Commit graph

1,308,922 commits

Author SHA1 Message Date
Christophe Leroy
c5aebb53b3 powerpc: Force inlining of arch_vmap_p{u/m}d_supported()
arch_vmap_pud_supported() and arch_vmap_pmd_supported() are
expected to constant-fold to false when RADIX is not enabled.

Force inlining in order to avoid following failure which
leads to unexpected call of non-existing pud_set_huge() and
pmd_set_huge() on powerpc 8xx.

In function 'pud_huge_tests',
    inlined from 'debug_vm_pgtable' at mm/debug_vm_pgtable.c:1399:2:
./arch/powerpc/include/asm/vmalloc.h:9:33: warning: inlining failed in call to 'arch_vmap_pud_supported.isra': call is unlikely and code size would grow [-Winline]
    9 | #define arch_vmap_pud_supported arch_vmap_pud_supported
      |                                 ^~~~~~~~~~~~~~~~~~~~~~~
./arch/powerpc/include/asm/vmalloc.h:10:20: note: in expansion of macro 'arch_vmap_pud_supported'
   10 | static inline bool arch_vmap_pud_supported(pgprot_t prot)
      |                    ^~~~~~~~~~~~~~~~~~~~~~~
./arch/powerpc/include/asm/vmalloc.h:9:33: note: called from here
    9 | #define arch_vmap_pud_supported arch_vmap_pud_supported
mm/debug_vm_pgtable.c:458:14: note: in expansion of macro 'arch_vmap_pud_supported'
  458 |         if (!arch_vmap_pud_supported(args->page_prot) ||
      |              ^~~~~~~~~~~~~~~~~~~~~~~

Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202402131836.OU1TDuoi-lkp@intel.com/
Fixes: 8309c9d717 ("powerpc: inline huge vmap supported functions")
Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://msgid.link/bbd84ad52bf377e8d3b5865a906f2dc5d99964ba.1707832677.git.christophe.leroy@csgroup.eu
2024-02-15 00:14:05 +11:00
Michael Ellerman
0875f1ceba powerpc/smp: Remap boot CPU onto core 0 if >= nr_cpu_ids
If nr_cpu_ids is too low to include the boot CPU, remap the boot CPU
onto logical core 0.

This is achieved in two stages. In early_init_dt_scan_cpus() the boot
CPU is renumbered to be on logical core 0, and the original boot core's
hardware ID is recorded.

Later in smp_setup_cpu_maps(), if the original boot core ID is set, the
logical CPU numbers on the 0th core are skipped in the normal device
tree search over CPU device tree nodes. Then the search is continued
until the device tree node matching the boot core is found, and those
CPUs are assigned the CPU numbers starting at 0.

This allows kdump kernels to be booted with low values for nr_cpu_ids
to conserve memory, while also allowing the crashing/boot CPU to be
any CPU.

Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Tested-by: Wen Xiong <wenxiong@us.ibm.com>
Link: https://msgid.link/20231229120107.2281153-5-mpe@ellerman.id.au
2024-02-15 00:14:05 +11:00
Michael Ellerman
9832de6544 powerpc/smp: Factor out assign_threads()
Factor out the for loop that assigns CPU numbers to threads of a core.
The function takes the next CPU number to use as input, and returns the
next available CPU number after the threads has been assigned.

This will allow a subsequent change to assign threads out of order.

Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://msgid.link/20231229120107.2281153-4-mpe@ellerman.id.au
2024-02-15 00:14:05 +11:00
Michael Ellerman
dca79603fb powerpc/smp: Lookup avail once per device tree node
The of_device_is_available() check only needs to be done once per device
node, there's no need to repeat it for each thread. Move it out of the
loop.

Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://msgid.link/20231229120107.2281153-3-mpe@ellerman.id.au
2024-02-15 00:14:04 +11:00
Michael Ellerman
777f81f0a9 powerpc/smp: Increase nr_cpu_ids to include the boot CPU
If nr_cpu_ids is too low to include the boot CPU adjust nr_cpu_ids
upward. Otherwise the kernel will BUG when trying to allocate a paca
for the boot CPU and fail to boot.

Cc: stable@vger.kernel.org
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://msgid.link/20231229120107.2281153-2-mpe@ellerman.id.au
2024-02-15 00:14:04 +11:00
Michael Ellerman
5580e96dad powerpc/smp: Adjust nr_cpu_ids to cover all threads of a core
If nr_cpu_ids is too low to include at least all the threads of a single
core adjust nr_cpu_ids upwards. This avoids triggering odd bugs in code
that assumes all threads of a core are available.

Cc: stable@vger.kernel.org
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://msgid.link/20231229120107.2281153-1-mpe@ellerman.id.au
2024-02-15 00:14:04 +11:00
Shivaprasad G Bhat
0846dd77c8 powerpc/iommu: Fix the missing iommu_group_put() during platform domain attach
The function spapr_tce_platform_iommu_attach_dev() is missing to call
iommu_group_put() when the domain is already set. This refcount leak
shows up with BUG_ON() during DLPAR remove operation as:

  KernelBug: Kernel bug in state 'None': kernel BUG at arch/powerpc/platforms/pseries/iommu.c:100!
  Oops: Exception in kernel mode, sig: 5 [#1]
  LE PAGE_SIZE=64K MMU=Radix SMP NR_CPUS=8192 NUMA pSeries
  <snip>
  Hardware name: IBM,9080-HEX POWER10 (raw) 0x800200 0xf000006 of:IBM,FW1060.00 (NH1060_016) hv:phyp pSeries
  NIP:  c0000000000ff4d4 LR: c0000000000ff4cc CTR: 0000000000000000
  REGS: c0000013aed5f840 TRAP: 0700   Tainted: G          I         (6.8.0-rc3-autotest-g99bd3cb0d12e)
  MSR:  8000000000029033 <SF,EE,ME,IR,DR,RI,LE>  CR: 44002402  XER: 20040000
  CFAR: c000000000a0d170 IRQMASK: 0
  ...
  NIP iommu_reconfig_notifier+0x94/0x200
  LR  iommu_reconfig_notifier+0x8c/0x200
  Call Trace:
    iommu_reconfig_notifier+0x8c/0x200 (unreliable)
    notifier_call_chain+0xb8/0x19c
    blocking_notifier_call_chain+0x64/0x98
    of_reconfig_notify+0x44/0xdc
    of_detach_node+0x78/0xb0
    ofdt_write.part.0+0x86c/0xbb8
    proc_reg_write+0xf4/0x150
    vfs_write+0xf8/0x488
    ksys_write+0x84/0x140
    system_call_exception+0x138/0x330
    system_call_vectored_common+0x15c/0x2ec

The patch adds the missing iommu_group_put() call.

Fixes: a8ca9fc913 ("powerpc/iommu: Do not do platform domain attach atctions after probe")
Reported-by: Venkat Rao Bagalkote <venkat88@linux.vnet.ibm.com>
Closes: https://lore.kernel.org/all/274e0d2b-b5cc-475e-94e6-8427e88e271d@linux.vnet.ibm.com/
Signed-off-by: Shivaprasad G Bhat <sbhat@linux.ibm.com>
Tested-by: Venkat Rao Bagalkote <venkat88@linux.vnet.ibm.com>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://msgid.link/170784021983.6249.10039296655906636112.stgit@linux.ibm.com
2024-02-14 23:59:23 +11:00
Maxime Jayat
2aa0a5e65e can: netlink: Fix TDCO calculation using the old data bittiming
The TDCO calculation was done using the currently applied data bittiming,
instead of the newly computed data bittiming, which means that the TDCO
had an invalid value unless setting the same data bittiming twice.

Fixes: d99755f71a ("can: netlink: add interface for CAN-FD Transmitter Delay Compensation (TDC)")
Signed-off-by: Maxime Jayat <maxime.jayat@mobile-devices.fr>
Reviewed-by: Vincent Mailhol <mailhol.vincent@wanadoo.fr>
Link: https://lore.kernel.org/all/40579c18-63c0-43a4-8d4c-f3a6c1c0b417@munic.io
Cc: stable@vger.kernel.org
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
2024-02-14 13:53:03 +01:00
Oleksij Rempel
efe7cf8280 can: j1939: Fix UAF in j1939_sk_match_filter during setsockopt(SO_J1939_FILTER)
Lock jsk->sk to prevent UAF when setsockopt(..., SO_J1939_FILTER, ...)
modifies jsk->filters while receiving packets.

Following trace was seen on affected system:
 ==================================================================
 BUG: KASAN: slab-use-after-free in j1939_sk_recv_match_one+0x1af/0x2d0 [can_j1939]
 Read of size 4 at addr ffff888012144014 by task j1939/350

 CPU: 0 PID: 350 Comm: j1939 Tainted: G        W  OE      6.5.0-rc5 #1
 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.13.0-1ubuntu1.1 04/01/2014
 Call Trace:
  print_report+0xd3/0x620
  ? kasan_complete_mode_report_info+0x7d/0x200
  ? j1939_sk_recv_match_one+0x1af/0x2d0 [can_j1939]
  kasan_report+0xc2/0x100
  ? j1939_sk_recv_match_one+0x1af/0x2d0 [can_j1939]
  __asan_load4+0x84/0xb0
  j1939_sk_recv_match_one+0x1af/0x2d0 [can_j1939]
  j1939_sk_recv+0x20b/0x320 [can_j1939]
  ? __kasan_check_write+0x18/0x20
  ? __pfx_j1939_sk_recv+0x10/0x10 [can_j1939]
  ? j1939_simple_recv+0x69/0x280 [can_j1939]
  ? j1939_ac_recv+0x5e/0x310 [can_j1939]
  j1939_can_recv+0x43f/0x580 [can_j1939]
  ? __pfx_j1939_can_recv+0x10/0x10 [can_j1939]
  ? raw_rcv+0x42/0x3c0 [can_raw]
  ? __pfx_j1939_can_recv+0x10/0x10 [can_j1939]
  can_rcv_filter+0x11f/0x350 [can]
  can_receive+0x12f/0x190 [can]
  ? __pfx_can_rcv+0x10/0x10 [can]
  can_rcv+0xdd/0x130 [can]
  ? __pfx_can_rcv+0x10/0x10 [can]
  __netif_receive_skb_one_core+0x13d/0x150
  ? __pfx___netif_receive_skb_one_core+0x10/0x10
  ? __kasan_check_write+0x18/0x20
  ? _raw_spin_lock_irq+0x8c/0xe0
  __netif_receive_skb+0x23/0xb0
  process_backlog+0x107/0x260
  __napi_poll+0x69/0x310
  net_rx_action+0x2a1/0x580
  ? __pfx_net_rx_action+0x10/0x10
  ? __pfx__raw_spin_lock+0x10/0x10
  ? handle_irq_event+0x7d/0xa0
  __do_softirq+0xf3/0x3f8
  do_softirq+0x53/0x80
  </IRQ>
  <TASK>
  __local_bh_enable_ip+0x6e/0x70
  netif_rx+0x16b/0x180
  can_send+0x32b/0x520 [can]
  ? __pfx_can_send+0x10/0x10 [can]
  ? __check_object_size+0x299/0x410
  raw_sendmsg+0x572/0x6d0 [can_raw]
  ? __pfx_raw_sendmsg+0x10/0x10 [can_raw]
  ? apparmor_socket_sendmsg+0x2f/0x40
  ? __pfx_raw_sendmsg+0x10/0x10 [can_raw]
  sock_sendmsg+0xef/0x100
  sock_write_iter+0x162/0x220
  ? __pfx_sock_write_iter+0x10/0x10
  ? __rtnl_unlock+0x47/0x80
  ? security_file_permission+0x54/0x320
  vfs_write+0x6ba/0x750
  ? __pfx_vfs_write+0x10/0x10
  ? __fget_light+0x1ca/0x1f0
  ? __rcu_read_unlock+0x5b/0x280
  ksys_write+0x143/0x170
  ? __pfx_ksys_write+0x10/0x10
  ? __kasan_check_read+0x15/0x20
  ? fpregs_assert_state_consistent+0x62/0x70
  __x64_sys_write+0x47/0x60
  do_syscall_64+0x60/0x90
  ? do_syscall_64+0x6d/0x90
  ? irqentry_exit+0x3f/0x50
  ? exc_page_fault+0x79/0xf0
  entry_SYSCALL_64_after_hwframe+0x6e/0xd8

 Allocated by task 348:
  kasan_save_stack+0x2a/0x50
  kasan_set_track+0x29/0x40
  kasan_save_alloc_info+0x1f/0x30
  __kasan_kmalloc+0xb5/0xc0
  __kmalloc_node_track_caller+0x67/0x160
  j1939_sk_setsockopt+0x284/0x450 [can_j1939]
  __sys_setsockopt+0x15c/0x2f0
  __x64_sys_setsockopt+0x6b/0x80
  do_syscall_64+0x60/0x90
  entry_SYSCALL_64_after_hwframe+0x6e/0xd8

 Freed by task 349:
  kasan_save_stack+0x2a/0x50
  kasan_set_track+0x29/0x40
  kasan_save_free_info+0x2f/0x50
  __kasan_slab_free+0x12e/0x1c0
  __kmem_cache_free+0x1b9/0x380
  kfree+0x7a/0x120
  j1939_sk_setsockopt+0x3b2/0x450 [can_j1939]
  __sys_setsockopt+0x15c/0x2f0
  __x64_sys_setsockopt+0x6b/0x80
  do_syscall_64+0x60/0x90
  entry_SYSCALL_64_after_hwframe+0x6e/0xd8

Fixes: 9d71dd0c70 ("can: add support of SAE J1939 protocol")
Reported-by: Sili Luo <rootlab@huawei.com>
Suggested-by: Sili Luo <rootlab@huawei.com>
Acked-by: Oleksij Rempel <o.rempel@pengutronix.de>
Cc: stable@vger.kernel.org
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Link: https://lore.kernel.org/all/20231020133814.383996-1-o.rempel@pengutronix.de
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
2024-02-14 13:53:03 +01:00
Ziqi Zhao
6cdedc18ba can: j1939: prevent deadlock by changing j1939_socks_lock to rwlock
The following 3 locks would race against each other, causing the
deadlock situation in the Syzbot bug report:

- j1939_socks_lock
- active_session_list_lock
- sk_session_queue_lock

A reasonable fix is to change j1939_socks_lock to an rwlock, since in
the rare situations where a write lock is required for the linked list
that j1939_socks_lock is protecting, the code does not attempt to
acquire any more locks. This would break the circular lock dependency,
where, for example, the current thread already locks j1939_socks_lock
and attempts to acquire sk_session_queue_lock, and at the same time,
another thread attempts to acquire j1939_socks_lock while holding
sk_session_queue_lock.

NOTE: This patch along does not fix the unregister_netdevice bug
reported by Syzbot; instead, it solves a deadlock situation to prepare
for one or more further patches to actually fix the Syzbot bug, which
appears to be a reference counting problem within the j1939 codebase.

Reported-by: <syzbot+1591462f226d9cbf0564@syzkaller.appspotmail.com>
Signed-off-by: Ziqi Zhao <astrajoan@yahoo.com>
Reviewed-by: Oleksij Rempel <o.rempel@pengutronix.de>
Acked-by: Oleksij Rempel <o.rempel@pengutronix.de>
Link: https://lore.kernel.org/all/20230721162226.8639-1-astrajoan@yahoo.com
[mkl: remove unrelated newline change]
Cc: stable@vger.kernel.org
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
2024-02-14 13:53:03 +01:00
Heiko Carstens
616c4ea9bc s390/vdso: remove unused ENTRY in linker scripts
When linking vdso64.so.dbg with ld.lld, there is a warning about not
finding _start for the starting address:

  ld.lld: warning: cannot find entry symbol _start; not setting start address

Fix this by removing the unused ENTRY in both vdso linker scripts. See
commit e247172854 ("powerpc/vdso: Remove unused ENTRY in linker
scripts"), which solved the same problem for powerpc, for further details.

Reviewed-by: Nathan Chancellor <nathan@kernel.org>
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
2024-02-14 13:50:53 +01:00
Arnd Bergmann
9b23fceb41 ethernet: cpts: fix function pointer cast warnings
clang-16 warns about the mismatched prototypes for the devm_* callbacks:

drivers/net/ethernet/ti/cpts.c:691:12: error: cast from 'void (*)(struct clk_hw *)' to 'void (*)(void *)' converts to incompatible function type [-Werror,-Wcast-function-type-strict]
  691 |                                        (void(*)(void *))clk_hw_unregister_mux,
      |                                        ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
include/linux/device.h:406:34: note: expanded from macro 'devm_add_action_or_reset'
  406 |         __devm_add_action_or_reset(dev, action, data, #action)
      |                                         ^~~~~~
drivers/net/ethernet/ti/cpts.c:703:12: error: cast from 'void (*)(struct device_node *)' to 'void (*)(void *)' converts to incompatible function type [-Werror,-Wcast-function-type-strict]
  703 |                                        (void(*)(void *))of_clk_del_provider,
      |                                        ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
include/linux/device.h:406:34: note: expanded from macro 'devm_add_action_or_reset'
  406 |         __devm_add_action_or_reset(dev, action, data, #action)

Use separate helper functions for this instead, using the expected prototypes
with a void* argument.

Fixes: a3047a81ba ("net: ethernet: ti: cpts: add support for ext rftclk selection")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
2024-02-14 12:50:53 +00:00
Heiko Carstens
3938490e78 s390/bug: remove entry size from __bug_table section
Commit e21f8baf8d ("s390/bug: add entry size to the __bug_table section")
changed the __EMIT_BUG() inline assembly to emit mergeable __bug_table
entries. This is at least currently not needed, but causes problems with
the upcoming s390 ld.lld support:

  ld.lld: error: drivers/nvme/host/fc.o:(__bug_table): writable SHF_MERGE section is not supported

Therefore revert the change for now.

Reported-by: Nathan Chancellor <nathan@kernel.org>
Closes: https://lore.kernel.org/all/20240207-s390-lld-and-orphan-warn-v1-0-8a665b3346ab@kernel.org/
Suggested-by: Nathan Chancellor <nathan@kernel.org>
Reviewed-by: Nathan Chancellor <nathan@kernel.org>
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
2024-02-14 13:50:53 +01:00
Nathan Chancellor
55cce52f1c s390: link vmlinux with '-z notext'
ld.bfd defaults to '-z notext' (although it is customizable with the
'--enable-textrel-check' configure option) but ld.lld defaults to '-z
text', which causes issues with building the kernel due to the presence
of dynamic relocations in sections that are not writable.

  ld.lld: error: relocation R_390_64 cannot be used against local symbol; recompile with -fPIC

Add '-z notext' to avoid these errors, as this is expected, which
matches other architectures.

Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Tested-by: Justin Stitt <justinstitt@google.com>
Reviewed-by: Fangrui Song <maskray@google.com>
Link: https://lore.kernel.org/r/20240207-s390-lld-and-orphan-warn-v1-11-8a665b3346ab@kernel.org
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
2024-02-14 13:50:53 +01:00
Nathan Chancellor
acb7c202ba s390: select CONFIG_ARCH_WANT_LD_ORPHAN_WARN
Now that all sections have been properly accounted for in the s390
linker scripts, select CONFIG_ARCH_WANT_LD_ORPHAN_WARN so that
'--orphan-handling' is added to LDFLAGS to catch any future sections
that are added without being described in linker scripts.

Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Tested-by: Justin Stitt <justinstitt@google.com>
Link: https://lore.kernel.org/r/20240207-s390-lld-and-orphan-warn-v1-10-8a665b3346ab@kernel.org
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
2024-02-14 13:50:53 +01:00
Nathan Chancellor
c0f98ea0e7 s390/boot: vmlinux.lds.S: handle commonly discarded sections
When building with CONFIG_LD_ORPHAN_WARN after selecting
CONFIG_ARCH_HAS_LD_ORPHAN_WARN, there are several series of warnings
from the various discardable sections that the kernel adds for build
purposes that are not needed at runtime:

  s390-linux-ld: warning: orphan section `.export_symbol' from `arch/s390/boot/decompressor.o' being placed in section `.export_symbol'
  s390-linux-ld: warning: orphan section `.discard.addressable' from `arch/s390/boot/decompressor.o' being placed in section `.discard.addressable'
  s390-linux-ld: warning: orphan section `.modinfo' from `arch/s390/boot/decompressor.o' being placed in section `.modinfo'

include/asm-generic/vmlinux.lds.h has a macro for easily discarding
these sections across the kernel named COMMON_DISCARDS, use it to clear
up the warnings.

Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Tested-by: Justin Stitt <justinstitt@google.com>
Link: https://lore.kernel.org/r/20240207-s390-lld-and-orphan-warn-v1-9-8a665b3346ab@kernel.org
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
2024-02-14 13:50:53 +01:00
Nathan Chancellor
6a4d37c886 s390/boot: vmlinux.lds.S: handle ELF required sections
When building with CONFIG_LD_ORPHAN_WARN after selecting
CONFIG_ARCH_HAS_LD_ORPHAN_WARN, there is a warning around the '.comment'
section for each file in arch/s390/boot

  s390-linux-ld: warning: orphan section `.comment' from `arch/s390/boot/als.o' being placed in section `.comment'
  s390-linux-ld: warning: orphan section `.comment' from `arch/s390/boot/startup.o' being placed in section `.comment'
  s390-linux-ld: warning: orphan section `.comment' from `arch/s390/boot/physmem_info.o' being placed in section `.comment'

include/asm-generic/vmlinux.lds.h has a macro for required ELF sections
not related to debugging named ELF_DETAILS, use it to clear up the
warnings.

Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Tested-by: Justin Stitt <justinstitt@google.com>
Link: https://lore.kernel.org/r/20240207-s390-lld-and-orphan-warn-v1-8-8a665b3346ab@kernel.org
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
2024-02-14 13:50:53 +01:00
Nathan Chancellor
ba6c26af1e s390/boot: vmlinux.lds.S: handle DWARF debug sections
When building with CONFIG_LD_ORPHAN_WARN after selecting
CONFIG_ARCH_HAS_LD_ORPHAN_WARN, there are several series of warnings for
each file in arch/s390/boot due to the boot linker script not handling
the DWARF debug sections:

  s390-linux-ld: warning: orphan section `.debug_line' from `arch/s390/boot/head.o' being placed in section `.debug_line'
  s390-linux-ld: warning: orphan section `.debug_info' from `arch/s390/boot/head.o' being placed in section `.debug_info'
  s390-linux-ld: warning: orphan section `.debug_abbrev' from `arch/s390/boot/head.o' being placed in section `.debug_abbrev'
  s390-linux-ld: warning: orphan section `.debug_aranges' from `arch/s390/boot/head.o' being placed in section `.debug_aranges'
  s390-linux-ld: warning: orphan section `.debug_str' from `arch/s390/boot/head.o' being placed in section `.debug_str'

include/asm-generic/vmlinux.lds.h has a macro for DWARF debug sections
named DWARF_DEBUG, use it to clear up the warnings.

Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Tested-by: Justin Stitt <justinstitt@google.com>
Reviewed-by: Fangrui Song <maskray@google.com>
Link: https://lore.kernel.org/r/20240207-s390-lld-and-orphan-warn-v1-7-8a665b3346ab@kernel.org
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
2024-02-14 13:50:52 +01:00
Nathan Chancellor
64d590a24f s390/boot: vmlinux.lds.S: handle '.rela' sections
When building with CONFIG_LD_ORPHAN_WARN after selecting
CONFIG_ARCH_HAS_LD_ORPHAN_WARN, there are several warnings from
arch/s390/boot/head.o due to the unhandled presence of '.rela' sections:

  s390-linux-ld: warning: orphan section `.rela.iplt' from `arch/s390/boot/head.o' being placed in section `.rela.dyn'
  s390-linux-ld: warning: orphan section `.rela.head.text' from `arch/s390/boot/head.o' being placed in section `.rela.dyn'
  s390-linux-ld: warning: orphan section `.rela.got' from `arch/s390/boot/head.o' being placed in section `.rela.dyn'
  s390-linux-ld: warning: orphan section `.rela.data' from `arch/s390/boot/head.o' being placed in section `.rela.dyn'
  s390-linux-ld: warning: orphan section `.rela.data.rel.ro' from `arch/s390/boot/head.o' being placed in section `.rela.dyn'
  s390-linux-ld: warning: orphan section `.rela.iplt' from `arch/s390/boot/head.o' being placed in section `.rela.dyn'
  s390-linux-ld: warning: orphan section `.rela.head.text' from `arch/s390/boot/head.o' being placed in section `.rela.dyn'
  s390-linux-ld: warning: orphan section `.rela.got' from `arch/s390/boot/head.o' being placed in section `.rela.dyn'
  s390-linux-ld: warning: orphan section `.rela.data' from `arch/s390/boot/head.o' being placed in section `.rela.dyn'
  s390-linux-ld: warning: orphan section `.rela.data.rel.ro' from `arch/s390/boot/head.o' being placed in section `.rela.dyn'

These sections are unneeded for the decompressor and they are not
emitted in the binary currently. In a manner similar to other
architectures, coalesce the sections into '.rela.dyn' and ensure it is
zero sized, which is a safe/tested approach versus full discard.

Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Tested-by: Justin Stitt <justinstitt@google.com>
Link: https://lore.kernel.org/r/20240207-s390-lld-and-orphan-warn-v1-6-8a665b3346ab@kernel.org
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
2024-02-14 13:50:52 +01:00
Nathan Chancellor
b23ab303dd s390/boot: vmlinux.lds.S: handle '.init.text'
When building with CONFIG_LD_ORPHAN_WARN after selecting
CONFIG_ARCH_HAS_LD_ORPHAN_WARN, there is a warning about the presence of
an '.init.text' section in arch/s390/boot:

  s390-linux-ld: warning: orphan section `.init.text' from `arch/s390/boot/sclp_early_core.o' being placed in section `.init.text'

arch/s390/boot/sclp_early_core.c includes a file from the main kernel
build, which picks up a usage of '__init' somewhere. For the
decompressed image, this section can just be coalesced into '.text'.

Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Tested-by: Justin Stitt <justinstitt@google.com>
Link: https://lore.kernel.org/r/20240207-s390-lld-and-orphan-warn-v1-5-8a665b3346ab@kernel.org
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
2024-02-14 13:50:52 +01:00
Nathan Chancellor
a691c8a6ef s390: vmlinux.lds.S: explicitly keep various sections
When building with CONFIG_LD_ORPHAN_WARN after selecting
CONFIG_ARCH_HAS_LD_ORPHAN_WARN, there are some warnings around certain
ELF sections:

  s390-linux-ld: warning: orphan section `.dynstr' from `arch/s390/kernel/head64.o' being placed in section `.dynstr'
  s390-linux-ld: warning: orphan section `.dynamic' from `arch/s390/kernel/head64.o' being placed in section `.dynamic'
  s390-linux-ld: warning: orphan section `.hash' from `arch/s390/kernel/head64.o' being placed in section `.hash'
  s390-linux-ld: warning: orphan section `.gnu.hash' from `arch/s390/kernel/head64.o' being placed in section `.gnu.hash'

Explicitly keep those sections like other architectures when
CONFIG_RELOCATABLE is enabled, which is always true for s390.

[hca@linux.ibm.com: keep sections instead of discarding]
Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Link: https://lore.kernel.org/r/20240207-s390-lld-and-orphan-warn-v1-4-8a665b3346ab@kernel.org
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
2024-02-14 13:50:52 +01:00
Nathan Chancellor
30226853d6 s390: vmlinux.lds.S: explicitly handle '.got' and '.plt' sections
When building with CONFIG_LD_ORPHAN_WARN after selecting
CONFIG_ARCH_HAS_LD_ORPHAN_WARN, there are a lot of warnings around the
GOT and PLT sections:

  s390-linux-ld: warning: orphan section `.plt' from `arch/s390/kernel/head64.o' being placed in section `.plt'
  s390-linux-ld: warning: orphan section `.got' from `arch/s390/kernel/head64.o' being placed in section `.got'
  s390-linux-ld: warning: orphan section `.got.plt' from `arch/s390/kernel/head64.o' being placed in section `.got.plt'
  s390-linux-ld: warning: orphan section `.iplt' from `arch/s390/kernel/head64.o' being placed in section `.iplt'
  s390-linux-ld: warning: orphan section `.igot.plt' from `arch/s390/kernel/head64.o' being placed in section `.igot.plt'

  s390-linux-ld: warning: orphan section `.iplt' from `arch/s390/boot/head.o' being placed in section `.iplt'
  s390-linux-ld: warning: orphan section `.igot.plt' from `arch/s390/boot/head.o' being placed in section `.igot.plt'
  s390-linux-ld: warning: orphan section `.got' from `arch/s390/boot/head.o' being placed in section `.got'

Currently, only the '.got' section is actually emitted in the final
binary. In a manner similar to other architectures, put the '.got'
section near the '.data' section and coalesce the PLT sections,
checking that the final section is zero sized, which is a safe/tested
approach versus full discard.

Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Tested-by: Justin Stitt <justinstitt@google.com>
Link: https://lore.kernel.org/r/20240207-s390-lld-and-orphan-warn-v1-3-8a665b3346ab@kernel.org
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
2024-02-14 13:50:52 +01:00
Nathan Chancellor
bdf2cd27a3 s390: vmlinux.lds.S: handle '.data.rel' sections explicitly
When building with CONFIG_LD_ORPHAN_WARN after selecting
CONFIG_ARCH_HAS_LD_ORPHAN_WARN, there are a lot of warnings around
'.data.rel' sections:

  s390-linux-ld: warning: orphan section `.data.rel' from `kernel/sched/build_utility.o' being placed in section `.data.rel'
  s390-linux-ld: warning: orphan section `.data.rel.local' from `kernel/sched/build_utility.o' being placed in section `.data.rel.local'
  s390-linux-ld: warning: orphan section `.data.rel.ro' from `kernel/sched/build_utility.o' being placed in section `.data.rel.ro'
  s390-linux-ld: warning: orphan section `.data.rel.ro.local' from `kernel/sched/build_utility.o' being placed in section `.data.rel.ro.local'

Describe these in vmlinux.lds.S so there is no more warning and the
sections are placed consistently between linkers.

Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Tested-by: Justin Stitt <justinstitt@google.com>
Link: https://lore.kernel.org/r/20240207-s390-lld-and-orphan-warn-v1-2-8a665b3346ab@kernel.org
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
2024-02-14 13:50:52 +01:00
Nathan Chancellor
2151fd9a6d s390/boot: add support for CONFIG_LD_ORPHAN_WARN
arch/s390/boot/vmlinux uses a different linker script and build rules
than the main vmlinux, so the '--orphan-handling' flag is not applied to
it. Add support for '--orphan-handling' so that all sections are
properly described in the linker script, which helps eliminate bugs
between linker implementations having different orphan section
heuristics.

Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Tested-by: Justin Stitt <justinstitt@google.com>
Link: https://lore.kernel.org/r/20240207-s390-lld-and-orphan-warn-v1-1-8a665b3346ab@kernel.org
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
2024-02-14 13:50:52 +01:00
Arnd Bergmann
5d07e432cb bnad: fix work_queue type mismatch
clang-16 warns about a function pointer cast:

drivers/net/ethernet/brocade/bna/bnad.c:1995:4: error: cast from 'void (*)(struct delayed_work *)' to 'work_func_t' (aka 'void (*)(struct work_struct *)') converts to incompatible function type [-Werror,-Wcast-function-type-strict]
 1995 |                         (work_func_t)bnad_tx_cleanup);
drivers/net/ethernet/brocade/bna/bnad.c:2252:4: error: cast from 'void (*)(void *)' to 'work_func_t' (aka 'void (*)(struct work_struct *)') converts to incompatible function type [-Werror,-Wcast-function-type-strict]
 2252 |                         (work_func_t)(bnad_rx_cleanup));

The problem here is mixing up work_struct and delayed_work, which relies
the former being the first member of the latter.

Change the code to use consistent types here to address the warning and
make it more robust against workqueue interface changes.

Side note: the use of a delayed workqueue for cleaning up TX descriptors
is probably a bad idea since this introduces a noticeable delay. The
driver currently does not appear to use BQL, but if one wanted to add
that, this would have to be changed as well.

Fixes: 01b54b1451 ("bna: tx rx cleanup fix")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
2024-02-14 12:49:47 +00:00
Marc Kleine-Budde
240335736b Merge patch series "can: tcan4x5x: support resume upon rx can frame"
Martin Hundebøll <martin@geanix.com> says:

This is the third iteration of the previous submitted patches [0] and
[1].

This revision replaces the "wake_source" function parameters to a flag
in the class device structure, and adds a patch to document the
"wakeup-source" device tree property.

Also, the previous revisions forgot to mention that the patches are
based on Markus' coalescing patches [2]. Those implements caching of
the enabled interrupts, which is handy when restoring the set of
interrupts in the resume path.

[0] https://lore.kernel.org/linux-can/20230912093807.1383720-1-martin@geanix.com
[1] https://lore.kernel.org/linux-can/20230919122841.3803289-1-martin@geanix.com
[2] https://lore.kernel.org/linux-can/20230929141304.3934380-1-msp@baylibre.com

Link: https://lore.kernel.org/all/20231113131452.214961-1-martin@geanix.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
2024-02-14 13:44:50 +01:00
Martin Hundebøll
b6b640c044 can: tcan4x5x: support resuming from rx interrupt signal
Implement the "wakeup-source" device tree property, so the chip is left
running when suspending, and its rx interrupt is used as a wakeup source
to resume operation.

Signed-off-by: Martin Hundebøll <martin@geanix.com>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
2024-02-14 13:44:02 +01:00
Martin Hundebøll
4a94d7e31c can: m_can: allow keeping the transceiver running in suspend
Add a flag to the device class structure that leaves the chip in a
running state with rx interrupt enabled, so that an m_can device driver
can configure and use the interrupt as a wakeup source.

Signed-off-by: Martin Hundebøll <martin@geanix.com>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
2024-02-14 13:44:02 +01:00
Martin Hundebøll
b00cf4f629 dt-bindings: can: tcan4x5x: Document the wakeup-source flag
Let it be known that the tcan4x5x device can now be configured to wake
the host from suspend when a can frame is received.

Signed-off-by: Martin Hundebøll <martin@geanix.com>
Acked-by: Conor Dooley <conor.dooley@microchip.com>
[mkl: make first the first patch]
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
2024-02-14 13:44:02 +01:00
Mark Brown
81ff296a81
ASoC: meson: aiu: fix function pointer type
Merge series from Jerome Brunet <jbrunet@baylibre.com>:

This patchset fixes 2 -Wcast-function-type-strict warning in amlogic
audio drivers with clang 16.
2024-02-14 12:17:04 +00:00
Catalin Popescu
d1d77120bc net: phy: dp83826: support TX data voltage tuning
DP83826 offers the possibility to tune the voltage of logical
levels of the MLT-3 encoded TX data. This is useful when there
is a voltage drop in between the PHY and the connector and we
want to increase the voltage levels to compensate for that drop.

Prior to PHY configuration, the driver SW resets the PHY which has
the same effect as the HW reset pin according to the datasheet.
Hence, there's no need to force update the VOD_CFG registers to make
sure they hold their reset values. VOD_CFG registers need to be
updated only if the DT has been configured with values other than
the reset ones.

Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Catalin Popescu <catalin.popescu@leica-geosystems.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2024-02-14 12:06:10 +00:00
Catalin Popescu
ed1d7dac08 dt-bindings: net: dp83826: support TX data voltage tuning
Add properties ti,cfg-dac-minus-one-bp/ti,cfg-dac-plus-one-bp
to support voltage tuning of logical levels -1/+1 of the MLT-3
encoded TX data.

Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Catalin Popescu <catalin.popescu@leica-geosystems.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2024-02-14 12:06:09 +00:00
Mark Brown
2775f88bec
ASoC: SOF: Extend ChainDMA and DSPless mode to LNL+
Merge series from Peter Ujfalusi <peter.ujfalusi@linux.intel.com>:

For both ChainDMA and DSPless mode the requirement is that the link must
be serviced by HD-DMA.
On pre Lunar Lake platforms this was only valid for HDAudio links but with
Lunar Lake all link types now serviced by HD-DMA.

This allows us to enable ChainDMA and DSPless mode for SoundWire links as
well.
2024-02-14 11:41:32 +00:00
David S. Miller
7c754e6a6c Merge branch 'dev_base_lock-remove'
Eric Dumazet says:

====================
net: complete dev_base_lock removal

Back in 2009 we started an effort to get rid of dev_base_lock
in favor of RCU.

It is time to finish this work.

Say goodbye to dev_base_lock !

v4: rebase, and move dev_addr_sem to net/core/dev.h in patch 06/13 (Jakub)

v3: I misread kbot reports, the issue was with dev->operstate (patch 10/13)
    So dev->reg_state is back to u8, and dev->operstate becomes an u32.
    Sorry for the noise.

v2: dev->reg_state must be a standard enum, some arches
    do not support cmpxchg() on u8.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2024-02-14 11:20:14 +00:00
Eric Dumazet
1b3ef46cb7 net: remove dev_base_lock
dev_base_lock is not needed anymore, all remaining users also hold RTNL.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2024-02-14 11:20:14 +00:00
Eric Dumazet
e51b962438 net: remove dev_base_lock from register_netdevice() and friends.
RTNL already protects writes to dev->reg_state, we no longer need to hold
dev_base_lock to protect the readers.

unlist_netdevice() second argument can be removed.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2024-02-14 11:20:14 +00:00
Eric Dumazet
2dd4d828d6 net: remove dev_base_lock from do_setlink()
We hold RTNL here, and dev->link_mode readers already
are using READ_ONCE().

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2024-02-14 11:20:14 +00:00
Eric Dumazet
6a2968ee1e net: add netdev_set_operstate() helper
dev_base_lock is going away, add netdev_set_operstate() helper
so that hsr does not have to know core internals.

Remove dev_base_lock acquisition from rfc2863_policy()

v3: use an "unsigned int" for dev->operstate,
    so that try_cmpxchg() can work on all arches.
        ( https://lore.kernel.org/oe-kbuild-all/202402081918.OLyGaea3-lkp@intel.com/ )

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2024-02-14 11:20:13 +00:00
Eric Dumazet
328771deab net: remove stale mentions of dev_base_lock in comments
Change comments incorrectly mentioning dev_base_lock.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2024-02-14 11:20:13 +00:00
Eric Dumazet
e154bb7a6e net-sysfs: convert netstat_show() to RCU
dev_get_stats() can be called from RCU, there is no need
to acquire dev_base_lock.

Change dev_isalive() comment to reflect we no longer use
dev_base_lock from net/core/net-sysfs.c

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2024-02-14 11:20:13 +00:00
Eric Dumazet
004d138364 net-sysfs: convert dev->operstate reads to lockless ones
operstate_show() can omit dev_base_lock acquisition only
to read dev->operstate.

Annotate accesses to dev->operstate.

Writers still acquire dev_base_lock for mutual exclusion.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2024-02-14 11:20:13 +00:00
Eric Dumazet
c7d52737e7 net-sysfs: use dev_addr_sem to remove races in address_show()
Using dev_base_lock is not preventing from reading garbage.

Use dev_addr_sem instead.

v4: place dev_addr_sem extern in net/core/dev.h (Jakub Kicinski)
 Link: https://lore.kernel.org/netdev/20240212175845.10f6680a@kernel.org/

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2024-02-14 11:20:13 +00:00
Eric Dumazet
12692e3df2 net-sysfs: convert netdev_show() to RCU
Make clear dev_isalive() can be called with RCU protection.

Then convert netdev_show() to RCU, to remove dev_base_lock
dependency.

Also add RCU to broadcast_show().

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2024-02-14 11:20:13 +00:00
Eric Dumazet
4d42b37def net: convert dev->reg_state to u8
Prepares things so that dev->reg_state reads can be lockless,
by adding WRITE_ONCE() on write side.

READ_ONCE()/WRITE_ONCE() do not support bitfields.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2024-02-14 11:20:13 +00:00
Eric Dumazet
a6473fe9b6 dev: annotate accesses to dev->link
Following patch will read dev->link locklessly,
annotate the write from do_setlink().

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2024-02-14 11:20:13 +00:00
Eric Dumazet
f694eee9e1 ip_tunnel: annotate data-races around t->parms.link
t->parms.link is read locklessly, annotate these reads
and opposite writes accordingly.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2024-02-14 11:20:13 +00:00
Eric Dumazet
1c07dbb0cc net: annotate data-races around dev->name_assign_type
name_assign_type_show() runs locklessly, we should annotate
accesses to dev->name_assign_type.

Alternative would be to grab devnet_rename_sem semaphore
from name_assign_type_show(), but this would not bring
more accuracy.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2024-02-14 11:20:13 +00:00
Leonard Göhrs
7fd195f01a ARM: dts: stm32: lxa-tac: reduce RGMII interface drive strength
This results in an improvement of EMI performance by reducing the
125MHz emission and its harmonics caused by the RGMII clock line.

Signed-off-by: Leonard Göhrs <l.goehrs@pengutronix.de>
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Alexandre Torgue <alexandre.torgue@foss.st.com>
2024-02-14 12:09:41 +01:00
David S. Miller
b7f9ef7278 Merge branch 'per-epoll-context-busy-poll'
Joe Damato says:

====================
Per epoll context busy poll support

Greetings:

Welcome to v8.

TL;DR This builds on commit bf3b9f6372 ("epoll: Add busy poll support to
epoll with socket fds.") by allowing user applications to enable
epoll-based busy polling, set a busy poll packet budget, and enable or
disable prefer busy poll on a per epoll context basis.

This makes epoll-based busy polling much more usable for user
applications than the current system-wide sysctl and hardcoded budget.

To allow for this, two ioctls have been added for epoll contexts for
getting and setting a new struct, struct epoll_params.

ioctl was chosen vs a new syscall after reviewing a suggestion by Willem
de Bruijn [1]. I am open to using a new syscall instead of an ioctl, but it
seemed that:
  - Busy poll affects all existing epoll_wait and epoll_pwait variants in
    the same way, so new verions of many syscalls might be needed. It
    seems much simpler for users to use the correct
    epoll_wait/epoll_pwait for their app and add a call to ioctl to enable
    or disable busy poll as needed. This also probably means less work to
    get an existing epoll app using busy poll.

  - previously added epoll_pwait2 helped to bring epoll closer to
    existing syscalls (like pselect and ppoll) and this busy poll change
    reflected as a new syscall would not have the same effect.

Note: patch 1/4 as of v4 uses an or (||) instead of an xor. I thought about
it some more and I realized that if the user enables both the per-epoll
context setting and the system wide sysctl, then busy poll should be
enabled and not disabled. Using xor doesn't seem to make much sense after
thinking through this a bit.

Longer explanation:

Presently epoll has support for a very useful form of busy poll based on
the incoming NAPI ID (see also: SO_INCOMING_NAPI_ID [2]).

This form of busy poll allows epoll_wait to drive NAPI packet processing
which allows for a few interesting user application designs which can
reduce latency and also potentially improve L2/L3 cache hit rates by
deferring NAPI until userland has finished its work.

The documentation available on this is, IMHO, a bit confusing so please
allow me to explain how one might use this:

1. Ensure each application thread has its own epoll instance mapping
1-to-1 with NIC RX queues. An n-tuple filter would likely be used to
direct connections with specific dest ports to these queues.

2. Optionally: Setup IRQ coalescing for the NIC RX queues where busy
polling will occur. This can help avoid the userland app from being
pre-empted by a hard IRQ while userland is running. Note this means that
userland must take care to call epoll_wait and not take too long in
userland since it now drives NAPI via epoll_wait.

3. Optionally: Consider using napi_defer_hard_irqs and gro_flush_timeout to
further restrict IRQ generation from the NIC. These settings are
system-wide so their impact must be carefully weighed against the running
applications.

4. Ensure that all incoming connections added to an epoll instance
have the same NAPI ID. This can be done with a BPF filter when
SO_REUSEPORT is used or getsockopt + SO_INCOMING_NAPI_ID when a single
accept thread is used which dispatches incoming connections to threads.

5. Lastly, busy poll must be enabled via a sysctl
(/proc/sys/net/core/busy_poll).

Please see Eric Dumazet's paper about busy polling [3] and a recent
academic paper about measured performance improvements of busy polling [4]
(albeit with a modification that is not currently present in the kernel)
for additional context.

The unfortunate part about step 5 above is that this enables busy poll
system-wide which affects all user applications on the system,
including epoll-based network applications which were not intended to
be used this way or applications where increased CPU usage for lower
latency network processing is unnecessary or not desirable.

If the user wants to run one low latency epoll-based server application
with epoll-based busy poll, but would like to run the rest of the
applications on the system (which may also use epoll) without busy poll,
this system-wide sysctl presents a significant problem.

This change preserves the system-wide sysctl, but adds a mechanism (via
ioctl) to enable or disable busy poll for epoll contexts as needed by
individual applications, making epoll-based busy poll more usable.

Note that this change includes an or (as of v4) instead of an xor. If the
user has enabled both the system-wide sysctl and also the per epoll-context
busy poll settings, then epoll should probably busy poll (vs being
disabled).

Thanks,
Joe

v7 -> v8:
   - Reviewed-by tag from Eric Dumazet applied to commit message of patch
     1/4.

   - patch 4/4:
     - EPIOCSPARAMS and EPIOCGPARAMS updated to use WRITE_ONCE and
       READ_ONCE, as requested by Eric Dumazet
     - Wrapped a long line (via netdev/checkpatch)

v6 -> v7:
   - Acked-by tags from Stanislav Fomichev applied to commit messages of
     all patches.
   - Reviewed-by tags from Jakub Kicinski, Eric Dumazet applied to commit
     messages of patches 2 and 3. Jiri Slaby's Reviewed-by applied to patch
     4.

   - patch 1/4:
     - busy_poll_usecs reduced from u64 to u32.
     - Unnecessary parens removed (via netdev/checkpatch)
     - Wrapped long line (via netdev/checkpatch)
     - Remove inline from busy_loop_ep_timeout as objdump suggests the
       function is already inlined
     - Moved struct eventpoll assignment to declaration
     - busy_loop_ep_timeout is moved within CONFIG_NET_RX_BUSY_POLL and the
       ifdefs internally have been removed as per Eric Dumazet's review
     - Removed ep_busy_loop_on from the !defined CONFIG_NET_RX_BUSY_POLL
       section as it is only called when CONFIG_NET_RX_BUSY_POLL is
       defined

   - patch 3/4:
     - Fix whitespace alignment issue (via netdev/checkpatch)

   - patch 4/4:
     - epoll_params.busy_poll_usecs has been reduced to u32
     - epoll_params.busy_poll_usecs is now checked to ensure it is <=
       S32_MAX
     - __pad has been reduced to a single u8
     - memchr_inv has been dropped and replaced with a simple check for the
       single __pad byte
     - Removed space after cast (via netdev/checkpatch)
     - Wrap long line (via netdev/checkpatch)
     - Move struct eventpoll *ep assignment to declaration as per Jiri
       Slaby's review
     - Remove unnecessary !! as per Jiri Slaby's review
     - Reorganized variables to be reverse christmas tree order

v5 -> v6:
  - patch 1/3 no functional change, but commit message corrected to explain
    that an or (||) is being used instead of xor.

  - patch 3/4 is a new patch which adds support for per epoll context
    prefer busy poll setting.

  - patch 4/4 updated to allow getting/setting per epoll context prefer
    busy poll setting; this setting is limited to either 0 or 1.

v4 -> v5:
  - patch 3/3 updated to use memchr_inv to ensure that __pad is zero for
    the EPIOCSPARAMS ioctl. Recommended by Greg K-H [5], Dave Chinner [6],
    and Jiri Slaby [7].

v3 -> v4:
  - patch 1/3 was updated to include an important functional change:
    ep_busy_loop_on was updated to use or (||) instead of xor (^). After
    thinking about it a bit more, I thought xor didn't make much sense.
    Enabling both the per-epoll context and the system-wide sysctl should
    probably enable busy poll, not disable it. So, or (||) makes more
    sense, I think.

  - patch 3/3 was updated:
    - to change the epoll_params fields to be __u64, __u16, and __u8 and
      to pad the struct to a multiple of 64bits. Suggested by Greg K-H [8]
      and Arnd Bergmann [9].
    - remove an unused pr_fmt, left over from the previous revision.
    - ioctl now returns -EINVAL when epoll_params.busy_poll_usecs >
      U32_MAX.

v2 -> v3:
  - cover letter updated to mention why ioctl seems (to me) like a better
    choice vs a new syscall.

  - patch 3/4 was modified in 3 ways:
    - when an unknown ioctl is received, -ENOIOCTLCMD is returned instead
      of -EINVAL as the ioctl documentation requires.
    - epoll_params.busy_poll_budget can only be set to a value larger than
      NAPI_POLL_WEIGHT if code is run by privileged (CAP_NET_ADMIN) users.
      Otherwise, -EPERM is returned.
    - busy poll specific ioctl code moved out to its own function. On
      kernels without busy poll support, -EOPNOTSUPP is returned. This also
      makes the kernel build robot happier without littering the code with
      more #ifdefs.

  - dropped patch 4/4 after Eric Dumazet's review of it when it was sent
    independently to the list [10].

v1 -> v2:
  - cover letter updated to make a mention of napi_defer_hard_irqs and
    gro_flush_timeout as an added step 3 and to cite both Eric Dumazet's
    busy polling paper and a paper from University of Waterloo for
    additional context. Specifically calling out the xor in patch 1/4
    incase it is missed by reviewers.

  - Patch 2/4 has its commit message updated, but no functional changes.
    Commit message now describes that allowing for a settable budget helps
    to improve throughput and is more consistent with other busy poll
    mechanisms that allow a settable budget via SO_BUSY_POLL_BUDGET.

  - Patch 3/4 was modified to check if the epoll_params.busy_poll_budget
    exceeds NAPI_POLL_WEIGHT. The larger value is allowed, but an error is
    printed. This was done for consistency with netif_napi_add_weight,
    which does the same.

  - Patch 3/4 the struct epoll_params was updated to fix the type of the
    data field; it was uint8_t and was changed to u8.

  - Patch 4/4 added to check if SO_BUSY_POLL_BUDGET exceeds
    NAPI_POLL_WEIGHT. The larger value is allowed, but an error is
    printed. This was done for consistency with netif_napi_add_weight,
    which does the same.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2024-02-14 11:01:01 +00:00
Joe Damato
18e2bf0edf eventpoll: Add epoll ioctl for epoll_params
Add an ioctl for getting and setting epoll_params. User programs can use
this ioctl to get and set the busy poll usec time, packet budget, and
prefer busy poll params for a specific epoll context.

Parameters are limited:
  - busy_poll_usecs is limited to <= s32_max
  - busy_poll_budget is limited to <= NAPI_POLL_WEIGHT by unprivileged
    users (!capable(CAP_NET_ADMIN))
  - prefer_busy_poll must be 0 or 1
  - __pad must be 0

Signed-off-by: Joe Damato <jdamato@fastly.com>
Acked-by: Stanislav Fomichev <sdf@google.com>
Reviewed-by: Jiri Slaby <jirislaby@kernel.org>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2024-02-14 11:01:01 +00:00