linux.git/drivers/net/ethernet/broadcom, branch vsnprintf

net: ethernet: bgmac-platform: fix an OF node reference leak

2024-12-17T12:22:05+00:00

The OF node obtained by of_parse_phandle() is not freed. Call
of_node_put() to balance the refcount.

This bug was found by an experimental static analysis tool that I am
developing.

Fixes: 1676aba5ef7e ("net: ethernet: bgmac: device tree phy enablement")
Signed-off-by: Joe Hattori 
Reviewed-by: Simon Horman 
Link: https://patch.msgid.link/20241214014912.2810315-1-joe@pf.is.s.u-tokyo.ac.jp
Signed-off-by: Paolo Abeni

bnxt_en: Fix aggregation ID mask to prevent oops on 5760X chips

2024-12-11T02:25:38+00:00

The 5760X (P7) chip's HW GRO/LRO interface is very similar to that of
the previous generation (5750X or P5).  However, the aggregation ID
fields in the completion structures on P7 have been redefined from
16 bits to 12 bits.  The freed up 4 bits are redefined for part of the
metadata such as the VLAN ID.  The aggregation ID mask was not modified
when adding support for P7 chips.  Including the extra 4 bits for the
aggregation ID can potentially cause the driver to store or fetch the
packet header of GRO/LRO packets in the wrong TPA buffer.  It may hit
the BUG() condition in __skb_pull() because the SKB contains no valid
packet header:

kernel BUG at include/linux/skbuff.h:2766!
Oops: invalid opcode: 0000 1 PREEMPT SMP NOPTI
CPU: 4 UID: 0 PID: 0 Comm: swapper/4 Kdump: loaded Tainted: G           OE      6.12.0-rc2+ #7
Tainted: [O]=OOT_MODULE, [E]=UNSIGNED_MODULE
Hardware name: Dell Inc. PowerEdge R760/0VRV9X, BIOS 1.0.1 12/27/2022
RIP: 0010:eth_type_trans+0xda/0x140
Code: 80 00 00 00 eb c1 8b 47 70 2b 47 74 48 8b 97 d0 00 00 00 83 f8 01 7e 1b 48 85 d2 74 06 66 83 3a ff 74 09 b8 00 04 00 00 eb a5 <0f> 0b b8 00 01 00 00 eb 9c 48 85 ff 74 eb 31 f6 b9 02 00 00 00 48
RSP: 0018:ff615003803fcc28 EFLAGS: 00010283
RAX: 00000000000022d2 RBX: 0000000000000003 RCX: ff2e8c25da334040
RDX: 0000000000000040 RSI: ff2e8c25c1ce8000 RDI: ff2e8c25869f9000
RBP: ff2e8c258c31c000 R08: ff2e8c25da334000 R09: 0000000000000001
R10: ff2e8c25da3342c0 R11: ff2e8c25c1ce89c0 R12: ff2e8c258e0990b0
R13: ff2e8c25bb120000 R14: ff2e8c25c1ce89c0 R15: ff2e8c25869f9000
FS:  0000000000000000(0000) GS:ff2e8c34be300000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 000055f05317e4c8 CR3: 000000108bac6006 CR4: 0000000000773ef0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe07f0 DR7: 0000000000000400
PKRU: 55555554
Call Trace:
 
 ? die+0x33/0x90
 ? do_trap+0xd9/0x100
 ? eth_type_trans+0xda/0x140
 ? do_error_trap+0x65/0x80
 ? eth_type_trans+0xda/0x140
 ? exc_invalid_op+0x4e/0x70
 ? eth_type_trans+0xda/0x140
 ? asm_exc_invalid_op+0x16/0x20
 ? eth_type_trans+0xda/0x140
 bnxt_tpa_end+0x10b/0x6b0 [bnxt_en]
 ? bnxt_tpa_start+0x195/0x320 [bnxt_en]
 bnxt_rx_pkt+0x902/0xd90 [bnxt_en]
 ? __bnxt_tx_int.constprop.0+0x89/0x300 [bnxt_en]
 ? kmem_cache_free+0x343/0x440
 ? __bnxt_tx_int.constprop.0+0x24f/0x300 [bnxt_en]
 __bnxt_poll_work+0x193/0x370 [bnxt_en]
 bnxt_poll_p5+0x9a/0x300 [bnxt_en]
 ? try_to_wake_up+0x209/0x670
 __napi_poll+0x29/0x1b0

Fix it by redefining the aggregation ID mask for P5_PLUS chips to be
12 bits.  This will work because the maximum aggregation ID is less
than 4096 on all P5_PLUS chips.

Fixes: 13d2d3d381ee ("bnxt_en: Add new P7 hardware interface definitions")
Reviewed-by: Damodharam Ammepalli 
Reviewed-by: Kalesh AP 
Reviewed-by: Andy Gospodarek 
Signed-off-by: Michael Chan 
Reviewed-by: Simon Horman 
Link: https://patch.msgid.link/20241209015448.1937766-1-michael.chan@broadcom.com
Signed-off-by: Jakub Kicinski

bnxt_en: Fix potential crash when dumping FW log coredump

2024-12-07T01:39:13+00:00

If the FW log context memory is retained after FW reset, the existing
code is not handling the condition correctly and zeroes out the data
structures.  This potentially will cause a division by zero crash
when the user runs ethtool -w.  The last_type is also not set
correctly when the context memory is retained.  This will cause errors
because the last_type signals to the FW that all context memory types
have been configured.

Oops: divide error: 0000 1 PREEMPT SMP NOPTI
CPU: 53 UID: 0 PID: 7019 Comm: ethtool Kdump: loaded Tainted: G           OE      6.12.0-rc7+ #1
Tainted: [O]=OOT_MODULE, [E]=UNSIGNED_MODULE
Hardware name: Supermicro SYS-621C-TN12R/X13DDW-A, BIOS 1.4 08/10/2023
RIP: 0010:__bnxt_copy_ctx_mem.constprop.0.isra.0+0x86/0x160 [bnxt_en]
Code: 0a 31 d2 4c 89 6c 24 10 45 8b a5 fc df ff ff 4c 8b 74 24 20 31 db 66 89 44 24 06 48 63 c5 c1 e5 09 4c 0f af e0 48 8b 44 24 30 <49> f7 f4 4c 89 64 24 08 48 63 c5 4d 89 ec 31 ed 48 89 44 24 18 49
RSP: 0018:ff480591603d78b8 EFLAGS: 00010206
RAX: 0000000000100000 RBX: 0000000000000000 RCX: 0000000000000000
RDX: 0000000000000000 RSI: ff23959e46740000 RDI: 0000000000000000
RBP: 0000000000000000 R08: 0000000000100000 R09: ff23959e46740000
R10: ff480591603d7a18 R11: 0000000000000010 R12: 0000000000000000
R13: ff23959e46742008 R14: 0000000000000000 R15: 0000000000000000
FS:  00007f04227c1740(0000) GS:ff2395adbf680000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 00007f04225b33a5 CR3: 000000108b9a4001 CR4: 0000000000773ef0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe07f0 DR7: 0000000000000400
PKRU: 55555554
Call Trace:
 
 ? die+0x33/0x90
 ? do_trap+0xd9/0x100
 ? __bnxt_copy_ctx_mem.constprop.0.isra.0+0x86/0x160 [bnxt_en]
 ? do_error_trap+0x65/0x80
 ? __bnxt_copy_ctx_mem.constprop.0.isra.0+0x86/0x160 [bnxt_en]
 ? exc_divide_error+0x36/0x50
 ? __bnxt_copy_ctx_mem.constprop.0.isra.0+0x86/0x160 [bnxt_en]
 ? asm_exc_divide_error+0x16/0x20
 ? __bnxt_copy_ctx_mem.constprop.0.isra.0+0x86/0x160 [bnxt_en]
 ? __bnxt_copy_ctx_mem.constprop.0.isra.0+0xda/0x160 [bnxt_en]
 bnxt_get_ctx_coredump.constprop.0+0x1ed/0x390 [bnxt_en]
 ? __memcg_slab_post_alloc_hook+0x21c/0x3c0
 ? __bnxt_get_coredump+0x473/0x4b0 [bnxt_en]
 __bnxt_get_coredump+0x473/0x4b0 [bnxt_en]
 ? security_file_alloc+0x74/0xe0
 ? cred_has_capability.isra.0+0x78/0x120
 bnxt_get_coredump_length+0x4b/0xf0 [bnxt_en]
 bnxt_get_dump_flag+0x40/0x60 [bnxt_en]
 __dev_ethtool+0x17e4/0x1fc0
 ? syscall_exit_to_user_mode+0xc/0x1d0
 ? do_syscall_64+0x85/0x150
 ? unmap_page_range+0x299/0x4b0
 ? vma_interval_tree_remove+0x215/0x2c0
 ? __kmalloc_cache_noprof+0x10a/0x300
 dev_ethtool+0xa8/0x170
 dev_ioctl+0x1b5/0x580
 ? sk_ioctl+0x4a/0x110
 sock_do_ioctl+0xab/0xf0
 sock_ioctl+0x1ca/0x2e0
 __x64_sys_ioctl+0x87/0xc0
 do_syscall_64+0x79/0x150

Fixes: 24d694aec139 ("bnxt_en: Allocate backing store memory for FW trace logs")
Signed-off-by: Hongguang Gao 
Signed-off-by: Michael Chan 
Link: https://patch.msgid.link/20241204215918.1692597-3-michael.chan@broadcom.com
Signed-off-by: Jakub Kicinski

bnxt_en: Fix GSO type for HW GRO packets on 5750X chips

2024-12-07T01:39:13+00:00

The existing code is using RSS profile to determine IPV4/IPV6 GSO type
on all chips older than 5760X.  This won't work on 5750X chips that may
be using modified RSS profiles.  This commit from 2018 has updated the
driver to not use RSS profile for HW GRO packets on newer chips:

50f011b63d8c ("bnxt_en: Update RSS setup and GRO-HW logic according to the latest spec.")

However, a recent commit to add support for the newest 5760X chip broke
the logic.  If the GRO packet needs to be re-segmented by the stack, the
wrong GSO type will cause the packet to be dropped.

Fix it to only use RSS profile to determine GSO type on the oldest
5730X/5740X chips which cannot use the new method and is safe to use the
RSS profiles.

Also fix the L3/L4 hash type for RX packets by not using the RSS
profile for the same reason.  Use the ITYPE field in the RX completion
to determine L3/L4 hash types correctly.

Fixes: a7445d69809f ("bnxt_en: Add support for new RX and TPA_START completion types for P7")
Reviewed-by: Colin Winegarden 
Reviewed-by: Somnath Kotur 
Reviewed-by: Kalesh AP 
Signed-off-by: Michael Chan 
Link: https://patch.msgid.link/20241204215918.1692597-2-michael.chan@broadcom.com
Signed-off-by: Jakub Kicinski

bnxt_en: handle tpa_info in queue API implementation

2024-12-05T03:23:35+00:00

Commit 7ed816be35ab ("eth: bnxt: use page pool for head frags") added a
page pool for header frags, which may be distinct from the existing pool
for the aggregation ring. Prior to this change, frags used in the TPA
ring rx_tpa were allocated from system memory e.g. napi_alloc_frag()
meaning their lifetimes were not associated with a page pool. They can
be returned at any time and so the queue API did not alloc or free
rx_tpa.

But now frags come from a separate head_pool which may be different to
page_pool. Without allocating and freeing rx_tpa, frags allocated from
the old head_pool may be returned to a different new head_pool which
causes a mismatch between the pp hold/release count.

Fix this problem by properly freeing and allocating rx_tpa in the queue
API implementation.

Signed-off-by: David Wei 
Reviewed-by: Michael Chan 
Link: https://patch.msgid.link/20241204041022.56512-4-dw@davidwei.uk
Signed-off-by: Jakub Kicinski

bnxt_en: refactor bnxt_alloc_rx_rings() to call bnxt_alloc_rx_agg_bmap()

2024-12-05T03:23:35+00:00

Refactor bnxt_alloc_rx_rings() to call bnxt_alloc_rx_agg_bmap() for
allocating rx_agg_bmap.

Reviewed-by: Somnath Kotur 
Signed-off-by: David Wei 
Reviewed-by: Michael Chan 
Link: https://patch.msgid.link/20241204041022.56512-3-dw@davidwei.uk
Signed-off-by: Jakub Kicinski

bnxt_en: refactor tpa_info alloc/free into helpers

2024-12-05T03:23:35+00:00

Refactor bnxt_rx_ring_info->tpa_info operations into helpers that work
on a single tpa_info in prep for queue API using them.

There are 2 pairs of operations:

* bnxt_alloc_one_tpa_info()
* bnxt_free_one_tpa_info()

These alloc/free the tpa_info array itself.

* bnxt_alloc_one_tpa_info_data()
* bnxt_free_one_tpa_info_data()

These alloc/free the frags stored in tpa_info array.

Reviewed-by: Somnath Kotur 
Signed-off-by: David Wei 
Reviewed-by: Michael Chan 
Link: https://patch.msgid.link/20241204041022.56512-2-dw@davidwei.uk
Signed-off-by: Jakub Kicinski

bnxt_en: ethtool: Supply ntuple rss context action

2024-11-30T22:16:12+00:00

Commit 2f4f9fe5bf5f ("bnxt_en: Support adding ntuple rules on RSS
contexts") added support for redirecting to an RSS context as an ntuple
rule action. However, it forgot to update the ETHTOOL_GRXCLSRULE
codepath. This caused `ethtool -n` to always report the action as
"Action: Direct to queue 0" which is wrong.

Fix by teaching bnxt driver to report the RSS context when applicable.

Fixes: 2f4f9fe5bf5f ("bnxt_en: Support adding ntuple rules on RSS contexts")
Reviewed-by: Michael Chan 
Signed-off-by: Daniel Xu 
Link: https://patch.msgid.link/2e884ae39e08dc5123be7c170a6089cefe6a78f7.1732748253.git.dxu@dxuuu.xyz
Signed-off-by: Jakub Kicinski

Merge tag 'net-6.13-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net

2024-11-28T18:15:20+00:00

Pull networking fixes from Paolo Abeni:
 "Including fixes from bluetooth.

  Current release - regressions:

   - rtnetlink: fix rtnl_dump_ifinfo() error path

   - bluetooth: remove the redundant sco_conn_put

  Previous releases - regressions:

   - netlink: fix false positive warning in extack during dumps

   - sched: sch_fq: don't follow the fast path if Tx is behind now

   - ipv6: delete temporary address if mngtmpaddr is removed or
     unmanaged

   - tcp: fix use-after-free of nreq in reqsk_timer_handler().

   - bluetooth: fix slab-use-after-free Read in set_powered_sync

   - l2tp: fix warning in l2tp_exit_net found

   - eth:
       - bnxt_en: fix receive ring space parameters when XDP is active
       - lan78xx: fix double free issue with interrupt buffer allocation
       - tg3: set coherent DMA mask bits to 31 for BCM57766 chipsets

  Previous releases - always broken:

   - ipmr: fix tables suspicious RCU usage

   - iucv: MSG_PEEK causes memory leak in iucv_sock_destruct()

   - eth:
       - octeontx2-af: fix low network performance
       - stmmac: dwmac-socfpga: set RX watchdog interrupt as broken
       - rtase: correct the speed for RTL907XD-V1

  Misc:

   - some documentation fixup"

* tag 'net-6.13-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (49 commits)
  ipmr: fix build with clang and DEBUG_NET disabled.
  Documentation: tls_offload: fix typos and grammar
  Fix spelling mistake
  ipmr: fix tables suspicious RCU usage
  ip6mr: fix tables suspicious RCU usage
  ipmr: add debug check for mr table cleanup
  selftests: rds: move test.py to TEST_FILES
  net_sched: sch_fq: don't follow the fast path if Tx is behind now
  tcp: Fix use-after-free of nreq in reqsk_timer_handler().
  net: phy: fix phy_ethtool_set_eee() incorrectly enabling LPI
  net: Comment copy_from_sockptr() explaining its behaviour
  rxrpc: Improve setsockopt() handling of malformed user input
  llc: Improve setsockopt() handling of malformed user input
  Bluetooth: SCO: remove the redundant sco_conn_put
  Bluetooth: MGMT: Fix possible deadlocks
  Bluetooth: MGMT: Fix slab-use-after-free Read in set_powered_sync
  bnxt_en: Unregister PTP during PCI shutdown and suspend
  bnxt_en: Refactor bnxt_ptp_init()
  bnxt_en: Fix receive ring space parameters when XDP is active
  bnxt_en: Fix queue start to update vnic RSS table
  ...

bnxt_en: Unregister PTP during PCI shutdown and suspend

2024-11-26T14:29:31+00:00

If we go through the PCI shutdown or suspend path, we shutdown the
NIC but PTP remains registered.  If the kernel continues to run for
a little bit, the periodic PTP .do_aux_work() function may be called
and it will read the PHC from the BAR register.  Since the device
has already been disabled, it will cause a PCIe completion timeout.
Fix it by calling bnxt_ptp_clear() in the PCI shutdown/suspend
handlers.  bnxt_ptp_clear() will unregister from PTP and
.do_aux_work() will be canceled.

In bnxt_resume(), we need to re-initialize PTP.

Fixes: a521c8a01d26 ("bnxt_en: Move bnxt_ptp_init() from bnxt_open() back to bnxt_init_one()")
Cc: Richard Cochran 
Reviewed-by: Somnath Kotur 
Reviewed-by: Pavan Chebbi 
Reviewed-by: Kalesh AP 
Signed-off-by: Michael Chan 
Signed-off-by: Paolo Abeni