Add parsing support by implementing struct mlx5e_tc_act
for this action.
Signed-off-by: Roi Dayan <roid@nvidia.com>
Reviewed-by: Oz Shlomo <ozsh@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Add parsing support by implementing struct mlx5e_tc_act
for this action.
Signed-off-by: Roi Dayan <roid@nvidia.com>
Reviewed-by: Oz Shlomo <ozsh@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Add parsing support by implementing struct mlx5e_tc_act
for this action.
Signed-off-by: Roi Dayan <roid@nvidia.com>
Reviewed-by: Oz Shlomo <ozsh@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Add parsing support by implementing struct mlx5e_tc_act
for this action.
Signed-off-by: Roi Dayan <roid@nvidia.com>
Reviewed-by: Oz Shlomo <ozsh@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Add parsing support by implementing struct mlx5e_tc_act
for this action.
Signed-off-by: Roi Dayan <roid@nvidia.com>
Reviewed-by: Oz Shlomo <ozsh@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Add an infrastructure to help parsing tc actions in a generic way.
Supporting an action parser means implementing struct mlx5e_tc_act
for that action.
The infrastructure will give the possibility to be generic when parsing tc
actions, i.e. parse_tc_nic_actions() and parse_tc_fdb_actions().
To parse tc actions a user needs to allocate a parse_state instance
and pass it when iterating over the tc actions parsers.
If a parser doesn't exists then a user can treat it as unsupported.
To add an action parser a user needs to implement two callbacks.
The can_offload() callback to quickly check if an action can be offloaded.
The parse_action() callback to do actual parsing and prepare for offload.
Add implementation for drop, trap, mark and accept action parsers with this
commit to act as examples and implement usage of the new infrastructure for
those actions.
Signed-off-by: Roi Dayan <roid@nvidia.com>
Reviewed-by: Oz Shlomo <ozsh@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Add support for QCM2290 including a few prep changes.
* icc-qcm2290
interconnect: icc-rpm: Define ICC device type
interconnect: icc-rpm: Add QNOC type QoS support
interconnect: icc-rpm: Support child NoC device probe
dt-bindings: interconnect: Add Qualcomm QCM2290 NoC support
interconnect: qcom: Add QCM2290 driver support
Link: https://lore.kernel.org/r/20211215002324.1727-1-shawn.guo@linaro.org
Signed-off-by: Georgi Djakov <djakov@kernel.org>
It adds interconnect driver support for QCM2290 platform. The topology
consists of 3 NoCs: BIMC, Config NoC (CNOC) and System NoC (SNOC). SNOC
is a QCOM_ICC_QNOC type device, as well as its 3 virtual child devices,
QUP, MMNRT and MMRT. QUP is owned by RPM and thus has no .regmap_cfg,
while the other 2 share the same .regmap_cfg with SNOC (parent).
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
Link: https://lore.kernel.org/r/20211215002324.1727-6-shawn.guo@linaro.org
Signed-off-by: Georgi Djakov <djakov@kernel.org>
As shown in downstream DT[1], the System NoC of QCM2290 is modelled
using 4 fab/noc devices: sys_noc + qup_virt + mmnrt_virt + mmrt_virt.
Among those 3 virtual devices, qup is owned by RPM and has no regmap
resource, while mmnrt and mmrt are owned by AP and share the same
regmap as sys_noc. So it's logical to represent these virtual devices
as child nodes of sys_noc in DT, so that such configuration can be
supported with a couple of changes on qnoc_probe():
- If there are child nodes, populate them.
- If the device descriptor has .regmap_cfg but there is no IOMEM
resource for the device, use parent's regmap.
[1] https://android.googlesource.com/kernel/msm-extra/devicetree/+/refs/tags/android-11.0.0_r0.56/qcom/scuba-bus.dtsi
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
Link: https://lore.kernel.org/r/20211215002324.1727-4-shawn.guo@linaro.org
Signed-off-by: Georgi Djakov <djakov@kernel.org>
Fix up unprivileged test case results for 'Dest pointer in r0' verifier tests
given they now need to reject R0 containing a pointer value, and add a couple
of new related ones with 32bit cmpxchg as well.
root@foo:~/bpf/tools/testing/selftests/bpf# ./test_verifier
#0/u invalid and of negative number OK
#0/p invalid and of negative number OK
[...]
#1268/p XDP pkt read, pkt_meta' <= pkt_data, bad access 1 OK
#1269/p XDP pkt read, pkt_meta' <= pkt_data, bad access 2 OK
#1270/p XDP pkt read, pkt_data <= pkt_meta', good access OK
#1271/p XDP pkt read, pkt_data <= pkt_meta', bad access 1 OK
#1272/p XDP pkt read, pkt_data <= pkt_meta', bad access 2 OK
Summary: 1900 PASSED, 0 SKIPPED, 0 FAILED
Acked-by: Brendan Jackman <jackmanb@google.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
The implementation of BPF_CMPXCHG on a high level has the following parameters:
.-[old-val] .-[new-val]
BPF_R0 = cmpxchg{32,64}(DST_REG + insn->off, BPF_R0, SRC_REG)
`-[mem-loc] `-[old-val]
Given a BPF insn can only have two registers (dst, src), the R0 is fixed and
used as an auxilliary register for input (old value) as well as output (returning
old value from memory location). While the verifier performs a number of safety
checks, it misses to reject unprivileged programs where R0 contains a pointer as
old value.
Through brute-forcing it takes about ~16sec on my machine to leak a kernel pointer
with BPF_CMPXCHG. The PoC is basically probing for kernel addresses by storing the
guessed address into the map slot as a scalar, and using the map value pointer as
R0 while SRC_REG has a canary value to detect a matching address.
Fix it by checking R0 for pointers, and reject if that's the case for unprivileged
programs.
Fixes: 5ffa25502b ("bpf: Add instructions for atomic_[cmp]xchg")
Reported-by: Ryota Shiga (Flatt Security)
Acked-by: Brendan Jackman <jackmanb@google.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Test whether unprivileged would be able to leak the spilled pointer either
by exporting the returned value from the atomic{32,64} operation or by reading
and exporting the value from the stack after the atomic operation took place.
Note that for unprivileged, the below atomic cmpxchg test case named "Dest
pointer in r0 - succeed" is failing. The reason is that in the dst memory
location (r10 -8) there is the spilled register r10:
0: R1=ctx(id=0,off=0,imm=0) R10=fp0
0: (bf) r0 = r10
1: R0_w=fp0 R1=ctx(id=0,off=0,imm=0) R10=fp0
1: (7b) *(u64 *)(r10 -8) = r0
2: R0_w=fp0 R1=ctx(id=0,off=0,imm=0) R10=fp0 fp-8_w=fp
2: (b7) r1 = 0
3: R0_w=fp0 R1_w=invP0 R10=fp0 fp-8_w=fp
3: (db) r0 = atomic64_cmpxchg((u64 *)(r10 -8), r0, r1)
4: R0_w=fp0 R1_w=invP0 R10=fp0 fp-8_w=mmmmmmmm
4: (79) r1 = *(u64 *)(r0 -8)
5: R0_w=fp0 R1_w=invP(id=0) R10=fp0 fp-8_w=mmmmmmmm
5: (b7) r0 = 0
6: R0_w=invP0 R1_w=invP(id=0) R10=fp0 fp-8_w=mmmmmmmm
6: (95) exit
However, allowing this case for unprivileged is a bit useless given an
update with a new pointer will fail anyway:
0: R1=ctx(id=0,off=0,imm=0) R10=fp0
0: (bf) r0 = r10
1: R0_w=fp0 R1=ctx(id=0,off=0,imm=0) R10=fp0
1: (7b) *(u64 *)(r10 -8) = r0
2: R0_w=fp0 R1=ctx(id=0,off=0,imm=0) R10=fp0 fp-8_w=fp
2: (db) r0 = atomic64_cmpxchg((u64 *)(r10 -8), r0, r10)
R10 leaks addr into mem
Acked-by: Brendan Jackman <jackmanb@google.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
The change in commit 37086bfdc7 ("bpf: Propagate stack bounds to registers
in atomics w/ BPF_FETCH") around check_mem_access() handling is buggy since
this would allow for unprivileged users to leak kernel pointers. For example,
an atomic fetch/and with -1 on a stack destination which holds a spilled
pointer will migrate the spilled register type into a scalar, which can then
be exported out of the program (since scalar != pointer) by dumping it into
a map value.
The original implementation of XADD was preventing this situation by using
a double call to check_mem_access() one with BPF_READ and a subsequent one
with BPF_WRITE, in both cases passing -1 as a placeholder value instead of
register as per XADD semantics since it didn't contain a value fetch. The
BPF_READ also included a check in check_stack_read_fixed_off() which rejects
the program if the stack slot is of __is_pointer_value() if dst_regno < 0.
The latter is to distinguish whether we're dealing with a regular stack spill/
fill or some arithmetical operation which is disallowed on non-scalars, see
also 6e7e63cbb0 ("bpf: Forbid XADD on spilled pointers for unprivileged
users") for more context on check_mem_access() and its handling of placeholder
value -1.
One minimally intrusive option to fix the leak is for the BPF_FETCH case to
initially check the BPF_READ case via check_mem_access() with -1 as register,
followed by the actual load case with non-negative load_reg to propagate
stack bounds to registers.
Fixes: 37086bfdc7 ("bpf: Propagate stack bounds to registers in atomics w/ BPF_FETCH")
Reported-by: <n4ke4mry@gmail.com>
Acked-by: Brendan Jackman <jackmanb@google.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Commit 0259d4498b ("bcache: move calc_cached_dev_sectors to proper
place on backing device detach") tries to fix calc_cached_dev_sectors
when bcache device detaches, but now we have:
cached_dev_detach_finish
...
bcache_device_detach(&dc->disk);
...
closure_put(&d->c->caching);
d->c = NULL; [*explicitly set dc->disk.c to NULL*]
list_move(&dc->list, &uncached_devices);
calc_cached_dev_sectors(dc->disk.c); [*passing a NULL pointer*]
...
Upper codeflows shows how bug happens, this patch fix the problem by
caching dc->disk.c beforehand, and cache_set won't be freed under us
because c->caching closure at least holds a reference count and closure
callback __cache_set_unregister only being called by bch_cache_set_stop
which using closure_queue(&c->caching), that means c->caching closure
callback for destroying cache_set won't be trigger by previous
closure_put(&d->c->caching).
So at this stage(while cached_dev_detach_finish is calling) it's safe to
access cache_set dc->disk.c.
Fixes: 0259d4498b ("bcache: move calc_cached_dev_sectors to proper place on backing device detach")
Signed-off-by: Lin Feng <linf@wangsu.com>
Signed-off-by: Coly Li <colyli@suse.de>
Link: https://lore.kernel.org/r/20211112053629.3437-2-colyli@suse.de
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The msm-id and board-id can be used to select the correct dtb when
multiple are provided to the bootloader.
Multiple DTBs can be provided on sdm845 devices using boot image header
v1 by appending them all to the kernel image before creating the boot
image. The bootloader then selects them like this:
Best match DTB tags 321/00000008/0x00000000/20001/20014/20115/20018/0/(offset)0x79998E27/(size)0x000173CD
Using pmic info 0x20014/0x20115/0x20018/0x0 for device 0x20014/0x20115/0x20018/0x0
Signed-off-by: Caleb Connolly <caleb.connolly@linaro.org>
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Link: https://lore.kernel.org/r/20211209225938.2427342-1-caleb.connolly@linaro.org
Add device tree bindings for global clock controller on SM8450 SoCs.
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Reviewed-by: Stephen Boyd <sboyd@kernel.org>
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Link: https://lore.kernel.org/r/20211207114003.100693-2-vkoul@kernel.org
Dexuan reports that he's seeing spikes of very heavy CPU utilization when
running 24 disks and using the 'none' scheduler. This happens off the
sched restart path, because SCSI requires the queue to be restarted async,
and hence we're hammering on mod_delayed_work_on() to ensure that the work
item gets run appropriately.
Avoid hammering on the timer and just use queue_work_on() if no delay
has been specified.
Reported-and-tested-by: Dexuan Cui <decui@microsoft.com>
Link: https://lore.kernel.org/linux-block/BYAPR21MB1270C598ED214C0490F47400BF719@BYAPR21MB1270.namprd21.prod.outlook.com/
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Add basic devicetree support for SDX65 platform and MTP board from
Qualcomm. The SDX65 platform features an ARM Cortex A7 CPU which forms
the Application Processor Sub System (APSS) along with standard Qualcomm
peripherals like GCC, TLMM, BLSP, QPIC, and BAM etc... Also, there
exists the networking parts such as IPA, MHI, PCIE-EP, EMAC, and Modem
etc..
This commit adds basic devicetree support that includes GCC, RPMh clock, INTC
and Debug UART.
Signed-off-by: Vamsi Krishna Lanka <quic_vamslank@quicinc.com>
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Link: https://lore.kernel.org/r/1635552125-16407-3-git-send-email-quic_vamslank@quicinc.com
Document the SDX65 platform binding and also the boards using it.
Signed-off-by: Vamsi Krishna Lanka <quic_vamslank@quicinc.com>
Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Acked-by: Rob Herring <robh@kernel.org>
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Link: https://lore.kernel.org/r/1635552125-16407-2-git-send-email-quic_vamslank@quicinc.com
-----BEGIN PGP SIGNATURE-----
iQJPBAABCAA5FiEEBd4DzF816k8JZtUlCx85Pw2ZrcUFAmG5WRMbHGJqb3JuLmFu
ZGVyc3NvbkBsaW5hcm8ub3JnAAoJEAsfOT8Nma3F5FEP/RzaaVFtW+MFJHPiVQrY
x/9JXkiyXzzQakzfuKtVYfch3qg7JI8AT80gPXe6dbjvMSuhtz3ZfHAnUMY0Al5k
5aPiSL1T5/9edZrQN3DjAps8qe8DR8vwz8iMTJzgKrS0LCpPT/YRj7hBCGiueafu
n0PrPs3QWZiLGEtm43hEeUMfWeIPgdFz4S79jeFNpJIy0rj8dMesqiDcUDh2UJcS
5x7T2yMoij14F+6kpvHWIEnxdHiKI4WxH8RP68nQC3+LE4xMpC9tGzndzTXuB6vd
6x+2kwEo0D+F01LzHKAxdL4+pbBn+RRril0qFQ/5HV9F5J8hR/06M3wGB//XVB8j
P7EMfGFCg1AEfL1SLhhqff3k7LKwpZ0VusXRE2MgyIFU5JtzbcxGb8xVOUPcLTnv
Z2ApPeupCP2Bif6nc2VXjP9ZvCiDTSgD5BSeHOhmDgNuaprTVtaZsuL/IJhgJMgT
L13peAxKsaCGiYEA3Z9gM5dkCSjqU6ntnexjftP9KQcLgi8ty6+M9lMns1KCT1vo
/e9yFTEyV5J+YEe/2CoFHiqOOz8MEZ2jcuQvUCPw8rMLnibxw+85vaOj+MJkcA58
5lYk66mZvCeNJBgWxtzNSapsRda7qdKjVaKInwGQv2tQuyFlf37kr4n6raP4xQR+
G1Smu5UOFVd4uUfUqXT5DyZP
=IxPk
-----END PGP SIGNATURE-----
Merge tag 'e15509b2b7c9b600ab38c5269d4fac609c077b5b.1638861860.git.quic_vamslank@quicinc.com' into dts-for-5.17
v5.16-rc1 + e15509b2b7c9b600ab38c5269d4fac609c077b5b.1638861860.git.quic_vamslank@quicinc.com
Merge the immutable branch containing the DT binding and clock
definitions to be used in the SDX65 dts files.
Mat Martineau says:
====================
mptcp: Fixes for ULP, a deadlock, and netlink docs
Two of the MPTCP fixes in this set are related to the TCP_ULP socket
option with MPTCP sockets operating in "fallback" mode (the connection
has reverted to regular TCP). The other issues are an observed deadlock
and missing parameter documentation in the MPTCP netlink API.
Patch 1 marks TCP_ULP as unsupported earlier in MPTCP setsockopt code,
so the fallback code path in the MPTCP layer does not pass the TCP_ULP
option down to the subflow TCP socket.
Patch 2 makes sure a TCP fallback socket returned to userspace by
accept()ing on a MPTCP listening socket does not allow use of the
"mptcp" TCP_ULP type. That ULP is intended only for use by in-kernel
MPTCP subflows.
Patch 3 fixes the possible deadlock when sending data and there are
socket option changes to sync to the subflows.
Patch 4 makes sure all MPTCP netlink event parameters are documented
in the MPTCP uapi header.
====================
Link: https://lore.kernel.org/r/20211214231604.211016-1-mathew.j.martineau@linux.intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
'loc_id' and 'rem_id' are set in all events linked to subflows but those
were missing in the events description in the comments.
Fixes: b911c97c7d ("mptcp: add netlink event support")
Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The mptcp ULP extension relies on sk->sk_sock_kern being set correctly:
It prevents setsockopt(fd, IPPROTO_TCP, TCP_ULP, "mptcp", 6); from
working for plain tcp sockets (any userspace-exposed socket).
But in case of fallback, accept() can return a plain tcp sk.
In such case, sk is still tagged as 'kernel' and setsockopt will work.
This will crash the kernel, The subflow extension has a NULL ctx->conn
mptcp socket:
BUG: KASAN: null-ptr-deref in subflow_data_ready+0x181/0x2b0
Call Trace:
tcp_data_ready+0xf8/0x370
[..]
Fixes: cf7da0d66c ("mptcp: Create SUBFLOW socket for incoming connections")
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
TCP_ULP setsockopt cannot be used for mptcp because its already
used internally to plumb subflow (tcp) sockets to the mptcp layer.
syzbot managed to trigger a crash for mptcp connections that are
in fallback mode:
KASAN: null-ptr-deref in range [0x0000000000000020-0x0000000000000027]
CPU: 1 PID: 1083 Comm: syz-executor.3 Not tainted 5.16.0-rc2-syzkaller #0
RIP: 0010:tls_build_proto net/tls/tls_main.c:776 [inline]
[..]
__tcp_set_ulp net/ipv4/tcp_ulp.c:139 [inline]
tcp_set_ulp+0x428/0x4c0 net/ipv4/tcp_ulp.c:160
do_tcp_setsockopt+0x455/0x37c0 net/ipv4/tcp.c:3391
mptcp_setsockopt+0x1b47/0x2400 net/mptcp/sockopt.c:638
Remove support for TCP_ULP setsockopt.
Fixes: d9e4c12918 ("mptcp: only admit explicitly supported sockopt")
Reported-by: syzbot+1fd9b69cde42967d1add@syzkaller.appspotmail.com
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Kurt Kanzenbach says:
====================
net: dsa: hellcreek: Fix handling of MGMT protocols
this series fixes some minor issues with regards to management protocols
such as PTP and STP in the hellcreek DSA driver. Configure static FDB
for these protocols. The end result is:
|root@tsn:~# mv88e6xxx_dump --atu
|Using device <platform/ff240000.switch>
|ATU:
|FID MAC 0123 Age OBT Pass Static Reprio Prio
| 0 01:1b:19:00:00:00 1100 1 X X 6
| 1 01:00:5e:00:01:81 1100 1 X X 6
| 2 33:33:00:00:01:81 1100 1 X X 6
| 3 01:80:c2:00:00:0e 1100 1 X X X 6
| 4 01:00:5e:00:00:6b 1100 1 X X X 6
| 5 33:33:00:00:00:6b 1100 1 X X X 6
| 6 01:80:c2:00:00:00 1100 1 X X X 6
Previous version:
* https://lore.kernel.org/r/20211213101810.121553-1-kurt@linutronix.de/
Changes since v1:
* Target net-next, as this never worked correctly and is not critical
* Add STP and PTP over UDP rules
* Use pass_blocked for PDelay messages only (Richard Cochran)
====================
Link: https://lore.kernel.org/r/20211214134508.57806-1-kurt@linutronix.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The switch supports PTP for UDP transport too. Therefore, add the missing static
FDB entries to ensure correct forwarding of these packets.
Fixes: ddd56dfe52 ("net: dsa: hellcreek: Add PTP clock support")
Signed-off-by: Kurt Kanzenbach <kurt@linutronix.de>
Acked-by: Richard Cochran <richardcochran@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Allow PTP peer delay measurements on blocked ports by STP. In case of topology
changes the PTP stack can directly start with the correct delays.
Fixes: ddd56dfe52 ("net: dsa: hellcreek: Add PTP clock support")
Signed-off-by: Kurt Kanzenbach <kurt@linutronix.de>
Reviewed-by: Vladimir Oltean <olteanv@gmail.com>
Acked-by: Richard Cochran <richardcochran@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Treat STP as management traffic. STP traffic is designated for the CPU port
only. In addition, STP traffic has to pass blocked ports.
Fixes: e4b27ebc78 ("net: dsa: Add DSA driver for Hirschmann Hellcreek switches")
Signed-off-by: Kurt Kanzenbach <kurt@linutronix.de>
Reviewed-by: Vladimir Oltean <olteanv@gmail.com>
Acked-by: Richard Cochran <richardcochran@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The insertion of static FDB entries ignores the pass_blocked bit. That bit is
evaluated with regards to STP. Add the missing functionality.
Fixes: e4b27ebc78 ("net: dsa: Add DSA driver for Hirschmann Hellcreek switches")
Signed-off-by: Kurt Kanzenbach <kurt@linutronix.de>
Reviewed-by: Vladimir Oltean <olteanv@gmail.com>
Acked-by: Richard Cochran <richardcochran@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Merge series from Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>:
We've been adding a 'deep buffer' PCM device to several SOF topologies
in order to reduce power consumption. The typical use-case would be
music playback over a headset: this additional PCM device provides
more buffering and longer latencies, leaving the rest of the system
sleep for longer periods. Notifications and 'regular' low-latency
audio playback would still use the 'normal' PCM device and be mixed
with the 'deep buffer' before rendering on the headphone endpoint. The
tentative direction would be to expose this alternate device to
PulseAudio/PipeWire/CRAS via the UCM SectionModifier definitions.
That seemed a straightforward topology change until our automated
validation stress tests started reporting issues on SoundWire
platforms, when e.g. two START triggers might be send and conversely
the STOP trigger is never sent. The SoundWire stream state management
flagged inconsistent states when the two 'normal' and 'deep buffer'
devices are used concurrently with rapid play/stop/pause monkey
testing.
Looking at the soc-pcm.c code, it seems that the BE state
management needs a lot of love.
a) there is no consistent protection for the BE state. In some parts
of the code, the state updates are protected by a spinlock but in the
trigger they are not. When we open/play/close the two PCM devices in
stress tests, we end-up testing a state that is being modified. That
can't be good.
b) there is a conceptual deadlock: on stop we check the FE states to
see if a shared BE can be stopped, but since we trigger the BE first
the FE states have not been modified yet, so the TRIGGER_STOP is never
sent.
This patchset suggests the removal of the dedicated 'dpcm_lock' and
follows the design suggested by Takashi Iwai. By default the
protection relies on the 'pcm_mutex', except for the FE and BE
triggers where the mutex cannot be used. In this case, the FE PCM
lock is used instead. In the cases where a BE is added/removed, the
pcm_mutex and FE PCM lock are both taken. In addition, the BE PCM
lock is used to serialize access to a shared BE.
With these patches I am able to run our entire validation suite
without any issues with this new 'deep buffer' topology, and no
regressions on existing solutions [1]. The tests were reproduced by
Bard Liao for SoundWire devices.
One might ask 'how come we didn't see this earlier'? The answer is
probably that the .trigger callbacks in most implementations seems to
perform DAPM operations, and sending the triggers multiple times is
not an issue. In the case of SoundWire, we do use the .trigger
callback to reconfigure the bus using the 'bank switch' mechanism. It
could be acceptable to tolerate a trigger multiple times, but the
deadlock on stop cannot be fixed at the SoundWire level alone.
Opens:
1) The issues reported by Nvidia on the RFCv3 may or may not be
present. We'd need test results to make sure the locking update does
not introduce a regression on Tegra.
2) There are other reports of kernel oopses [2] that seem related to
the lack of protection. I'd be good to confirm if this patchset solve
these problems as well.
[1] https://github.com/thesofproject/linux/pull/3146
[2] https://lore.kernel.org/alsa-devel/002f01d7b4f5$c030f4a0$4092dde0$@samsung.com/
changes since RFCv3:
Used two patches from Takashi. We now use the pcm_mutex, the FE stream
lock when adding and deleting a BE, and the BE stream lock to handle
concurrency between streams using the same BE.
Added a patch to use GFP_ATOMIC for the DPCM structure.
Fixed PAUSE_RELEASE transition (GitHub comment from Kai Vehmanen)
changes since RFCv2:
Removal of dpcm_lock to use FE PCM locks (credits to Takashi Iwai for
the suggestion). The FE PCM lock is now used before each use of
for_each_dpcm_be() - with the exception of the trigger where the lock
is already taken. This change is also applied in drivers which make
use of this loop (compress, SH, FSL).
Addition of BE PCM lock to deal with mutual exclusion between triggers
for the same BE.
Alignment of the BE atomicity on the FE on connections, this is
required to avoid sleeping in atomic context.
Additional cleanups (indentation, static functions)
changes since RFC v1:
Removed unused function
Removed exported symbols only used in soc-pcm.c, used static instead
Use a mutex instead of a spinlock
Protect all for_each_dpcm_be() loops
Fix bugs introduced in the refcount
Pierre-Louis Bossart (4):
ASoC: soc-pcm: use GFP_ATOMIC for dpcm structure
ASoC: soc-pcm: align BE 'atomicity' with that of the FE
ASoC: soc-pcm: test refcount before triggering
ASoC: soc-pcm: fix BE handling of PAUSE_RELEASE
Takashi Iwai (2):
ASoC: soc-pcm: Fix and cleanup DPCM locking
ASoC: soc-pcm: serialize BE triggers
include/sound/soc-dpcm.h | 2 +
include/sound/soc.h | 2 -
sound/soc/soc-core.c | 1 -
sound/soc/soc-pcm.c | 351 +++++++++++++++++++++++++++------------
4 files changed, 246 insertions(+), 110 deletions(-)
--
2.25.1
kyber turns on IO statistics when it is loaded on a queue, which means
that even if kyber is then later unloaded, we're still stuck with stats
enabled on the queue.
Change the account enabled from a bool to an int, and pair the enable call
with the equivalent disable call. This ensures that stats gets turned off
again appropriately.
Reviewed-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
In preparation for FORTIFY_SOURCE performing compile-time and run-time
field bounds checking for memset(), avoid intentionally writing across
neighboring fields.
Use memset_startat() so memset() doesn't get confused about writing beyond
the destination member that is intended to be the starting point of
zeroing through the end of the struct.
Link: https://lore.kernel.org/r/20211213223331.135412-15-keescook@chromium.org
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
In preparation for FORTIFY_SOURCE performing compile-time and run-time
field bounds checking for memset(), avoid intentionally writing across
neighboring fields.
Use memset_startat() so memset() doesn't get confused about writing beyond
the destination member that is intended to be the starting point of
zeroing through the end of the struct. Additionally, since everything
appears to perform a roundup (including allocation), just change the size
of the struct itself and add a build-time check to validate the expected
size.
Link: https://lore.kernel.org/r/20211213223331.135412-13-keescook@chromium.org
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
In preparation for FORTIFY_SOURCE performing compile-time and run-time
field bounds checking for memset(), avoid intentionally writing across
neighboring fields.
Use memset_after() to zero the end of struct mlx5_ib_mr that should be
initialized.
Link: https://lore.kernel.org/r/20211213223331.135412-10-keescook@chromium.org
Acked-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
-----BEGIN PGP SIGNATURE-----
iQFSBAABCAA8FiEEq68RxlopcLEwq+PEeb4+QwBBGIYFAmG2fU0eHHRvcnZhbGRz
QGxpbnV4LWZvdW5kYXRpb24ub3JnAAoJEHm+PkMAQRiGC7EH/3R7Rt+OD8Wn8Ss3
w8V+dBxVwa2u2oMTyUHPxaeOXZ7bi38XlUdLFPOK/76bGwO0a5TmYZqsWdRbGyT0
HfcYjHsQ0lbJXk/nh2oM47oJxJXVpThIHXJEk0FZ0Y5t+DYjIYlNHzqZymUyhLem
St74zgWcyT+MXuqY34vB827FJDUnOxhhhi85tObeunaSPAomy9aiYidSC1ARREnz
iz2VUntP/QnRnKVvL2nUZNzcz1xL5vfCRSKsRGRSv3qW1Y/1M71ylt6JVmSftWq+
VmMdFxFhdrb1OK/1ct/930Un/UP2NG9EJsWxote2XYlnVSZHzDqH7lUhbqgdCcLz
1m2tVNY=
=7wRd
-----END PGP SIGNATURE-----
Merge tag 'v5.16-rc5' into rdma.git for-next
Required due to dependencies in following patches.
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
Variables allocated by kvmalloc_array() should not be freed by kfree.
Because they may be allocated by vmalloc. So we replace kfree() with
kvfree() here.
Fixes: 6fd610c573 ("RDMA/hns: Support 0 hop addressing for SRQ buffer")
Link: https://lore.kernel.org/r/20211210094234.5829-1-billsjc@sjtu.edu.cn
Signed-off-by: Jiacheng Shi <billsjc@sjtu.edu.cn>
Acked-by: Wenpeng Liang <liangwenpeng@huawei.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
Currently, when cma_resolve_ib_dev() searches for a matching GID it will
stop searching after encountering the first empty GID table entry. This
behavior is wrong since neither IB nor RoCE spec enforce tightly packed
GID tables.
For example, when the matching valid GID entry exists at index N, and if a
GID entry is empty at index N-1, cma_resolve_ib_dev() will fail to find
the matching valid entry.
Fix it by making cma_resolve_ib_dev() continue searching even after
encountering missing entries.
Fixes: f17df3b0de ("RDMA/cma: Add support for AF_IB to rdma_resolve_addr()")
Link: https://lore.kernel.org/r/b7346307e3bb396c43d67d924348c6c496493991.1639055490.git.leonro@nvidia.com
Signed-off-by: Avihai Horon <avihaih@nvidia.com>
Reviewed-by: Mark Zhang <markzhang@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
Currently, ib_find_gid() will stop searching after encountering the first
empty GID table entry. This behavior is wrong since neither IB nor RoCE
spec enforce tightly packed GID tables.
For example, when a valid GID entry exists at index N, and if a GID entry
is empty at index N-1, ib_find_gid() will fail to find the valid entry.
Fix it by making ib_find_gid() continue searching even after encountering
missing entries.
Fixes: 5eb620c81c ("IB/core: Add helpers for uncached GID and P_Key searches")
Link: https://lore.kernel.org/r/e55d331b96cecfc2cf19803d16e7109ea966882d.1639055490.git.leonro@nvidia.com
Signed-off-by: Avihai Horon <avihaih@nvidia.com>
Reviewed-by: Mark Zhang <markzhang@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
