Pull vfs mount updates from Christian Brauner:
"This contains the work to retrieve detailed information about mounts
via two new system calls. This is hopefully the beginning of the end
of the saga that started with fsinfo() years ago.
The LWN articles in [1] and [2] can serve as a summary so we can avoid
rehashing everything here.
At LSFMM in May 2022 we got into a room and agreed on what we want to
do about fsinfo(). Basically, split it into pieces. This is the first
part of that agreement. Specifically, it is concerned with retrieving
information about mounts. So this only concerns the mount information
retrieval, not the mount table change notification, or the extended
filesystem specific mount option work. That is separate work.
Currently mounts have a 32bit id. Mount ids are already in heavy use
by libmount and other low-level userspace but they can't be relied
upon because they're recycled very quickly. We agreed that mounts
should carry a unique 64bit id by which they can be referenced
directly. This is now implemented as part of this work.
The new 64bit mount id is exposed in statx() through the new
STATX_MNT_ID_UNIQUE flag. If the flag isn't raised the old mount id is
returned. If it is raised and the kernel supports the new 64bit mount
id the flag is raised in the result mask and the new 64bit mount id is
returned. New and old mount ids do not overlap so they cannot be
conflated.
Two new system calls are introduced that operate on the 64bit mount
id: statmount() and listmount(). A summary of the api and usage can be
found on LWN as well (cf. [3]) but of course, I'll provide a summary
here as well.
Both system calls rely on struct mnt_id_req. Which is the request
struct used to pass the 64bit mount id identifying the mount to
operate on. It is extensible to allow for the addition of new
parameters and for future use in other apis that make use of mount
ids.
statmount() mimicks the semantics of statx() and exposes a set flags
that userspace may raise in mnt_id_req to request specific information
to be retrieved. A statmount() call returns a struct statmount filled
in with information about the requested mount. Supported requests are
indicated by raising the request flag passed in struct mnt_id_req in
the @mask argument in struct statmount.
Currently we do support:
- STATMOUNT_SB_BASIC:
Basic filesystem info
- STATMOUNT_MNT_BASIC
Mount information (mount id, parent mount id, mount attributes etc)
- STATMOUNT_PROPAGATE_FROM
Propagation from what mount in current namespace
- STATMOUNT_MNT_ROOT
Path of the root of the mount (e.g., mount --bind /bla /mnt returns /bla)
- STATMOUNT_MNT_POINT
Path of the mount point (e.g., mount --bind /bla /mnt returns /mnt)
- STATMOUNT_FS_TYPE
Name of the filesystem type as the magic number isn't enough due to submounts
The string options STATMOUNT_MNT_{ROOT,POINT} and STATMOUNT_FS_TYPE
are appended to the end of the struct. Userspace can use the offsets
in @fs_type, @mnt_root, and @mnt_point to reference those strings
easily.
The struct statmount reserves quite a bit of space currently for
future extensibility. This isn't really a problem and if this bothers
us we can just send a follow-up pull request during this cycle.
listmount() is given a 64bit mount id via mnt_id_req just as
statmount(). It takes a buffer and a size to return an array of the
64bit ids of the child mounts of the requested mount. Userspace can
thus choose to either retrieve child mounts for a mount in batches or
iterate through the child mounts. For most use-cases it will be
sufficient to just leave space for a few child mounts. But for big
mount tables having an iterator is really helpful. Iterating through a
mount table works by setting @param in mnt_id_req to the mount id of
the last child mount retrieved in the previous listmount() call"
Link: https://lwn.net/Articles/934469 [1]
Link: https://lwn.net/Articles/829212 [2]
Link: https://lwn.net/Articles/950569 [3]
* tag 'vfs-6.8.mount' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs:
add selftest for statmount/listmount
fs: keep struct mnt_id_req extensible
wire up syscalls for statmount/listmount
add listmount(2) syscall
statmount: simplify string option retrieval
statmount: simplify numeric option retrieval
add statmount(2) syscall
namespace: extract show_path() helper
mounts: keep list of mounts in an rbtree
add unique mount ID
Pull misc vfs updates from Christian Brauner:
"This contains the usual miscellaneous features, cleanups, and fixes
for vfs and individual fses.
Features:
- Add Jan Kara as VFS reviewer
- Show correct device and inode numbers in proc/<pid>/maps for vma
files on stacked filesystems. This is now easily doable thanks to
the backing file work from the last cycles. This comes with
selftests
Cleanups:
- Remove a redundant might_sleep() from wait_on_inode()
- Initialize pointer with NULL, not 0
- Clarify comment on access_override_creds()
- Rework and simplify eventfd_signal() and eventfd_signal_mask()
helpers
- Process aio completions in batches to avoid needless wakeups
- Completely decouple struct mnt_idmap from namespaces. We now only
keep the actual idmapping around and don't stash references to
namespaces
- Reformat maintainer entries to indicate that a given subsystem
belongs to fs/
- Simplify fput() for files that were never opened
- Get rid of various pointless file helpers
- Rename various file helpers
- Rename struct file members after SLAB_TYPESAFE_BY_RCU switch from
last cycle
- Make relatime_need_update() return bool
- Use GFP_KERNEL instead of GFP_USER when allocating superblocks
- Replace deprecated ida_simple_*() calls with their current ida_*()
counterparts
Fixes:
- Fix comments on user namespace id mapping helpers. They aren't
kernel doc comments so they shouldn't be using /**
- s/Retuns/Returns/g in various places
- Add missing parameter documentation on can_move_mount_beneath()
- Rename i_mapping->private_data to i_mapping->i_private_data
- Fix a false-positive lockdep warning in pipe_write() for watch
queues
- Improve __fget_files_rcu() code generation to improve performance
- Only notify writer that pipe resizing has finished after setting
pipe->max_usage otherwise writers are never notified that the pipe
has been resized and hang
- Fix some kernel docs in hfsplus
- s/passs/pass/g in various places
- Fix kernel docs in ntfs
- Fix kcalloc() arguments order reported by gcc 14
- Fix uninitialized value in reiserfs"
* tag 'vfs-6.8.misc' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs: (36 commits)
reiserfs: fix uninit-value in comp_keys
watch_queue: fix kcalloc() arguments order
ntfs: dir.c: fix kernel-doc function parameter warnings
fs: fix doc comment typo fs tree wide
selftests/overlayfs: verify device and inode numbers in /proc/pid/maps
fs/proc: show correct device and inode numbers in /proc/pid/maps
eventfd: Remove usage of the deprecated ida_simple_xx() API
fs: super: use GFP_KERNEL instead of GFP_USER for super block allocation
fs/hfsplus: wrapper.c: fix kernel-doc warnings
fs: add Jan Kara as reviewer
fs/inode: Make relatime_need_update return bool
pipe: wakeup wr_wait after setting max_usage
file: remove __receive_fd()
file: stop exposing receive_fd_user()
fs: replace f_rcuhead with f_task_work
file: remove pointless wrapper
file: s/close_fd_get_file()/file_close_fd()/g
Improve __fget_files_rcu() code generation (and thus __fget_light())
file: massage cleanup of files that failed to open
fs/pipe: Fix lockdep false-positive in watchqueue pipe_write()
...
KVM x86 Hyper-V changes for 6.8:
- Guard KVM-on-HyperV's range-based TLB flush hooks with an #ifdef on
CONFIG_HYPERV as a minor optimization, and to self-document the code.
- Add CONFIG_KVM_HYPERV to allow disabling KVM support for HyperV "emulation"
at build time.
KVM/arm64 updates for Linux 6.8
- LPA2 support, adding 52bit IPA/PA capability for 4kB and 16kB
base granule sizes. Branch shared with the arm64 tree.
- Large Fine-Grained Trap rework, bringing some sanity to the
feature, although there is more to come. This comes with
a prefix branch shared with the arm64 tree.
- Some additional Nested Virtualization groundwork, mostly
introducing the NV2 VNCR support and retargetting the NV
support to that version of the architecture.
- A small set of vgic fixes and associated cleanups.
ASoC: Updates for v6.8
This is a relatively quiet release, there's a lot of driver specific
changes and the usual high level of activity in the SOF core but the
one big core change was Mormioto-san's work to support more N:M
CPU:CODEC mapping cases. Highlights include:
- Enhanced support for N:M CPU:CODEC mappings in the core and in
audio-graph-card2.
- Support for falling back to older SOF IPC versions where firmware for
new versions is not available.
- Support for notification of control changes generated by SOF firmware
with IPC4.
- Device tree support for describing parts of the card which can be
active over suspend (for very low power playback or wake word use
cases).
- ACPI parsing support for the ES83xx driver, reducing the number of
quirks neede for x86 systems.
- Support for more AMD and Intel systems, NXP i.MX8m MICFIL, Qualcomm
SM8250, SM8550, SM8650 and X1E80100.
- Removal of Freescale MPC8610 support, the SoC is no longer supported
by Linux.
Commit ba24ea1291 ("net/sched: Retire ipt action") removed the ipt action
but not the testcases. This patch removes the outstanding tdc tests.
Signed-off-by: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The expression "source ../lib.sh" added to net/forwarding/lib.sh in commit
25ae948b44 ("selftests/net: add lib.sh") does not work for tests outside
net/forwarding which source net/forwarding/lib.sh (1). It also does not
work in some cases where only a subset of tests are exported (2).
Avoid the problems mentioned above by replacing the faulty expression with
a copy of the content from net/lib.sh which is used by files under
net/forwarding.
A more thorough solution which avoids duplicating content between
net/lib.sh and net/forwarding/lib.sh has been posted here:
https://lore.kernel.org/netdev/20231222135836.992841-1-bpoirier@nvidia.com/
The approach in the current patch is a stopgap solution to avoid submitting
large changes at the eleventh hour of this development cycle.
Example of problem 1)
tools/testing/selftests/drivers/net/bonding$ ./dev_addr_lists.sh
./net_forwarding_lib.sh: line 41: ../lib.sh: No such file or directory
TEST: bonding cleanup mode active-backup [ OK ]
TEST: bonding cleanup mode 802.3ad [ OK ]
TEST: bonding LACPDU multicast address to slave (from bond down) [ OK ]
TEST: bonding LACPDU multicast address to slave (from bond up) [ OK ]
An error message is printed but since the test does not use functions from
net/lib.sh, the test results are not affected.
Example of problem 2)
tools/testing/selftests$ make install TARGETS="net/forwarding"
tools/testing/selftests$ cd kselftest_install/net/forwarding/
tools/testing/selftests/kselftest_install/net/forwarding$ ./pedit_ip.sh veth{0..3}
lib.sh: line 41: ../lib.sh: No such file or directory
TEST: ping [ OK ]
TEST: ping6 [ OK ]
./pedit_ip.sh: line 135: busywait: command not found
TEST: dev veth1 ingress pedit ip src set 198.51.100.1 [FAIL]
Expected to get 10 packets, but got .
./pedit_ip.sh: line 135: busywait: command not found
TEST: dev veth2 egress pedit ip src set 198.51.100.1 [FAIL]
Expected to get 10 packets, but got .
./pedit_ip.sh: line 135: busywait: command not found
TEST: dev veth1 ingress pedit ip dst set 198.51.100.1 [FAIL]
Expected to get 10 packets, but got .
./pedit_ip.sh: line 135: busywait: command not found
TEST: dev veth2 egress pedit ip dst set 198.51.100.1 [FAIL]
Expected to get 10 packets, but got .
./pedit_ip.sh: line 135: busywait: command not found
TEST: dev veth1 ingress pedit ip6 src set 2001:db8:2::1 [FAIL]
Expected to get 10 packets, but got .
./pedit_ip.sh: line 135: busywait: command not found
TEST: dev veth2 egress pedit ip6 src set 2001:db8:2::1 [FAIL]
Expected to get 10 packets, but got .
./pedit_ip.sh: line 135: busywait: command not found
TEST: dev veth1 ingress pedit ip6 dst set 2001:db8:2::1 [FAIL]
Expected to get 10 packets, but got .
./pedit_ip.sh: line 135: busywait: command not found
TEST: dev veth2 egress pedit ip6 dst set 2001:db8:2::1 [FAIL]
Expected to get 10 packets, but got .
In this case, the test results are affected.
Fixes: 25ae948b44 ("selftests/net: add lib.sh")
Suggested-by: Ido Schimmel <idosch@nvidia.com>
Suggested-by: Petr Machata <petrm@nvidia.com>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Tested-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Benjamin Poirier <bpoirier@nvidia.com>
Reviewed-by: Hangbin Liu <liuhangbin@gmail.com>
Link: https://lore.kernel.org/r/20240104141109.100672-1-bpoirier@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Daniel Borkmann says:
====================
pull-request: bpf-next 2024-01-05
We've added 40 non-merge commits during the last 2 day(s) which contain
a total of 73 files changed, 1526 insertions(+), 951 deletions(-).
The main changes are:
1) Fix a memory leak when streaming AF_UNIX sockets were inserted
into multiple sockmap slots/maps, from John Fastabend.
2) Fix gotol in s390 BPF JIT with large offsets, from Ilya Leoshkevich.
3) Fix reattachment branch in bpf_tracing_prog_attach() and reject
the request if there is no valid attach_btf, from Jiri Olsa.
4) Remove deprecated bpfilter kernel leftovers given the project
is developed in user space (https://github.com/facebook/bpfilter),
from Quentin Deslandes.
5) Relax tracing BPF program recursive attach rules given right now
it is not possible to create tracing program call cycles,
from Dmitrii Dolgov.
6) Fix excessive memory consumption for the bpf_global_percpu_ma
for systems with a large number of CPUs, from Yonghong Song.
7) Small x86 BPF JIT cleanup to reuse emit_nops instead of open-coding
memcpy of x86_nops, from Leon Hwang.
8) Follow-up for libbpf to support __arg_ctx global function argument tag
semantics to complement the merged kernel side, from Andrii Nakryiko.
9) Introduce "volatile compare" macros for BPF selftests in order
to make the latter more robust against compiler optimization,
from Alexei Starovoitov.
10) Small simplification in verifier's size checking of helper accesses
along with additional selftests, from Andrei Matei.
* tag 'for-netdev' of https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next: (40 commits)
selftests/bpf: Test re-attachment fix for bpf_tracing_prog_attach
bpf: Fix re-attachment branch in bpf_tracing_prog_attach
selftests/bpf: Add test for recursive attachment of tracing progs
bpf: Relax tracing prog recursive attach rules
bpf, x86: Use emit_nops to replace memcpy x86_nops
selftests/bpf: Test gotol with large offsets
selftests/bpf: Double the size of test_loader log
s390/bpf: Fix gotol with large offsets
bpfilter: remove bpfilter
bpf: Remove unnecessary cpu == 0 check in memalloc
selftests/bpf: add __arg_ctx BTF rewrite test
selftests/bpf: add arg:ctx cases to test_global_funcs tests
libbpf: implement __arg_ctx fallback logic
libbpf: move BTF loading step after relocation step
libbpf: move exception callbacks assignment logic into relocation step
libbpf: use stable map placeholder FDs
libbpf: don't rely on map->fd as an indicator of map being created
libbpf: use explicit map reuse flag to skip map creation steps
libbpf: make uniform use of btf__fd() accessor inside libbpf
selftests/bpf: Add a selftest with > 512-byte percpu allocation size
...
====================
Link: https://lore.kernel.org/r/20240105170105.21070-1-daniel@iogearbox.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Currently there is no test which checks that IPv6 extension header packets
successfully coalesce. This commit adds a test, which verifies two IPv6
packets with HBH extension headers do coalesce, and another test which
checks that packets with different extension header data do not coalesce
in GRO.
I changed the receive socket filter to accept a packet with one extension
header. This change exposed a bug in the fragment test -- the old BPF did
not accept the fragment packet. I updated correct_num_packets in the
fragment test accordingly.
Signed-off-by: Richard Gobert <richardbgobert@gmail.com>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Link: https://lore.kernel.org/r/69282fed-2415-47e8-b3d3-34939ec3eb56@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Pull networking fixes from Jakub Kicinski:
"Including fixes from wireless and netfilter.
We haven't accumulated much over the break. If it wasn't for the
uninterrupted stream of fixes for Intel drivers this PR would be very
slim. There was a handful of user reports, however, either they stood
out because of the lower traffic or users have had more time to test
over the break. The ones which are v6.7-relevant should be wrapped up.
Current release - regressions:
- Revert "net: ipv6/addrconf: clamp preferred_lft to the minimum
required", it caused issues on networks where routers send prefixes
with preferred_lft=0
- wifi:
- iwlwifi: pcie: don't synchronize IRQs from IRQ, prevent deadlock
- mac80211: fix re-adding debugfs entries during reconfiguration
Current release - new code bugs:
- tcp: print AO/MD5 messages only if there are any keys
Previous releases - regressions:
- virtio_net: fix missing dma unmap for resize, prevent OOM
Previous releases - always broken:
- mptcp: prevent tcp diag from closing listener subflows
- nf_tables:
- set transport header offset for egress hook, fix IPv4 mangling
- skip set commit for deleted/destroyed sets, avoid double deactivation
- nat: make sure action is set for all ct states, fix openvswitch
matching on ICMP packets in related state
- eth: mlxbf_gige: fix receive hang under heavy traffic
- eth: r8169: fix PCI error on system resume for RTL8168FP
- net: add missing getsockopt(SO_TIMESTAMPING_NEW) and cmsg handling"
* tag 'net-6.7-rc9' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (52 commits)
net/tcp: Only produce AO/MD5 logs if there are any keys
net: Implement missing SO_TIMESTAMPING_NEW cmsg support
bnxt_en: Remove mis-applied code from bnxt_cfg_ntp_filters()
net: ravb: Wait for operating mode to be applied
asix: Add check for usbnet_get_endpoints
octeontx2-af: Re-enable MAC TX in otx2_stop processing
octeontx2-af: Always configure NIX TX link credits based on max frame size
net/smc: fix invalid link access in dumping SMC-R connections
net/qla3xxx: fix potential memleak in ql_alloc_buffer_queues
virtio_net: fix missing dma unmap for resize
igc: Fix hicredit calculation
ice: fix Get link status data length
i40e: Restore VF MSI-X state during PCI reset
i40e: fix use-after-free in i40e_aqc_add_filters()
net: Save and restore msg_namelen in sock_sendmsg
netfilter: nft_immediate: drop chain reference counter on error
netfilter: nf_nat: fix action not being set for all ct states
net: bcmgenet: Fix FCS generation for fragmented skbuffs
mptcp: prevent tcp diag from closing listener subflows
MAINTAINERS: add Geliang as reviewer for MPTCP
...
Testing long jumps requires having >32k instructions. That many
instructions require the verifier log buffer of 2 megabytes.
The regular test_progs run doesn't need an increased buffer, since
gotol test with 40k instructions doesn't request a log,
but test_progs -v will set the verifier log level.
Hence to avoid breaking gotol test with -v increase the buffer size.
Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Acked-by: Yonghong Song <yonghong.song@linux.dev>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/r/20240102193531.3169422-3-iii@linux.ibm.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
* for-next/selftests:
kselftest/arm64: Don't probe the current VL for unsupported vector types
kselftest/arm64: Log SVCR when the SME tests barf
kselftest/arm64: Improve output for skipped TPIDR2 ABI test
Add a few extra cases of global funcs with context arguments. This time
rely on "arg:ctx" decl_tag (__arg_ctx macro), but put it next to
"classic" cases where context argument has to be of an exact type that
BPF verifier expects (e.g., bpf_user_pt_regs_t for kprobe/uprobe).
Colocating all these cases separately from other global func args that
rely on arg:xxx decl tags (in verifier_global_subprogs.c) allows for
simpler backwards compatibility testing on old kernels. All the cases in
test_global_func_ctx_args.c are supposed to work on older kernels, which
was manually validated during development.
Acked-by: Jiri Olsa <jolsa@kernel.org>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/r/20240104013847.3875810-9-andrii@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
In the previous patch, the maximum data size for bpf_global_percpu_ma
is 512 bytes. This breaks selftest test_bpf_ma. The test is adjusted
in two aspects:
- Since the maximum allowed data size for bpf_global_percpu_ma is
512, remove all tests beyond that, names sizes 1024, 2048 and 4096.
- Previously the percpu data size is bucket_size - 8 in order to
avoid percpu allocation into the next bucket. This patch removed
such data size adjustment thanks to Patch 1.
Also, a better way to generate BTF type is used than adding
a member to the value struct.
Acked-by: Hou Tao <houtao1@huawei.com>
Signed-off-by: Yonghong Song <yonghong.song@linux.dev>
Link: https://lore.kernel.org/r/20231222031807.1292853-1-yonghong.song@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
The patch set [1] added a general lib.sh in net selftests, and converted
several test scripts to source the lib.sh.
unicast_extensions.sh (converted in [1]) and pmtu.sh (converted in [2])
have a /bin/sh shebang which may point to various shells in different
distributions, but "source" is only available in some of them. For
example, "source" is a built-it function in bash, but it cannot be
used in dash.
Refer to other scripts that were converted together, simply change the
shebang to bash to fix the following issues when the default /bin/sh
points to other shells.
not ok 51 selftests: net: unicast_extensions.sh # exit=1
v1 -> v2:
- Fix pmtu.sh which has the same issue as unicast_extensions.sh,
suggested by Hangbin
- Change the style of the "source" line to be consistent with other
tests, suggested by Hangbin
Link: https://lore.kernel.org/all/20231202020110.362433-1-liuhangbin@gmail.com/ [1]
Link: https://lore.kernel.org/all/20231219094856.1740079-1-liuhangbin@gmail.com/ [2]
Reported-by: kernel test robot <oliver.sang@intel.com>
Fixes: 378f082eaf ("selftests/net: convert pmtu.sh to run it in unique namespace")
Fixes: 0f4765d0b4 ("selftests/net: convert unicast_extensions.sh to run it in unique namespace")
Signed-off-by: Yujie Liu <yujie.liu@intel.com>
Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com>
Reviewed-by: Hangbin Liu <liuhangbin@gmail.com>
Reviewed-by: Muhammad Usama Anjum <usama.anjum@collabora.com>
Link: https://lore.kernel.org/r/20231229131931.3961150-1-yujie.liu@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
bpf_nop_mov(var) asm macro emits nop register move: rX = rX.
If 'var' is a scalar and not a fixed constant the verifier will assign ID to it.
If it's later spilled the stack slot will carry that ID as well.
Hence the range refining comparison "if rX < const" will update all copies
including spilled slot.
This macro is a temporary workaround until the verifier gets smarter.
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20231226191148.48536-6-alexei.starovoitov@gmail.com
Convert exceptions_assert.c to bpf_cmp_unlikely() macro.
Since
bpf_assert(bpf_cmp_unlikely(var, ==, 100));
other code;
will generate assembly code:
if r1 == 100 goto L2;
r0 = 0
call bpf_throw
L1:
other code;
...
L2: goto L1;
LLVM generates redundant basic block with extra goto. LLVM will be fixed eventually.
Right now it's less efficient than __bpf_assert(var, ==, 100) macro that produces:
if r1 == 100 goto L1;
r0 = 0
call bpf_throw
L1:
other code;
But extra goto doesn't hurt the verification process.
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Acked-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/bpf/20231226191148.48536-4-alexei.starovoitov@gmail.com
Compilers optimize conditional operators at will, but often bpf programmers
want to force compilers to keep the same operator in asm as it's written in C.
Introduce bpf_cmp_likely/unlikely(var1, conditional_op, var2) macros that can be used as:
- if (seen >= 1000)
+ if (bpf_cmp_unlikely(seen, >=, 1000))
The macros take advantage of BPF assembly that is C like.
The macros check the sign of variable 'seen' and emits either
signed or unsigned compare.
For example:
int a;
bpf_cmp_unlikely(a, >, 0) will be translated to 'if rX s> 0 goto' in BPF assembly.
unsigned int a;
bpf_cmp_unlikely(a, >, 0) will be translated to 'if rX > 0 goto' in BPF assembly.
C type conversions coupled with comparison operator are tricky.
int i = -1;
unsigned int j = 1;
if (i < j) // this is false.
long i = -1;
unsigned int j = 1;
if (i < j) // this is true.
Make sure BPF program is compiled with -Wsign-compare then the macros will catch
the mistake.
The macros check LHS (left hand side) only to figure out the sign of compare.
'if 0 < rX goto' is not allowed in the assembly, so the users
have to use a variable on LHS anyway.
The patch updates few tests to demonstrate the use of the macros.
The macro allows to use BPF_JSET in C code, since LLVM doesn't generate it at
present. For example:
if (i & j) compiles into r0 &= r1; if r0 == 0 goto
while
if (bpf_cmp_unlikely(i, &, j)) compiles into if r0 & r1 goto
Note that the macros has to be careful with RHS assembly predicate.
Since:
u64 __rhs = 1ull << 42;
asm goto("if r0 < %[rhs] goto +1" :: [rhs] "ri" (__rhs));
LLVM will silently truncate 64-bit constant into s32 imm.
Note that [lhs] "r"((short)LHS) the type cast is a workaround for LLVM issue.
When LHS is exactly 32-bit LLVM emits redundant <<=32, >>=32 to zero upper 32-bits.
When LHS is 64 or 16 or 8-bit variable there are no shifts.
When LHS is 32-bit the (u64) cast doesn't help. Hence use (short) cast.
It does _not_ truncate the variable before it's assigned to a register.
Traditional likely()/unlikely() macros that use __builtin_expect(!!(x), 1 or 0)
have no effect on these macros, hence macros implement the logic manually.
bpf_cmp_unlikely() macro preserves compare operator as-is while
bpf_cmp_likely() macro flips the compare.
Consider two cases:
A.
for() {
if (foo >= 10) {
bar += foo;
}
other code;
}
B.
for() {
if (foo >= 10)
break;
other code;
}
It's ok to use either bpf_cmp_likely or bpf_cmp_unlikely macros in both cases,
but consider that 'break' is effectively 'goto out_of_the_loop'.
Hence it's better to use bpf_cmp_unlikely in the B case.
While 'bar += foo' is better to keep as 'fallthrough' == likely code path in the A case.
When it's written as:
A.
for() {
if (bpf_cmp_likely(foo, >=, 10)) {
bar += foo;
}
other code;
}
B.
for() {
if (bpf_cmp_unlikely(foo, >=, 10))
break;
other code;
}
The assembly will look like:
A.
for() {
if r1 < 10 goto L1;
bar += foo;
L1:
other code;
}
B.
for() {
if r1 >= 10 goto L2;
other code;
}
L2:
The bpf_cmp_likely vs bpf_cmp_unlikely changes basic block layout, hence it will
greatly influence the verification process. The number of processed instructions
will be different, since the verifier walks the fallthrough first.
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Acked-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/bpf/20231226191148.48536-3-alexei.starovoitov@gmail.com
GCC's -Wall includes -Wsign-compare while clang does not.
Since BPF programs are built with clang we need to add this flag explicitly
to catch problematic comparisons like:
int i = -1;
unsigned int j = 1;
if (i < j) // this is false.
long i = -1;
unsigned int j = 1;
if (i < j) // this is true.
C standard for reference:
- If either operand is unsigned long the other shall be converted to unsigned long.
- Otherwise, if one operand is a long int and the other unsigned int, then if a
long int can represent all the values of an unsigned int, the unsigned int
shall be converted to a long int; otherwise both operands shall be converted to
unsigned long int.
- Otherwise, if either operand is long, the other shall be converted to long.
- Otherwise, if either operand is unsigned, the other shall be converted to unsigned.
Unfortunately clang's -Wsign-compare is very noisy.
It complains about (s32)a == (u32)b which is safe and doen't have surprising behavior.
This patch fixes some of the issues. It needs a follow up to fix the rest.
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Acked-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/bpf/20231226191148.48536-2-alexei.starovoitov@gmail.com
This patch simplifies the verification of size arguments associated to
pointer arguments to helpers and kfuncs. Many helpers take a pointer
argument followed by the size of the memory access performed to be
performed through that pointer. Before this patch, the handling of the
size argument in check_mem_size_reg() was confusing and wasteful: if the
size register's lower bound was 0, then the verification was done twice:
once considering the size of the access to be the lower-bound of the
respective argument, and once considering the upper bound (even if the
two are the same). The upper bound checking is a super-set of the
lower-bound checking(*), except: the only point of the lower-bound check
is to handle the case where zero-sized-accesses are explicitly not
allowed and the lower-bound is zero. This static condition is now
checked explicitly, replacing a much more complex, expensive and
confusing verification call to check_helper_mem_access().
Error messages change in this patch. Before, messages about illegal
zero-size accesses depended on the type of the pointer and on other
conditions, and sometimes the message was plain wrong: in some tests
that changed you'll see that the old message was something like "R1 min
value is outside of the allowed memory range", where R1 is the pointer
register; the error was wrongly claiming that the pointer was bad
instead of the size being bad. Other times the information that the size
came for a register with a possible range of values was wrong, and the
error presented the size as a fixed zero. Now the errors refer to the
right register. However, the old error messages did contain useful
information about the pointer register which is now lost; recovering
this information was deemed not important enough.
(*) Besides standing to reason that the checks for a bigger size access
are a super-set of the checks for a smaller size access, I have also
mechanically verified this by reading the code for all types of
pointers. I could convince myself that it's true for all but
PTR_TO_BTF_ID (check_ptr_to_btf_access). There, simply looking
line-by-line does not immediately prove what we want. If anyone has any
qualms, let me know.
Signed-off-by: Andrei Matei <andreimatei1@gmail.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20231221232225.568730-2-andreimatei1@gmail.com
The "count" parameter associated with the 'cpus' parameter of the
hwprobe syscall is the size in bytes of 'cpus'. Naming it 'cpu_count'
may mislead users (it did me) to think it's the number of CPUs that
are or can be represented by 'cpus' instead. This is particularly
easy (IMO) to get wrong since 'cpus' is documented to be defined by
CPU_SET(3) and CPU_SET(3) also documents a CPU_COUNT() (the number
of CPUs in set) macro. CPU_SET(3) refers to the size of cpu sets
with 'setsize'. Adopt 'cpusetsize' for the hwprobe parameter and
specifically state it is in bytes in Documentation/riscv/hwprobe.rst
to clarify.
Reviewed-by: Palmer Dabbelt <palmer@rivosinc.com>
Reviewed-by: Conor Dooley <conor.dooley@microchip.com>
Signed-off-by: Andrew Jones <ajones@ventanamicro.com>
Link: https://lore.kernel.org/r/20231122164700.127954-7-ajones@ventanamicro.com
Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com>
KVM/riscv changes for 6.8 part #1
- KVM_GET_REG_LIST improvement for vector registers
- Generate ISA extension reg_list using macros in get-reg-list selftest
- Steal time account support along with selftest
Similar to commit be80942465 ("selftests: bonding: do not set port down
before adding to bond"). The bond-arp-interval-causes-panic test failed
after commit a4abfa627c ("net: rtnetlink: Enslave device before bringing
it up") as the kernel will set the port down _after_ adding to bond if setting
port down specifically.
Fix it by removing the link down operation when adding to bond.
Fixes: 2ffd57327f ("selftests: bonding: cause oops in bond_rr_gen_slave_id")
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Tested-by: Benjamin Poirier <benjamin.poirier@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch adds a new helper chk_msk_cestab() to check the current
established connections counter MIB_CURRESTAB in diag.sh. Invoke it
to check the counter during the connection after every chk_msk_inuse().
Signed-off-by: Geliang Tang <geliang.tang@linux.dev>
Reviewed-by: Matthieu Baerts <matttbe@kernel.org>
Signed-off-by: Matthieu Baerts <matttbe@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch adds a new helper chk_cestab_nr() to check the current
established connections counter MIB_CURRESTAB. Set the newly added
variables cestab_ns1 and cestab_ns2 to indicate how many connections
are expected in ns1 or ns2.
Invoke check_cestab() to check the counter during the connection in
do_transfer() and invoke chk_cestab_nr() to re-check it when the
connection closed. These checks are embedded in add_tests().
Signed-off-by: Geliang Tang <geliang.tang@linux.dev>
Acked-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Matthieu Baerts <matttbe@kernel.org>
Signed-off-by: Matthieu Baerts <matttbe@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
