Simon Wunderlich [off-list ref] writes:

Hey Kalle,

quoted

Simon Wunderlich [off-list ref] writes:

quoted

quoted

quoted

+struct fft_sample_ath10k {
+	struct fft_sample_tlv tlv;
+	u8 chan_width_mhz;
+	__be16 freq1;
+	__be16 freq2;
+	__be16 noise;
+	__be16 max_magnitude;
+	__be16 total_gain_db;
+	__be16 base_pwr_db;
+	__be64 tsf;
+	s8 max_index;
+	u8 rssi;
+	u8 relpwr_db;
+	u8 avgpwr_db;
+	u8 max_exp;
+
+	u8 data[0];
+} __packed;

__be16, that's a first. Just making sure that this really is big endian?

As the __le32 you use in other ath10k structs, this is just for checking
- at least sparse should check that, maybe other tools as well.

Sorry, I didn't understand your comment here. But basically I was asking
is the fft sample really in big endian? I assume it would be little
endian as everything else coming from the firmware.

Yeah that is intended - we are also using big endian in ath9k to send fft 
samples to userspace, because that is the network byte order and we then just 
use ntohs() and friends in userspace to read samples from any system. 
Therefore we intent to use the same encoding in ath10k. :)

Ah, this is the struct going to user space? I thought it's coming from
the firmware. Then forget my question :)

-- 
Kalle Valo