linux-stable.git/drivers/net/wireless/broadcom, branch v5.4.4

brcmfmac: disable PCIe interrupts before bus reset

2019-12-17T18:55:52+00:00

commit 5d26a6a6150c486f51ea2aaab33af04db02f63b8 upstream.

Keeping interrupts on could result in brcmfmac freeing some resources
and then IRQ handlers trying to use them. That was obviously a straight
path for crashing a kernel.

Example:
CPU0                           CPU1
----                           ----
brcmf_pcie_reset
  brcmf_pcie_bus_console_read
  brcmf_detach
    ...
    brcmf_fweh_detach
    brcmf_proto_detach
                               brcmf_pcie_isr_thread
                                 ...
                                 brcmf_proto_msgbuf_rx_trigger
                                   ...
                                   drvr->proto->pd
    brcmf_pcie_release_irq

[  363.789218] Unable to handle kernel NULL pointer dereference at virtual address 00000038
[  363.797339] pgd = c0004000
[  363.800050] [00000038] *pgd=00000000
[  363.803635] Internal error: Oops: 17 [#1] SMP ARM
(...)
[  364.029209] Backtrace:
[  364.031725] [] (brcmf_proto_msgbuf_rx_trigger [brcmfmac]) from [] (brcmf_pcie_isr_thread+0x228/0x274 [brcmfmac])
[  364.043662]  r7:00000001 r6:c8ca0000 r5:00010000 r4:c7b4f800

Fixes: 4684997d9eea ("brcmfmac: reset PCIe bus on a firmware crash")
Cc: stable@vger.kernel.org # v5.2+
Signed-off-by: Rafał Miłecki 
Signed-off-by: Kalle Valo 
Signed-off-by: Greg Kroah-Hartman

brcmsmac: Use DIV_ROUND_CLOSEST directly to make it readable

2019-09-13T13:45:17+00:00

The kernel.h macro DIV_ROUND_CLOSEST performs the computation (x + d/2)/d
but is perhaps more readable.

Signed-off-by: zhong jiang 
Signed-off-by: Kalle Valo

brcmfmac: don't realloc wiphy during PCIe reset

2019-09-13T13:42:34+00:00

Providing a new wiphy on every PCIe reset was confusing and was causing
configuration problems for some users (supplicant and authenticators).
Sticking to the existing wiphy should make error recovery much simpler
and more reliable.

Signed-off-by: Rafał Miłecki 
Acked-by: Arend van Spriel 
Signed-off-by: Kalle Valo

brcmfmac: split brcmf_attach() and brcmf_detach() functions

2019-09-13T13:42:33+00:00

Move code allocating/freeing wiphy out of above functions. This will
allow reinitializing the driver (e.g. on some error) without allocating
a new wiphy.

Signed-off-by: Rafał Miłecki 
Acked-by: Arend van Spriel 
Signed-off-by: Kalle Valo

brcmfmac: move "cfg80211_ops" pointer to another struct

2019-09-13T13:42:32+00:00

This moves "ops" pointer from "struct brcmf_cfg80211_info" to the
"struct brcmf_pub". This movement makes it possible to allocate wiphy
without attaching cfg80211 (brcmf_cfg80211_attach()). It's required for
later separation of wiphy allocation and driver initialization.

While at it fix also an unlikely memory leak in the brcmf_attach().

Signed-off-by: Rafał Miłecki 
Signed-off-by: Kalle Valo

brcmfmac: use %*ph to print small buffer

2019-09-06T14:14:45+00:00

Use %*ph format to print small buffer as hex string.

Signed-off-by: Andy Shevchenko 
Signed-off-by: Kalle Valo

brcmfmac: add "reset" debugfs entry for testing reset

2019-09-06T14:08:57+00:00

This is a trivial debugfs entry for triggering reset just like in case
of firmware crash. It works by writing 1 to it:
echo 1 > reset

Signed-off-by: Rafał Miłecki 
Signed-off-by: Kalle Valo

brcmfmac: add stub version of brcmf_debugfs_get_devdir()

2019-09-06T14:08:56+00:00

In case of compiling driver without DEBUG expose a stub function to make
writing debug code much simpler (no extra conditions). This will allow
e.g. using debugfs_create_file() without any magic if or #ifdef.

Signed-off-by: Rafał Miłecki 
Signed-off-by: Kalle Valo

brcmfmac: get chip's default RAM info during PCIe setup

2019-09-03T13:44:56+00:00

Getting RAM info just once per driver's lifetime (during chip
recognition) is not enough as it may get adjusted later (depending on
the used firmware). Subsequent inits may load different firmwares so a
full RAM recognition is required on every PCIe setup. This is especially
important since implementing hardware reset on a firmware crash.

Moreover calling brcmf_chip_get_raminfo() makes sure that RAM core is
up. It's important as having BCMA_CORE_SYS_MEM down on BCM4366 was
resulting in firmware failing to initialize and following error:
[   65.657546] brcmfmac 0000:01:00.0: brcmf_pcie_download_fw_nvram: Invalid shared RAM address 0x04000001

This change makes brcmf_chip_get_raminfo() call during chip recognition
redundant for PCIe devices but SDIO and USB still need it and it's a
very small overhead anyway.

Fixes: 4684997d9eea ("brcmfmac: reset PCIe bus on a firmware crash")
Signed-off-by: Rafał Miłecki 
Signed-off-by: Kalle Valo

brcmfmac: replace strncpy() by strscpy()

2019-09-03T13:41:53+00:00

The strncpy() may truncate the copied string,
replace it by the safer strscpy().

To avoid below compile warning with gcc 8.2:

drivers/net/wireless/broadcom/brcm80211/brcmfmac/cfg80211.c:In function 'brcmf_vndr_ie':
drivers/net/wireless/broadcom/brcm80211/brcmfmac/cfg80211.c:4227:2:
warning: 'strncpy' output truncated before terminating nul copying 3 bytes from a string of the same length [-Wstringop-truncation]
  strncpy(iebuf, add_del_cmd, VNDR_IE_CMD_LEN - 1);
  ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Signed-off-by: Xulin Sun 
Acked-by: Arend van Spriel 
Signed-off-by: Kalle Valo