The xfs_log_item flags were converted to atomic bitops as of commit
22525c17ed ("xfs: log item flags are racy"). The assert check for
AIL presence in xfs_buf_item_relse() still uses the old value based
check. This likely went unnoticed as XFS_LI_IN_AIL evaluates to 0
and causes the assert to unconditionally pass. Fix up the check.
Signed-off-by: Brian Foster <bfoster@redhat.com>
Fixes: 22525c17ed ("xfs: log item flags are racy")
Reviewed-by: Eric Sandeen <sandeen@redhat.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Fix yet another printf warning for %llu specifier on ppc64le. This time size_t
casting won't work, so cast to verbose `unsigned long long`.
Fixes: 166750bc1d ("libbpf: Support libbpf-provided extern variables")
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20191219052103.3515-1-andriin@fb.com
Naresh pointed out that libbpf builds fail on 32-bit architectures because
rlimit.rlim_cur is defined as 'unsigned long long' on those architectures.
Fix this by using %zu in printf and casting to size_t.
Fixes: dc3a2d2547 ("libbpf: Print hint about ulimit when getting permission denied error")
Reported-by: Naresh Kamboju <naresh.kamboju@linaro.org>
Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20191219090236.905059-1-toke@redhat.com
Maxim Mikityanskiy says:
====================
This series addresses the issue described in the commit message of the
first patch: lack of synchronization between XSK wakeup and destroying
the resources used by XSK wakeup. The idea is similar to napi_synchronize.
The series contains fixes for the drivers that implement XSK.
v2 incorporates changes suggested by Björn:
1. Call synchronize_rcu in Intel drivers only if the XDP program is
being unloaded.
2. Don't forget rcu_read_lock when wakeup is called from xsk_poll.
3. Use xs->zc as the condition to call ndo_xsk_wakeup.
====================
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Use synchronize_rcu to wait until the XSK wakeup function finishes
before destroying the resources it uses:
1. ixgbe_down already calls synchronize_rcu after setting __IXGBE_DOWN.
2. After switching the XDP program, call synchronize_rcu to let
ixgbe_xsk_wakeup exit before the XDP program is freed.
3. Changing the number of channels brings the interface down.
4. Disabling UMEM sets __IXGBE_TX_DISABLED before closing hardware
resources and resetting xsk_umem. Check that bit in ixgbe_xsk_wakeup to
avoid using the XDP ring when it's already destroyed. synchronize_rcu is
called from ixgbe_txrx_ring_disable.
Signed-off-by: Maxim Mikityanskiy <maximmi@mellanox.com>
Signed-off-by: Björn Töpel <bjorn.topel@intel.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20191217162023.16011-5-maximmi@mellanox.com
Use synchronize_rcu to wait until the XSK wakeup function finishes
before destroying the resources it uses:
1. i40e_down already calls synchronize_rcu. On i40e_down either
__I40E_VSI_DOWN or __I40E_CONFIG_BUSY is set. Check the latter in
i40e_xsk_wakeup (the former is already checked there).
2. After switching the XDP program, call synchronize_rcu to let
i40e_xsk_wakeup exit before the XDP program is freed.
3. Changing the number of channels brings the interface down (see
i40e_prep_for_reset and i40e_pf_quiesce_all_vsi).
4. Disabling UMEM sets __I40E_CONFIG_BUSY, too.
Signed-off-by: Maxim Mikityanskiy <maximmi@mellanox.com>
Signed-off-by: Björn Töpel <bjorn.topel@intel.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20191217162023.16011-4-maximmi@mellanox.com
After disabling resources necessary for XSK (the XDP program, channels,
XSK queues), use synchronize_rcu to wait until the XSK wakeup function
finishes, before freeing the resources.
Suspend XSK wakeups during switching channels. If the XDP program is
being removed, synchronize_rcu before closing the old channels to allow
XSK wakeup to complete.
Signed-off-by: Maxim Mikityanskiy <maximmi@mellanox.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20191217162023.16011-3-maximmi@mellanox.com
The XSK wakeup callback in drivers makes some sanity checks before
triggering NAPI. However, some configuration changes may occur during
this function that affect the result of those checks. For example, the
interface can go down, and all the resources will be destroyed after the
checks in the wakeup function, but before it attempts to use these
resources. Wrap this callback in rcu_read_lock to allow driver to
synchronize_rcu before actually destroying the resources.
xsk_wakeup is a new function that encapsulates calling ndo_xsk_wakeup
wrapped into the RCU lock. After this commit, xsk_poll starts using
xsk_wakeup and checks xs->zc instead of ndo_xsk_wakeup != NULL to decide
ndo_xsk_wakeup should be called. It also fixes a bug introduced with the
need_wakeup feature: a non-zero-copy socket may be used with a driver
supporting zero-copy, and in this case ndo_xsk_wakeup should not be
called, so the xs->zc check is the correct one.
Fixes: 77cd0d7b3f ("xsk: add support for need_wakeup flag in AF_XDP rings")
Signed-off-by: Maxim Mikityanskiy <maximmi@mellanox.com>
Signed-off-by: Björn Töpel <bjorn.topel@intel.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20191217162023.16011-2-maximmi@mellanox.com
Fix two issues in test_attach_probe:
1. it was not able to parse /proc/self/maps beyond the first line,
since %s means parse string until white space.
2. offset has to be accounted for otherwise uprobed address is incorrect.
Fixes: 1e8611bbdf ("selftests/bpf: add kprobe/uprobe selftests")
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Yonghong Song <yhs@fb.com>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Link: https://lore.kernel.org/bpf/20191219020442.1922617-1-ast@kernel.org
Björn Töpel says:
====================
This series contain one non-critical fix, support for far jumps, and
some optimizations for the BPF JIT.
Previously, the JIT only supported 12b branch targets for conditional
branches, and 21b for unconditional branches. Starting with this
series, 32b branching is supported.
As part of supporting far jumps, branch relaxation was introduced. The
idea is to start with a pessimistic jump (e.g. auipc/jalr) and for
each pass the JIT will have an opportunity to pick a better
instruction (e.g. jal) and shrink the image. Instead of two passes,
the JIT requires more passes. It typically converges after 3 passes.
The optimizations mentioned in the subject are for calls and tail
calls. In the tail call generation we can save one instruction by
using the offset in jalr. Calls are optimized by doing (auipc)/jal(r)
relative jumps instead of loading the entire absolute address and
doing jalr. This required that the JIT image allocator was made RISC-V
specific, so we can ensure that the JIT image and the kernel text are
in range (32b).
The last two patches of the series is not critical to the series, but
are two UAPI build issues for BPF events. A closer look from the
RV-folks would be much appreciated.
The test_bpf.ko module, selftests/bpf/test_verifier and
selftests/seccomp/seccomp_bpf pass all tests.
RISC-V is still missing proper kprobe and tracepoint support, so a lot
of BPF selftests cannot be run.
v1->v2: [1]
* Removed unused function parameter from emit_branch()
* Added patch to support far branch in tail call emit
[1] https://lore.kernel.org/bpf/20191209173136.29615-1-bjorn.topel@gmail.com/
====================
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Add missing uapi header the BPF_PROG_TYPE_PERF_EVENT programs by
exporting struct user_regs_struct instead of struct pt_regs which is
in-kernel only.
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20191216091343.23260-9-bjorn.topel@gmail.com
Instead of using emit_imm() and emit_jalr() which can expand to six
instructions, start using jal or auipc+jalr.
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20191216091343.23260-8-bjorn.topel@gmail.com
This commit makes sure that the JIT images is kept close to the kernel
text, so BPF calls can use relative calling with auipc/jalr or jal
instead of loading the full 64-bit address and jalr.
The BPF JIT image region is 128 MB before the kernel text.
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20191216091343.23260-7-bjorn.topel@gmail.com
This commit add support for far (offset > 21b) jumps and exits.
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Luke Nelson <lukenels@cs.washington.edu>
Link: https://lore.kernel.org/bpf/20191216091343.23260-5-bjorn.topel@gmail.com
This commit adds branch relaxation to the BPF JIT, and with that
support for far (offset greater than 12b) branching.
The branch relaxation requires more than two passes to converge. For
most programs it is three passes, but for larger programs it can be
more.
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Luke Nelson <lukenels@cs.washington.edu>
Link: https://lore.kernel.org/bpf/20191216091343.23260-3-bjorn.topel@gmail.com
The BPF JIT incorrectly clobbered the a0 register, and did not flag
usage of s5 register when BPF stack was being used.
Fixes: 2353ecc6f9 ("bpf, riscv: add BPF JIT for RV64G")
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20191216091343.23260-2-bjorn.topel@gmail.com
The erratum A-009204 workaround patch was reverted because of
incorrect implementation.
8b6dc6b mmc: sdhci-of-esdhc: Revert "mmc: sdhci-of-esdhc: add
erratum A-009204 support"
This patch is to re-implement the workaround (add a 5 ms delay
before setting SYSCTL[RSTD] to make sure all the DMA transfers
are finished).
Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
Link: https://lore.kernel.org/r/20191219032335.26528-1-yangbo.lu@nxp.com
Fixes: 5dd1955225 ("mmc: sdhci-of-esdhc: add erratum A-009204 support")
Cc: stable@vger.kernel.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Mark the msm8998 cpu CX gdsc as votable and use the hw control to avoid
corner cases with SMMU per hardware documentation.
Fixes: 3f7df5baa2 ("clk: qcom: Add MSM8998 GPU Clock Controller (GPUCC) driver")
Signed-off-by: Jeffrey Hugo <jeffrey.l.hugo@gmail.com>
Link: https://lkml.kernel.org/r/20191217171905.5619-1-jeffrey.l.hugo@gmail.com
Signed-off-by: Stephen Boyd <sboyd@kernel.org>
Andrii Nakryiko says:
====================
Based on latest feedback and discussions, this patch set implements the
following changes:
- Kconfig-provided externs have to be in .kconfig section, for which
bpf_helpers.h provides convenient __kconfig macro (Daniel);
- instead of allowing to override Kconfig file path, switch this to ability to
extend and override system Kconfig with user-provided custom values (Alexei);
- BTF is required when externs are used.
====================
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Instead of all or nothing approach of overriding Kconfig file location, allow
to extend it with extra values and override chosen subset of values though
optional user-provided extra config, passed as a string through open options'
.kconfig option. If same config key is present in both user-supplied config
and Kconfig, user-supplied one wins. This allows applications to more easily
test various conditions despite host kernel's real configuration. If all of
BPF object's __kconfig externs are satisfied from user-supplied config, system
Kconfig won't be read at all.
Simplify selftests by not needing to create temporary Kconfig files.
Suggested-by: Alexei Starovoitov <ast@fb.com>
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20191219002837.3074619-3-andriin@fb.com
There are cases in which BPF resource (program, map, etc) has to outlive
userspace program that "installed" it in the system in the first place.
When BPF program is attached, libbpf returns bpf_link object, which
is supposed to be destroyed after no longer necessary through
bpf_link__destroy() API. Currently, bpf_link destruction causes both automatic
detachment and frees up any resources allocated to for bpf_link in-memory
representation. This is inconvenient for the case described above because of
coupling of detachment and resource freeing.
This patch introduces bpf_link__disconnect() API call, which marks bpf_link as
disconnected from its underlying BPF resouces. This means that when bpf_link
is destroyed later, all its memory resources will be freed, but BPF resource
itself won't be detached.
This design allows to follow strict and resource-leak-free design by default,
while giving easy and straightforward way for user code to opt for keeping BPF
resource attached beyond lifetime of a bpf_link. For some BPF programs (i.e.,
FS-based tracepoints, kprobes, raw tracepoint, etc), user has to make sure to
pin BPF program to prevent kernel to automatically detach it on process exit.
This should typically be achived by pinning BPF program (or map in some cases)
in BPF FS.
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Link: https://lore.kernel.org/bpf/20191218225039.2668205-1-andriin@fb.com
-----BEGIN PGP SIGNATURE-----
iJYEABYIAD4WIQRE6pSOnaBC00OEHEIaerohdGur0gUCXfrLASAcamFya2tvLnNh
a2tpbmVuQGxpbnV4LmludGVsLmNvbQAKCRAaerohdGur0pZfAQD9F5Vjdqp3fWk+
pxt+eD9+xaD2MYuSVO2AEVBC949vdQD/TP7xnb66w7n9YtMtm9MgvysHAakJYeAe
l4XsHAiPHgI=
=CFIs
-----END PGP SIGNATURE-----
Merge tag 'tpmdd-next-20191219' of git://git.infradead.org/users/jjs/linux-tpmdd
Pull tpm fixes from Jarkko Sakkinen:
"Bunch of fixes for rc3"
* tag 'tpmdd-next-20191219' of git://git.infradead.org/users/jjs/linux-tpmdd:
tpm/tpm_ftpm_tee: add shutdown call back
tpm: selftest: cleanup after unseal with wrong auth/policy test
tpm: selftest: add test covering async mode
tpm: fix invalid locking in NONBLOCKING mode
security: keys: trusted: fix lost handle flush
tpm_tis: reserve chip for duration of tpm_tis_core_init
KEYS: asymmetric: return ENOMEM if akcipher_request_alloc() fails
KEYS: remove CONFIG_KEYS_COMPAT
allow to pass skb's mark field into bpf_prog_test_run ctx
for BPF_PROG_TYPE_SCHED_CLS prog type. that would allow
to test bpf programs which are doing decision based on this
field
Signed-off-by: Nikita V. Shirokov <tehnerd@tehnerd.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Work-around what appears to be a bug in rst2man convertion tool, used to
create man pages out of reStructureText-formatted documents. If text line
starts with dot, rst2man will put it in resulting man file verbatim. This
seems to cause man tool to interpret it as a directive/command (e.g., `.bs`), and
subsequently not render entire line because it's unrecognized one.
Enclose '.xxx' words in extra formatting to work around.
Fixes: cb21ac5885 ("bpftool: Add gen subcommand manpage")
Reported-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Yonghong Song <yhs@fb.com
Link: https://lore.kernel.org/bpf/20191218221707.2552199-1-andriin@fb.com
Change format string referring to just single argument out of two available.
Some versions of libc can reject such format string.
Reported-by: Nikita Shirokov <tehnerd@tehnerd.com>
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/20191218214314.2403729-1-andriin@fb.com
Add shutdown call back to close existing session with fTPM TA
to support kexec scenario.
Add parentheses to function names in comments as specified in kdoc.
Signed-off-by: Thirupathaiah Annapureddy <thiruan@microsoft.com>
Signed-off-by: Pavel Tatashin <pasha.tatashin@soleen.com>
Reviewed-by: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com>
Tested-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com>
The driver forgets to call component_del in remove to match component_add
in probe.
Add the missed call to fix it.
Signed-off-by: Chuhong Yuan <hslester96@gmail.com>
Signed-off-by: Inki Dae <inki.dae@samsung.net>
For ARCHs that don't support 64 bits division we need to use the
helpers.
Fixes: b60189e039 ("net: stmmac: Integrate EST with TAPRIO scheduler API")
Signed-off-by: Jose Abreu <Jose.Abreu@synopsys.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
A typical backtrace acquired from ftraced function currently looks like
the following (e.g. for "path_openat"):
arch_stack_walk+0x15c/0x2d8
stack_trace_save+0x50/0x68
stack_trace_call+0x15a/0x3b8
ftrace_graph_caller+0x0/0x1c
0x3e0007e3c98 <- ftraced function caller (should be do_filp_open+0x7c/0xe8)
do_open_execat+0x70/0x1b8
__do_execve_file.isra.0+0x7d8/0x860
__s390x_sys_execve+0x56/0x68
system_call+0xdc/0x2d8
Note random "0x3e0007e3c98" stack value as ftraced function caller. This
value causes either imprecise unwinder result or unwinding failure.
That "0x3e0007e3c98" comes from r14 of ftraced function stack frame, which
it haven't had a chance to initialize since the very first instruction
calls ftrace code ("ftrace_caller"). (ftraced function might never
save r14 as well). Nevertheless according to s390 ABI any function
is called with stack frame allocated for it and r14 contains return
address. "ftrace_caller" itself is called with "brasl %r0,ftrace_caller".
So, to fix this issue simply always save traced function caller onto
ftraced function stack frame.
Reported-by: Sven Schnelle <svens@linux.ibm.com>
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
Consider reaching user mode pt_regs at the bottom of irq stack graceful
unwinder termination. This is the case when irq/mcck/ext interrupt arrives
while in user mode.
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
the purgatory must not rely on functions from the "old" kernel,
so we must disable kasan and friends. We also need to have a
separate copy of string.c as the default does not build memcmp
with KASAN.
Reported-by: kbuild test robot <lkp@intel.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Vasily Gorbik <gor@linux.ibm.com>
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
Since we link purgatory with -r aka we enable "incremental linking"
no checks for unresolved symbols are done while linking the purgatory.
This commit adds an extra check for unresolved symbols by calling ld
without -r before running objcopy to generate purgatory.ro.
This will help us catch missing symbols in the purgatory sooner.
Note this commit also removes --no-undefined from LDFLAGS_purgatory
as that has no effect.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Link: https://lore.kernel.org/lkml/20191212205304.191610-1-hdegoede@redhat.com
Tested-by: Philipp Rudo <prudo@linux.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
The following sequence triggers a kernel stack overflow on s390x:
mount -t tracefs tracefs /sys/kernel/tracing
cd /sys/kernel/tracing
echo function_graph > current_tracer
[crash]
This is because preempt_count_{add,sub} are in the list of traced
functions, which can be demonstrated by:
echo preempt_count_add >set_ftrace_filter
echo function_graph > current_tracer
[crash]
The stack overflow happens because get_tod_clock_monotonic() gets called
by ftrace but itself calls preempt_{disable,enable}(), which leads to a
endless recursion. Fix this by using preempt_{disable,enable}_notrace().
Fixes: 011620688a ("s390/time: ensure get_clock_monotonic() returns monotonic values")
Signed-off-by: Sven Schnelle <svens@linux.ibm.com>
Reviewed-by: Vasily Gorbik <gor@linux.ibm.com>
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
Petr Machata says:
====================
Add a new Qdisc, ETS
The IEEE standard 802.1Qaz (and 802.1Q-2014) specifies four principal
transmission selection algorithms: strict priority, credit-based shaper,
ETS (bandwidth sharing), and vendor-specific. All these have their
corresponding knobs in DCB. But DCB does not have interfaces to configure
RED and ECN, unlike Qdiscs.
In the Qdisc land, strict priority is implemented by PRIO. Credit-based
transmission selection algorithm can then be modeled by having e.g. TBF or
CBS Qdisc below some of the PRIO bands. ETS would then be modeled by
placing a DRR Qdisc under the last PRIO band.
The problem with this approach is that DRR on its own, as well as the
combination of PRIO and DRR, are tricky to configure and tricky to offload
to 802.1Qaz-compliant hardware. This is due to several reasons:
- As any classful Qdisc, DRR supports adding classifiers to decide in which
class to enqueue packets. Unlike PRIO, there's however no fallback in the
form of priomap. A way to achieve classification based on packet priority
is e.g. like this:
# tc filter add dev swp1 root handle 1: \
basic match 'meta(priority eq 0)' flowid 1:10
Expressing the priomap in this manner however forces drivers to deep dive
into the classifier block to parse the individual rules.
A possible solution would be to extend the classes with a "defmap" a la
split / defmap mechanism of CBQ, and introduce this as a last resort
classification. However, unlike priomap, this doesn't have the guarantee
of covering all priorities. Traffic whose priority is not covered is
dropped by DRR as unclassified. But ASICs tend to implement dropping in
the ACL block, not in scheduling pipelines. The need to treat these
configurations correctly (if only to decide to not offload at all)
complicates a driver.
It's not clear how to retrofit priomap with all its benefits to DRR
without changing it beyond recognition.
- The interplay between PRIO and DRR is also causing problems. 802.1Qaz has
all ETS TCs as a last resort. Switch ASICs that support ETS at all are
likely to handle ETS traffic this way as well. However, the Linux model
is more generic, allowing the DRR block in any band. Drivers would need
to be careful to handle this case correctly, otherwise the offloaded
model might not match the slow-path one.
In a similar vein, PRIO and DRR need to agree on the list of priorities
assigned to DRR. This is doubly problematic--the user needs to take care
to keep the two in sync, and the driver needs to watch for any holes in
DRR coverage and treat the traffic correctly, as discussed above.
Note that at the time that DRR Qdisc is added, it has no classes, and
thus any priorities assigned to that PRIO band are not covered. Thus this
case is surprisingly rather common, and needs to be handled gracefully by
the driver.
- Similarly due to DRR flexibility, when a Qdisc (such as RED) is attached
below it, it is not immediately clear which TC the class represents. This
is unlike PRIO with its straightforward classid scheme. When DRR is
combined with PRIO, the relationship between classes and TCs gets even
more murky.
This is a problem for users as well: the TC mapping is rather important
for (devlink) shared buffer configuration and (ethtool) counters.
So instead, this patch set introduces a new Qdisc, which is based on
802.1Qaz wording. It is PRIO-like in how it is configured, meaning one
needs to specify how many bands there are, how many are strict and how many
are ETS, quanta for the latter, and priomap.
The new Qdisc operates like the PRIO / DRR combo would when configured as
per the standard. The strict classes, if any, are tried for traffic first.
When there's no traffic in any of the strict queues, the ETS ones (if any)
are treated in the same way as in DRR.
The chosen interface makes the overall system both reasonably easy to
configure, and reasonably easy to offload. The extra code to support ETS in
mlxsw (which already supports PRIO) is about 150 lines, of which perhaps 20
lines is bona fide new business logic.
Credit-based shaping transmission selection algorithm can be configured by
adding a CBS Qdisc under one of the strict bands (e.g. TBF can be used to a
similar effect as well). As a non-work-conserving Qdisc, CBS can't be
hooked under the ETS bands. This is detected and handled identically to DRR
Qdisc at runtime. Note that offloading CBS is not subject of this patchset.
The patchset proceeds in four stages:
- Patches #1-#3 are cleanups.
- Patches #4 and #5 contain the new Qdisc.
- Patches #6 and #7 update mlxsw to offload the new Qdisc.
- Patches #8-#10 add selftests for ETS.
Examples:
- Add a Qdisc with 6 bands, 3 strict and 3 ETS with 45%-30%-25% weights:
# tc qdisc add dev swp1 root handle 1: \
ets strict 3 quanta 4500 3000 2500 priomap 0 1 1 1 2 3 4 5
# tc qdisc sh dev swp1
qdisc ets 1: root refcnt 2 bands 6 strict 3 quanta 4500 3000 2500 priomap 0 1 1 1 2 3 4 5 5 5 5 5 5 5 5 5
- Tweak quantum of one of the classes of the previous Qdisc:
# tc class ch dev swp1 classid 1:4 ets quantum 1000
# tc qdisc sh dev swp1
qdisc ets 1: root refcnt 2 bands 6 strict 3 quanta 1000 3000 2500 priomap 0 1 1 1 2 3 4 5 5 5 5 5 5 5 5 5
# tc class ch dev swp1 classid 1:3 ets quantum 1000
Error: Strict bands do not have a configurable quantum.
- Purely strict Qdisc with 1:1 mapping between priorities and TCs:
# tc qdisc add dev swp1 root handle 1: \
ets strict 8 priomap 7 6 5 4 3 2 1 0
# tc qdisc sh dev swp1
qdisc ets 1: root refcnt 2 bands 8 strict 8 priomap 7 6 5 4 3 2 1 0 7 7 7 7 7 7 7 7
- Use "bands" to specify number of bands explicitly. Underspecified bands
are implicitly ETS and their quantum is taken from MTU. The following
thus gives each band the same weight:
# tc qdisc add dev swp1 root handle 1: \
ets bands 8 priomap 7 6 5 4 3 2 1 0
# tc qdisc sh dev swp1
qdisc ets 1: root refcnt 2 bands 8 quanta 1514 1514 1514 1514 1514 1514 1514 1514 priomap 7 6 5 4 3 2 1 0 7 7 7 7 7 7 7 7
v2:
- This addresses points raised by David Miller.
- Patch #4:
- sch_ets.c: Add a comment with description of the Qdisc and the
dequeuing algorithm.
- Kconfig: Add a high-level description to the help blurb.
v1:
- No changes, first upstream submission after RFC.
v3 (internal):
- This addresses review from Jiri Pirko.
- Patch #3:
- Rename to _HR_ instead of to _HIERARCHY_.
- Patch #4:
- pkt_sched.h: Keep all the TCA_ETS_ constants in one enum.
- pkt_sched.h: Rename TCA_ETS_BANDS to _NBANDS, _STRICT to _NSTRICT,
_BAND_QUANTUM to _QUANTA_BAND and _PMAP_BAND to _PRIOMAP_BAND.
- sch_ets.c: Update to reflect the above changes. Add a new policy,
ets_class_policy, which is used when parsing class changes.
Currently that policy is the same as the quanta policy, but that
might change.
- sch_ets.c: Move MTU handling from ets_quantum_parse() to the one
caller that makes use of it.
- sch_ets.c: ets_qdisc_priomap_parse(): WARN_ON_ONCE on invalid
attribute instead of returning an extack.
- Patch #6:
- __mlxsw_sp_qdisc_ets_replace(): Pass the weights argument to this
function in this patch already. Drop the weight computation.
- mlxsw_sp_qdisc_prio_replace(): Rename "quanta" to "zeroes" and
pass for the abovementioned "weights".
- mlxsw_sp_qdisc_prio_graft(): Convert to a wrapper around
__mlxsw_sp_qdisc_ets_graft(), instead of invoking the latter
directly from mlxsw_sp_setup_tc_prio().
- Update to follow the _HIERARCHY_ -> _HR_ renaming.
- Patch #7:
- __mlxsw_sp_qdisc_ets_replace(): The "weights" argument passing and
weight computation removal are now done in a previous patch.
- mlxsw_sp_setup_tc_ets(): Drop case TC_ETS_REPLACE, which is handled
earlier in the function.
- Patch #3 (iproute2):
- Add an example output to the commit message.
- tc-ets.8: Fix output of two examples.
- tc-ets.8: Describe default values of "bands", "quanta".
- q_ets.c: A number of fixes in error messages.
- q_ets.c: Comment formatting: /*padding*/ -> /* padding */
- q_ets.c: parse_nbands: Move duplicate checking to callers.
- q_ets.c: Don't accept both "quantum" and "quanta" as equivalent.
v2 (internal):
- This addresses review from Ido Schimmel and comments from Alexander
Kushnarov.
- Patch #2:
- s/coment/comment in the commit message.
- Patch #4:
- sch_ets: ets_class_is_strict(), ets_class_id(): Constify an argument
- ets_class_find(): RXTify
- Patch #3 (iproute2):
- tc-ets.8: some spelling fixes
- tc-ets.8: add another example
- tc.8: add an ETS to "CLASSFUL QDISCS" section
v1 (internal):
- This addresses RFC reviews from Ido Schimmel and Roman Mashak, bugs found
by Alexander Petrovskiy and myself, and other improvements.
- Patch #2:
- Expand the explanation with an explicit example.
- Patch #4:
- Kconfig: s/sch_drr/sch_ets/
- sch_ets: Reorder includes to be in alphabetical order
- sch_ets: ets_quantum_parse(): Rename the return-pointer argument
from pquantum to quantum, and use it directly, not going through a
local temporary.
- sch_ets: ets_qdisc_quanta_parse(): Convert syntax of function
argument "quanta" from an array to a pointer.
- sch_ets: ets_qdisc_priomap_parse(): Likewise with "priomap".
- sch_ets: ets_qdisc_quanta_parse(), ets_qdisc_priomap_parse(): Invoke
__nla_validate_nested directly instead of nl80211_validate_nested().
- sch_ets: ets_qdisc_quanta_parse(): WARN_ON_ONCE on invalid attribute
instead of returning an extack.
- sch_ets: ets_qdisc_change(): Make the last band the default one for
unmentioned priomap priorities.
- sch_ets: Fix a panic when an offloaded child in a bandwidth-sharing
band notified its ETS parent.
- sch_ets: When ungrafting, add the newly-created invisible FIFO to
the Qdisc hash
- Patch #5:
- pkt_cls.h: Note that quantum=0 signifies a strict band.
- Fix error path handling when ets_offload_dump() fails.
- Patch #6:
- __mlxsw_sp_qdisc_ets_replace(): Convert syntax of function arguments
"quanta" and "priomap" from arrays to pointers.
- Patch #7:
- __mlxsw_sp_qdisc_ets_replace(): Convert syntax of function argument
"weights" from an array to a pointer.
- Patch #9:
- mlxsw/sch_ets.sh: Add a comment explaining packet prioritization.
- Adjust the whole suite to allow testing of traffic classifiers
in addition to testing priomap.
- Patch #10:
- Add a number of new tests to test default priomap band, overlarge
number of bands, zeroes in quanta, and altogether missing quanta.
- Patch #1 (iproute2):
- State motivation for inclusion of this patch in the patcheset in the
commit message.
- Patch #3 (iproute2):
- tc-ets.8: it is now December
- tc-ets.8: explain inactivity WRT using non-WC Qdiscs under ETS band
- tc-ets.8: s/flow/band in explanation of quantum
- tc-ets.8: explain what happens with priorities not covered by priomap
- tc-ets.8: default priomap band is now the last one
- q_ets.c: ets_parse_opt(): Remove unnecessary initialization of
priomap and quanta.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Add TDC coverage for the new ETS Qdisc.
Signed-off-by: Petr Machata <petrm@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This tests the newly-added ETS Qdisc. It runs two to three streams of
traffic, each with a different priority. ETS Qdisc is supposed to allocate
bandwidth according to the DRR algorithm and given weights. After running
the traffic for a while, counters are compared for each stream to check
that the expected ratio is in fact observed.
In order for the DRR process to kick in, a traffic bottleneck must exist in
the first place. In slow path, such bottleneck can be implemented by
wrapping the ETS Qdisc inside a TBF or other shaper. This might however
make the configuration unoffloadable. Instead, on HW datapath, the
bottleneck would be set up by lowering port speed and configuring shared
buffer suitably.
Therefore the test is structured as a core component that implements the
testing, with two wrapper scripts that implement the details of slow path
resp. fast path configuration.
Signed-off-by: Petr Machata <petrm@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
These two functions are used for starting several streams of traffic, and
then stopping them later. They will be handy for the test coverage of ETS
Qdisc. Move them from mlxsw-specific qos_lib.sh to the generic lib.sh.
Signed-off-by: Petr Machata <petrm@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Handle TC_SETUP_QDISC_ETS, add a new ops structure for the ETS Qdisc.
Invoke the extended prio handlers implemented in the previous patch. For
stats ops, invoke directly the prio callbacks, which are not sensitive to
differences between PRIO and ETS.
Signed-off-by: Petr Machata <petrm@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Thanks to the similarity between PRIO and ETS it is possible to simply
reuse most of the code for offloading PRIO Qdisc. Extract the common
functionality into separate functions, making the current PRIO handlers
thin API adapters.
Extend the new functions to pass quanta for individual bands, which allows
configuring a subset of bands as WRR. Invoke mlxsw_sp_port_ets_set() as
appropriate to de/configure WRR-ness and weight of individual bands.
Signed-off-by: Petr Machata <petrm@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add hooks at appropriate points to make it possible to offload the ETS
Qdisc.
Signed-off-by: Petr Machata <petrm@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Introduces a new Qdisc, which is based on 802.1Q-2014 wording. It is
PRIO-like in how it is configured, meaning one needs to specify how many
bands there are, how many are strict and how many are dwrr, quanta for the
latter, and priomap.
The new Qdisc operates like the PRIO / DRR combo would when configured as
per the standard. The strict classes, if any, are tried for traffic first.
When there's no traffic in any of the strict queues, the ETS ones (if any)
are treated in the same way as in DRR.
Signed-off-by: Petr Machata <petrm@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
These enums want to be named MLXSW_REG_QEEC_HIERARCHY_, but due to a typo
lack the second H. That is confusing and complicates searching.
But actually the enumerators should be named _HR_, because that is how
their enum type is called. So rename them as appropriate.
Signed-off-by: Petr Machata <petrm@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
