platform/kernel/linux-starfive.git
2 years agoptp: ocp: constify selectors
Jonathan Lemon [Thu, 19 May 2022 21:21:49 +0000 (14:21 -0700)]
ptp: ocp: constify selectors

The ocp selectors are all constant, so label them as such.

Signed-off-by: Jonathan Lemon <jonathan.lemon@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoptp: ocp: parameterize input/output sma selectors
Jonathan Lemon [Thu, 19 May 2022 21:21:48 +0000 (14:21 -0700)]
ptp: ocp: parameterize input/output sma selectors

Group the sma input/output tables together and select the correct
group from the bp information.  This allows adding new groups with
different sma mappings.

Signed-off-by: Jonathan Lemon <jonathan.lemon@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoptp: ocp: revise firmware display
Jonathan Lemon [Thu, 19 May 2022 21:21:47 +0000 (14:21 -0700)]
ptp: ocp: revise firmware display

Preparse the firmware image information into loader/tag/version,
and set the fw capabilities based on the tag/version.

Signed-off-by: Jonathan Lemon <jonathan.lemon@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoptp: ocp: add Celestica timecard PCI ids
Vadim Fedorenko [Thu, 19 May 2022 21:21:46 +0000 (14:21 -0700)]
ptp: ocp: add Celestica timecard PCI ids

Celestica is producing card with their own vendor id and device id.
Add these ids to driver to support this card.

Signed-off-by: Vadim Fedorenko <vadfed@fb.com>
Signed-off-by: Jonathan Lemon <jonathan.lemon@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoptp: ocp: Remove #ifdefs around PCI IDs
Jonathan Lemon [Thu, 19 May 2022 21:21:45 +0000 (14:21 -0700)]
ptp: ocp: Remove #ifdefs around PCI IDs

These #ifdefs are not required, so remove them.

Suggested-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Jonathan Lemon <jonathan.lemon@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoptp: ocp: 32-bit fixups for pci start address
Jonathan Lemon [Thu, 19 May 2022 21:21:44 +0000 (14:21 -0700)]
ptp: ocp: 32-bit fixups for pci start address

Use 'resource_size_t' instead of 'unsigned long' when computing the
pci start address, for the benefit of 32-bit platforms.

Signed-off-by: Jonathan Lemon <jonathan.lemon@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoRevert "net/smc: fix listen processing for SMC-Rv2"
liuyacan [Tue, 24 May 2022 09:02:30 +0000 (17:02 +0800)]
Revert "net/smc: fix listen processing for SMC-Rv2"

This reverts commit 8c3b8dc5cc9bf6d273ebe18b16e2d6882bcfb36d.

Some rollback issue will be fixed in other patches in the future.

Link: https://lore.kernel.org/all/20220523055056.2078994-1-liuyacan@corp.netease.com/
Fixes: 8c3b8dc5cc9b ("net/smc: fix listen processing for SMC-Rv2")
Signed-off-by: liuyacan <liuyacan@corp.netease.com>
Link: https://lore.kernel.org/r/20220524090230.2140302-1-liuyacan@corp.netease.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoath6kl: Use cc-disable-warning to disable -Wdangling-pointer
Nathan Chancellor [Tue, 24 May 2022 14:56:55 +0000 (07:56 -0700)]
ath6kl: Use cc-disable-warning to disable -Wdangling-pointer

Clang does not support this option so the build fails:

  error: unknown warning option '-Wno-dangling-pointer' [-Werror,-Wunknown-warning-option]

Use cc-disable-warning so that the option is only added when it is
supported.

Fixes: bd1d129daa3e ("wifi: ath6k: silence false positive -Wno-dangling-pointer warning on GCC 12")
Reported-by: "kernelci.org bot" <bot@kernelci.org>
Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Reviewed-by: Tom Rix <trix@redhat.com>
Link: https://lore.kernel.org/r/20220524145655.869822-1-nathan@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net
Jakub Kicinski [Tue, 24 May 2022 04:19:17 +0000 (21:19 -0700)]
Merge git://git./linux/kernel/git/netdev/net

drivers/net/ethernet/cadence/macb_main.c
  5cebb40bc955 ("net: macb: Fix PTP one step sync support")
  138badbc21a0 ("net: macb: use NAPI for TX completion path")
https://lore.kernel.org/all/20220523111021.31489367@canb.auug.org.au/

net/smc/af_smc.c
  75c1edf23b95 ("net/smc: postpone sk_refcnt increment in connect()")
  3aba103006bc ("net/smc: align the connect behaviour with TCP")
https://lore.kernel.org/all/20220524114408.4bf1af38@canb.auug.org.au/

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 [Mon, 23 May 2022 23:07:13 +0000 (16:07 -0700)]
Merge https://git./linux/kernel/git/bpf/bpf-next

Daniel Borkmann says:

====================
pull-request: bpf-next 2022-05-23

We've added 113 non-merge commits during the last 26 day(s) which contain
a total of 121 files changed, 7425 insertions(+), 1586 deletions(-).

The main changes are:

1) Speed up symbol resolution for kprobes multi-link attachments, from Jiri Olsa.

2) Add BPF dynamic pointer infrastructure e.g. to allow for dynamically sized ringbuf
   reservations without extra memory copies, from Joanne Koong.

3) Big batch of libbpf improvements towards libbpf 1.0 release, from Andrii Nakryiko.

4) Add BPF link iterator to traverse links via seq_file ops, from Dmitrii Dolgov.

5) Add source IP address to BPF tunnel key infrastructure, from Kaixi Fan.

6) Refine unprivileged BPF to disable only object-creating commands, from Alan Maguire.

7) Fix JIT blinding of ld_imm64 when they point to subprogs, from Alexei Starovoitov.

8) Add BPF access to mptcp_sock structures and their meta data, from Geliang Tang.

9) Add new BPF helper for access to remote CPU's BPF map elements, from Feng Zhou.

10) Allow attaching 64-bit cookie to BPF link of fentry/fexit/fmod_ret, from Kui-Feng Lee.

11) Follow-ups to typed pointer support in BPF maps, from Kumar Kartikeya Dwivedi.

12) Add busy-poll test cases to the XSK selftest suite, from Magnus Karlsson.

13) Improvements in BPF selftest test_progs subtest output, from Mykola Lysenko.

14) Fill bpf_prog_pack allocator areas with illegal instructions, from Song Liu.

15) Add generic batch operations for BPF map-in-map cases, from Takshak Chahande.

16) Make bpf_jit_enable more user friendly when permanently on 1, from Tiezhu Yang.

17) Fix an array overflow in bpf_trampoline_get_progs(), from Yuntao Wang.

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

Link: https://lore.kernel.org/r/20220523223805.27931-1-daniel@iogearbox.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoMerge tag 'linux-can-next-for-5.19-20220523' of git://git.kernel.org/pub/scm/linux...
Jakub Kicinski [Mon, 23 May 2022 22:59:41 +0000 (15:59 -0700)]
Merge tag 'linux-can-next-for-5.19-20220523' of git://git./linux/kernel/git/mkl/linux-can-next

Marc Kleine-Budde says:

====================
pull-request: can-next 2022-05-23

this is a pull request of 3 patches for net-next/master.

The first patch is by Julia Lawall and fixes a typo in the peak_usb
driver.

Jakub Kicinski contributes a patch for the kvaser_usb driver to
silence a GCC 12 -Warray-bounds warning.

The last patch is by me, target the ctucanfd platform driver and adds
a missing dependency to HAS_IOMEM.

* tag 'linux-can-next-for-5.19-20220523' of git://git.kernel.org/pub/scm/linux/kernel/git/mkl/linux-can-next:
  can: ctucanfd: platform: add missing dependency to HAS_IOMEM
  can: kvaser_usb: silence a GCC 12 -Warray-bounds warning
  can: peak_usb: fix typo in comment
====================

Link: https://lore.kernel.org/r/20220523201045.1708855-1-mkl@pengutronix.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoMerge tag 'for-net-next-2022-05-23' of git://git.kernel.org/pub/scm/linux/kernel...
Jakub Kicinski [Mon, 23 May 2022 21:36:36 +0000 (14:36 -0700)]
Merge tag 'for-net-next-2022-05-23' of git://git./linux/kernel/git/bluetooth/bluetooth-next

Luiz Augusto von Dentz says:

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

 - Add support for Realtek 8761BUV
 - Add HCI_QUIRK_BROKEN_ENHANCED_SETUP_SYNC_CONN quirk
 - Add support for RTL8852C
 - Add a new PID/VID 0489/e0c8 for MT7921
 - Add support for Qualcomm WCN785x

* tag 'for-net-next-2022-05-23' of git://git.kernel.org/pub/scm/linux/kernel/git/bluetooth/bluetooth-next: (26 commits)
  Bluetooth: hci_sync: use hci_skb_event() helper
  Bluetooth: eir: Add helpers for managing service data
  Bluetooth: hci_sync: Fix attempting to suspend with unfiltered passive scan
  Bluetooth: MGMT: Add conditions for setting HCI_CONN_FLAG_REMOTE_WAKEUP
  Bluetooth: btmtksdio: fix the reset takes too long
  Bluetooth: btmtksdio: fix possible FW initialization failure
  Bluetooth: btmtksdio: fix use-after-free at btmtksdio_recv_event
  Bluetooth: btbcm: Add entry for BCM4373A0 UART Bluetooth
  Bluetooth: btusb: Add a new PID/VID 0489/e0c8 for MT7921
  Bluetooth: btusb: Add 0x0bda:0x8771 Realtek 8761BUV devices
  Bluetooth: btusb: Set HCI_QUIRK_BROKEN_ERR_DATA_REPORTING for QCA
  Bluetooth: core: Fix missing power_on work cancel on HCI close
  Bluetooth: btusb: add support for Qualcomm WCN785x
  Bluetooth: protect le accept and resolv lists with hdev->lock
  Bluetooth: use hdev lock for accept_list and reject_list in conn req
  Bluetooth: use hdev lock in activate_scan for hci_is_adv_monitoring
  Bluetooth: btrtl: Add support for RTL8852C
  Bluetooth: btusb: Set HCI_QUIRK_BROKEN_ENHANCED_SETUP_SYNC_CONN for QCA
  Bluetooth: Print broken quirks
  Bluetooth: HCI: Add HCI_QUIRK_BROKEN_ENHANCED_SETUP_SYNC_CONN quirk
  ...
====================

Link: https://lore.kernel.org/r/20220523204151.3327345-1-luiz.dentz@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoMerge tag 'for-net-2022-05-23' of git://git.kernel.org/pub/scm/linux/kernel/git/bluet...
Jakub Kicinski [Mon, 23 May 2022 21:33:38 +0000 (14:33 -0700)]
Merge tag 'for-net-2022-05-23' of git://git./linux/kernel/git/bluetooth/bluetooth

Luiz Augusto von Dentz says:

====================
bluetooth pull request for net:

 - Fix crash when an LE Connection fails to be established.

* tag 'for-net-2022-05-23' of git://git.kernel.org/pub/scm/linux/kernel/git/bluetooth/bluetooth:
  Bluetooth: hci_conn: Fix hci_connect_le_sync
====================

Link: https://lore.kernel.org/r/20220523200349.3322806-1-luiz.dentz@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoMerge branch 'Dynamic pointers'
Andrii Nakryiko [Mon, 23 May 2022 21:30:18 +0000 (14:30 -0700)]
Merge branch 'Dynamic pointers'

Joanne Koong says:

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

This patchset implements the basics of dynamic pointers in bpf.

A dynamic pointer (struct bpf_dynptr) is a pointer that stores extra metadata
alongside the address it points to. This abstraction is useful in bpf given
that every memory access in a bpf program must be safe. The verifier and bpf
helper functions can use the metadata to enforce safety guarantees for things
such as dynamically sized strings and kernel heap allocations.

From the program side, the bpf_dynptr is an opaque struct and the verifier
will enforce that its contents are never written to by the program.
It can only be written to through specific bpf helper functions.

There are several uses cases for dynamic pointers in bpf programs. Some
examples include: dynamically sized ringbuf reservations without extra
memcpys, dynamic string parsing and memory comparisons, dynamic memory
allocations that can be persisted in maps, and dynamic + ergonomic parsing of
sk_buff and xdp_md packet data.

At a high-level, the patches are as follows:
1/6 - Adds verifier support for dynptrs
2/6 - Adds bpf_dynptr_from_mem (local dynptr)
3/6 - Adds dynptr support for ring buffers
4/6 - Adds bpf_dynptr_read and bpf_dynptr_write
5/6 - Adds dynptr data slices (ptr to the dynptr data)
6/6 - Tests to check that the verifier rejects invalid cases and passes valid ones

This is the first dynptr patchset in a larger series. The next series of
patches will add dynptrs that support dynamic memory allocations that can also
be persisted in maps, support for parsing packet data through dynptrs, convenience
helpers for using dynptrs as iterators, and more helper functions for interacting
with strings and memory dynamically.

Changelog:
----------
v5 -> v6:
v5:
https://lore.kernel.org/bpf/20220520044245.3305025-1-joannelkoong@gmail.com/
* enforce PTR_TO_MAP_VALUE for bpf_dynptr_from_mem data in check_helper_call
instead of using DYNPTR_TYPE_LOCAL when checking func arg compatiblity
* remove MEM_DYNPTR modifier

v4 -> v5:
v4:
https://lore.kernel.org/bpf/20220509224257.3222614-1-joannelkoong@gmail.com/
* Remove malloc dynptr; this will be part of the 2nd patchset while we
figure out memory accounting
* For data slices, only set the register's ref_obj_id to dynptr_id (Alexei)
* Tidying (eg remove "inline", move offset checking to "check_func_arg_reg_off")
(David)
* Add a few new test cases, remove malloc-only ones.

v3 -> v4:
v3:
https://lore.kernel.org/bpf/20220428211059.4065379-1-joannelkoong@gmail.com/
1/6 - Change mem ptr + size check to use more concise inequality expression
(David + Andrii)
2/6 - Add check for meta->uninit_dynptr_regno not already set (Andrii)
      Move DYNPTR_TYPE_FLAG_MASK to include/linux/bpf.h (Andrii)
3/6 - Remove four underscores for invoking BPF_CALL (Andrii)
      Add __BPF_TYPE_FLAG_MAX and use it for __BPF_TYPE_LAST_FLAG (Andrii)
4/6 - Fix capacity to be bpf_dynptr size value in check_off_len (Andrii)
      Change -EINVAL to -E2BIG if len + offset is out of bounds (Andrii)
5/6 - Add check for only 1 dynptr arg for dynptr data function (Andrii)
6/6 - For ringbuf map, set max_entries from userspace (Andrii)
      Use err ?: ... for interactring with dynptr APIs (Andrii)
      Define array_map2 for add_dynptr_to_map2 test where value is a struct
with an embedded dynptr
      Remove ref id from missing_put_callback message, since on different
environments, ref id is not always = 1

v2 -> v3:
v2:
https://lore.kernel.org/bpf/20220416063429.3314021-1-joannelkoong@gmail.com/
* Reorder patches (move ringbuffer patch to be right after the verifier +
* malloc
dynptr patchset)
* Remove local type dynptrs (Andrii + Alexei)
* Mark stack slot as STACK_MISC after any writes into a dynptr instead of
* explicitly prohibiting writes (Alexei)
* Pass number of slots, not memory size to is_spi_bounds_valid (Kumar)
* Check reference leaks by adding dynptr id to state->refs instead of checking
stack slots (Alexei)

v1 -> v2:
v1: https://lore.kernel.org/bpf/20220402015826.3941317-1-joannekoong@fb.com/
1/7 -
    * Remove ARG_PTR_TO_MAP_VALUE_UNINIT alias and use
      ARG_PTR_TO_MAP_VALUE | MEM_UNINIT directly (Andrii)
    * Drop arg_type_is_mem_ptr() wrapper function (Andrii)
2/7 -
    * Change name from MEM_RELEASE to OBJ_RELEASE (Andrii)
    * Use meta.release_ref instead of ref_obj_id != 0 to determine whether
      to release reference (Kumar)
    * Drop type_is_release_mem() wrapper function (Andrii)
3/7 -
    * Add checks for nested dynptrs edge-cases, which could lead to corrupt
    * writes of the dynptr stack variable.
    * Add u64 flags to bpf_dynptr_from_mem() and bpf_dynptr_alloc() (Andrii)
    * Rename from bpf_malloc/bpf_free to bpf_dynptr_alloc/bpf_dynptr_put
      (Alexei)
    * Support alloc flag __GFP_ZERO (Andrii)
    * Reserve upper 8 bits in dynptr size and offset fields instead of
      reserving just the upper 4 bits (Andrii)
    * Allow dynptr zero-slices (Andrii)
    * Use the highest bit for is_rdonly instead of the 28th bit (Andrii)
    * Rename check_* functions to is_* functions for better readability
      (Andrii)
    * Add comment for code that checks the spi bounds (Andrii)
4/7 -
    * Fix doc description for bpf_dynpt_read (Toke)
    * Move bpf_dynptr_check_off_len() from function patch 1 to here (Andrii)
5/7 -
    * When finding the id for the dynptr to associate the data slice with,
      look for dynptr arg instead of assuming it is BPF_REG_1.
6/7 -
    * Add __force when casting from unsigned long to void * (kernel test
    * robot)
    * Expand on docs for ringbuf dynptr APIs (Andrii)
7/7 -
    * Use table approach for defining test programs and error messages
    * (Andrii)
    * Print out full log if there’s an error (Andrii)
    * Use bpf_object__find_program_by_name() instead of specifying
      program name as a string (Andrii)
    * Add 6 extra cases: invalid_nested_dynptrs1, invalid_nested_dynptrs2,
      invalid_ref_mem1, invalid_ref_mem2, zero_slice_access,
      and test_alloc_zero_bytes
    * Add checking for edge cases (eg allocing with invalid flags)
====================

Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
2 years agoselftests/bpf: Dynptr tests
Joanne Koong [Mon, 23 May 2022 21:07:12 +0000 (14:07 -0700)]
selftests/bpf: Dynptr tests

This patch adds tests for dynptrs, which include cases that the
verifier needs to reject (for example, a bpf_ringbuf_reserve_dynptr
without a corresponding bpf_ringbuf_submit/discard_dynptr) as well
as cases that should successfully pass.

Signed-off-by: Joanne Koong <joannelkoong@gmail.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20220523210712.3641569-7-joannelkoong@gmail.com
2 years agobpf: Add dynptr data slices
Joanne Koong [Mon, 23 May 2022 21:07:11 +0000 (14:07 -0700)]
bpf: Add dynptr data slices

This patch adds a new helper function

void *bpf_dynptr_data(struct bpf_dynptr *ptr, u32 offset, u32 len);

which returns a pointer to the underlying data of a dynptr. *len*
must be a statically known value. The bpf program may access the returned
data slice as a normal buffer (eg can do direct reads and writes), since
the verifier associates the length with the returned pointer, and
enforces that no out of bounds accesses occur.

Signed-off-by: Joanne Koong <joannelkoong@gmail.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/20220523210712.3641569-6-joannelkoong@gmail.com
2 years agobpf: Add bpf_dynptr_read and bpf_dynptr_write
Joanne Koong [Mon, 23 May 2022 21:07:10 +0000 (14:07 -0700)]
bpf: Add bpf_dynptr_read and bpf_dynptr_write

This patch adds two helper functions, bpf_dynptr_read and
bpf_dynptr_write:

long bpf_dynptr_read(void *dst, u32 len, struct bpf_dynptr *src, u32 offset);

long bpf_dynptr_write(struct bpf_dynptr *dst, u32 offset, void *src, u32 len);

The dynptr passed into these functions must be valid dynptrs that have
been initialized.

Signed-off-by: Joanne Koong <joannelkoong@gmail.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20220523210712.3641569-5-joannelkoong@gmail.com
2 years agobpf: Dynptr support for ring buffers
Joanne Koong [Mon, 23 May 2022 21:07:09 +0000 (14:07 -0700)]
bpf: Dynptr support for ring buffers

Currently, our only way of writing dynamically-sized data into a ring
buffer is through bpf_ringbuf_output but this incurs an extra memcpy
cost. bpf_ringbuf_reserve + bpf_ringbuf_commit avoids this extra
memcpy, but it can only safely support reservation sizes that are
statically known since the verifier cannot guarantee that the bpf
program won’t access memory outside the reserved space.

The bpf_dynptr abstraction allows for dynamically-sized ring buffer
reservations without the extra memcpy.

There are 3 new APIs:

long bpf_ringbuf_reserve_dynptr(void *ringbuf, u32 size, u64 flags, struct bpf_dynptr *ptr);
void bpf_ringbuf_submit_dynptr(struct bpf_dynptr *ptr, u64 flags);
void bpf_ringbuf_discard_dynptr(struct bpf_dynptr *ptr, u64 flags);

These closely follow the functionalities of the original ringbuf APIs.
For example, all ringbuffer dynptrs that have been reserved must be
either submitted or discarded before the program exits.

Signed-off-by: Joanne Koong <joannelkoong@gmail.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: David Vernet <void@manifault.com>
Link: https://lore.kernel.org/bpf/20220523210712.3641569-4-joannelkoong@gmail.com
2 years agobpf: Add bpf_dynptr_from_mem for local dynptrs
Joanne Koong [Mon, 23 May 2022 21:07:08 +0000 (14:07 -0700)]
bpf: Add bpf_dynptr_from_mem for local dynptrs

This patch adds a new api bpf_dynptr_from_mem:

long bpf_dynptr_from_mem(void *data, u32 size, u64 flags, struct bpf_dynptr *ptr);

which initializes a dynptr to point to a bpf program's local memory. For now
only local memory that is of reg type PTR_TO_MAP_VALUE is supported.

Signed-off-by: Joanne Koong <joannelkoong@gmail.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20220523210712.3641569-3-joannelkoong@gmail.com
2 years agobpf: Add verifier support for dynptrs
Joanne Koong [Mon, 23 May 2022 21:07:07 +0000 (14:07 -0700)]
bpf: Add verifier support for dynptrs

This patch adds the bulk of the verifier work for supporting dynamic
pointers (dynptrs) in bpf.

A bpf_dynptr is opaque to the bpf program. It is a 16-byte structure
defined internally as:

struct bpf_dynptr_kern {
    void *data;
    u32 size;
    u32 offset;
} __aligned(8);

The upper 8 bits of *size* is reserved (it contains extra metadata about
read-only status and dynptr type). Consequently, a dynptr only supports
memory less than 16 MB.

There are different types of dynptrs (eg malloc, ringbuf, ...). In this
patchset, the most basic one, dynptrs to a bpf program's local memory,
is added. For now only local memory that is of reg type PTR_TO_MAP_VALUE
is supported.

In the verifier, dynptr state information will be tracked in stack
slots. When the program passes in an uninitialized dynptr
(ARG_PTR_TO_DYNPTR | MEM_UNINIT), the stack slots corresponding
to the frame pointer where the dynptr resides at are marked
STACK_DYNPTR. For helper functions that take in initialized dynptrs (eg
bpf_dynptr_read + bpf_dynptr_write which are added later in this
patchset), the verifier enforces that the dynptr has been initialized
properly by checking that their corresponding stack slots have been
marked as STACK_DYNPTR.

The 6th patch in this patchset adds test cases that the verifier should
successfully reject, such as for example attempting to use a dynptr
after doing a direct write into it inside the bpf program.

Signed-off-by: Joanne Koong <joannelkoong@gmail.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: David Vernet <void@manifault.com>
Link: https://lore.kernel.org/bpf/20220523210712.3641569-2-joannelkoong@gmail.com
2 years agobpf: Suppress 'passing zero to PTR_ERR' warning
Kumar Kartikeya Dwivedi [Sat, 21 May 2022 13:26:20 +0000 (18:56 +0530)]
bpf: Suppress 'passing zero to PTR_ERR' warning

Kernel Test Robot complains about passing zero to PTR_ERR for the said
line, suppress it by using PTR_ERR_OR_ZERO.

Fixes: c0a5a21c25f3 ("bpf: Allow storing referenced kptr in map")
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20220521132620.1976921-1-memxor@gmail.com
2 years agobpf: Introduce bpf_arch_text_invalidate for bpf_prog_pack
Song Liu [Fri, 20 May 2022 23:57:53 +0000 (16:57 -0700)]
bpf: Introduce bpf_arch_text_invalidate for bpf_prog_pack

Introduce bpf_arch_text_invalidate and use it to fill unused part of the
bpf_prog_pack with illegal instructions when a BPF program is freed.

Fixes: 57631054fae6 ("bpf: Introduce bpf_prog_pack allocator")
Fixes: 33c9805860e5 ("bpf: Introduce bpf_jit_binary_pack_[alloc|finalize|free]")
Reported-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Song Liu <song@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20220520235758.1858153-4-song@kernel.org
2 years agox86/alternative: Introduce text_poke_set
Song Liu [Fri, 20 May 2022 23:57:52 +0000 (16:57 -0700)]
x86/alternative: Introduce text_poke_set

Introduce a memset like API for text_poke. This will be used to fill the
unused RX memory with illegal instructions.

Suggested-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Song Liu <song@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/bpf/20220520235758.1858153-3-song@kernel.org
2 years agobpf: Fill new bpf_prog_pack with illegal instructions
Song Liu [Fri, 20 May 2022 23:57:51 +0000 (16:57 -0700)]
bpf: Fill new bpf_prog_pack with illegal instructions

bpf_prog_pack enables sharing huge pages among multiple BPF programs.
These pages are marked as executable before the JIT engine fill it with
BPF programs. To make these pages safe, fill the hole bpf_prog_pack with
illegal instructions before making it executable.

Fixes: 57631054fae6 ("bpf: Introduce bpf_prog_pack allocator")
Fixes: 33c9805860e5 ("bpf: Introduce bpf_jit_binary_pack_[alloc|finalize|free]")
Reported-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Song Liu <song@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20220520235758.1858153-2-song@kernel.org
2 years agocan: ctucanfd: platform: add missing dependency to HAS_IOMEM
Marc Kleine-Budde [Mon, 23 May 2022 12:33:44 +0000 (14:33 +0200)]
can: ctucanfd: platform: add missing dependency to HAS_IOMEM

The kernel test robot noticed that the ctucanfd platform driver fails
during modpost on platforms that don't support IOMEM.

| ERROR: modpost: "devm_ioremap_resource" [drivers/net/can/ctucanfd/ctucanfd_platform.ko] undefined!

This patch adds the missing HAS_IOMEM dependency.

Link: https://lore.kernel.org/all/20220523123720.1656611-1-mkl@pengutronix.de
Fixes: e8f0c23a2415 ("can: ctucanfd: CTU CAN FD open-source IP core - platform/SoC support.")
Reported-by: kernel test robot <lkp@intel.com>
Acked-by: Pavel Pisa <pisa@cmp.felk.cvut.cz>
Cc: Ondrej Ille <ondrej.ille@gmail.com>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
2 years agocan: kvaser_usb: silence a GCC 12 -Warray-bounds warning
Jakub Kicinski [Fri, 20 May 2022 19:46:59 +0000 (12:46 -0700)]
can: kvaser_usb: silence a GCC 12 -Warray-bounds warning

This driver does a lot of casting of smaller buffers to
struct kvaser_cmd_ext, GCC 12 does not like that:

| drivers/net/can/usb/kvaser_usb/kvaser_usb_hydra.c:489:65: warning: array subscript ‘struct kvaser_cmd_ext[0]’ is partly outside array bounds of ‘unsigned char[32]’ [-Warray-bounds]
| drivers/net/can/usb/kvaser_usb/kvaser_usb_hydra.c:489:23: note: in expansion of macro ‘le16_to_cpu’
|   489 |                 ret = le16_to_cpu(((struct kvaser_cmd_ext *)cmd)->len);
|       |                       ^~~~~~~~~~~

Temporarily silence this warning (move it to W=1 builds).

Link: https://lore.kernel.org/all/20220520194659.2356903-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Tested-by: Marc Kleine-Budde <mkl@pengutronix.de>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
2 years agocan: peak_usb: fix typo in comment
Julia Lawall [Sat, 21 May 2022 11:10:34 +0000 (13:10 +0200)]
can: peak_usb: fix typo in comment

Spelling mistake (triple letters) in comment.
Detected with the help of Coccinelle.

Link: https://lore.kernel.org/all/20220521111145.81697-24-Julia.Lawall@inria.fr
Signed-off-by: Julia Lawall <Julia.Lawall@inria.fr>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
2 years agoselftests/bpf: Fix spelling mistake: "unpriviliged" -> "unprivileged"
Colin Ian King [Mon, 23 May 2022 11:56:04 +0000 (12:56 +0100)]
selftests/bpf: Fix spelling mistake: "unpriviliged" -> "unprivileged"

There are spelling mistakes in ASSERT messages. Fix these.

Signed-off-by: Colin Ian King <colin.i.king@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20220523115604.49942-1-colin.i.king@gmail.com
2 years agoselftests/bpf: fix btf_dump/btf_dump due to recent clang change
Yonghong Song [Mon, 23 May 2022 15:20:44 +0000 (08:20 -0700)]
selftests/bpf: fix btf_dump/btf_dump due to recent clang change

Latest llvm-project upstream had a change of behavior
related to qualifiers on function return type ([1]).
This caused selftests btf_dump/btf_dump failure.
The following example shows what changed.

  $ cat t.c
  typedef const char * const (* const (* const fn_ptr_arr2_t[5])())(char * (*)(int));
  struct t {
    int a;
    fn_ptr_arr2_t l;
  };
  int foo(struct t *arg) {
    return arg->a;
  }

Compiled with latest upstream llvm15,
  $ clang -O2 -g -target bpf -S -emit-llvm t.c
The related generated debuginfo IR looks like:
  !16 = !DIDerivedType(tag: DW_TAG_typedef, name: "fn_ptr_arr2_t", file: !1, line: 1, baseType: !17)
  !17 = !DICompositeType(tag: DW_TAG_array_type, baseType: !18, size: 320, elements: !32)
  !18 = !DIDerivedType(tag: DW_TAG_const_type, baseType: !19)
  !19 = !DIDerivedType(tag: DW_TAG_pointer_type, baseType: !20, size: 64)
  !20 = !DISubroutineType(types: !21)
  !21 = !{!22, null}
  !22 = !DIDerivedType(tag: DW_TAG_pointer_type, baseType: !23, size: 64)
  !23 = !DISubroutineType(types: !24)
  !24 = !{!25, !28}
  !25 = !DIDerivedType(tag: DW_TAG_pointer_type, baseType: !26, size: 64)
  !26 = !DIDerivedType(tag: DW_TAG_const_type, baseType: !27)
  !27 = !DIBasicType(name: "char", size: 8, encoding: DW_ATE_signed_char)
You can see two intermediate const qualifier to pointer are dropped in debuginfo IR.

With llvm14, we have following debuginfo IR:
  !16 = !DIDerivedType(tag: DW_TAG_typedef, name: "fn_ptr_arr2_t", file: !1, line: 1, baseType: !17)
  !17 = !DICompositeType(tag: DW_TAG_array_type, baseType: !18, size: 320, elements: !34)
  !18 = !DIDerivedType(tag: DW_TAG_const_type, baseType: !19)
  !19 = !DIDerivedType(tag: DW_TAG_pointer_type, baseType: !20, size: 64)
  !20 = !DISubroutineType(types: !21)
  !21 = !{!22, null}
  !22 = !DIDerivedType(tag: DW_TAG_const_type, baseType: !23)
  !23 = !DIDerivedType(tag: DW_TAG_pointer_type, baseType: !24, size: 64)
  !24 = !DISubroutineType(types: !25)
  !25 = !{!26, !30}
  !26 = !DIDerivedType(tag: DW_TAG_const_type, baseType: !27)
  !27 = !DIDerivedType(tag: DW_TAG_pointer_type, baseType: !28, size: 64)
  !28 = !DIDerivedType(tag: DW_TAG_const_type, baseType: !29)
  !29 = !DIBasicType(name: "char", size: 8, encoding: DW_ATE_signed_char)
All const qualifiers are preserved.

To adapt the selftest to both old and new llvm, this patch removed
the intermediate const qualifier in const-to-ptr types, to make the
test succeed again.

  [1] https://reviews.llvm.org/D125919

Reported-by: Mykola Lysenko <mykolal@fb.com>
Signed-off-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/r/20220523152044.3905809-1-yhs@fb.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2 years agoBluetooth: hci_conn: Fix hci_connect_le_sync
Luiz Augusto von Dentz [Fri, 20 May 2022 18:37:13 +0000 (11:37 -0700)]
Bluetooth: hci_conn: Fix hci_connect_le_sync

The handling of connection failures shall be handled by the request
completion callback as already done by hci_cs_le_create_conn, also make
sure to use hci_conn_failed instead of hci_le_conn_failed as the later
don't actually call hci_conn_del to cleanup.

Link: https://github.com/bluez/bluez/issues/340
Fixes: 8e8b92ee60de5 ("Bluetooth: hci_sync: Add hci_le_create_conn_sync")
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
2 years agos390/bpf: Fix typo in comment
Julia Lawall [Sat, 21 May 2022 11:11:34 +0000 (13:11 +0200)]
s390/bpf: Fix typo in comment

Spelling mistake (triple letters) in comment.
Detected with the help of Coccinelle.

Signed-off-by: Julia Lawall <Julia.Lawall@inria.fr>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Ilya Leoshkevich <iii@linux.ibm.com>
Link: https://lore.kernel.org/bpf/20220521111145.81697-84-Julia.Lawall@inria.fr
2 years agolibbpf: Fix typo in comment
Julia Lawall [Sat, 21 May 2022 11:11:21 +0000 (13:11 +0200)]
libbpf: Fix typo in comment

Spelling mistake (triple letters) in comment.
Detected with the help of Coccinelle.

Signed-off-by: Julia Lawall <Julia.Lawall@inria.fr>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Daniel Müller <deso@posteo.net>
Link: https://lore.kernel.org/bpf/20220521111145.81697-71-Julia.Lawall@inria.fr
2 years agoMAINTAINERS: Add maintainer to AF_XDP
Magnus Karlsson [Mon, 23 May 2022 08:32:54 +0000 (10:32 +0200)]
MAINTAINERS: Add maintainer to AF_XDP

Maciej Fijalkowski has gracefully accepted to become the third
maintainer for the AF_XDP code. Thank you Maciej!

Signed-off-by: Magnus Karlsson <magnus.karlsson@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Björn Töpel <bjorn@kernel.org>
Acked-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Link: https://lore.kernel.org/bpf/20220523083254.32285-1-magnus.karlsson@gmail.com
2 years agoBluetooth: hci_sync: use hci_skb_event() helper
Ahmad Fatoum [Sat, 21 May 2022 20:10:50 +0000 (22:10 +0200)]
Bluetooth: hci_sync: use hci_skb_event() helper

This instance is the only opencoded version of the macro, so have it
follow suit.

Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
2 years agonet: dsa: OF-ware slave_mii_bus
Luiz Angelo Daros de Luca [Mon, 23 May 2022 01:32:34 +0000 (22:32 -0300)]
net: dsa: OF-ware slave_mii_bus

If found, register the DSA internally allocated slave_mii_bus with an OF
"mdio" child object. It can save some drivers from creating their
custom internal MDIO bus.

Signed-off-by: Luiz Angelo Daros de Luca <luizluca@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoeth: de4x5: remove support for Generic DECchip & DIGITAL EtherWORKS PCI/EISA
Jakub Kicinski [Thu, 19 May 2022 03:13:45 +0000 (20:13 -0700)]
eth: de4x5: remove support for Generic DECchip & DIGITAL EtherWORKS PCI/EISA

Looks like almost all changes to this driver had been tree-wide
refactoring since git era begun. There is one commit from Al
15 years ago which could potentially be fixing a real bug.

The driver is using virt_to_bus() and is a real magnet for pointless
cleanups. It seems unlikely to have real users. Let's try to shed
this maintenance burden.

Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Acked-by: Michael Ellerman <mpe@ellerman.id.au> (powerpc)
Acked-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: ethernet: mtk_eth_soc: fix error code in mtk_flow_offload_replace()
Dan Carpenter [Thu, 19 May 2022 14:08:00 +0000 (17:08 +0300)]
net: ethernet: mtk_eth_soc: fix error code in mtk_flow_offload_replace()

Preserve the error code from mtk_foe_entry_commit().  Do not return
success.

Fixes: c4f033d9e03e ("net: ethernet: mtk_eth_soc: rework hardware flow table management")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoMerge branch 'dsa-multi-cpu-port-part-two'
David S. Miller [Mon, 23 May 2022 09:40:10 +0000 (10:40 +0100)]
Merge branch 'dsa-multi-cpu-port-part-two'

Vladimir Oltean says:

====================
DSA changes for multiple CPU ports (part 2)

As explained in part 1:
https://patchwork.kernel.org/project/netdevbpf/cover/20220511095020.562461-1-vladimir.oltean@nxp.com/
I am trying to enable the second internal port pair from the NXP LS1028A
Felix switch for DSA-tagged traffic via "ocelot-8021q". This series
represents part 2 (of an unknown number) of that effort.

This series deals only with a minor bug fix (first patch) and with code
reorganization in the Felix DSA driver and in the Ocelot switch library.
Hopefully this will lay the ground for a clean introduction of new UAPI
for changing the DSA master of a user port in part 3.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: dsa: felix: tag_8021q preparation for multiple CPU ports
Vladimir Oltean [Sat, 21 May 2022 21:37:43 +0000 (00:37 +0300)]
net: dsa: felix: tag_8021q preparation for multiple CPU ports

Update the VCAP filters to support multiple tag_8021q CPU ports.

TX works using a filter for VLAN ID on the ingress of the CPU port, with
a redirect and a VLAN pop action. This can be updated trivially by
amending the ingress port mask of this rule to match on all tag_8021q
CPU ports.

RX works using a filter for ingress port on the egress of the CPU port,
with a VLAN push action. Here we need to replicate these filters for
each tag_8021q CPU port, and let them all have the same action.
This means that the OCELOT_VCAP_ES0_TAG_8021Q_RXVLAN() cookie needs to
encode a unique value for every {user port, CPU port} pair it's given.
Do this by encoding the CPU port in the upper 16 bits of the cookie, and
the user port in the lower 16 bits.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: mscc: ocelot: switch from {,un}set to {,un}assign for tag_8021q CPU ports
Vladimir Oltean [Sat, 21 May 2022 21:37:42 +0000 (00:37 +0300)]
net: mscc: ocelot: switch from {,un}set to {,un}assign for tag_8021q CPU ports

There is a desire for the felix driver to gain support for multiple
tag_8021q CPU ports, but the current model prevents it.

This is because ocelot_apply_bridge_fwd_mask() only takes into
consideration whether a port is a tag_8021q CPU port, but not whose CPU
port it is.

We need a model where we can have a direct affinity between an ocelot
port and a tag_8021q CPU port. This serves as the basis for multiple CPU
ports.

Declare a "dsa_8021q_cpu" backpointer in struct ocelot_port which
encodes that affinity. Repurpose the "ocelot_set_dsa_8021q_cpu" API to
"ocelot_assign_dsa_8021q_cpu" to express the change of paradigm.

Note that this change makes the first practical use of the new
ocelot_port->index field in ocelot_port_unassign_dsa_8021q_cpu(), where
we need to remove the old tag_8021q CPU port from the reserved VLAN range.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: dsa: felix: directly call ocelot_port_{set,unset}_dsa_8021q_cpu
Vladimir Oltean [Sat, 21 May 2022 21:37:41 +0000 (00:37 +0300)]
net: dsa: felix: directly call ocelot_port_{set,unset}_dsa_8021q_cpu

Absorb the final details of calling ocelot_port_{,un}set_dsa_8021q_cpu(),
i.e. the need to lock &ocelot->fwd_domain_lock, into the callee, to
simplify the caller and permit easier code reuse later.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: dsa: felix: update bridge fwd mask from ocelot lib when changing tag_8021q CPU
Vladimir Oltean [Sat, 21 May 2022 21:37:40 +0000 (00:37 +0300)]
net: dsa: felix: update bridge fwd mask from ocelot lib when changing tag_8021q CPU

Add more logic to ocelot_port_{,un}set_dsa_8021q_cpu() from the ocelot
switch lib by encapsulating the ocelot_apply_bridge_fwd_mask() call that
felix used to have.

This is necessary because the CPU port change procedure will also need
to do this, and it's good to reduce code duplication by having an entry
point in the ocelot switch lib that does all that is needed.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: dsa: felix: move the updating of PGID_CPU to the ocelot lib
Vladimir Oltean [Sat, 21 May 2022 21:37:39 +0000 (00:37 +0300)]
net: dsa: felix: move the updating of PGID_CPU to the ocelot lib

PGID_CPU must be updated every time a port is configured or unconfigured
as a tag_8021q CPU port. The ocelot switch lib already has a hook for
that operation, so move the updating of PGID_CPU to those hooks.

These bits are pretty specific to DSA, so normally I would keep them out
of the common switch lib, but when tag_8021q is in use, this has
implications upon the forwarding mask determined by
ocelot_apply_bridge_fwd_mask() and called extensively by the switch lib.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: dsa: fix missing adjustment of host broadcast flooding
Vladimir Oltean [Sat, 21 May 2022 21:37:38 +0000 (00:37 +0300)]
net: dsa: fix missing adjustment of host broadcast flooding

PGID_BC is configured statically by ocelot_init() to flood towards the
CPU port module, and dynamically by ocelot_port_set_bcast_flood()
towards all user ports.

When the tagging protocol changes, the intention is to turn off flooding
towards the old pipe towards the host, and to turn it on towards the new
pipe.

Due to a recent change which removed the adjustment of PGID_BC from
felix_set_host_flood(), 3 things happen.

- when we change from NPI to tag_8021q mode: in this mode, the CPU port
  module is accessed via registers, and used to read PTP packets with
  timestamps. We fail to disable broadcast flooding towards the CPU port
  module, and to enable broadcast flooding towards the physical port
  that serves as a DSA tag_8021q CPU port.

- from tag_8021q to NPI mode: in this mode, the CPU port module is
  redirected to a physical port. We fail to disable broadcast flooding
  towards the physical tag_8021q CPU port, and to enable it towards the
  CPU port module at ocelot->num_phys_ports.

- when the ports are put in promiscuous mode, we also fail to update
  PGID_BC towards the host pipe of the current protocol.

First issue means that felix_check_xtr_pkt() has to do extra work,
because it will not see only PTP packets, but also broadcasts. It needs
to dequeue these packets just to drop them.

Third issue is inconsequential, since PGID_BC is allocated from the
nonreserved multicast PGID space, and these PGIDs are conveniently
initialized to 0x7f (i.e. flood towards all ports except the CPU port
module). Broadcasts reach the NPI port via ocelot_init(), and reach the
tag_8021q CPU port via the hardware defaults.

Second issue is also inconsequential, because we fail both at disabling
and at enabling broadcast flooding on a port, so the defaults mentioned
above are preserved, and they are fine except for the performance impact.

Fixes: 7a29d220f4c0 ("net: dsa: felix: reimplement tagging protocol change with function pointers")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet/smc: fix listen processing for SMC-Rv2
liuyacan [Mon, 23 May 2022 05:50:56 +0000 (13:50 +0800)]
net/smc: fix listen processing for SMC-Rv2

In the process of checking whether RDMAv2 is available, the current
implementation first sets ini->smcrv2.ib_dev_v2, and then allocates
smc buf desc, but the latter may fail. Unfortunately, the caller
will only check the former. In this case, a NULL pointer reference
will occur in smc_clc_send_confirm_accept() when accessing
conn->rmb_desc.

This patch does two things:
1. Use the return code to determine whether V2 is available.
2. If the return code is NODEV, continue to check whether V1 is
available.

Fixes: e49300a6bf62 ("net/smc: add listen processing for SMC-Rv2")
Signed-off-by: liuyacan <liuyacan@corp.netease.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet/smc: postpone sk_refcnt increment in connect()
liuyacan [Mon, 23 May 2022 04:57:07 +0000 (12:57 +0800)]
net/smc: postpone sk_refcnt increment in connect()

Same trigger condition as commit 86434744. When setsockopt runs
in parallel to a connect(), and switch the socket into fallback
mode. Then the sk_refcnt is incremented in smc_connect(), but
its state stay in SMC_INIT (NOT SMC_ACTIVE). This cause the
corresponding sk_refcnt decrement in __smc_release() will not be
performed.

Fixes: 86434744fedf ("net/smc: add fallback check to connect()")
Signed-off-by: liuyacan <liuyacan@corp.netease.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoMerge branch 'fix-silence-gcc-12-warnings-in-drivers-net-wireless'
Jakub Kicinski [Mon, 23 May 2022 00:23:15 +0000 (17:23 -0700)]
Merge branch 'fix-silence-gcc-12-warnings-in-drivers-net-wireless'

Jakub Kicinski says:

====================
Fix/silence GCC 12 warnings in drivers/net/wireless/

as mentioned off list we'd like to get GCC 12 warnings quashed.
This set takes care of the warnings we have in drivers/net/wireless/
mostly by relegating them to W=1/W=2 builds.
====================

Link: https://lore.kernel.org/r/20220520194320.2356236-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agowifi: carl9170: silence a GCC 12 -Warray-bounds warning
Jakub Kicinski [Fri, 20 May 2022 19:43:20 +0000 (12:43 -0700)]
wifi: carl9170: silence a GCC 12 -Warray-bounds warning

carl9170 has a big union (struct carl9170_cmd) with all the command
types in it. But it allocates buffers only large enough for a given
command. This upsets GCC 12:

drivers/net/wireless/ath/carl9170/cmd.c:125:30: warning: array subscript ‘struct carl9170_cmd[0]’ is partly outside array bounds of ‘unsigned char[8]’ [-Warray-bounds]
  125 |                 tmp->hdr.cmd = cmd;
      |                 ~~~~~~~~~~~~~^~~~~

Punt the warning to W=1 for now. Hopefully GCC will learn to
recognize which fields are in-bounds.

Acked-by: Christian Lamparter <chunkeey@gmail.com>
Acked-by: Kalle Valo <kvalo@kernel.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agowifi: brcmfmac: work around a GCC 12 -Warray-bounds warning
Jakub Kicinski [Fri, 20 May 2022 19:43:18 +0000 (12:43 -0700)]
wifi: brcmfmac: work around a GCC 12 -Warray-bounds warning

GCC 12 really doesn't like partial struct allocations:

drivers/net/wireless/broadcom/brcm80211/brcmfmac/cfg80211.c:2202:32: warning: array subscript ‘struct brcmf_ext_join_params_le[0]’ is partly outside array bounds of ‘void[70]’ [-Warray-bounds]
 2202 |                 ext_join_params->scan_le.passive_time =
      |                                ^~

brcmfmac is trying to save 2 bytes at the end by either allocating
or not allocating a channel member. Let's keep @join_params_size
the "right" size but kmalloc() the full structure.

Acked-by: Kalle Valo <kvalo@kernel.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agowifi: iwlwifi: use unsigned to silence a GCC 12 warning
Jakub Kicinski [Fri, 20 May 2022 19:43:17 +0000 (12:43 -0700)]
wifi: iwlwifi: use unsigned to silence a GCC 12 warning

GCC 12 says:

drivers/net/wireless/intel/iwlwifi/mvm/sta.c:1076:37: warning: array subscript -1 is below array bounds of ‘struct iwl_mvm_tid_data[9]’ [-Warray-bounds]
 1076 |                 if (mvmsta->tid_data[tid].state != IWL_AGG_OFF)
      |                     ~~~~~~~~~~~~~~~~^~~~~

Whatever, tid is a bit from for_each_set_bit(), it's clearly unsigned.

Acked-by: Kalle Valo <kvalo@kernel.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agowifi: ath6k: silence false positive -Wno-dangling-pointer warning on GCC 12
Jakub Kicinski [Fri, 20 May 2022 19:43:16 +0000 (12:43 -0700)]
wifi: ath6k: silence false positive -Wno-dangling-pointer warning on GCC 12

For some reason GCC 12 decided to complain about the common
pattern of queuing an object onto a list on the stack in ath6k:

    inlined from ‘ath6kl_htc_mbox_tx’ at drivers/net/wireless/ath/ath6kl/htc_mbox.c:1142:3:
include/linux/list.h:74:19: warning: storing the address of local variable ‘queue’ in ‘*&packet_15(D)->list.prev’ [-Wdangling-pointer=]
   74 |         new->prev = prev;
      |         ~~~~~~~~~~^~~~~~

Move the warning to W=1, hopefully it goes away with a compiler
update.

Acked-by: Kalle Valo <kvalo@kernel.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agowifi: rtlwifi: remove always-true condition pointed out by GCC 12
Jakub Kicinski [Fri, 20 May 2022 19:43:15 +0000 (12:43 -0700)]
wifi: rtlwifi: remove always-true condition pointed out by GCC 12

The .value is a two-dim array, not a pointer.

struct iqk_matrix_regs {
bool iqk_done;
        long value[1][IQK_MATRIX_REG_NUM];
};

Acked-by: Kalle Valo <kvalo@kernel.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agowifi: ath9k: silence array-bounds warning on GCC 12
Jakub Kicinski [Fri, 20 May 2022 19:43:14 +0000 (12:43 -0700)]
wifi: ath9k: silence array-bounds warning on GCC 12

GCC 12 says:

drivers/net/wireless/ath/ath9k/mac.c: In function ‘ath9k_hw_resettxqueue’:
drivers/net/wireless/ath/ath9k/mac.c:373:22: warning: array subscript 32 is above array bounds of ‘struct ath9k_tx_queue_info[10]’ [-Warray-bounds]
  373 |         qi = &ah->txq[q];
      |               ~~~~~~~^~~

I don't know where it got the 32 from, relegate the warning to W=1+.

Acked-by: Kalle Valo <kvalo@kernel.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agowifi: plfxlc: remove redundant NULL-check for GCC 12
Jakub Kicinski [Fri, 20 May 2022 19:43:13 +0000 (12:43 -0700)]
wifi: plfxlc: remove redundant NULL-check for GCC 12

GCC is upset that we check the return value of plfxlc_usb_dev()
even tho it can't be NULL:

drivers/net/wireless/purelifi/plfxlc/usb.c: In function ‘resume’:
drivers/net/wireless/purelifi/plfxlc/usb.c:840:20: warning: the comparison will always evaluate as ‘true’ for the address of ‘dev’ will never be NULL [-Waddress]
  840 |         if (!pl || !plfxlc_usb_dev(pl))
      |                    ^

plfxlc_usb_dev() returns an address of one of the members of pl,
so it's safe to drop these checks.

Acked-by: Kalle Valo <kvalo@kernel.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agodt-bindings: net: toshiba,visconti-dwmac: Update the common clock properties
Nobuhiro Iwamatsu [Fri, 20 May 2022 08:48:23 +0000 (17:48 +0900)]
dt-bindings: net: toshiba,visconti-dwmac: Update the common clock properties

The clock for this driver switched to the common clock controller driver.
Therefore, update common clock properties for ethernet device in the binding
document.

Signed-off-by: Nobuhiro Iwamatsu <nobuhiro1.iwamatsu@toshiba.co.jp>
Acked-by: Rob Herring <robh@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: dsa: restrict SMSC_LAN9303_I2C kconfig
Randy Dunlap [Fri, 20 May 2022 05:15:23 +0000 (22:15 -0700)]
net: dsa: restrict SMSC_LAN9303_I2C kconfig

Since kconfig 'select' does not follow dependency chains, if symbol KSA
selects KSB, then KSA should also depend on the same symbols that KSB
depends on, in order to prevent Kconfig warnings and possible build
errors.

Change NET_DSA_SMSC_LAN9303_I2C and NET_DSA_SMSC_LAN9303_MDIO so that
they are limited to VLAN_8021Q if the latter is enabled. This prevents
the Kconfig warning:

WARNING: unmet direct dependencies detected for NET_DSA_SMSC_LAN9303
  Depends on [m]: NETDEVICES [=y] && NET_DSA [=y] && (VLAN_8021Q [=m] || VLAN_8021Q [=m]=n)
  Selected by [y]:
  - NET_DSA_SMSC_LAN9303_I2C [=y] && NETDEVICES [=y] && NET_DSA [=y] && I2C [=y]

Fixes: 430065e26719 ("net: dsa: lan9303: add VLAN IDs to master device")
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Cc: Andrew Lunn <andrew@lunn.ch>
Cc: Vivien Didelot <vivien.didelot@gmail.com>
Cc: Florian Fainelli <f.fainelli@gmail.com>
Cc: Vladimir Oltean <olteanv@gmail.com>
Cc: Juergen Borleis <jbe@pengutronix.de>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Eric Dumazet <edumazet@google.com>
Cc: Jakub Kicinski <kuba@kernel.org>
Cc: Paolo Abeni <pabeni@redhat.com>
Cc: Mans Rullgard <mans@mansr.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 agonet: fddi: skfp: smt: Remove extra parameters to vararg macro
Tom Rix [Sun, 22 May 2022 13:36:27 +0000 (09:36 -0400)]
net: fddi: skfp: smt: Remove extra parameters to vararg macro

cppcheck reports
[drivers/net/fddi/skfp/smt.c:750]: (warning) printf format string requires 0 parameters but 2 are given.

DB_SBAN is a vararg macro, like DB_ESSN.  Remove the extra args and the nl.

Signed-off-by: Tom Rix <trix@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoMerge branch 'mt7986-support'
David S. Miller [Sun, 22 May 2022 21:24:33 +0000 (22:24 +0100)]
Merge branch 'mt7986-support'

Jakub Kicinski says:

====================
introduce mt7986 ethernet support

Add support for mt7986-eth driver available on mt7986 soc.

Changes since v2:
- rely on GFP_KERNEL whenever possible
- define mtk_reg_map struct to introduce soc register map and avoid macros
- improve comments

Changes since v1:
- drop SRAM option
- convert ring->dma to void
- convert scratch_ring to void
- enable port4
- fix irq dts bindings
- drop gmac1 support from mt7986a-rfb dts for the moment
====================

Acked-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: ethernet: mtk_eth_soc: introduce support for mt7986 chipset
Lorenzo Bianconi [Fri, 20 May 2022 18:11:39 +0000 (20:11 +0200)]
net: ethernet: mtk_eth_soc: introduce support for mt7986 chipset

Add support for mt7986-eth driver available on mt7986 soc.

Tested-by: Sam Shih <sam.shih@mediatek.com>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: ethernet: mtk_eth_soc: convert scratch_ring pointer to void
Lorenzo Bianconi [Fri, 20 May 2022 18:11:38 +0000 (20:11 +0200)]
net: ethernet: mtk_eth_soc: convert scratch_ring pointer to void

Simplify the code converting scratch_ring pointer to void

Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: ethernet: mtk_eth_soc: convert ring dma pointer to void
Lorenzo Bianconi [Fri, 20 May 2022 18:11:37 +0000 (20:11 +0200)]
net: ethernet: mtk_eth_soc: convert ring dma pointer to void

Simplify the code converting {tx,rx} ring dma pointer to void

Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: ethernet: mtk_eth_soc: introduce MTK_NETSYS_V2 support
Lorenzo Bianconi [Fri, 20 May 2022 18:11:36 +0000 (20:11 +0200)]
net: ethernet: mtk_eth_soc: introduce MTK_NETSYS_V2 support

Introduce MTK_NETSYS_V2 support. MTK_NETSYS_V2 defines 32B TX/RX DMA
descriptors.
This is a preliminary patch to add mt7986 ethernet support.

Tested-by: Sam Shih <sam.shih@mediatek.com>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: ethernet: mtk_eth_soc: introduce device register map
Lorenzo Bianconi [Fri, 20 May 2022 18:11:35 +0000 (20:11 +0200)]
net: ethernet: mtk_eth_soc: introduce device register map

Introduce reg_map structure to add the capability to support different
register definitions. Move register definitions in mtk_regmap structure.
This is a preliminary patch to introduce mt7986 ethernet support.

Tested-by: Sam Shih <sam.shih@mediatek.com>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: ethernet: mtk_eth_soc: rely on rxd_size field in mtk_rx_alloc/mtk_rx_clean
Lorenzo Bianconi [Fri, 20 May 2022 18:11:34 +0000 (20:11 +0200)]
net: ethernet: mtk_eth_soc: rely on rxd_size field in mtk_rx_alloc/mtk_rx_clean

Remove mtk_rx_dma structure layout dependency in mtk_rx_alloc/mtk_rx_clean.
Initialize to 0 rxd3 and rxd4 in mtk_rx_alloc.
This is a preliminary patch to add mt7986 ethernet support.

Tested-by: Sam Shih <sam.shih@mediatek.com>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: ethernet: mtk_eth_soc: rely on txd_size field in mtk_poll_tx/mtk_poll_rx
Lorenzo Bianconi [Fri, 20 May 2022 18:11:33 +0000 (20:11 +0200)]
net: ethernet: mtk_eth_soc: rely on txd_size field in mtk_poll_tx/mtk_poll_rx

This is a preliminary to ad mt7986 ethernet support.

Tested-by: Sam Shih <sam.shih@mediatek.com>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: ethernet: mtk_eth_soc: add rxd_size to mtk_soc_data
Lorenzo Bianconi [Fri, 20 May 2022 18:11:32 +0000 (20:11 +0200)]
net: ethernet: mtk_eth_soc: add rxd_size to mtk_soc_data

Similar to tx counterpart, introduce rxd_size in mtk_soc_data data
structure.
This is a preliminary patch to add mt7986 ethernet support.

Tested-by: Sam Shih <sam.shih@mediatek.com>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: ethernet: mtk_eth_soc: rely on txd_size in txd_to_idx
Lorenzo Bianconi [Fri, 20 May 2022 18:11:31 +0000 (20:11 +0200)]
net: ethernet: mtk_eth_soc: rely on txd_size in txd_to_idx

This is a preliminary patch to add mt7986 ethernet support.

Tested-by: Sam Shih <sam.shih@mediatek.com>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: ethernet: mtk_eth_soc: rely on txd_size in mtk_desc_to_tx_buf
Lorenzo Bianconi [Fri, 20 May 2022 18:11:30 +0000 (20:11 +0200)]
net: ethernet: mtk_eth_soc: rely on txd_size in mtk_desc_to_tx_buf

This is a preliminary patch to add mt7986 ethernet support.

Tested-by: Sam Shih <sam.shih@mediatek.com>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: ethernet: mtk_eth_soc: rely on txd_size in mtk_tx_alloc/mtk_tx_clean
Lorenzo Bianconi [Fri, 20 May 2022 18:11:29 +0000 (20:11 +0200)]
net: ethernet: mtk_eth_soc: rely on txd_size in mtk_tx_alloc/mtk_tx_clean

This is a preliminary patch to add mt7986 ethernet support.

Tested-by: Sam Shih <sam.shih@mediatek.com>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: ethernet: mtk_eth_soc: add txd_size to mtk_soc_data
Lorenzo Bianconi [Fri, 20 May 2022 18:11:28 +0000 (20:11 +0200)]
net: ethernet: mtk_eth_soc: add txd_size to mtk_soc_data

In order to remove mtk_tx_dma size dependency, introduce txd_size in
mtk_soc_data data structure. Rely on txd_size in mtk_init_fq_dma() and
mtk_dma_free() routines.
This is a preliminary patch to add mt7986 ethernet support.

Tested-by: Sam Shih <sam.shih@mediatek.com>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: ethernet: mtk_eth_soc: move tx dma desc configuration in mtk_tx_set_dma_desc
Lorenzo Bianconi [Fri, 20 May 2022 18:11:27 +0000 (20:11 +0200)]
net: ethernet: mtk_eth_soc: move tx dma desc configuration in mtk_tx_set_dma_desc

Move tx dma descriptor configuration in mtk_tx_set_dma_desc routine.
This is a preliminary patch to introduce mt7986 ethernet support since
it relies on a different tx dma descriptor layout.

Tested-by: Sam Shih <sam.shih@mediatek.com>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: ethernet: mtk_eth_soc: rely on GFP_KERNEL for dma_alloc_coherent whenever possible
Lorenzo Bianconi [Fri, 20 May 2022 18:11:26 +0000 (20:11 +0200)]
net: ethernet: mtk_eth_soc: rely on GFP_KERNEL for dma_alloc_coherent whenever possible

Rely on GFP_KERNEL for dma descriptors mappings in mtk_tx_alloc(),
mtk_rx_alloc() and mtk_init_fq_dma() since they are run in non-irq
context.

Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agodt-bindings: net: mediatek,net: add mt7986-eth binding
Lorenzo Bianconi [Fri, 20 May 2022 18:11:25 +0000 (20:11 +0200)]
dt-bindings: net: mediatek,net: add mt7986-eth binding

Introduce dts bindings for mt7986 soc in mediatek,net.yaml.

Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoarm64: dts: mediatek: mt7986: introduce ethernet nodes
Lorenzo Bianconi [Fri, 20 May 2022 18:11:24 +0000 (20:11 +0200)]
arm64: dts: mediatek: mt7986: introduce ethernet nodes

Introduce ethernet nodes in mt7986 bindings in order to
enable mt7986a/mt7986b ethernet support.

Co-developed-by: Sam Shih <sam.shih@mediatek.com>
Signed-off-by: Sam Shih <sam.shih@mediatek.com>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoMerge branch 'net-gcc12-warnings'
David S. Miller [Sun, 22 May 2022 21:21:11 +0000 (22:21 +0100)]
Merge branch 'net-gcc12-warnings'

Jakub Kicinski says:

====================
eth: silence the GCC 12 array-bounds warnings

Silence the array-bounds warnings in Ethernet drivers.
v2 uses -Wno-array-bounds directly.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoeth: tg3: silence the GCC 12 array-bounds warning
Jakub Kicinski [Fri, 20 May 2022 19:56:05 +0000 (12:56 -0700)]
eth: tg3: silence the GCC 12 array-bounds warning

GCC 12 currently generates a rather inconsistent warning:

drivers/net/ethernet/broadcom/tg3.c:17795:51: warning: array subscript 5 is above array bounds of ‘struct tg3_napi[5]’ [-Warray-bounds]
17795 |                 struct tg3_napi *tnapi = &tp->napi[i];
      |                                           ~~~~~~~~^~~

i is guaranteed < tp->irq_max which in turn is either 1 or 5.
There are more loops like this one in the driver, but strangely
GCC 12 dislikes only this single one.

Silence this silliness for now.

Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoeth: ice: silence the GCC 12 array-bounds warning
Jakub Kicinski [Fri, 20 May 2022 19:56:04 +0000 (12:56 -0700)]
eth: ice: silence the GCC 12 array-bounds warning

GCC 12 gets upset because driver allocates partial
struct ice_aqc_sw_rules_elem buffers. The writes are
within bounds.

Silence these warnings for now, our build bot runs GCC 12
so we won't allow any new instances.

Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoeth: mtk_eth_soc: silence the GCC 12 array-bounds warning
Jakub Kicinski [Fri, 20 May 2022 19:56:03 +0000 (12:56 -0700)]
eth: mtk_eth_soc: silence the GCC 12 array-bounds warning

GCC 12 gets upset because in mtk_foe_entry_commit_subflow()
this driver allocates a partial structure. The writes are
within bounds.

Silence these warnings for now, our build bot runs GCC 12
so we won't allow any new instances.

Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoMerge branch 'dpaa2-swtso-fixes'
David S. Miller [Sun, 22 May 2022 21:18:05 +0000 (22:18 +0100)]
Merge branch 'dpaa2-swtso-fixes'

Ioana Ciornei says:

====================
dpaa2-eth: software TSO fixes

This patch fixes the software TSO feature in dpaa2-eth.

There are multiple errors that I made in the initial submission of the
code, which I didn't caught since I was always running with passthough
IOMMU.

The bug report came in bugzilla:
https://bugzilla.kernel.org/show_bug.cgi?id=215886

The bugs are in the Tx confirmation path, where I was trying to retrieve
a virtual address after DMA unmapping the area. Besides that, another
dma_unmap call was made with the wrong size.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agodpaa2-eth: unmap the SGT buffer before accessing its contents
Ioana Ciornei [Sun, 22 May 2022 12:52:51 +0000 (15:52 +0300)]
dpaa2-eth: unmap the SGT buffer before accessing its contents

DMA unmap the Scatter/Gather table before going through the array to
unmap and free each of the header and data chunks. This is so we do not
touch the data between the dma_map and dma_unmap calls.

Fixes: 3dc709e0cd47 ("dpaa2-eth: add support for software TSO")
Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agodpaa2-eth: use the correct software annotation field
Ioana Ciornei [Sun, 22 May 2022 12:52:50 +0000 (15:52 +0300)]
dpaa2-eth: use the correct software annotation field

The incorrect software annotation field was being used, swa->sg.sgt_size
instead of swa->tso.sgt_size, which meant that the SGT buffer was
unmapped with a wrong size.
This is also confirmed by the DMA API debug prints which showed the
following:

[   38.962434] DMA-API: fsl_dpaa2_eth dpni.2: device driver frees DMA memory with different size [device address=0x0000fffffafba740] [map size=224 bytes] [unmap size=0 bytes]
[   38.980496] WARNING: CPU: 11 PID: 1131 at kernel/dma/debug.c:973 check_unmap+0x58c/0x9b0
[   38.988586] Modules linked in:
[   38.991631] CPU: 11 PID: 1131 Comm: iperf3 Not tainted 5.18.0-rc7-00117-g59130eeb2b8f #1972
[   38.999970] Hardware name: NXP Layerscape LX2160ARDB (DT)

Fixes: 3dc709e0cd47 ("dpaa2-eth: add support for software TSO")
Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agodpaa2-eth: retrieve the virtual address before dma_unmap
Ioana Ciornei [Sun, 22 May 2022 12:52:49 +0000 (15:52 +0300)]
dpaa2-eth: retrieve the virtual address before dma_unmap

The TSO header was DMA unmapped before the virtual address was retrieved
and then used to free the buffer. This meant that we were actually
removing the DMA map and then trying to search for it to help in
retrieving the virtual address. This lead to a invalid virtual address
being used in the kfree call.

Fix this by calling dpaa2_iova_to_virt() prior to the dma_unmap call.

[  487.231819] Unable to handle kernel paging request at virtual address fffffd9807000008

(...)

[  487.354061] Hardware name: SolidRun LX2160A Honeycomb (DT)
[  487.359535] pstate: a0400005 (NzCv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--)
[  487.366485] pc : kfree+0xac/0x304
[  487.369799] lr : kfree+0x204/0x304
[  487.373191] sp : ffff80000c4eb120
[  487.376493] x29: ffff80000c4eb120 x28: ffff662240c46400 x27: 0000000000000001
[  487.383621] x26: 0000000000000001 x25: ffff662246da0cc0 x24: ffff66224af78000
[  487.390748] x23: ffffad184f4ce008 x22: ffffad1850185000 x21: ffffad1838d13cec
[  487.397874] x20: ffff6601c0000000 x19: fffffd9807000000 x18: 0000000000000000
[  487.405000] x17: ffffb910cdc49000 x16: ffffad184d7d9080 x15: 0000000000004000
[  487.412126] x14: 0000000000000008 x13: 000000000000ffff x12: 0000000000000000
[  487.419252] x11: 0000000000000004 x10: 0000000000000001 x9 : ffffad184d7d927c
[  487.426379] x8 : 0000000000000000 x7 : 0000000ffffffd1d x6 : ffff662240a94900
[  487.433505] x5 : 0000000000000003 x4 : 0000000000000009 x3 : ffffad184f4ce008
[  487.440632] x2 : ffff662243eec000 x1 : 0000000100000100 x0 : fffffc0000000000
[  487.447758] Call trace:
[  487.450194]  kfree+0xac/0x304
[  487.453151]  dpaa2_eth_free_tx_fd.isra.0+0x33c/0x3e0 [fsl_dpaa2_eth]
[  487.459507]  dpaa2_eth_tx_conf+0x100/0x2e0 [fsl_dpaa2_eth]
[  487.464989]  dpaa2_eth_poll+0xdc/0x380 [fsl_dpaa2_eth]

Fixes: 3dc709e0cd47 ("dpaa2-eth: add support for software TSO")
Link: https://bugzilla.kernel.org/show_bug.cgi?id=215886
Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: mscc: ocelot: offload tc action "ok" using an empty action vector
Vladimir Oltean [Sun, 22 May 2022 09:27:01 +0000 (12:27 +0300)]
net: mscc: ocelot: offload tc action "ok" using an empty action vector

The "ok" tc action is useful when placed in front of a more generic
filter to exclude some more specific rules from matching it.

The ocelot switches can offload this tc action by creating an empty
action vector (no _ENA fields set to 1). This makes sense for all of
VCAP IS1, IS2 and ES0 (but not for PSFP).

Add support for this action. Note that this makes the
gact_drop_and_ok_test() selftest pass, where "action ok" is used in
front of an "action drop" rule, both offloaded to VCAP IS2.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoMerge branch 'ocelot-selftests'
David S. Miller [Sun, 22 May 2022 21:14:13 +0000 (22:14 +0100)]
Merge branch 'ocelot-selftests'

Vladimir Oltean says:

====================
Streamline Ocelot tc-chains selftest

This series changes the output and the argument format of the Ocelot
switch selftest so that it is more similar to what can be found in
tools/testing/selftests/net/forwarding/.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoselftests: ocelot: tc_flower_chains: reorder interfaces
Vladimir Oltean [Sun, 22 May 2022 09:50:40 +0000 (12:50 +0300)]
selftests: ocelot: tc_flower_chains: reorder interfaces

Use the standard interface order h1, swp1, swp2, h2 that is used by the
forwarding selftest framework. The previous order was confusing even
with the ASCII drawing. That isn't needed anymore.

This also drops the fixed MAC addresses and uses STABLE_MAC_ADDRS, which
ensures the MAC addresses are unique.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoselftests: ocelot: tc_flower_chains: use conventional interface names
Vladimir Oltean [Sun, 22 May 2022 09:50:39 +0000 (12:50 +0300)]
selftests: ocelot: tc_flower_chains: use conventional interface names

This is a robotic rename as follows:

eth0 -> swp1
eth1 -> swp2
eth2 -> h2
eth3 -> h1

This brings the selftest more in line with the other forwarding
selftests, where h1 is connected to swp1, and h2 to swp2.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoselftests: ocelot: tc_flower_chains: streamline test output
Vladimir Oltean [Sun, 22 May 2022 09:50:38 +0000 (12:50 +0300)]
selftests: ocelot: tc_flower_chains: streamline test output

Bring this driver-specific selftest output in line with the other
selftests.

Before:

Testing VLAN pop..                      OK
Testing VLAN push..                     OK
Testing ingress VLAN modification..             OK
Testing egress VLAN modification..              OK
Testing frame prioritization..          OK

After:

TEST: VLAN pop                                                      [ OK ]
TEST: VLAN push                                                     [ OK ]
TEST: Ingress VLAN modification                                     [ OK ]
TEST: Egress VLAN modification                                      [ OK ]
TEST: Frame prioritization                                          [ OK ]

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: wrap the wireless pointers in struct net_device in an ifdef
Jakub Kicinski [Thu, 19 May 2022 20:20:54 +0000 (13:20 -0700)]
net: wrap the wireless pointers in struct net_device in an ifdef

Most protocol-specific pointers in struct net_device are under
a respective ifdef. Wireless is the notable exception. Since
there's a sizable number of custom-built kernels for datacenter
workloads which don't build wireless it seems reasonable to
ifdefy those pointers as well.

While at it move IPv4 and IPv6 pointers up, those are special
for obvious reasons.

Acked-by: Johannes Berg <johannes@sipsolutions.net>
Acked-by: Stefan Schmidt <stefan@datenfreihafen.org> # ieee802154
Acked-by: Sven Eckelmann <sven@narfation.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: fec: Do proper error checking for enet_out clk
Uwe Kleine-König [Fri, 20 May 2022 06:26:50 +0000 (08:26 +0200)]
net: fec: Do proper error checking for enet_out clk

An error code returned by devm_clk_get() might have other meanings than
"This clock doesn't exist". So use devm_clk_get_optional() and handle
all remaining errors as fatal.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agohinic: Avoid some over memory allocation
Christophe JAILLET [Sat, 21 May 2022 06:33:01 +0000 (08:33 +0200)]
hinic: Avoid some over memory allocation

'prod_idx' (atomic_t) is larger than 'shadow_idx' (u16), so some memory is
over-allocated.

Fixes: b15a9f37be2b ("net-next/hinic: Add wq")
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: fec: Do proper error checking for optional clks
Uwe Kleine-König [Sat, 21 May 2022 08:34:25 +0000 (10:34 +0200)]
net: fec: Do proper error checking for optional clks

An error code returned by devm_clk_get() might have other meanings than
"This clock doesn't exist". So use devm_clk_get_optional() and handle
all remaining errors as fatal.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: phy: DP83822: enable rgmii mode if phy_interface_is_rgmii
Tommaso Merciai [Fri, 20 May 2022 23:58:46 +0000 (01:58 +0200)]
net: phy: DP83822: enable rgmii mode if phy_interface_is_rgmii

RGMII mode can be enable from dp83822 straps, and also writing bit 9
of register 0x17 - RMII and Status Register (RCSR).
When phy_interface_is_rgmii rgmii mode must be enabled, same for
contrary, this prevents malconfigurations of hw straps

References:
 - https://www.ti.com/lit/gpn/dp83822i p66

Signed-off-by: Tommaso Merciai <tommaso.merciai@amarulasolutions.com>
Co-developed-by: Michael Trimarchi <michael@amarulasolutions.com>
Suggested-by: Alberto Bianchi <alberto.bianchi@amarulasolutions.com>
Tested-by: Tommaso Merciai <tommaso.merciai@amarulasolutions.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: selftests: Add stress_reuseport_listen to .gitignore
Muhammad Usama Anjum [Sat, 21 May 2022 09:37:06 +0000 (14:37 +0500)]
net: selftests: Add stress_reuseport_listen to .gitignore

Add newly added stress_reuseport_listen object to .gitignore file.

Fixes: ec8cb4f617a2 ("net: selftests: Stress reuseport listen")
Signed-off-by: Muhammad Usama Anjum <usama.anjum@collabora.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoMerge branch 'rxrpc-fixes'
David S. Miller [Sun, 22 May 2022 20:30:53 +0000 (21:30 +0100)]
Merge branch 'rxrpc-fixes'

David Howells says:

====================
rxrpc: Miscellaneous fixes

Here are some fixes for AF_RXRPC:

 (1) Fix listen() allowing preallocation to overrun the prealloc buffer.

 (2) Prevent resending the request if we've seen the reply starting to
     arrive.

 (3) Fix accidental sharing of ACK state between transmission and
     reception.

 (4) Ignore ACKs in which ack.previousPacket regresses.  This indicates the
     highest DATA number so far seen, so should not be seen to go
     backwards.

 (5) Fix the determination of when to generate an IDLE-type ACK,
     simplifying it so that we generate one if we have more than two DATA
     packets that aren't hard-acked (consumed) or soft-acked (in the rx
     buffer, but could be discarded and re-requested).
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agorxrpc: Fix decision on when to generate an IDLE ACK
David Howells [Sat, 21 May 2022 08:03:31 +0000 (09:03 +0100)]
rxrpc: Fix decision on when to generate an IDLE ACK

Fix the decision on when to generate an IDLE ACK by keeping a count of the
number of packets we've received, but not yet soft-ACK'd, and the number of
packets we've processed, but not yet hard-ACK'd, rather than trying to keep
track of which DATA sequence numbers correspond to those points.

We then generate an ACK when either counter exceeds 2.  The counters are
both cleared when we transcribe the information into any sort of ACK packet
for transmission.  IDLE and DELAY ACKs are skipped if both counters are 0
(ie. no change).

Fixes: 805b21b929e2 ("rxrpc: Send an ACK after every few DATA packets we receive")
Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agorxrpc: Don't let ack.previousPacket regress
David Howells [Sat, 21 May 2022 08:03:24 +0000 (09:03 +0100)]
rxrpc: Don't let ack.previousPacket regress

The previousPacket field in the rx ACK packet should never go backwards -
it's now the highest DATA sequence number received, not the last on
received (it used to be used for out of sequence detection).

Fixes: 248f219cb8bc ("rxrpc: Rewrite the data and ack handling code")
Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agorxrpc: Fix overlapping ACK accounting
David Howells [Sat, 21 May 2022 08:03:18 +0000 (09:03 +0100)]
rxrpc: Fix overlapping ACK accounting

Fix accidental overlapping of Rx-phase ACK accounting with Tx-phase ACK
accounting through variables shared between the two.  call->acks_* members
refer to ACKs received in the Tx phase and call->ackr_* members to ACKs
sent/to be sent during the Rx phase.

Fixes: 1a2391c30c0b ("rxrpc: Fix detection of out of order acks")
Signed-off-by: David Howells <dhowells@redhat.com>
cc: Jeffrey Altman <jaltman@auristor.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agorxrpc: Don't try to resend the request if we're receiving the reply
David Howells [Sat, 21 May 2022 08:03:11 +0000 (09:03 +0100)]
rxrpc: Don't try to resend the request if we're receiving the reply

rxrpc has a timer to trigger resending of unacked data packets in a call.
This is not cancelled when a client call switches to the receive phase on
the basis that most calls don't last long enough for it to ever expire.
However, if it *does* expire after we've started to receive the reply, we
shouldn't then go into trying to retransmit or pinging the server to find
out if an ack got lost.

Fix this by skipping the resend code if we're into receiving the reply to a
client call.

Fixes: 17926a79320a ("[AF_RXRPC]: Provide secure RxRPC sockets for use by userspace and kernel both")
Signed-off-by: David Howells <dhowells@redhat.com>
cc: linux-afs@lists.infradead.org
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agorxrpc: Fix listen() setting the bar too high for the prealloc rings
David Howells [Sat, 21 May 2022 08:03:04 +0000 (09:03 +0100)]
rxrpc: Fix listen() setting the bar too high for the prealloc rings

AF_RXRPC's listen() handler lets you set the backlog up to 32 (if you bump
up the sysctl), but whilst the preallocation circular buffers have 32 slots
in them, one of them has to be a dead slot because we're using CIRC_CNT().

This means that listen(rxrpc_sock, 32) will cause an oops when the socket
is closed because rxrpc_service_prealloc_one() allocated one too many calls
and rxrpc_discard_prealloc() won't then be able to get rid of them because
it'll think the ring is empty.  rxrpc_release_calls_on_socket() then tries
to abort them, but oopses because call->peer isn't yet set.

Fix this by setting the maximum backlog to RXRPC_BACKLOG_MAX - 1 to match
the ring capacity.

 BUG: kernel NULL pointer dereference, address: 0000000000000086
 ...
 RIP: 0010:rxrpc_send_abort_packet+0x73/0x240 [rxrpc]
 Call Trace:
  <TASK>
  ? __wake_up_common_lock+0x7a/0x90
  ? rxrpc_notify_socket+0x8e/0x140 [rxrpc]
  ? rxrpc_abort_call+0x4c/0x60 [rxrpc]
  rxrpc_release_calls_on_socket+0x107/0x1a0 [rxrpc]
  rxrpc_release+0xc9/0x1c0 [rxrpc]
  __sock_release+0x37/0xa0
  sock_close+0x11/0x20
  __fput+0x89/0x240
  task_work_run+0x59/0x90
  do_exit+0x319/0xaa0

Fixes: 00e907127e6f ("rxrpc: Preallocate peers, conns and calls for incoming service requests")
Reported-by: Marc Dionne <marc.dionne@auristor.com>
Signed-off-by: David Howells <dhowells@redhat.com>
cc: linux-afs@lists.infradead.org
Link: https://lists.infradead.org/pipermail/linux-afs/2022-March/005079.html
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoMerge branch 'rxrpc-misc'
David S. Miller [Sun, 22 May 2022 20:03:02 +0000 (21:03 +0100)]
Merge branch 'rxrpc-misc'

David Howells says:

====================
rxrpc: Miscellaneous changes

Here are some miscellaneous changes for AF_RXRPC:

 (1) Allow the list of local endpoints to be viewed through /proc.

 (2) Switch to using refcount_t for refcounting.

 (3) Fix a locking issue found by lockdep.

 (4) Autogenerate tracing symbol enums from symbol->string maps to make it
     easier to keep them in sync.

 (5) Return an error to sendmsg() if a call it tried to set up failed.
     Because it failed at this point, no notification will be generated for
     recvmsg to pick up - but userspace still needs to know about the
     failure.

 (6) Fix the selection of abort codes generated by internal events.  In
     particular, rxrpc and kafs shouldn't be generating RX_USER_ABORT
     unless it's because userspace did something to cancel a call.

 (7) Adjust the interpretation and handling of certain ACK types to try and
     detect NAT changes causing a call to seem to start mid-flow from a
     different peer.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>