Introduce BPF_F_LOCK flag for map_lookup and map_update syscall commands
and for map_update() helper function.
In all these cases take a lock of existing element (which was provided
in BTF description) before copying (in or out) the rest of map value.
Implementation details that are part of uapi:
Array:
The array map takes the element lock for lookup/update.
Hash:
hash map also takes the lock for lookup/update and tries to avoid the bucket lock.
If old element exists it takes the element lock and updates the element in place.
If element doesn't exist it allocates new one and inserts into hash table
while holding the bucket lock.
In rare case the hashmap has to take both the bucket lock and the element lock
to update old value in place.
Cgroup local storage:
It is similar to array. update in place and lookup are done with lock taken.
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
add bpf_spin_lock C based test that requires latest llvm with BTF support
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
add bpf_spin_lock tests to test_verifier.c that don't require
latest llvm with BTF support
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Introduce 'struct bpf_spin_lock' and bpf_spin_lock/unlock() helpers to let
bpf program serialize access to other variables.
Example:
struct hash_elem {
int cnt;
struct bpf_spin_lock lock;
};
struct hash_elem * val = bpf_map_lookup_elem(&hash_map, &key);
if (val) {
bpf_spin_lock(&val->lock);
val->cnt++;
bpf_spin_unlock(&val->lock);
}
Restrictions and safety checks:
- bpf_spin_lock is only allowed inside HASH and ARRAY maps.
- BTF description of the map is mandatory for safety analysis.
- bpf program can take one bpf_spin_lock at a time, since two or more can
cause dead locks.
- only one 'struct bpf_spin_lock' is allowed per map element.
It drastically simplifies implementation yet allows bpf program to use
any number of bpf_spin_locks.
- when bpf_spin_lock is taken the calls (either bpf2bpf or helpers) are not allowed.
- bpf program must bpf_spin_unlock() before return.
- bpf program can access 'struct bpf_spin_lock' only via
bpf_spin_lock()/bpf_spin_unlock() helpers.
- load/store into 'struct bpf_spin_lock lock;' field is not allowed.
- to use bpf_spin_lock() helper the BTF description of map value must be
a struct and have 'struct bpf_spin_lock anyname;' field at the top level.
Nested lock inside another struct is not allowed.
- syscall map_lookup doesn't copy bpf_spin_lock field to user space.
- syscall map_update and program map_update do not update bpf_spin_lock field.
- bpf_spin_lock cannot be on the stack or inside networking packet.
bpf_spin_lock can only be inside HASH or ARRAY map value.
- bpf_spin_lock is available to root only and to all program types.
- bpf_spin_lock is not allowed in inner maps of map-in-map.
- ld_abs is not allowed inside spin_lock-ed region.
- tracing progs and socket filter progs cannot use bpf_spin_lock due to
insufficient preemption checks
Implementation details:
- cgroup-bpf class of programs can nest with xdp/tc programs.
Hence bpf_spin_lock is equivalent to spin_lock_irqsave.
Other solutions to avoid nested bpf_spin_lock are possible.
Like making sure that all networking progs run with softirq disabled.
spin_lock_irqsave is the simplest and doesn't add overhead to the
programs that don't use it.
- arch_spinlock_t is used when its implemented as queued_spin_lock
- archs can force their own arch_spinlock_t
- on architectures where queued_spin_lock is not available and
sizeof(arch_spinlock_t) != sizeof(__u32) trivial lock is used.
- presence of bpf_spin_lock inside map value could have been indicated via
extra flag during map_create, but specifying it via BTF is cleaner.
It provides introspection for map key/value and reduces user mistakes.
Next steps:
- allow bpf_spin_lock in other map types (like cgroup local storage)
- introduce BPF_F_LOCK flag for bpf_map_update() syscall and helper
to request kernel to grab bpf_spin_lock before rewriting the value.
That will serialize access to map elements.
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Building with W=1 reveals some bitrot:
CC kernel/bpf/cgroup.o
kernel/bpf/cgroup.c:238: warning: Function parameter or member 'flags' not described in '__cgroup_bpf_attach'
kernel/bpf/cgroup.c:367: warning: Function parameter or member 'unused_flags' not described in '__cgroup_bpf_detach'
Add a kerneldoc line for 'flags'.
Fixing the warning for 'unused_flags' is best approached by
removing the unused parameter on the function call.
Signed-off-by: Valdis Kletnieks <valdis.kletnieks@vt.edu>
Acked-by: Song Liu <songliubraving@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Compiling with W=1 generates warnings:
CC kernel/bpf/core.o
kernel/bpf/core.c:721:12: warning: no previous prototype for ?bpf_jit_alloc_exec_limit? [-Wmissing-prototypes]
721 | u64 __weak bpf_jit_alloc_exec_limit(void)
| ^~~~~~~~~~~~~~~~~~~~~~~~
kernel/bpf/core.c:757:14: warning: no previous prototype for ?bpf_jit_alloc_exec? [-Wmissing-prototypes]
757 | void *__weak bpf_jit_alloc_exec(unsigned long size)
| ^~~~~~~~~~~~~~~~~~
kernel/bpf/core.c:762:13: warning: no previous prototype for ?bpf_jit_free_exec? [-Wmissing-prototypes]
762 | void __weak bpf_jit_free_exec(void *addr)
| ^~~~~~~~~~~~~~~~~
All three are weak functions that archs can override, provide
proper prototypes for when a new arch provides their own.
Signed-off-by: Valdis Kletnieks <valdis.kletnieks@vt.edu>
Acked-by: Song Liu <songliubraving@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Stanislav Fomichev says:
====================
If test_maps/test_verifier is running against the kernel which doesn't
have _all_ BPF features enabled, it fails with an error. This patch
series tries to probe kernel support for each failed test and skip
it instead. This lets users run BPF selftests in the not-all-bpf-yes
environments and received correct PASS/NON-PASS result.
See https://www.spinics.net/lists/netdev/msg539331.html for more
context.
The series goes like this:
* patch #1 skips sockmap tests in test_maps.c if BPF_MAP_TYPE_SOCKMAP
map is not supported (if bpf_create_map fails, we probe the kernel
for support)
* patch #2 skips verifier tests if test->prog_type is not supported (if
bpf_verify_program fails, we probe the kernel for support)
* patch #3 skips verifier tests if test fixup map is not supported (if
create_map fails, we probe the kernel for support)
* next patches fix various small issues that arise from the first four:
* patch #4 sets "unknown func bpf_trace_printk#6" prog_type to
BPF_PROG_TYPE_TRACEPOINT so it is correctly skipped in
CONFIG_BPF_EVENTS=n case
* patch #5 exposes BPF_PROG_TYPE_CGROUP_{SKB,SOCK,SOCK_ADDR} only when
CONFIG_CGROUP_BPF=y, this makes verifier correctly skip appropriate
tests
v3 changes:
* rebased on top of Quentin's series which adds probes to libbpf
v2 changes:
* don't sprinkle "ifdef CONFIG_CGROUP_BPF" all around net/core/filter.c,
doing it only in the bpf_types.h is enough to disable
BPF_PROG_TYPE_CGROUP_{SKB,SOCK,SOCK_ADDR} prog types for non-cgroup
enabled kernels
====================
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
There is no way to exercise appropriate attach points without cgroups
enabled. This lets test_verifier correctly skip tests for these
prog_types if kernel was compiled without BPF cgroup support.
Signed-off-by: Stanislav Fomichev <sdf@google.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
We don't have this helper if the kernel was compiled without
CONFIG_BPF_EVENTS. Setting prog_type to BPF_PROG_TYPE_TRACEPOINT
let's verifier correctly skip this test based on the missing
prog_type support in the kernel.
Signed-off-by: Stanislav Fomichev <sdf@google.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Use recently introduced bpf_probe_map_type() to skip tests in the
test_verifier if map creation (create_map) fails. It's handled
explicitly for each fixup, i.e. if bpf_create_map returns negative fd,
we probe the kernel for the appropriate map support and skip the
test is map type is not supported.
Signed-off-by: Stanislav Fomichev <sdf@google.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Use recently introduced bpf_probe_prog_type() to skip tests in the
test_verifier() if bpf_verify_program() fails. The skipped test is
indicated in the output.
Example:
...
679/p bpf_get_stack return R0 within range SKIP (unsupported program
type 5)
680/p ld_abs: invalid op 1 OK
...
Summary: 863 PASSED, 165 SKIPPED, 3 FAILED
Signed-off-by: Stanislav Fomichev <sdf@google.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Use recently introduced bpf_probe_map_type() to skip test_sockmap()
if map creation fails. The skipped test is indicated in the output.
Example:
test_sockmap SKIP (unsupported map type BPF_MAP_TYPE_SOCKMAP)
Fork 1024 tasks to 'test_update_delete'
...
test_sockmap SKIP (unsupported map type BPF_MAP_TYPE_SOCKMAP)
Fork 1024 tasks to 'test_update_delete'
...
test_maps: OK, 2 SKIPPED
Signed-off-by: Stanislav Fomichev <sdf@google.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Huazhong Tan says:
====================
code optimizations & bugfixes for HNS3 driver
This patchset includes bugfixes and code optimizations for the HNS3
ethernet controller driver
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
In orginal codes, driver always enables flow director when
intializing. When user disable flow director with command
ethtool -K, the flow director will be enabled again after
resetting.
This patch fixes it by only enabling it when first initialzing.
Fixes: 6871af29b3 ("net: hns3: Add reset handle for flow director")
Signed-off-by: Jian Shen <shenjian15@huawei.com>
Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When VF is resetting, it can't communicate to PF with mailbox msg.
This patch adds reset state checking before sending keep alive msg
to PF.
Fixes: a6d818e31d ("net: hns3: Add vport alive state checking support")
Signed-off-by: Jian Shen <shenjian15@huawei.com>
Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
HNS3 VF driver support NIC and Roce, hdev stores NIC
handle and Roce handle, should use correct parameter for
container_of.
Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
While hclge_init_umv_space() failed in the hclge_init_ae_dev(),
we should undo all the operation which has been done successfully,
the last success operation maybe hclge_mac_mdio_config(), so if
hclge_init_umv_space() failed, we also need to undo it.
Fixes: 288475b2ad01 ("{topost} net: hns3: refine umv space allocation")
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The rss result is more uniform when use recommended hash key from
microsoft, instead of the one generated by netdev_rss_key_fill().
Also using hash algorithm "xor" is better than "toeplitz".
This patch modifies the default hash key and hash algorithm.
Signed-off-by: Jian Shen <shenjian15@huawei.com>
Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When the driver is unloading, if a global reset occurs,
unmap_ring_from_vector() in the hns3_nic_uninit_vector_data() will
fail, and hns3_nic_uninit_vector_data() just return. There may be
some netif_napi_del() not be done.
Since hardware will unmap all ring while resetting, so
hns3_nic_uninit_vector_data() should ignore this error, and do the
rest uninitialization.
Fixes: 76ad4f0ee7 ("net: hns3: Add support of HNS3 Ethernet Driver for hip08 SoC")
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Alexei Starovoitov
Valdis Kletnieks
Daniel Borkmann
Stanislav Fomichev
David S. Miller
Jian Shen
Peng Li
Huazhong Tan