[PATCH v2] wifi: ath12k: Support DMAC Reset Stats

[Jeff Johnson]

On 10/5/2024 3:52 AM, Roopni Devanathan wrote:
> From: Rajat Soni <quic_rajson at quicinc.com>
> 
> Add support to request DMAC reset stats from firmware through HTT stats
> type 45. These stats give debug SoC error stats such as reset count, reset
> time, engage time and count, disengage time and count and destination
> ring mask.
> 
> Sample output:
> -------------
> echo 45 > /sys/kernel/debug/ath12k/pci-0000\:06\:00.0/mac0/htt_stats_type
> cat /sys/kernel/debug/ath12k/pci-0000\:06\:00.0/mac0/htt_stats
> HTT_DMAC_RESET_STATS_TLV:
> reset_count = 1
> reset_time_ms = 8036717648
> disengage_time_ms = 8036717648
> engage_time_ms = 8036717649
> disengage_count = 1
> engage_count = 1
> drain_dest_ring_mask = 0x0
> 
> Tested-on: QCN9274 hw2.0 PCI WLAN.WBE.1.1.1-00214-QCAHKSWPL_SILICONZ-1
> Tested-on: WCN7850 hw2.0 PCI WLAN.HMT.1.0-03427-QCAHMTSWPL_V1.0_V2.0_SILICONZ-1.15378.4
> 
> Signed-off-by: Rajat Soni <quic_rajson at quicinc.com>
> Signed-off-by: Roopni Devanathan <quic_rdevanat at quicinc.com>
> ---
> v2:
>  - Updated dependencies. No change in code.
> ---
> Depends-on:
> [PATCH v2] wifi: ath12k: Modify print_array_to_buf() to support arrays with 1-based semantics
> Link: https://lore.kernel.org/all/20241004085915.1788951-1-quic_rdevanat@quicinc.com/
> 
> [PATCH v2 0/4] wifi: ath12k: Support Ring, SFM, Transmit MU, SelfGen stats, CCA stats
> Link: https://lore.kernel.org/ath12k/20241005101816.3314728-1-quic_rdevanat@quicinc.com/
> 
> [PATCH v2] wifi: ath12k: Support Pdev OBSS Stats
> Link: https://lore.kernel.org/ath12k/20241005104206.3327143-1-quic_rdevanat@quicinc.com/
> ---
> ---
>  .../wireless/ath/ath12k/debugfs_htt_stats.c   | 41 +++++++++++++++++++
>  .../wireless/ath/ath12k/debugfs_htt_stats.h   | 15 +++++++
>  2 files changed, 56 insertions(+)
> 
> diff --git a/drivers/net/wireless/ath/ath12k/debugfs_htt_stats.c b/drivers/net/wireless/ath/ath12k/debugfs_htt_stats.c
> index d890679a3f16..2debb253185c 100644
> --- a/drivers/net/wireless/ath/ath12k/debugfs_htt_stats.c
> +++ b/drivers/net/wireless/ath/ath12k/debugfs_htt_stats.c
> @@ -2502,6 +2502,44 @@ ath12k_htt_print_pdev_obss_pd_stats_tlv(const void *tag_buf, u16 tag_len,
>  	stats_req->buf_len = len;
>  }
>  
> +static void
> +ath12k_htt_print_dmac_reset_stats_tlv(const void *tag_buf, u16 tag_len,
> +				      struct debug_htt_stats_req *stats_req)
> +{
> +	const struct ath12k_htt_dmac_reset_stats_tlv *htt_stats_buf = tag_buf;
> +	u8 *buf = stats_req->buf;
> +	u32 len = stats_req->buf_len;
> +	u32 buf_len = ATH12K_HTT_STATS_BUF_SIZE;
> +	unsigned long time;
> +
> +	if (tag_len < sizeof(*htt_stats_buf))
> +		return;
> +
> +	len += scnprintf(buf + len, buf_len - len, "HTT_DMAC_RESET_STATS_TLV:\n");
> +	len += scnprintf(buf + len, buf_len - len, "reset_count = %u\n",
> +			 le32_to_cpu(htt_stats_buf->reset_count));
> +	time = ((unsigned long)le32_to_cpu(htt_stats_buf->reset_time_hi_ms) << 32) |
> +		le32_to_cpu(htt_stats_buf->reset_time_lo_ms);

The v1 version was flagged by the Intel 0-DAY CI Kernel Test Service:
https://lore.kernel.org/all/202410040200.Mwb85JzQ-lkp@intel.com/

I believe you need to use u64 instead of unsigned long since unsigned long
will only be 32 bits on a 32-bit system. Also suggest you refactor into a
helper function to remove any ambiguity about the order of operations between
(typecast) and <<, i.e.:

/* untested */
static u64 ath12k_le32hilo_to_u64(__le32 hi, __le32 lo)
{
	u64 hi64 = le32_to_cpu(hi);
	u64 lo64 = le32_to_cpu(lo);

	return (hi64 << 32) | lo64;
}