Merge tag 'wireless-drivers-next-for-davem-2017-12-18' of git://git.kernel.org/pub/scm/linux/kernel/git/kvalo/wireless-drivers-next

The drivers/net/wireless/intel/iwlwifi/pcie/drv.c conflict was
resolved using a diff provided by Kalle in his pull request.

Kalle Valo says:

====================
wireless-drivers-next patches for 4.16

A bigger pull request this time, the most visible change being the new
driver mt76. But there's also Kconfig refactoring in ath9k and ath10k,
work beginning in iwlwifi to have rate scaling in firmware/hardware,
wcn3990 support getting closer in ath10k and lots of smaller changes.

mt76

* a new driver for MT76x2e, a 2x2 PCIe 802.11ac chipset by MediaTek

ath10k

* enable multiqueue support for all hw using mac80211 wake_tx_queue op

* new Kconfig option ATH10K_SPECTRAL to save RAM

* show tx stats on QCA9880

* new qcom,ath10k-calibration-variant DT entry

* WMI layer support for wcn3990

ath9k

* new Kconfig option ATH9K_COMMON_SPECTRAL to save RAM

wcn36xx

* hardware scan offload support

wil6210

* run-time PM support when interface is down

iwlwifi

* initial work for rate-scaling offload

* Support for new FW API version 36

* Rename the temporary hw name A000 to 22000

ssb

* make SSB a menuconfig to ease disabling it all

mwl8k

* enable non-DFS 5G channels 149-165
====================

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

1 files changed, 14 insertions, 1 deletions

diff --git a/drivers/net/wireless/intel/iwlwifi/fw/dbg.c b/drivers/net/wireless/intel/iwlwifi/fw/dbg.c
index 8106fd4be996..67aefc8fc9ac 100644
--- a/drivers/net/wireless/intel/iwlwifi/fw/dbg.c
+++ b/drivers/net/wireless/intel/iwlwifi/fw/dbg.c
@@ -964,7 +964,20 @@ int iwl_fw_dbg_collect_desc(struct iwl_fw_runtime *fwrt,
 	if (trigger)
 		delay = msecs_to_jiffies(le32_to_cpu(trigger->stop_delay));
 
-	if (WARN(fwrt->trans->state == IWL_TRANS_NO_FW,
+	/*
+	 * If the loading of the FW completed successfully, the next step is to
+	 * get the SMEM config data. Thus, if fwrt->smem_cfg.num_lmacs is non
+	 * zero, the FW was already loaded successully. If the state is "NO_FW"
+	 * in such a case - WARN and exit, since FW may be dead. Otherwise, we
+	 * can try to collect the data, since FW might just not be fully
+	 * loaded (no "ALIVE" yet), and the debug data is accessible.
+	 *
+	 * Corner case: got the FW alive but crashed before getting the SMEM
+	 *	config. In such a case, due to HW access problems, we might
+	 *	collect garbage.
+	 */
+	if (WARN((fwrt->trans->state == IWL_TRANS_NO_FW) &&
+		 fwrt->smem_cfg.num_lmacs,
 		 "Can't collect dbg data when FW isn't alive\n"))
 		return -EIO;