Re: [PATCH iwl-next v8 11/11] igc: add support to get frame preemption statistics via ethtool

["Abdul Rahim, Faizal" <faizal.abdul.rahim@linux.intel.com>]

On 6/3/2025 8:48 am, Vladimir Oltean wrote:
> On Wed, Mar 05, 2025 at 08:00:26AM -0500, Faizal Rahim wrote:
>> +/* Received out of order packets with SMD-C */
>> +#define IGC_PRMEXCPRCNT_OOO_SMDC			0x000000FF
>> +/* Received out of order packets with SMD-C and wrong Frame CNT */
>> +#define IGC_PRMEXCPRCNT_OOO_FRAME_CNT			0x0000FF00
>> +/* Received out of order packets with SMD-C and wrong Frag CNT */
>> +#define IGC_PRMEXCPRCNT_OOO_FRAG_CNT			0x00FF0000
>> +/* Received packets with SMD-S and wrong Frag CNT and Frame CNT */
>> +#define IGC_PRMEXCPRCNT_MISS_FRAME_FRAG_CNT		0xFF000000
>>   
>> +/**
>> + * igc_ethtool_get_frame_ass_error - Get the frame assembly error count.
>> + * @reg_value: Register value for IGC_PRMEXCPRCNT
>> + * Return: The count of frame assembly errors.
>> + */
>> +static u64 igc_ethtool_get_frame_ass_error(u32 reg_value)
>> +{
>> +	u32 ooo_frame_cnt, ooo_frag_cnt; /* Out of order statistics */
>> +	u32 miss_frame_frag_cnt;
>> +
>> +	ooo_frame_cnt = FIELD_GET(IGC_PRMEXCPRCNT_OOO_FRAME_CNT, reg_value);
>> +	ooo_frag_cnt = FIELD_GET(IGC_PRMEXCPRCNT_OOO_FRAG_CNT, reg_value);
>> +	miss_frame_frag_cnt = FIELD_GET(IGC_PRMEXCPRCNT_MISS_FRAME_FRAG_CNT, reg_value);
>> +
>> +	return ooo_frame_cnt + ooo_frag_cnt + miss_frame_frag_cnt;
>> +}
> 
> These counters are quite small (8 bits each). What is their behavior
> once they reach 255? Saturate? Truncate? Do they clear on read?
> 
Hi Vladimir,

These are part of the statistic registers, which in IGC, reset upon read. 
When they reach their maximum value, each field remain at 0xFF.