platform/kernel/linux-rpi.git
2 years agobpf: Add check_func_arg_reg_off function
Kumar Kartikeya Dwivedi [Fri, 4 Mar 2022 22:46:38 +0000 (04:16 +0530)]
bpf: Add check_func_arg_reg_off function

Lift the list of register types allowed for having fixed and variable
offsets when passed as helper function arguments into a common helper,
so that they can be reused for kfunc checks in later commits. Keeping a
common helper aids maintainability and allows us to follow the same
consistent rules across helpers and kfuncs. Also, convert check_func_arg
to use this function.

Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20220304224645.3677453-2-memxor@gmail.com
2 years agoMerge branch 'libbpf: support custom SEC() handlers'
Alexei Starovoitov [Sat, 5 Mar 2022 17:38:15 +0000 (09:38 -0800)]
Merge branch 'libbpf: support custom SEC() handlers'

Andrii Nakryiko says:

====================

Add ability for user applications and libraries to register custom BPF program
SEC() handlers. See patch #2 for examples where this is useful.

Patch #1 does some preliminary refactoring to allow exponsing program
init, preload, and attach callbacks as public API. It also establishes
a protocol to allow optional auto-attach behavior. This will also help the
case of sometimes auto-attachable uprobes.

v4->v5:
  - API documentation improvements (Daniel);
v3->v4:
  - init_fn -> prog_setup_fn, preload_fn -> prog_prepare_load_fn (Alexei);
v2->v3:
  - moved callbacks and cookie into OPTS struct (Alan);
  - added more test scenarios (Alan);
  - address most of Alan's feedback, but kept API name;
v1->v2:
  - resubmitting due to git send-email screw up.

Cc: Alan Maguire <alan.maguire@oracle.com>
====================

Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2 years agoselftests/bpf: Add custom SEC() handling selftest
Andrii Nakryiko [Sat, 5 Mar 2022 01:01:29 +0000 (17:01 -0800)]
selftests/bpf: Add custom SEC() handling selftest

Add a selftest validating various aspects of libbpf's handling of custom
SEC() handlers. It also demonstrates how libraries can ensure very early
callbacks registration and unregistration using
__attribute__((constructor))/__attribute__((destructor)) functions.

Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Tested-by: Alan Maguire <alan.maguire@oracle.com>
Reviewed-by: Alan Maguire <alan.maguire@oracle.com>
Link: https://lore.kernel.org/bpf/20220305010129.1549719-4-andrii@kernel.org
2 years agolibbpf: Support custom SEC() handlers
Andrii Nakryiko [Sat, 5 Mar 2022 01:01:28 +0000 (17:01 -0800)]
libbpf: Support custom SEC() handlers

Allow registering and unregistering custom handlers for BPF program.
This allows user applications and libraries to plug into libbpf's
declarative SEC() definition handling logic. This allows to offload
complex and intricate custom logic into external libraries, but still
provide a great user experience.

One such example is USDT handling library, which has a lot of code and
complexity which doesn't make sense to put into libbpf directly, but it
would be really great for users to be able to specify BPF programs with
something like SEC("usdt/<path-to-binary>:<usdt_provider>:<usdt_name>")
and have correct BPF program type set (BPF_PROGRAM_TYPE_KPROBE, as it is
uprobe) and even support BPF skeleton's auto-attach logic.

In some cases, it might be even good idea to override libbpf's default
handling, like for SEC("perf_event") programs. With custom library, it's
possible to extend logic to support specifying perf event specification
right there in SEC() definition without burdening libbpf with lots of
custom logic or extra library dependecies (e.g., libpfm4). With current
patch it's possible to override libbpf's SEC("perf_event") handling and
specify a completely custom ones.

Further, it's possible to specify a generic fallback handling for any
SEC() that doesn't match any other custom or standard libbpf handlers.
This allows to accommodate whatever legacy use cases there might be, if
necessary.

See doc comments for libbpf_register_prog_handler() and
libbpf_unregister_prog_handler() for detailed semantics.

This patch also bumps libbpf development version to v0.8 and adds new
APIs there.

Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Tested-by: Alan Maguire <alan.maguire@oracle.com>
Reviewed-by: Alan Maguire <alan.maguire@oracle.com>
Link: https://lore.kernel.org/bpf/20220305010129.1549719-3-andrii@kernel.org
2 years agolibbpf: Allow BPF program auto-attach handlers to bail out
Andrii Nakryiko [Sat, 5 Mar 2022 01:01:27 +0000 (17:01 -0800)]
libbpf: Allow BPF program auto-attach handlers to bail out

Allow some BPF program types to support auto-attach only in subste of
cases. Currently, if some BPF program type specifies attach callback, it
is assumed that during skeleton attach operation all such programs
either successfully attach or entire skeleton attachment fails. If some
program doesn't support auto-attachment from skeleton, such BPF program
types shouldn't have attach callback specified.

This is limiting for cases when, depending on how full the SEC("")
definition is, there could either be enough details to support
auto-attach or there might not be and user has to use some specific API
to provide more details at runtime.

One specific example of such desired behavior might be SEC("uprobe"). If
it's specified as just uprobe auto-attach isn't possible. But if it's
SEC("uprobe/<some_binary>:<some_func>") then there are enough details to
support auto-attach. Note that there is a somewhat subtle difference
between auto-attach behavior of BPF skeleton and using "generic"
bpf_program__attach(prog) (which uses the same attach handlers under the
cover). Skeleton allow some programs within bpf_object to not have
auto-attach implemented and doesn't treat that as an error. Instead such
BPF programs are just skipped during skeleton's (optional) attach step.
bpf_program__attach(), on the other hand, is called when user *expects*
auto-attach to work, so if specified program doesn't implement or
doesn't support auto-attach functionality, that will be treated as an
error.

Another improvement to the way libbpf is handling SEC()s would be to not
require providing dummy kernel function name for kprobe. Currently,
SEC("kprobe/whatever") is necessary even if actual kernel function is
determined by user at runtime and bpf_program__attach_kprobe() is used
to specify it. With changes in this patch, it's possible to support both
SEC("kprobe") and SEC("kprobe/<actual_kernel_function"), while only in
the latter case auto-attach will be performed. In the former one, such
kprobe will be skipped during skeleton attach operation.

Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Tested-by: Alan Maguire <alan.maguire@oracle.com>
Reviewed-by: Alan Maguire <alan.maguire@oracle.com>
Link: https://lore.kernel.org/bpf/20220305010129.1549719-2-andrii@kernel.org
2 years agoMerge branch 'bnxt_en-updates'
David S. Miller [Sat, 5 Mar 2022 11:16:56 +0000 (11:16 +0000)]
Merge branch 'bnxt_en-updates'

Michael Chan says:

====================
bnxt_en: Updates.

This patch series contains mainly NVRAM related features.  More
NVRAM error checking and logging are added when installing firmware
packages.  A new devlink hw health report is now added to report
and diagnose NVRAM issues.  Other miscellaneous patches include
reporting correctly cards that don't support link pause, adding
an internal unknown link state, and avoiding unnecessary link
toggle during firmware reset.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agobnxt_en: add an nvm test for hw diagnose
Vikas Gupta [Sat, 5 Mar 2022 08:54:42 +0000 (03:54 -0500)]
bnxt_en: add an nvm test for hw diagnose

Add an NVM test function for devlink hw reporter.
In this function an NVM VPD area is read followed by
a write. Test result is cached and if it is successful then
the next test can be conducted only after HW_RETEST_MIN_TIME to
avoid frequent writes to the NVM.

Reviewed-by: Edwin Peer <edwin.peer@broadcom.com>
Signed-off-by: Vikas Gupta <vikas.gupta@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agobnxt_en: implement hw health reporter
Kalesh AP [Sat, 5 Mar 2022 08:54:41 +0000 (03:54 -0500)]
bnxt_en: implement hw health reporter

This reporter will report NVM errors which are non-fatal.
When we receive these NVM error events, we'll report it
through this new hw health reporter.

Reviewed-by: Edwin Peer <edwin.peer@broadcom.com>
Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agobnxt_en: Do not destroy health reporters during reset
Edwin Peer [Sat, 5 Mar 2022 08:54:40 +0000 (03:54 -0500)]
bnxt_en: Do not destroy health reporters during reset

Health reporter state should be maintained over resets. Previously
reporters were destroyed if the device capabilities changed, but
since none of the reporters depend on capabilities anymore, this
logic should be removed.

Signed-off-by: Edwin Peer <edwin.peer@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agobnxt_en: Eliminate unintended link toggle during FW reset
Michael Chan [Sat, 5 Mar 2022 08:54:39 +0000 (03:54 -0500)]
bnxt_en: Eliminate unintended link toggle during FW reset

If the flow control settings have been changed, a subsequent FW reset
may cause the ethernet link to toggle unnecessarily.  This link toggle
will increase the down time by a few seconds.

The problem is caused by bnxt_update_phy_setting() detecting a false
mismatch in the flow control settings between the stored software
settings and the current FW settings after the FW reset.  This mismatch
is caused by the AUTONEG bit added to link_info->req_flow_ctrl in an
inconsistent way in bnxt_set_pauseparam() in autoneg mode.  The AUTONEG
bit should not be added to link_info->req_flow_ctrl.

Reviewed-by: Colin Winegarden <colin.winegarden@broadcom.com>
Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agobnxt_en: Properly report no pause support on some cards
Michael Chan [Sat, 5 Mar 2022 08:54:38 +0000 (03:54 -0500)]
bnxt_en: Properly report no pause support on some cards

Some cards are configured to never support link pause or PFC.  Discover
these cards and properly report no pause support to ethtool.  Disable
PFC settings from DCBNL if PFC is unsupported.

Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agobnxt_en: introduce initial link state of unknown
Edwin Peer [Sat, 5 Mar 2022 08:54:37 +0000 (03:54 -0500)]
bnxt_en: introduce initial link state of unknown

This will force link state to always be logged for initial NIC open.

Signed-off-by: Edwin Peer <edwin.peer@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agobnxt_en: parse result field when NVRAM package install fails
Kalesh AP [Sat, 5 Mar 2022 08:54:36 +0000 (03:54 -0500)]
bnxt_en: parse result field when NVRAM package install fails

Instead of always returning -ENOPKG, decode the firmware error
code further when the HWRM_NVM_INSTALL_UPDATE firmware call fails.
Return a more suitable error code to userspace and log an error
in dmesg.

Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com>
Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agobnxt_en: add more error checks to HWRM_NVM_INSTALL_UPDATE
Kalesh AP [Sat, 5 Mar 2022 08:54:35 +0000 (03:54 -0500)]
bnxt_en: add more error checks to HWRM_NVM_INSTALL_UPDATE

FW returns error code "NVM_INSTALL_UPDATE_CMD_ERR_CODE_ANTI_ROLLBACK"
in the response to indicate that HWRM_NVM_INSTALL_UPDATE command has
failed due to Anti-rollback feature. Parse the error and return an
appropriate error code to the user.

Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com>
Reviewed-by: Edwin Peer <edwin.peer@broadcom.com>
Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agobnxt_en: refactor error handling of HWRM_NVM_INSTALL_UPDATE
Kalesh AP [Sat, 5 Mar 2022 08:54:34 +0000 (03:54 -0500)]
bnxt_en: refactor error handling of HWRM_NVM_INSTALL_UPDATE

This is in anticipation of handling more "cmd_err" from FW in the next
patch.

Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com>
Reviewed-by: Edwin Peer <edwin.peer@broadcom.com>
Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoMerge branch 'axienet-napi-gro-support'
David S. Miller [Sat, 5 Mar 2022 11:12:09 +0000 (11:12 +0000)]
Merge branch 'axienet-napi-gro-support'

Robert Hancock says:

====================
NAPI/GRO support for axienet driver

Add support for NAPI and GRO receive in the Xilinx AXI Ethernet driver,
and some other related cleanups.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: axienet: add coalesce timer ethtool configuration
Robert Hancock [Sat, 5 Mar 2022 02:24:43 +0000 (20:24 -0600)]
net: axienet: add coalesce timer ethtool configuration

Add the ability to configure the RX/TX coalesce timer with ethtool.
Change default setting to scale with the clock rate rather than being a
fixed number of clock cycles.

Signed-off-by: Robert Hancock <robert.hancock@calian.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: axienet: reduce default RX interrupt threshold to 1
Robert Hancock [Sat, 5 Mar 2022 02:24:42 +0000 (20:24 -0600)]
net: axienet: reduce default RX interrupt threshold to 1

Now that NAPI has been implemented, the hardware interrupt mitigation
mechanism is not needed to avoid excessive interrupt load in most cases.
Reduce the default RX interrupt threshold to 1 to reduce introduced
latency. This can be increased with ethtool if desired if some applications
still want to reduce interrupts.

Signed-off-by: Robert Hancock <robert.hancock@calian.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: axienet: implement NAPI and GRO receive
Robert Hancock [Sat, 5 Mar 2022 02:24:41 +0000 (20:24 -0600)]
net: axienet: implement NAPI and GRO receive

Implement NAPI and GRO receive. In addition to better performance, this
also avoids handling RX packets in hard IRQ context, which reduces the
IRQ latency impact to other devices.

Signed-off-by: Robert Hancock <robert.hancock@calian.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: axienet: don't set IRQ timer when IRQ delay not used
Robert Hancock [Sat, 5 Mar 2022 02:24:40 +0000 (20:24 -0600)]
net: axienet: don't set IRQ timer when IRQ delay not used

When the RX or TX coalesce count is set to 1, there's no point in
setting the delay timer value since an interrupt will already be raised
on every packet, and the delay interrupt just causes extra pointless
interrupts.

Signed-off-by: Robert Hancock <robert.hancock@calian.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: axienet: Clean up DMA start/stop and error handling
Robert Hancock [Sat, 5 Mar 2022 02:24:39 +0000 (20:24 -0600)]
net: axienet: Clean up DMA start/stop and error handling

Simplify the DMA error handling process, and remove some duplicated code
between the DMA error handling and the stop function.

Signed-off-by: Robert Hancock <robert.hancock@calian.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: axienet: Clean up device used for DMA calls
Robert Hancock [Sat, 5 Mar 2022 02:24:38 +0000 (20:24 -0600)]
net: axienet: Clean up device used for DMA calls

Instead of using lp->ndev.parent to find the correct device to use for
DMA API calls, just use the dev attribute in the device structure.

Signed-off-by: Robert Hancock <robert.hancock@calian.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: axienet: fix RX ring refill allocation failure handling
Robert Hancock [Sat, 5 Mar 2022 02:24:37 +0000 (20:24 -0600)]
net: axienet: fix RX ring refill allocation failure handling

If a memory allocation error occurred during an attempt to refill a slot
in the RX ring after the packet was received, the hardware tail pointer
would still have been updated to point to or past the slot which remained
marked as previously completed. This would likely result in the DMA engine
raising an error when it eventually tried to use that slot again.

If a slot cannot be refilled, then just stop processing and do not move
the tail pointer past it. On the next attempt, we should skip receiving
the packet from the empty slot and just try to refill it again.

This failure mode has not actually been observed, but was found as part
of other driver updates.

Fixes: 8a3b7a252dca ("drivers/net/ethernet/xilinx: added Xilinx AXI Ethernet driver")
Signed-off-by: Robert Hancock <robert.hancock@calian.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoMerge branch 'dsa-realtek-add-rtl8_4t-tags'
David S. Miller [Sat, 5 Mar 2022 11:04:25 +0000 (11:04 +0000)]
Merge branch 'dsa-realtek-add-rtl8_4t-tags'

Luiz Angelo Daros de Luca says:

====================
net: dsa: realtek: add rtl8_4t tag

This patch series adds support for rtl8_4t tag. It is a variant of
rtl8_4 tag, with identical values but placed at the end of the packet
(before CRC).

It forces checksum in software before adding the tag as those extra
bytes at the end of the packet would be summed together with the rest of
the payload. When the switch removes the tag before sending the packet
to the network, that checksum will not match.

It might be useful to diagnose or avoid checksum offload issues. With an
ethertype tag like rtl8_4, the cpu port ethernet driver must work with
cksum_start and chksum_offset to correctly calculate checksums. If not,
the checksum field will be broken (it will contain the fake ip header
sum).  In those cases, using 'rtl8_4t' might be an alternative way to
avoid checksum offload, either using runtime or device-tree property.

Regards,

Luiz

v4-v5)
- tags in alphabetical order in dsa_port.yaml
- remove ret var from rtl8365mb_change_tag_protocol
- Comment typos fixes

v3-v4)
- added rtl8_4 and rtl8_4t to dsa_port.yaml
- removed generic considerations about checksum problems with DSA tags.
  They belong to Documentation/networking/dsa/dsa.rst

v2-v3)
- updated tag documentation (file header)
- do not remove position and format from rtl8365mb_cpu
- reinstate cpu to rtl8365mb
- moved rtl8365mb_change_tag_protocol after rtl8365mb_cpu_config
- do not modify rtl8365mb_cpu_config() logic
- remove cpu arg from rtl8365mb_cpu_config(); get it from priv
- dropped tag_protocol from rtl8365mb. It is now derived from
  cpu->position.
- init cpu struct before dsa_register as default tag must be already
  defined before dsa_register()
- fix formatting issues

v1-v2)
- remove mention to tail tagger, use trailing tagger.
- use void* instead of char* for pointing to tag beginning
- use memcpy to avoid problems with unaligned tags
- calculate checksum if it still pending
- keep in-use tag protocol in memory instead of reading from switch
  register
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: dsa: realtek: rtl8365mb: add support for rtl8_4t
Luiz Angelo Daros de Luca [Thu, 3 Mar 2022 01:52:35 +0000 (22:52 -0300)]
net: dsa: realtek: rtl8365mb: add support for rtl8_4t

The trailing tag is also supported by this family. The default is still
rtl8_4 but now the switch supports changing the tag to rtl8_4t.

Reintroduce the dropped cpu in struct rtl8365mb (removed by 6147631).

Signed-off-by: Luiz Angelo Daros de Luca <luizluca@gmail.com>
Reviewed-by: Alvin Šipraga <alsi@bang-olufsen.dk>
Reviewed-by: Vladimir Oltean <olteanv@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: dsa: tag_rtl8_4: add rtl8_4t trailing variant
Luiz Angelo Daros de Luca [Thu, 3 Mar 2022 01:52:34 +0000 (22:52 -0300)]
net: dsa: tag_rtl8_4: add rtl8_4t trailing variant

Realtek switches supports the same tag both before ethertype or between
payload and the CRC.

Signed-off-by: Luiz Angelo Daros de Luca <luizluca@gmail.com>
Reviewed-by: Alvin Šipraga <alsi@bang-olufsen.dk>
Reviewed-by: Vladimir Oltean <olteanv@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agodt-bindings: net: dsa: add rtl8_4 and rtl8_4t tag formats
Luiz Angelo Daros de Luca [Thu, 3 Mar 2022 01:52:33 +0000 (22:52 -0300)]
dt-bindings: net: dsa: add rtl8_4 and rtl8_4t tag formats

Realtek rtl8365mb DSA driver can use these two tag formats.

Cc: devicetree@vger.kernel.org
Signed-off-by: Luiz Angelo Daros de Luca <luizluca@gmail.com>
Reviewed-by: Vladimir Oltean <olteanv@gmail.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoMerge branch 'mptcp-selftest-refinements-and-a-new-test'
Jakub Kicinski [Sat, 5 Mar 2022 05:54:33 +0000 (21:54 -0800)]
Merge branch 'mptcp-selftest-refinements-and-a-new-test'

Mat Martineau says:

====================
mptcp: Selftest refinements and a new test

Patches 1 and 11 improve the printed output of the mptcp_join.sh
selftest.

Patches 2-8 add a test for the MP_FASTCLOSE option, including
prerequisite changes like additional MPTCP MIBs.

Patches 9-10 add some groundwork for upcoming tests.
====================

Link: https://lore.kernel.org/r/20220304193636.219315-1-mathew.j.martineau@linux.intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoselftests: mptcp: update output info of chk_rm_nr
Geliang Tang [Fri, 4 Mar 2022 19:36:36 +0000 (11:36 -0800)]
selftests: mptcp: update output info of chk_rm_nr

This patch updated the output info of chk_rm_nr. Renamed 'sf' to 'rmsf',
which means 'remove subflow'. Added the display of whether the inverted
namespaces has been used to check the mib counters.

The new output looks like this:

 002 remove multiple subflows          syn[ ok ] - synack[ ok ] - ack[ ok ]
                                       rm [ ok ] - rmsf  [ ok ]
 003 remove single address             syn[ ok ] - synack[ ok ] - ack[ ok ]
                                       add[ ok ] - echo  [ ok ]
                                       rm [ ok ] - rmsf  [ ok ]   invert

Signed-off-by: Geliang Tang <geliang.tang@suse.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoselftests: mptcp: add more arguments for chk_join_nr
Geliang Tang [Fri, 4 Mar 2022 19:36:35 +0000 (11:36 -0800)]
selftests: mptcp: add more arguments for chk_join_nr

This patch added five more arguments for chk_join_nr(). The default
values of them are all zero.

The first two, csum_ns1 and csum_ns1, are passed to chk_csum_nr(), to
check the mib counters of the checksum errors in ns1 and ns2. A '+'
can be added into this two arguments to represent that multiple
checksum errors are allowed when doing this check. For example,

        chk_csum_nr "" +2 +2

indicates that two or more checksum errors are allowed in both ns1 and
ns2.

The remaining two, fail_nr and rst_nr, are passed to chk_fail_nr() and
chk_rst_nr() respectively, to check the sending and receiving mib
counters of MP_FAIL and MP_RST.

Also did some cleanups in chk_fail_nr(), renamed two local variables
and updated the output message.

Signed-off-by: Geliang Tang <geliang.tang@suse.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoselftests: mptcp: add invert check in check_transfer
Geliang Tang [Fri, 4 Mar 2022 19:36:34 +0000 (11:36 -0800)]
selftests: mptcp: add invert check in check_transfer

This patch added the invert bytes check for the output data in
check_transfer().

Instead of the file mismatch error:

  [ FAIL ] file received by server does not match (in, out):
  -rw------- 1 root root 45643832 Jan 16 15:04 /tmp/tmp.9xpM6Paivv
  Trailing bytes are:
  MPTCP_TEST_FILE_END_MARKER
  -rw------- 1 root root 45643832 Jan 16 15:04 /tmp/tmp.wnz1Yp4u7Z
  Trailing bytes are:
  MPTCP_TEST_FILE_END_MARKER

Print out the inverted bytes like this:

  file received by server has inverted byte at 7454789
  file received by server has inverted byte at 7454790
  file received by server has inverted byte at 7454791
  file received by server has inverted byte at 7454792

Signed-off-by: Geliang Tang <geliang.tang@suse.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoselftests: mptcp: add fastclose testcase
Geliang Tang [Fri, 4 Mar 2022 19:36:33 +0000 (11:36 -0800)]
selftests: mptcp: add fastclose testcase

This patch added the self test for MP_FASTCLOSE. Reused the argument
addr_nr_ns2 of do_transfer() to pass the extra arguments '-I 2' to
mptcp_connect commands. Then mptcp_connect disconnected the
connections to trigger the MP_FASTCLOSE sending and receiving. Used
chk_fclose_nr to check the MP_FASTCLOSE mibs and used chk_rst_nr to
check the MP_RST mibs. This test used the test_linkfail value to make
1024KB test files.

The output looks like this:

Created /tmp/tmp.XB8sfv1hJ0 (size 1024 KB) containing data sent by client
Created /tmp/tmp.RtTDbzqrXI (size 1024 KB) containing data sent by server
001 fastclose test                syn[ ok ] - synack[ ok ] - ack[ ok ]
                                  ctx[ ok ] - fclzrx[ ok ]
                                  rtx[ ok ] - rstrx [ ok ]   invert

Signed-off-by: Geliang Tang <geliang.tang@suse.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoselftests: mptcp: reuse linkfail to make given size files
Geliang Tang [Fri, 4 Mar 2022 19:36:32 +0000 (11:36 -0800)]
selftests: mptcp: reuse linkfail to make given size files

This patch reused the test_linkfail values above 2 to make test files with
the given sizes (KB) for both the client side and the server side. It's
useful for the test cases using different file sizes.

Signed-off-by: Geliang Tang <geliang.tang@suse.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoselftests: mptcp: add extra_args in do_transfer
Geliang Tang [Fri, 4 Mar 2022 19:36:31 +0000 (11:36 -0800)]
selftests: mptcp: add extra_args in do_transfer

Instead of using a global variable mptcp_connect, this patch added
a new local variable extra_args in do_transfer() to store the extra
arguments passing to the mptcp_connect commands.

This patch also renamed the speed level 'least' to 'speed_*'. This
more flexible way can avoid the need to add new speed levels in the
future.

Signed-off-by: Geliang Tang <geliang.tang@suse.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoselftests: mptcp: add the MP_RST mibs check
Geliang Tang [Fri, 4 Mar 2022 19:36:30 +0000 (11:36 -0800)]
selftests: mptcp: add the MP_RST mibs check

This patch added a new function chk_rst_nr() to check the numbers
of the MP_RST sending and receiving mibs.

Showed in the output whether the inverted namespaces check order is used.
Since if we pass -Cz to mptcp_join.sh, the MP_RST information is showed
twice.

Signed-off-by: Geliang Tang <geliang.tang@suse.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agomptcp: add the mibs for MP_RST
Geliang Tang [Fri, 4 Mar 2022 19:36:29 +0000 (11:36 -0800)]
mptcp: add the mibs for MP_RST

This patch added two more mibs for MP_RST, MPTCP_MIB_MPRSTTX for
the MP_RST sending and MPTCP_MIB_MPRSTRX for the MP_RST receiving.

Signed-off-by: Geliang Tang <geliang.tang@suse.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoselftests: mptcp: add the MP_FASTCLOSE mibs check
Geliang Tang [Fri, 4 Mar 2022 19:36:28 +0000 (11:36 -0800)]
selftests: mptcp: add the MP_FASTCLOSE mibs check

This patch added a new function chk_fclose_nr() to check the numbers
of the MP_FASTCLOSE sending and receiving mibs.

Signed-off-by: Geliang Tang <geliang.tang@suse.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agomptcp: add the mibs for MP_FASTCLOSE
Geliang Tang [Fri, 4 Mar 2022 19:36:27 +0000 (11:36 -0800)]
mptcp: add the mibs for MP_FASTCLOSE

This patch added two more mibs for MP_FASTCLOSE, MPTCP_MIB_MPFASTCLOSETX
for the MP_FASTCLOSE sending and MPTCP_MIB_MPFASTCLOSERX for receiving.

Also added a debug log for MP_FASTCLOSE receiving, printed out the recv_key
of MP_FASTCLOSE in mptcp_parse_option to show that MP_RST is received.

Signed-off-by: Geliang Tang <geliang.tang@suse.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoselftests: mptcp: adjust output alignment for more tests
Geliang Tang [Fri, 4 Mar 2022 19:36:26 +0000 (11:36 -0800)]
selftests: mptcp: adjust output alignment for more tests

The number of self tests in mptcp_join.sh will soon be more than 100, the
output alignment is no longer OK. This patch adjusted it.

Signed-off-by: Geliang Tang <geliang.tang@suse.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agobcm63xx_enet: Use platform_get_irq() to get the interrupt
Meng Tang [Thu, 3 Mar 2022 10:08:15 +0000 (18:08 +0800)]
bcm63xx_enet: Use platform_get_irq() to get the interrupt

platform_get_resource(pdev, IORESOURCE_IRQ, ..) relies on
static allocation of IRQ resources in DT core code, this
causes an issue when using hierarchical interrupt domains
using "interrupts" property in the node as this bypassed
the hierarchical setup and messed up the irq chaining.

In preparation for removal of static setup of IRQ resource
from DT core code use platform_get_irq().

Signed-off-by: Meng Tang <tangmeng@uniontech.com>
Link: https://lore.kernel.org/r/20220303100815.25605-1-tangmeng@uniontech.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoptp: ocp: Add serial port information to the debug summary
Jonathan Lemon [Fri, 4 Mar 2022 05:46:15 +0000 (21:46 -0800)]
ptp: ocp: Add serial port information to the debug summary

On the debug summary page, show the /dev/ttyS<port> mapping.

Signed-off-by: Jonathan Lemon <jonathan.lemon@gmail.com>
Link: https://lore.kernel.org/r/20220304054615.1737-1-jonathan.lemon@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agonet: sparx5: Fix initialization of variables on stack
Horatiu Vultur [Fri, 4 Mar 2022 14:09:18 +0000 (15:09 +0100)]
net: sparx5: Fix initialization of variables on stack

The variables 'res' inside the functions sparx5_ptp_get_1ppm and
sparx5_ptp_get_nominal_value was not initialized. So in case of the default
case of the switch after, it would return an uninitialized variable.
This makes also the clang builds to failed.

Fixes: 0933bd04047c3b ("net: sparx5: Add support for ptp clocks")
Signed-off-by: Horatiu Vultur <horatiu.vultur@microchip.com>
Link: https://lore.kernel.org/r/20220304140918.3356873-1-horatiu.vultur@microchip.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoMerge tag 'for-net-next-2022-03-04' of git://git.kernel.org/pub/scm/linux/kernel...
Jakub Kicinski [Sat, 5 Mar 2022 03:42:04 +0000 (19:42 -0800)]
Merge tag 'for-net-next-2022-03-04' of git://git./linux/kernel/git/bluetooth/bluetooth-next

Luiz Augusto von Dentz says:

====================
bluetooth-next pull request for net-next:

 - Add new PID/VID (0x13d3/0x3567) for MT7921
 - Add new PID/VID (0x2550/0x8761) for Realtek 8761BU
 - Add support for LG LGSBWAC02 (MT7663BUN)
 - Add support for BCM43430A0 and BCM43430A1
 - Add support for Intel Madison Peak (MsP2)

* tag 'for-net-next-2022-03-04' of git://git.kernel.org/pub/scm/linux/kernel/git/bluetooth/bluetooth-next: (21 commits)
  Bluetooth: btusb: Add another Realtek 8761BU
  Bluetooth: hci_bcm: add BCM43430A0 & BCM43430A1
  Bluetooth: use memset avoid memory leaks
  Bluetooth: btmtksdio: Fix kernel oops when sdio suspend.
  Bluetooth: btusb: Add a new PID/VID 13d3/3567 for MT7921
  Bluetooth: move adv_instance_cnt read within the device lock
  Bluetooth: hci_event: Add missing locking on hdev in hci_le_ext_adv_term_evt
  Bluetooth: btusb: Make use of of BIT macro to declare flags
  Bluetooth: Fix not checking for valid hdev on bt_dev_{info,warn,err,dbg}
  Bluetooth: mediatek: fix the conflict between mtk and msft vendor event
  Bluetooth: mt7921s: support bluetooth reset mechanism
  Bluetooth: make array bt_uuid_any static const
  Bluetooth: 6lowpan: No need to clear memory twice
  Bluetooth: btusb: Improve stability for QCA devices
  Bluetooth: btusb: add support for LG LGSBWAC02 (MT7663BUN)
  Bluetooth: btusb: Add support for Intel Madison Peak (MsP2) device
  Bluetooth: Improve skb handling in mgmt_device_connected()
  Bluetooth: Fix skb allocation in mgmt_remote_name() & mgmt_device_connected()
  Bluetooth: mgmt: Remove unneeded variable
  Bluetooth: hci_sync: fix undefined return of hci_disconnect_all_sync()
  ...
====================

Link: https://lore.kernel.org/r/20220304193919.649815-1-luiz.dentz@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoMerge https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next
Jakub Kicinski [Sat, 5 Mar 2022 03:28:17 +0000 (19:28 -0800)]
Merge https://git./linux/kernel/git/bpf/bpf-next

Daniel Borkmann says:

====================
pull-request: bpf-next 2022-03-04

We've added 32 non-merge commits during the last 14 day(s) which contain
a total of 59 files changed, 1038 insertions(+), 473 deletions(-).

The main changes are:

1) Optimize BPF stackmap's build_id retrieval by caching last valid build_id,
   as consecutive stack frames are likely to be in the same VMA and therefore
   have the same build id, from Hao Luo.

2) Several improvements to arm64 BPF JIT, that is, support for JITing
   the atomic[64]_fetch_add, atomic[64]_[fetch_]{and,or,xor} and lastly
   atomic[64]_{xchg|cmpxchg}. Also fix the BTF line info dump for JITed
   programs, from Hou Tao.

3) Optimize generic BPF map batch deletion by only enforcing synchronize_rcu()
   barrier once upon return to user space, from Eric Dumazet.

4) For kernel build parse DWARF and generate BTF through pahole with enabled
   multithreading, from Kui-Feng Lee.

5) BPF verifier usability improvements by making log info more concise and
   replacing inv with scalar type name, from Mykola Lysenko.

6) Two follow-up fixes for BPF prog JIT pack allocator, from Song Liu.

7) Add a new Kconfig to allow for loading kernel modules with non-matching
   BTF type info; their BTF info is then removed on load, from Connor O'Brien.

8) Remove reallocarray() usage from bpftool and switch to libbpf_reallocarray()
   in order to fix compilation errors for older glibc, from Mauricio Vásquez.

9) Fix libbpf to error on conflicting name in BTF when type declaration
   appears before the definition, from Xu Kuohai.

10) Fix issue in BPF preload for in-kernel light skeleton where loaded BPF
    program fds prevent init process from setting up fd 0-2, from Yucong Sun.

11) Fix libbpf reuse of pinned perf RB map when max_entries is auto-determined
    by libbpf, from Stijn Tintel.

12) Several cleanups for libbpf and a fix to enforce perf RB map #pages to be
    non-zero, from Yuntao Wang.

* https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next: (32 commits)
  bpf: Small BPF verifier log improvements
  libbpf: Add a check to ensure that page_cnt is non-zero
  bpf, x86: Set header->size properly before freeing it
  x86: Disable HAVE_ARCH_HUGE_VMALLOC on 32-bit x86
  bpf, test_run: Fix overflow in XDP frags bpf_test_finish
  selftests/bpf: Update btf_dump case for conflicting names
  libbpf: Skip forward declaration when counting duplicated type names
  bpf: Add some description about BPF_JIT_ALWAYS_ON in Kconfig
  bpf, docs: Add a missing colon in verifier.rst
  bpf: Cache the last valid build_id
  libbpf: Fix BPF_MAP_TYPE_PERF_EVENT_ARRAY auto-pinning
  bpf, selftests: Use raw_tp program for atomic test
  bpf, arm64: Support more atomic operations
  bpftool: Remove redundant slashes
  bpf: Add config to allow loading modules with BTF mismatches
  bpf, arm64: Feed byte-offset into bpf line info
  bpf, arm64: Call build_prologue() first in first JIT pass
  bpf: Fix issue with bpf preload module taking over stdout/stdin of kernel.
  bpftool: Bpf skeletons assert type sizes
  bpf: Cleanup comments
  ...
====================

Link: https://lore.kernel.org/r/20220304164313.31675-1-daniel@iogearbox.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoBluetooth: btusb: Add another Realtek 8761BU
Helmut Grohne [Sat, 26 Feb 2022 15:22:56 +0000 (16:22 +0100)]
Bluetooth: btusb: Add another Realtek 8761BU

This device is sometimes wrapped with a label "EDUP".

T:  Bus=01 Lev=02 Prnt=02 Port=02 Cnt=03 Dev#=107 Spd=12   MxCh= 0
D:  Ver= 1.10 Cls=e0(wlcon) Sub=01 Prot=01 MxPS=64 #Cfgs=  1
P:  Vendor=2550 ProdID=8761 Rev= 2.00
S:  Manufacturer=Realtek
S:  Product=Bluetooth Radio
S:  SerialNumber=00E04C239987
C:* #Ifs= 2 Cfg#= 1 Atr=e0 MxPwr=500mA
I:* If#= 0 Alt= 0 #EPs= 3 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=81(I) Atr=03(Int.) MxPS=  16 Ivl=1ms
E:  Ad=02(O) Atr=02(Bulk) MxPS=  64 Ivl=0ms
E:  Ad=82(I) Atr=02(Bulk) MxPS=  64 Ivl=0ms
I:* If#= 1 Alt= 0 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=03(O) Atr=01(Isoc) MxPS=   0 Ivl=1ms
E:  Ad=83(I) Atr=01(Isoc) MxPS=   0 Ivl=1ms
I:  If#= 1 Alt= 1 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=03(O) Atr=01(Isoc) MxPS=   9 Ivl=1ms
E:  Ad=83(I) Atr=01(Isoc) MxPS=   9 Ivl=1ms
I:  If#= 1 Alt= 2 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=03(O) Atr=01(Isoc) MxPS=  17 Ivl=1ms
E:  Ad=83(I) Atr=01(Isoc) MxPS=  17 Ivl=1ms
I:  If#= 1 Alt= 3 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=03(O) Atr=01(Isoc) MxPS=  25 Ivl=1ms
E:  Ad=83(I) Atr=01(Isoc) MxPS=  25 Ivl=1ms
I:  If#= 1 Alt= 4 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=03(O) Atr=01(Isoc) MxPS=  33 Ivl=1ms
E:  Ad=83(I) Atr=01(Isoc) MxPS=  33 Ivl=1ms
I:  If#= 1 Alt= 5 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=03(O) Atr=01(Isoc) MxPS=  49 Ivl=1ms
E:  Ad=83(I) Atr=01(Isoc) MxPS=  49 Ivl=1ms

Signed-off-by: Helmut Grohne <helmut@subdivi.de>
Link: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1955351
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
2 years agoBluetooth: hci_bcm: add BCM43430A0 & BCM43430A1
Luca Weiss [Fri, 25 Feb 2022 20:41:38 +0000 (21:41 +0100)]
Bluetooth: hci_bcm: add BCM43430A0 & BCM43430A1

Add the compatible for BCM43430A0 bluetooth used in lg-lenok and
BCM43430A1 used in asus-sparrow.

Signed-off-by: Luca Weiss <luca@z3ntu.xyz>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
2 years agoBluetooth: use memset avoid memory leaks
Minghao Chi (CGEL ZTE) [Fri, 25 Feb 2022 07:41:52 +0000 (07:41 +0000)]
Bluetooth: use memset avoid memory leaks

Use memset to initialize structs to prevent memory leaks
in l2cap_ecred_connect

Reported-by: Zeal Robot <zealci@zte.com.cn>
Signed-off-by: Minghao Chi (CGEL ZTE) <chi.minghao@zte.com.cn>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
2 years agoBluetooth: btmtksdio: Fix kernel oops when sdio suspend.
Yake Yang [Tue, 22 Feb 2022 23:56:00 +0000 (07:56 +0800)]
Bluetooth: btmtksdio: Fix kernel oops when sdio suspend.

Do not kfree the skb when the skb failed to send with __hci_cmd_sync

2435.241210] Call trace:
[ 2435.241224]  kfree_skb+0x20/0x140
[ 2435.241240]  btmtksdio_sdio_wakeup+0x8c/0xcc
[ 2435.241360]  hci_suspend_notifier+0x108/0x184
[ 2435.241377]  notifier_call_chain_robust+0x58/0xd8
[ 2435.241390]  blocking_notifier_call_chain_robust+0x54/0x84
[ 2435.241402]  pm_notifier_call_chain_robust+0x2c/0x4c
[ 2435.241414]  suspend_prepare+0x40/0x238
[ 2435.241425]  enter_state+0xec/0x3a8
[ 2435.241436]  pm_suspend+0x60/0xcc
[ 2435.241446]  state_store+0xb8/0x114
[ 2435.241460]  kobj_attr_store+0x18/0x2c
[ 2435.241473]  sysfs_kf_write+0x44/0x58
[ 2435.241486]  kernfs_fop_write_iter+0xf4/0x190
[ 2435.241500]  vfs_write+0x2b0/0x2e4
[ 2435.241511]  ksys_write+0x80/0xec
[ 2435.241523]  __arm64_sys_write+0x24/0x30
[ 2435.241538]  el0_svc_common+0xf0/0x1d8
[ 2435.241550]  do_el0_svc_compat+0x28/0x54
[ 2435.241564]  el0_svc_compat+0x10/0x1c
[ 2435.241574]  el0_sync_compat_handler+0xa8/0xcc
[ 2435.241586]  el0_sync_compat+0x188/0x1c0
[ 2435.241606] Code: a9014ff4 910003fd b40002c0 aa0003f3 (b84d4c08)
[ 2435.241619] ---[ end trace d496539b850baf14 ]---

Fixes: ce64b3e94919  ("Bluetooth: mt7921s: Support wake on bluetooth")
Co-developed-by: Sean Wang <sean.wang@mediatek.com>
Signed-off-by: Sean Wang <sean.wang@mediatek.com>
Signed-off-by: Yake Yang <yake.yang@mediatek.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
2 years agoBluetooth: btusb: Add a new PID/VID 13d3/3567 for MT7921
Yake Yang [Tue, 22 Feb 2022 23:55:59 +0000 (07:55 +0800)]
Bluetooth: btusb: Add a new PID/VID 13d3/3567 for MT7921

Add VID 13D3 & PID 3567 for MediaTek MT7921 USB Bluetooth chip.

The information in /sys/kernel/debug/usb/devices about the Bluetooth
device is listed as the below.

T:  Bus=05 Lev=01 Prnt=01 Port=00 Cnt=01 Dev#=  2 Spd=480  MxCh= 0
D:  Ver= 2.10 Cls=ef(misc ) Sub=02 Prot=01 MxPS=64 #Cfgs=  1
P:  Vendor=13d3 ProdID=3567 Rev= 1.00
S:  Manufacturer=MediaTek Inc.
S:  Product=Wireless_Device
S:  SerialNumber=000000000
C:* #Ifs= 3 Cfg#= 1 Atr=e0 MxPwr=100mA
A:  FirstIf#= 0 IfCount= 3 Cls=e0(wlcon) Sub=01 Prot=01
I:* If#= 0 Alt= 0 #EPs= 3 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=81(I) Atr=03(Int.) MxPS=  16 Ivl=125us
E:  Ad=82(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=02(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
I:* If#= 1 Alt= 0 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=83(I) Atr=01(Isoc) MxPS=   0 Ivl=1ms
E:  Ad=03(O) Atr=01(Isoc) MxPS=   0 Ivl=1ms
I:  If#= 1 Alt= 1 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=83(I) Atr=01(Isoc) MxPS=   9 Ivl=1ms
E:  Ad=03(O) Atr=01(Isoc) MxPS=   9 Ivl=1ms
I:  If#= 1 Alt= 2 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=83(I) Atr=01(Isoc) MxPS=  17 Ivl=1ms
E:  Ad=03(O) Atr=01(Isoc) MxPS=  17 Ivl=1ms
I:  If#= 1 Alt= 3 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=83(I) Atr=01(Isoc) MxPS=  25 Ivl=1ms
E:  Ad=03(O) Atr=01(Isoc) MxPS=  25 Ivl=1ms
I:  If#= 1 Alt= 4 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=83(I) Atr=01(Isoc) MxPS=  33 Ivl=1ms
E:  Ad=03(O) Atr=01(Isoc) MxPS=  33 Ivl=1ms
I:  If#= 1 Alt= 5 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=83(I) Atr=01(Isoc) MxPS=  49 Ivl=1ms
E:  Ad=03(O) Atr=01(Isoc) MxPS=  49 Ivl=1ms
I:  If#= 1 Alt= 6 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=83(I) Atr=01(Isoc) MxPS=  63 Ivl=1ms
E:  Ad=03(O) Atr=01(Isoc) MxPS=  63 Ivl=1ms
I:* If#= 2 Alt= 0 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=(none)
E:  Ad=8a(I) Atr=03(Int.) MxPS=  64 Ivl=125us
E:  Ad=0a(O) Atr=03(Int.) MxPS=  64 Ivl=125us
I:  If#= 2 Alt= 1 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=(none)
E:  Ad=8a(I) Atr=03(Int.) MxPS=  64 Ivl=125us
E:  Ad=0a(O) Atr=03(Int.) MxPS=  64 Ivl=125us

Co-developed-by: Sean Wang <sean.wang@mediatek.com>
Signed-off-by: Sean Wang <sean.wang@mediatek.com>
Signed-off-by: Yake Yang <yake.yang@mediatek.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
2 years agoBluetooth: move adv_instance_cnt read within the device lock
Niels Dossche [Sat, 12 Feb 2022 23:14:30 +0000 (00:14 +0100)]
Bluetooth: move adv_instance_cnt read within the device lock

The field adv_instance_cnt is always accessed within a device lock,
except in the function add_advertising. A concurrent remove of an
advertisement with adding another one could result in the if check
"if a new instance was actually added" to not trigger, resulting
in not triggering the "advertising added event".

Signed-off-by: Niels Dossche <niels.dossche@ugent.be>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
2 years agoBluetooth: hci_event: Add missing locking on hdev in hci_le_ext_adv_term_evt
Niels Dossche [Wed, 9 Feb 2022 14:08:51 +0000 (15:08 +0100)]
Bluetooth: hci_event: Add missing locking on hdev in hci_le_ext_adv_term_evt

Both hci_find_adv_instance and hci_remove_adv_instance have a comment
above their function definition saying that these two functions require
the caller to hold the hdev->lock lock. However, hci_le_ext_adv_term_evt
does not acquire that lock and neither does its caller hci_le_meta_evt
(hci_le_meta_evt calls hci_le_ext_adv_term_evt via an indirect function
call because of the lookup in hci_le_ev_table).

The other event handlers all acquire and release the hdev->lock and they
follow the rule that hci_find_adv_instance and hci_remove_adv_instance
must be called while holding the hdev->lock lock.

The solution is to make sure hci_le_ext_adv_term_evt also acquires and
releases the hdev->lock lock. The check on ev->status which logs a
warning and does an early return is not covered by the lock because
other functions also access ev->status without holding the lock.

Signed-off-by: Niels Dossche <niels.dossche@ugent.be>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
2 years agoBluetooth: btusb: Make use of of BIT macro to declare flags
Luiz Augusto von Dentz [Thu, 3 Mar 2022 22:17:07 +0000 (14:17 -0800)]
Bluetooth: btusb: Make use of of BIT macro to declare flags

This makes use of BIT macro when declaring BTUSB flags which makes it
understand how many bits are in use and many are left.

Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
2 years agoBluetooth: Fix not checking for valid hdev on bt_dev_{info,warn,err,dbg}
Luiz Augusto von Dentz [Thu, 3 Mar 2022 21:11:57 +0000 (13:11 -0800)]
Bluetooth: Fix not checking for valid hdev on bt_dev_{info,warn,err,dbg}

This fixes attemting to print hdev->name directly which causes them to
print an error:

kernel: read_version:367: (efault): sock 000000006a3008f2

Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
2 years agoBluetooth: mediatek: fix the conflict between mtk and msft vendor event
Sean Wang [Tue, 8 Feb 2022 18:17:41 +0000 (02:17 +0800)]
Bluetooth: mediatek: fix the conflict between mtk and msft vendor event

There is a conflict between MediaTek wmt event and msft vendor extension
logic in the core layer since 145373cb1b1f ("Bluetooth: Add framework for
Microsoft vendor extension") was introduced because we changed the type of
mediatek wmt event to the type of msft vendor event in the driver.

But the purpose we reported mediatek event to the core layer is for the
diagnostic purpose with that we are able to see the full packet trace via
monitoring socket with btmon. Thus, it is harmless we keep the original
type of mediatek vendor event here to avoid breaking the msft extension
function especially they can be supported by Mediatek chipset like MT7921
, MT7922 devices and future devices.

Signed-off-by: Sean Wang <sean.wang@mediatek.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
2 years agoBluetooth: mt7921s: support bluetooth reset mechanism
Chih-Ying Chiang [Tue, 8 Feb 2022 18:17:40 +0000 (02:17 +0800)]
Bluetooth: mt7921s: support bluetooth reset mechanism

Implement .cmd_timeout to reset the MT7921s device via a dedicated GPIO
pin when the firmware hang or the command has no response.

Co-developed-by: Mark Chen <mark-yw.chen@mediatek.com>
Signed-off-by: Mark Chen <mark-yw.chen@mediatek.com>
Co-developed-by: Sean Wang <sean.wang@mediatek.com>
Signed-off-by: Sean Wang <sean.wang@mediatek.com>
Signed-off-by: Chih-Ying Chiang <chih-yin.chiang@mediatek.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
2 years agoBluetooth: make array bt_uuid_any static const
Colin Ian King [Mon, 14 Feb 2022 21:51:30 +0000 (21:51 +0000)]
Bluetooth: make array bt_uuid_any static const

Don't populate the read-only array bt_uuid_any on the stack but
instead make it static const. Also makes the object code a little
smaller.

Signed-off-by: Colin Ian King <colin.i.king@gmail.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
2 years agoBluetooth: 6lowpan: No need to clear memory twice
Christophe JAILLET [Sun, 13 Feb 2022 21:30:47 +0000 (22:30 +0100)]
Bluetooth: 6lowpan: No need to clear memory twice

'peer_addr' is a structure embedded in 'struct lowpan_peer'. So there is no
need to explicitly call memset(0) on it. It is already zeroed by kzalloc()
when 'peer' is allocated.

Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
2 years agoBluetooth: btusb: Improve stability for QCA devices
Zijun Hu [Mon, 14 Feb 2022 10:43:39 +0000 (18:43 +0800)]
Bluetooth: btusb: Improve stability for QCA devices

WCN6855 2.1 will reset to apply firmware downloaded, so wait
a moment for reset done then go ahead to improve stability.

Signed-off-by: Zijun Hu <quic_zijuhu@quicinc.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
2 years agoBluetooth: btusb: add support for LG LGSBWAC02 (MT7663BUN)
Piotr Dymacz [Fri, 11 Feb 2022 09:31:03 +0000 (10:31 +0100)]
Bluetooth: btusb: add support for LG LGSBWAC02 (MT7663BUN)

The LG LGSBWAC02 (FCC ID: BEJLGSBWAC02) is a combo Wi-Fi/BT module which
can be found in several different LG TV models. This module is based on
already supported MediaTek MT7663U series.

Device from /sys/kernel/debug/usb/devices:

T:  Bus=01 Lev=02 Prnt=02 Port=03 Cnt=02 Dev#=  4 Spd=480  MxCh= 0
D:  Ver= 2.10 Cls=ef(misc ) Sub=02 Prot=01 MxPS=64 #Cfgs=  1
P:  Vendor=043e ProdID=310c Rev= 1.00
S:  Manufacturer=MediaTek Inc.
S:  Product=Wireless_Device
S:  SerialNumber=000000000
C:* #Ifs= 3 Cfg#= 1 Atr=a0 MxPwr=100mA
A:  FirstIf#= 0 IfCount= 2 Cls=e0(wlcon) Sub=01 Prot=01
I:* If#= 0 Alt= 0 #EPs= 3 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=81(I) Atr=03(Int.) MxPS=  16 Ivl=125us
E:  Ad=82(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=02(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
I:* If#= 1 Alt= 0 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=83(I) Atr=01(Isoc) MxPS=   0 Ivl=1ms
E:  Ad=03(O) Atr=01(Isoc) MxPS=   0 Ivl=1ms
I:  If#= 1 Alt= 1 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=83(I) Atr=01(Isoc) MxPS=   9 Ivl=1ms
E:  Ad=03(O) Atr=01(Isoc) MxPS=   9 Ivl=1ms
I:  If#= 1 Alt= 2 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=83(I) Atr=01(Isoc) MxPS=  17 Ivl=1ms
E:  Ad=03(O) Atr=01(Isoc) MxPS=  17 Ivl=1ms
I:  If#= 1 Alt= 3 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=83(I) Atr=01(Isoc) MxPS=  25 Ivl=1ms
E:  Ad=03(O) Atr=01(Isoc) MxPS=  25 Ivl=1ms
I:  If#= 1 Alt= 4 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=83(I) Atr=01(Isoc) MxPS=  33 Ivl=1ms
E:  Ad=03(O) Atr=01(Isoc) MxPS=  33 Ivl=1ms
I:  If#= 1 Alt= 5 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=83(I) Atr=01(Isoc) MxPS=  49 Ivl=1ms
E:  Ad=03(O) Atr=01(Isoc) MxPS=  49 Ivl=1ms
I:  If#= 1 Alt= 6 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=83(I) Atr=01(Isoc) MxPS=  63 Ivl=1ms
E:  Ad=03(O) Atr=01(Isoc) MxPS=  63 Ivl=1ms
I:* If#= 2 Alt= 0 #EPs= 8 Cls=ff(vend.) Sub=ff Prot=ff Driver=mt7663u
E:  Ad=84(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=85(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=08(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=04(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=05(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=06(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=07(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=09(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms

Signed-off-by: Piotr Dymacz <pepe2k@gmail.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
2 years agoBluetooth: btusb: Add support for Intel Madison Peak (MsP2) device
Kiran K [Mon, 31 Jan 2022 16:32:35 +0000 (22:02 +0530)]
Bluetooth: btusb: Add support for Intel Madison Peak (MsP2) device

$ lsusb
Bus 003 Device 003: ID 8087:0035 Intel Corp.

$ cat /sys/kernel/debug/usb/devices
T:  Bus=03 Lev=01 Prnt=01 Port=09 Cnt=02 Dev#=  3 Spd=12   MxCh= 0
D:  Ver= 2.01 Cls=e0(wlcon) Sub=01 Prot=01 MxPS=64 #Cfgs=  1
P:  Vendor=8087 ProdID=0035 Rev= 0.00
C:* #Ifs= 2 Cfg#= 1 Atr=e0 MxPwr=100mA
I:* If#= 0 Alt= 0 #EPs= 3 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=81(I) Atr=03(Int.) MxPS=  64 Ivl=1ms
E:  Ad=02(O) Atr=02(Bulk) MxPS=  64 Ivl=0ms
E:  Ad=82(I) Atr=02(Bulk) MxPS=  64 Ivl=0ms
I:* If#= 1 Alt= 0 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=03(O) Atr=01(Isoc) MxPS=   0 Ivl=1ms
E:  Ad=83(I) Atr=01(Isoc) MxPS=   0 Ivl=1ms
I:  If#= 1 Alt= 1 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=03(O) Atr=01(Isoc) MxPS=   9 Ivl=1ms
E:  Ad=83(I) Atr=01(Isoc) MxPS=   9 Ivl=1ms
I:  If#= 1 Alt= 2 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=03(O) Atr=01(Isoc) MxPS=  17 Ivl=1ms
E:  Ad=83(I) Atr=01(Isoc) MxPS=  17 Ivl=1ms
I:  If#= 1 Alt= 3 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=03(O) Atr=01(Isoc) MxPS=  25 Ivl=1ms
E:  Ad=83(I) Atr=01(Isoc) MxPS=  25 Ivl=1ms
I:  If#= 1 Alt= 4 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=03(O) Atr=01(Isoc) MxPS=  33 Ivl=1ms
E:  Ad=83(I) Atr=01(Isoc) MxPS=  33 Ivl=1ms
I:  If#= 1 Alt= 5 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=03(O) Atr=01(Isoc) MxPS=  49 Ivl=1ms
E:  Ad=83(I) Atr=01(Isoc) MxPS=  49 Ivl=1ms
I:  If#= 1 Alt= 6 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=03(O) Atr=01(Isoc) MxPS=  63 Ivl=1ms
E:  Ad=83(I) Atr=01(Isoc) MxPS=  63 Ivl=1ms

Signed-off-by: Kiran K <kiran.k@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
2 years agoBluetooth: Improve skb handling in mgmt_device_connected()
Radoslaw Biernacki [Tue, 1 Feb 2022 20:10:33 +0000 (20:10 +0000)]
Bluetooth: Improve skb handling in mgmt_device_connected()

This patch introduce eir_skb_put_data() that can be used to simplify
operations on eir in goal of eliminating the necessity of intermediary
buffers.
eir_skb_put_data() is in pair to what eir_append_data() does with help of
eir_len, but without awkwardness when passing return value to skb_put() (as
it returns updated offset not size).

Signed-off-by: Radoslaw Biernacki <rad@semihalf.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
2 years agoBluetooth: Fix skb allocation in mgmt_remote_name() & mgmt_device_connected()
Radoslaw Biernacki [Tue, 1 Feb 2022 20:10:32 +0000 (20:10 +0000)]
Bluetooth: Fix skb allocation in mgmt_remote_name() & mgmt_device_connected()

This patch fixes skb allocation, as lack of space for ev might push skb
tail beyond its end.
Also introduce eir_precalc_len() that can be used instead of magic
numbers for similar eir operations on skb.

Fixes: cf1bce1de7eeb ("Bluetooth: mgmt: Make use of mgmt_send_event_skb in MGMT_EV_DEVICE_FOUND")
Fixes: e96741437ef0a ("Bluetooth: mgmt: Make use of mgmt_send_event_skb in MGMT_EV_DEVICE_CONNECTED")
Signed-off-by: Angela Czubak <acz@semihalf.com>
Signed-off-by: Marek Maslanka <mm@semihalf.com>
Signed-off-by: Radoslaw Biernacki <rad@semihalf.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
2 years agoBluetooth: mgmt: Remove unneeded variable
Minghao Chi [Tue, 18 Jan 2022 07:50:33 +0000 (07:50 +0000)]
Bluetooth: mgmt: Remove unneeded variable

Return value from mgmt_cmd_complete() directly instead
of taking this in another redundant variable.

Reported-by: Zeal Robot <zealci@zte.com.cn>
Signed-off-by: Minghao Chi <chi.minghao@zte.com.cn>
Signed-off-by: CGEL ZTE <cgel.zte@gmail.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
2 years agoBluetooth: hci_sync: fix undefined return of hci_disconnect_all_sync()
Tom Rix [Tue, 1 Feb 2022 17:42:56 +0000 (09:42 -0800)]
Bluetooth: hci_sync: fix undefined return of hci_disconnect_all_sync()

clang static analysis reports this problem
hci_sync.c:4428:2: warning: Undefined or garbage value
  returned to caller
        return err;
        ^~~~~~~~~~

If there are no connections this function is a noop but
err is never set and a false error could be reported.
Return 0 as other hci_* functions do.

Fixes: 182ee45da083 ("Bluetooth: hci_sync: Rework hci_suspend_notifier")
Signed-off-by: Tom Rix <trix@redhat.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
2 years agoBluetooth: mgmt: Replace zero-length array with flexible-array member
Changcheng Deng [Fri, 21 Jan 2022 01:35:08 +0000 (01:35 +0000)]
Bluetooth: mgmt: Replace zero-length array with flexible-array member

There is a regular need in the kernel to provide a way to declare having
a dynamically sized set of trailing elements in a structure. Kernel code
should always use "flexible array members" for these cases. The older
style of one-element or zero-length arrays should no longer be used.
Reference:
https://www.kernel.org/doc/html/latest/process/deprecated.html#zero-length-and-one-element-arrays

Reported-by: Zeal Robot <zealci@zte.com.cn>
Signed-off-by: Changcheng Deng <deng.changcheng@zte.com.cn>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
2 years agonet: ethernet: sun: Remove redundant code
Jiapeng Chong [Fri, 4 Mar 2022 08:36:53 +0000 (16:36 +0800)]
net: ethernet: sun: Remove redundant code

Since the starting value in the for loop is greater than or equal to 1,
the restriction is CAS_FLAG_REG_PLUS is in the file cassini.h is
defined as 0x1 by macro, and the for loop and if condition is not
satisfied, so the code here is redundant.

Clean up the following smatch warning:

drivers/net/ethernet/sun/cassini.c:3513 cas_start_dma() warn: we never
enter this loop.

drivers/net/ethernet/sun/cassini.c:1239 cas_init_rx_dma() warn: we never
enter this loop.

drivers/net/ethernet/sun/cassini.c:1247 cas_init_rx_dma() warn: we never
enter this loop.

Reported-by: Abaci Robot <abaci@linux.alibaba.com>
Signed-off-by: Jiapeng Chong <jiapeng.chong@linux.alibaba.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoMerge branch 'nfp-AF_XDP-zero-copy'
David S. Miller [Fri, 4 Mar 2022 13:06:12 +0000 (13:06 +0000)]
Merge branch 'nfp-AF_XDP-zero-copy'

Simon Horman says:

====================
Add AF_XDP zero-copy support for NFP

Niklas Söderlund says:

This series adds AF_XDP zero-copy support for the NFP driver. The series
is based on previous work done by Jakub Kicinski.

Patch 1/5 and 2/5 prepares the driver for AF_XDP support by refactoring
functions that will act differently once AF_XDP is active or not making
the driver easier to read and by preparing some functions to be reused
outside the local file scope. Patch 3/5 and 4/5 prepares the driver for
dealing the UMEM while finally patch 5/5 adds AF_XDP support.

Based on work by Jakub Kicinski.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonfp: xsk: add AF_XDP zero-copy Rx and Tx support
Niklas Söderlund [Fri, 4 Mar 2022 10:22:14 +0000 (11:22 +0100)]
nfp: xsk: add AF_XDP zero-copy Rx and Tx support

This patch adds zero-copy Rx and Tx support for AF_XDP sockets. It do so
by adding a separate NAPI poll function that is attached to a each
channel when the XSK socket is attached with XDP_SETUP_XSK_POOL, and
restored when the XSK socket is terminated, this is done per channel.

Support for XDP_TX is implemented and the XDP buffer can safely be moved
from the Rx to the Tx queue and correctly freed and returned to the XSK
pool once it's transmitted.

Note that when AF_XDP zero-copy is enabled, the XDP action XDP_PASS
will allocate a new buffer and copy the zero-copy frame prior
passing it to the kernel stack.

This patch is based on previous work by Jakub Kicinski.

Signed-off-by: Niklas Söderlund <niklas.soderlund@corigine.com>
Signed-off-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonfp: xsk: add configuration check for XSK socket chunk size
Niklas Söderlund [Fri, 4 Mar 2022 10:22:13 +0000 (11:22 +0100)]
nfp: xsk: add configuration check for XSK socket chunk size

In preparation for adding AF_XDP support add a configuration check to
make sure the buffer size can not be set to a larger value then the XSK
socket chunk size.

Signed-off-by: Niklas Söderlund <niklas.soderlund@corigine.com>
Signed-off-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonfp: xsk: add an array of xsk buffer pools to each data path
Niklas Söderlund [Fri, 4 Mar 2022 10:22:12 +0000 (11:22 +0100)]
nfp: xsk: add an array of xsk buffer pools to each data path

Each data path needs an array of xsk pools to track if an xsk socket is
in use. Add this array and make sure it's handled correctly when the
data path is duplicated.

Signed-off-by: Niklas Söderlund <niklas.soderlund@corigine.com>
Signed-off-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonfp: wrap napi add/del logic
Jakub Kicinski [Fri, 4 Mar 2022 10:22:11 +0000 (11:22 +0100)]
nfp: wrap napi add/del logic

There will be more NAPI register logic once AF_XDP support is
added, wrap our already conditional napi add/del into helpers.

Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: Niklas Söderlund <niklas.soderlund@corigine.com>
Signed-off-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonfp: expose common functions to be used for AF_XDP
Niklas Söderlund [Fri, 4 Mar 2022 10:22:10 +0000 (11:22 +0100)]
nfp: expose common functions to be used for AF_XDP

There are some common functionality that can be reused in the upcoming
AF_XDP support. Expose those functions in the header. While at it mark
some arguments of nfp_net_rx_csum() as const.

Signed-off-by: Niklas Söderlund <niklas.soderlund@corigine.com>
Signed-off-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoMerge branch 'sparx5-ptp'
David S. Miller [Fri, 4 Mar 2022 13:03:10 +0000 (13:03 +0000)]
Merge branch 'sparx5-ptp'

Horatiu Vultur says:

====================
net: sparx5: Add PTP Hardware Clock support

Add support for PTP Hardware Clock (PHC) for sparx5.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: sparx5: Implement get_ts_info
Horatiu Vultur [Fri, 4 Mar 2022 11:09:00 +0000 (12:09 +0100)]
net: sparx5: Implement get_ts_info

Implement the function get_ts_info in ethtool_ops which is needed to get
the HW capabilities for timestamping.

Signed-off-by: Horatiu Vultur <horatiu.vultur@microchip.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: sparx5: Add support for ptp interrupts
Horatiu Vultur [Fri, 4 Mar 2022 11:08:59 +0000 (12:08 +0100)]
net: sparx5: Add support for ptp interrupts

When doing 2-step timestamping the HW will generate an interrupt when it
managed to timestamp a frame. It is the SW responsibility to read it
from the FIFO.

Signed-off-by: Horatiu Vultur <horatiu.vultur@microchip.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: sparx5: Update extraction/injection for timestamping
Horatiu Vultur [Fri, 4 Mar 2022 11:08:58 +0000 (12:08 +0100)]
net: sparx5: Update extraction/injection for timestamping

Update both the extraction and injection to do timestamping of the
frames. The extraction is always doing the timestamping while for
injection is doing the timestamping only if it is configured.

Signed-off-by: Horatiu Vultur <horatiu.vultur@microchip.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: sparx5: Implement SIOCSHWTSTAMP and SIOCGHWTSTAMP
Horatiu Vultur [Fri, 4 Mar 2022 11:08:57 +0000 (12:08 +0100)]
net: sparx5: Implement SIOCSHWTSTAMP and SIOCGHWTSTAMP

Implement the ioctl callbacks SIOCSHWTSTAMP and SIOCGHWTSTAMP to allow
to configure the ports to enable/disable timestamping for TX. The RX
timestamping is always enabled. The HW is capable to run both 1-step
timestamping and 2-step timestamping.

Signed-off-by: Horatiu Vultur <horatiu.vultur@microchip.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: sparx5: Add support for ptp clocks
Horatiu Vultur [Fri, 4 Mar 2022 11:08:56 +0000 (12:08 +0100)]
net: sparx5: Add support for ptp clocks

The sparx5 has 3 PHC. Enable each of them, for now all the
timestamping is happening on the first PHC.

Signed-off-by: Horatiu Vultur <horatiu.vultur@microchip.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: sparx5: Add registers that are used by ptp functionality
Horatiu Vultur [Fri, 4 Mar 2022 11:08:55 +0000 (12:08 +0100)]
net: sparx5: Add registers that are used by ptp functionality

Add the registers that will be used to configure the PHC in the HW.

Signed-off-by: Horatiu Vultur <horatiu.vultur@microchip.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agodts: sparx5: Enable ptp interrupt
Horatiu Vultur [Fri, 4 Mar 2022 11:08:54 +0000 (12:08 +0100)]
dts: sparx5: Enable ptp interrupt

Add support for ptp interrupt. This interrupt is used when using 2-step
timestamping. For each timestamp that is added in a queue, an interrupt
is generated.

Signed-off-by: Horatiu Vultur <horatiu.vultur@microchip.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agodt-bindings: net: sparx5: Extend with the ptp interrupt
Horatiu Vultur [Fri, 4 Mar 2022 11:08:53 +0000 (12:08 +0100)]
dt-bindings: net: sparx5: Extend with the ptp interrupt

Extend dt-bindings for sparx5 with ptp interrupt. This is generated
when doing 2-step timestamping and the timestamp can be read from the
FIFO.

Signed-off-by: Horatiu Vultur <horatiu.vultur@microchip.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: sparx5: Move ifh from port to local variable
Horatiu Vultur [Fri, 4 Mar 2022 11:08:52 +0000 (12:08 +0100)]
net: sparx5: Move ifh from port to local variable

Currently the ifh is not changed, it is fixed for each frame for each
port that is sent out. Move this on the stack because this ifh needs to
be change based on the frames that are send out. This is needed for PTP
frames.

Signed-off-by: Horatiu Vultur <horatiu.vultur@microchip.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoMerge branch 'lan937x-t1-phy-driver'
David S. Miller [Fri, 4 Mar 2022 12:45:19 +0000 (12:45 +0000)]
Merge branch 'lan937x-t1-phy-driver'

Arun Ramadoss says:

====================
Add support for LAN937x T1 Phy Driver

LAN937x is a Multi-port 100Base-T1 Switch and it internally uses LAN87xx
T1 Phy.  This series of patch update the initialization routine for the
LAN87xx phy and also add LAN937x part support. Added the T1 Phy
master-slave configuration through ethtool.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: phy: added ethtool master-slave configuration support
Arun Ramadoss [Fri, 4 Mar 2022 09:44:01 +0000 (15:14 +0530)]
net: phy: added ethtool master-slave configuration support

To configure the T1 phy as master or slave using the ethtool -s <dev>
master-slave <forced-master/forced-slave>, the config_aneg and read
status functions are added.

Signed-off-by: Prasanna Vengateshan <prasanna.vengateshan@microchip.com>
Signed-off-by: Arun Ramadoss <arun.ramadoss@microchip.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: phy: added the LAN937x phy support
Arun Ramadoss [Fri, 4 Mar 2022 09:44:00 +0000 (15:14 +0530)]
net: phy: added the LAN937x phy support

LAN937x T1 switch is based on LAN87xx Phy, so reusing the init script of
the LAN87xx. There is a workaround in accessing the DSP bank register
for LAN937x Phy. Whenever there is a bank switch to DSP registers, then
we need a one dummy read access before proceeding to the actual register
access.

Signed-off-by: Prasanna Vengateshan <prasanna.vengateshan@microchip.com>
Signed-off-by: Arun Ramadoss <arun.ramadoss@microchip.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: phy: updated the initialization routine for LAN87xx
Arun Ramadoss [Fri, 4 Mar 2022 09:43:59 +0000 (15:13 +0530)]
net: phy: updated the initialization routine for LAN87xx

The new initialization sequence is the improvement to the existing init
routine. Init routine does soft reset, run init script and set
Hw_init. Added the new access_smi_poll_timeout() for polling smi
bank write.

Signed-off-by: Prasanna Vengateshan <prasanna.vengateshan@microchip.com>
Signed-off-by: Arun Ramadoss <arun.ramadoss@microchip.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: phy: removed empty lines in LAN87XX
Arun Ramadoss [Fri, 4 Mar 2022 09:43:58 +0000 (15:13 +0530)]
net: phy: removed empty lines in LAN87XX

Removed the empty lines in struct phy_drivers.

Signed-off-by: Arun Ramadoss <arun.ramadoss@microchip.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: phy: used the PHY_ID_MATCH_MODEL macro for LAN87XX
Arun Ramadoss [Fri, 4 Mar 2022 09:43:57 +0000 (15:13 +0530)]
net: phy: used the PHY_ID_MATCH_MODEL macro for LAN87XX

Used the PHY_ID_MATCH_MODEL MACRO for describing the phy_id and
phy_id_mask.

Signed-off-by: Arun Ramadoss <arun.ramadoss@microchip.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: phy: used genphy_soft_reset for phy reset in LAN87xx
Arun Ramadoss [Fri, 4 Mar 2022 09:43:56 +0000 (15:13 +0530)]
net: phy: used genphy_soft_reset for phy reset in LAN87xx

Replaced the current code of resetting of LAN87xx phy to
genphy_soft_reset library function.

Signed-off-by: Arun Ramadoss <arun.ramadoss@microchip.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoMerge branch 'lan8814-1588-support'
David S. Miller [Fri, 4 Mar 2022 12:43:01 +0000 (12:43 +0000)]
Merge branch 'lan8814-1588-support'

Divya Koppera says:

====================
Add support for 1588 in LAN8814

The following patch series contains:
- Fix for concurrent register access, which provides
  atomic access to extended page register reads/writes.
- Provides dt-bindings related to latency and timestamping
  that are required for LAN8814 phy.
- 1588 hardware timestamping support in LAN8814 phy.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: phy: micrel: 1588 support for LAN8814 phy
Divya Koppera [Fri, 4 Mar 2022 09:34:18 +0000 (15:04 +0530)]
net: phy: micrel: 1588 support for LAN8814 phy

Add support for 1588 in LAN8814 phy driver.
It supports 1-step and 2-step timestamping.

Co-developed-by: Horatiu Vultur <horatiu.vultur@microchip.com>
Signed-off-by: Divya Koppera <Divya.Koppera@microchip.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agodt-bindings: net: micrel: Configure latency values and timestamping check for LAN8814 phy
Divya Koppera [Fri, 4 Mar 2022 09:34:17 +0000 (15:04 +0530)]
dt-bindings: net: micrel: Configure latency values and timestamping check for LAN8814 phy

Supports configuring latency values and also adds
check for phy timestamping feature.

Signed-off-by: Divya Koppera<Divya.Koppera@microchip.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: phy: micrel: Fix concurrent register access
Divya Koppera [Fri, 4 Mar 2022 09:34:16 +0000 (15:04 +0530)]
net: phy: micrel: Fix concurrent register access

Make Extended page register accessing atomic,
to overcome unexpected output from register
reads/writes.

Fixes: 7c2dcfa295b1 ("net: phy: micrel: Add support for LAN8804 PHY")
Signed-off-by: Divya Koppera<Divya.Koppera@microchip.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoMerge branch 'skb-drop-reasons'
David S. Miller [Fri, 4 Mar 2022 12:17:11 +0000 (12:17 +0000)]
Merge branch 'skb-drop-reasons'

Menglong Dong says:

====================
net: dev: add skb drop reasons to net/core/dev.c

In the commit c504e5c2f964 ("net: skb: introduce kfree_skb_reason()"),
we added the support of reporting the reasons of skb drops to kfree_skb
tracepoint. And in this series patches, reasons for skb drops are added
to the link layer, which means that 'net/core/dev.c' is our target.

Following functions are processed:

sch_handle_egress()
__dev_xmit_skb()
enqueue_to_backlog()
do_xdp_generic()
sch_handle_ingress()
__netif_receive_skb_core()

and following new drop reasons are added (what they mean can be see in
the document of them):

SKB_DROP_REASON_QDISC_EGRESS
SKB_DROP_REASON_QDISC_DROP
SKB_DROP_REASON_CPU_BACKLOG
SKB_DROP_REASON_XDP
SKB_DROP_REASON_QDISC_INGRESS
SKB_DROP_REASON_PTYPE_ABSENT

In order to add skb drop reasons to kfree_skb_list(), the function
kfree_skb_list_reason() is introduced in the 2th patch, which will be
used in __dev_xmit_skb() in the 3th patch.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: dev: use kfree_skb_reason() for __netif_receive_skb_core()
Menglong Dong [Fri, 4 Mar 2022 06:00:46 +0000 (14:00 +0800)]
net: dev: use kfree_skb_reason() for __netif_receive_skb_core()

Add reason for skb drops to __netif_receive_skb_core() when packet_type
not found to handle the skb. For this purpose, the drop reason
SKB_DROP_REASON_PTYPE_ABSENT is introduced. Take ether packets for
example, this case mainly happens when L3 protocol is not supported.

Signed-off-by: Menglong Dong <imagedong@tencent.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: dev: use kfree_skb_reason() for sch_handle_ingress()
Menglong Dong [Fri, 4 Mar 2022 06:00:45 +0000 (14:00 +0800)]
net: dev: use kfree_skb_reason() for sch_handle_ingress()

Replace kfree_skb() used in sch_handle_ingress() with
kfree_skb_reason(). Following drop reasons are introduced:

SKB_DROP_REASON_TC_INGRESS

Signed-off-by: Menglong Dong <imagedong@tencent.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: dev: use kfree_skb_reason() for do_xdp_generic()
Menglong Dong [Fri, 4 Mar 2022 06:00:44 +0000 (14:00 +0800)]
net: dev: use kfree_skb_reason() for do_xdp_generic()

Replace kfree_skb() used in do_xdp_generic() with kfree_skb_reason().
The drop reason SKB_DROP_REASON_XDP is introduced for this case.

Signed-off-by: Menglong Dong <imagedong@tencent.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: dev: use kfree_skb_reason() for enqueue_to_backlog()
Menglong Dong [Fri, 4 Mar 2022 06:00:43 +0000 (14:00 +0800)]
net: dev: use kfree_skb_reason() for enqueue_to_backlog()

Replace kfree_skb() used in enqueue_to_backlog() with
kfree_skb_reason(). The skb rop reason SKB_DROP_REASON_CPU_BACKLOG is
introduced for the case of failing to enqueue the skb to the per CPU
backlog queue. The further reason can be backlog queue full or RPS
flow limition, and I think we needn't to make further distinctions.

Signed-off-by: Menglong Dong <imagedong@tencent.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: dev: add skb drop reasons to __dev_xmit_skb()
Menglong Dong [Fri, 4 Mar 2022 06:00:42 +0000 (14:00 +0800)]
net: dev: add skb drop reasons to __dev_xmit_skb()

Add reasons for skb drops to __dev_xmit_skb() by replacing
kfree_skb_list() with kfree_skb_list_reason(). The drop reason of
SKB_DROP_REASON_QDISC_DROP is introduced for qdisc enqueue fails.

Signed-off-by: Menglong Dong <imagedong@tencent.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: skb: introduce the function kfree_skb_list_reason()
Menglong Dong [Fri, 4 Mar 2022 06:00:41 +0000 (14:00 +0800)]
net: skb: introduce the function kfree_skb_list_reason()

To report reasons of skb drops, introduce the function
kfree_skb_list_reason() and make kfree_skb_list() an inline call to
it. This function will be used in the next commit in
__dev_xmit_skb().

Signed-off-by: Menglong Dong <imagedong@tencent.com>
Signed-off-by: David S. Miller <davem@davemloft.net>