rmmod ath10k_pci hangs on stock wireless-testing kernel

[Ben Greear]

I'm doing a bit of testing on stock kernels for once.  I was using a
10.1 firmware that was constantly crashing, and then I tried to rmmod ath10k_pci.

rmmod hangs.  Here is the stack trace from /proc/[foo]/stack
for rmmod.

Regardless of how the FW crashes, rmmod should not hang.

[root at lf0313-6477 ~]# cat rmmod_hang.txt
[<0>] msleep+0x30/0x40
[<0>] ath10k_pci_hif_stop+0x6a/0xc0 [ath10k_pci]
[<0>] ath10k_core_stop+0x48/0x80 [ath10k_core]
[<0>] ath10k_halt+0x117/0x1b0 [ath10k_core]
[<0>] ath10k_stop+0x31/0x70 [ath10k_core]
[<0>] drv_stop+0x53/0x2c0 [mac80211]
[<0>] ieee80211_do_stop+0x4ce/0x850 [mac80211]
[<0>] ieee80211_stop+0x11/0x20 [mac80211]
[<0>] __dev_close_many+0x97/0xf0
[<0>] dev_close_many+0x78/0x120
[<0>] dev_close.part.91+0x2f/0x50
[<0>] cfg80211_shutdown_all_interfaces+0x3e/0xc0 [cfg80211]
[<0>] ieee80211_remove_interfaces+0x3d/0x1c0 [mac80211]
[<0>] ieee80211_unregister_hw+0x41/0x130 [mac80211]
[<0>] ath10k_mac_unregister+0x10/0x60 [ath10k_core]
[<0>] ath10k_core_unregister+0x35/0x90 [ath10k_core]
[<0>] ath10k_pci_remove+0x29/0x70 [ath10k_pci]
[<0>] pci_device_remove+0x31/0xb0
[<0>] device_release_driver_internal+0x15b/0x230
[<0>] driver_detach+0x35/0x70
[<0>] bus_remove_driver+0x53/0xd0
[<0>] pci_unregister_driver+0x1e/0x80
[<0>] ath10k_pci_exit+0xc/0x214 [ath10k_pci]
[<0>] __x64_sys_delete_module+0x14d/0x250
[<0>] do_syscall_64+0x55/0x1a0
[<0>] entry_SYSCALL_64_after_hwframe+0x49/0xbe
[<0>] 0xffffffffffffffff


(gdb) l *(ath10k_pci_hif_stop+0x6a)
0x332a is in ath10k_pci_hif_stop (/home/greearb/git/linux.wireless-testing/arch/x86/include/asm/bitops.h:328).
323	}
324	
325	static __always_inline bool constant_test_bit(long nr, const volatile unsigned long *addr)
326	{
327		return ((1UL << (nr & (BITS_PER_LONG-1))) &
328			(addr[nr >> _BITOPS_LONG_SHIFT])) != 0;
329	}
330	
331	static __always_inline bool variable_test_bit(long nr, volatile const unsigned long *addr)
332	{
(gdb)


I wouldn't be surprised if this was NAPI related hang similar to what has been reported
before, probably exacerbated by the firmware crashing.

Thanks,
Ben

-- 
Ben Greear <greearb at candelatech.com>
Candela Technologies Inc  http://www.candelatech.com