to an arbitrary offset
and carry some TX metadata in the headroom. For copy mode, there
is no way currently to populate skb metadata.
Introduce new tx_metadata_len umem config option that indicates how many
bytes to treat as metadata. Metadata bytes come prior to tx_desc address
(same as in RX case).
The size of the metadata has mostly the same constraints as XDP:
- less than 256 bytes
- 8-byte aligned (compared to 4-byte alignment on xdp, due to 8-byte
timestamp in the completion)
- non-zero
This data is not interpreted in any way right now.
Reviewed-by: Song Yoong Siang <yoong.siang.song@intel.com>
Signed-off-by: Stanislav Fomichev <sdf@google.com>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20231127190319.1190813-2-sdf@google.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Userspace applications indicate their multi-buffer capability to xsk
using XSK_USE_SG socket bind flag. For sockets using shared umem the
bind flag may contain XSK_USE_SG only for the first socket. For any
subsequent socket the only option supported is XDP_SHARED_UMEM.
Add option XDP_UMEM_SG_FLAG in umem config flags to store the
multi-buffer handling capability when indicated by XSK_USE_SG option in
bing flag by the first socket. Use this to derive multi-buffer capability
for subsequent sockets in xsk core.
Signed-off-by: Tirthendu Sarkar <tirthendu.sarkar@intel.com>
Fixes: 81470b5c3c ("xsk: introduce XSK_USE_SG bind flag for xsk socket")
Acked-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Link: https://lore.kernel.org/r/20230907035032.2627879-1-tirthendu.sarkar@intel.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Given that skb_shared_info relies on skb_frag_t, in order to support
xskb chaining, introduce xdp_buff_xsk::xskb_list_node and
xsk_buff_pool::xskb_list.
This is needed so ZC drivers can add frags as xskb nodes which will make
it possible to handle it both when producing AF_XDP Rx descriptors as
well as freeing/recycling all the frags that a single frame carries.
Speaking of latter, update xsk_buff_free() to take care of list nodes.
For the former (adding as frags), introduce xsk_buff_add_frag() for ZC
drivers usage that is going to be used to add a frag to xskb list from
pool.
xsk_buff_get_frag() will be utilized by XDP_TX and, on contrary, will
return xdp_buff.
One of the previous patches added a wrapper for ZC Rx so implement xskb
list walk and production of Rx descriptors there.
On bind() path, bail out if socket wants to use ZC multi-buffer but
underlying netdev does not support it.
Signed-off-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Link: https://lore.kernel.org/r/20230719132421.584801-12-maciej.fijalkowski@intel.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
When a XSK pool gets mapped, xp_check_dma_contiguity() adds bit 0x1
to pages' DMA addresses that go in ascending order and at 4K stride.
The problem is that the bit does not get cleared before doing unmap.
As a result, a lot of warnings from iommu_dma_unmap_page() are seen
in dmesg, which indicates that lookups by iommu_iova_to_phys() fail.
Fixes: 2b43470add ("xsk: Introduce AF_XDP buffer allocation API")
Signed-off-by: Ivan Malov <ivan.malov@oktetlabs.ru>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Magnus Karlsson <magnus.karlsson@intel.com>
Link: https://lore.kernel.org/bpf/20220628091848.534803-1-ivan.malov@oktetlabs.ru
For the case when xp_alloc_batch() is used but the batched allocation
cannot be used, there is a slow path that uses the non-batched
xp_alloc(). When it fails to allocate an entry, it returns NULL. The
current code wrote this NULL into the entry of the provided results
array (pointer to the driver SW ring usually) and returned. This might
not be what the driver expects and to make things simpler, just write
successfully allocated xdp_buffs into the SW ring,. The driver might
have information in there that is still important after an allocation
failure.
Note that at this point in time, there are no drivers using
xp_alloc_batch() that could trigger this slow path. But one might get
added.
Fixes: 47e4075df3 ("xsk: Batched buffer allocation for the pool")
Signed-off-by: Magnus Karlsson <magnus.karlsson@intel.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20220328142123.170157-2-maciej.fijalkowski@intel.com
Move desc_array from the driver to the pool. The reason behind this is
that we can then reuse this array as a temporary storage for descriptors
in all zero-copy drivers that use the batched interface. This will make
it easier to add batching to more drivers.
i40e is the only driver that has a batched Tx zero-copy
implementation, so no need to touch any other driver.
Signed-off-by: Magnus Karlsson <magnus.karlsson@intel.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Alexander Lobakin <alexandr.lobakin@intel.com>
Link: https://lore.kernel.org/bpf/20220125160446.78976-6-maciej.fijalkowski@intel.com
This commit initialises the xskb's free_list_node when the xskb is
allocated. This prevents a potential false negative returned from a call
to list_empty for that node, such as the one introduced in commit
199d983bc0 ("xsk: Fix crash on double free in buffer pool")
In my environment this issue caused packets to not be received by
the xdpsock application if the traffic was running prior to application
launch. This happened when the first batch of packets failed the xskmap
lookup and XDP_PASS was returned from the bpf program. This action is
handled in the i40e zc driver (and others) by allocating an skbuff,
freeing the xdp_buff and adding the associated xskb to the
xsk_buff_pool's free_list if it hadn't been added already. Without this
fix, the xskb is not added to the free_list because the check to determine
if it was added already returns an invalid positive result. Later, this
caused allocation errors in the driver and the failure to receive packets.
Fixes: 199d983bc0 ("xsk: Fix crash on double free in buffer pool")
Fixes: 2b43470add ("xsk: Introduce AF_XDP buffer allocation API")
Signed-off-by: Ciara Loftus <ciara.loftus@intel.com>
Acked-by: Magnus Karlsson <magnus.karlsson@intel.com>
Link: https://lore.kernel.org/r/20211220155250.2746-1-ciara.loftus@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Fix a crash in the buffer pool allocator when a buffer is double
freed. It is possible to trigger this behavior not only from a faulty
driver, but also from user space like this: Create a zero-copy AF_XDP
socket. Load an XDP program that will issue XDP_DROP for all
packets. Put the same umem buffer into the fill ring multiple times,
then bind the socket and send some traffic. This will crash the kernel
as the XDP_DROP action triggers one call to xsk_buff_free()/xp_free()
for every packet dropped. Each call will add the corresponding buffer
entry to the free_list and increase the free_list_cnt. Some entries
will have been added multiple times due to the same buffer being
freed. The buffer allocation code will then traverse this broken list
and since the same buffer is in the list multiple times, it will try
to delete the same buffer twice from the list leading to a crash.
The fix for this is just to test that the buffer has not been added
before in xp_free(). If it has been, just return from the function and
do not put it in the free_list a second time.
Note that this bug was not present in the code before the commit
referenced in the Fixes tag. That code used one list entry per
allocated buffer, so multiple frees did not have any side effects. But
the commit below optimized the usage of the pool and only uses a
single entry per buffer in the umem, meaning that multiple
allocations/frees of the same buffer will also only use one entry,
thus leading to the problem.
Fixes: 47e4075df3 ("xsk: Batched buffer allocation for the pool")
Signed-off-by: Magnus Karlsson <magnus.karlsson@intel.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Björn Töpel <bjorn@kernel.org>
Link: https://lore.kernel.org/bpf/20211111075707.21922-1-magnus.karlsson@gmail.com
Optimize for the aligned case by precomputing the parameter values of
the xdp_buff_xsk and xdp_buff structures in the heads array. We can do
this as the heads array size is equal to the number of chunks in the
umem for the aligned case. Then every entry in this array will reflect
a certain chunk/frame and can therefore be prepopulated with the
correct values and we can drop the use of the free_heads stack. Note
that it is not possible to allocate more buffers than what has been
allocated in the aligned case since each chunk can only contain a
single buffer.
We can unfortunately not do this in the unaligned case as one chunk
might contain multiple buffers. In this case, we keep the old scheme
of populating a heads entry every time it is used and using
the free_heads stack.
Also move xp_release() and xp_get_handle() to xsk_buff_pool.h. They
were for some reason in xsk.c even though they are buffer pool
operations.
Signed-off-by: Magnus Karlsson <magnus.karlsson@intel.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20210922075613.12186-7-magnus.karlsson@gmail.com
Add a new driver interface xsk_buff_alloc_batch() offering batched
buffer allocations to improve performance. The new interface takes
three arguments: the buffer pool to allocated from, a pointer to an
array of struct xdp_buff pointers which will contain pointers to the
allocated xdp_buffs, and an unsigned integer specifying the max number
of buffers to allocate. The return value is the actual number of
buffers that the allocator managed to allocate and it will be in the
range 0 <= N <= max, where max is the third parameter to the function.
u32 xsk_buff_alloc_batch(struct xsk_buff_pool *pool, struct xdp_buff **xdp,
u32 max);
A second driver interface is also introduced that need to be used in
conjunction with xsk_buff_alloc_batch(). It is a helper that sets the
size of struct xdp_buff and is used by the NIC Rx irq routine when
receiving a packet. This helper sets the three struct members data,
data_meta, and data_end. The two first ones is in the xsk_buff_alloc()
case set in the allocation routine and data_end is set when a packet
is received in the receive irq function. This unfortunately leads to
worse performance since the xdp_buff is touched twice with a long time
period in between leading to an extra cache miss. Instead, we fill out
the xdp_buff with all 3 fields at one single point in time in the
driver, when the size of the packet is known. Hence this helper. Note
that the driver has to use this helper (or set all three fields
itself) when using xsk_buff_alloc_batch(). xsk_buff_alloc() works as
before and does not require this.
void xsk_buff_set_size(struct xdp_buff *xdp, u32 size);
Signed-off-by: Magnus Karlsson <magnus.karlsson@intel.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20210922075613.12186-3-magnus.karlsson@gmail.com
