netdev - Re: [net-next] amd-xgbe: add hardware PTP timestamping support

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <d0489a7b-9af9-4c6e-9397-c3a8062f2bfc@amd.com>
Date: Sat, 12 Jul 2025 15:52:08 +0530
From: "Rangoju, Raju" <raju.rangoju@....com>
To: Vadim Fedorenko <vadim.fedorenko@...ux.dev>, netdev@...r.kernel.org
Cc: andrew+netdev@...n.ch, davem@...emloft.net, edumazet@...gle.com,
 kuba@...nel.org, pabeni@...hat.com, Shyam-sundar.S-k@....com
Subject: Re: [net-next] amd-xgbe: add hardware PTP timestamping support



On 7/8/2025 4:24 PM, Vadim Fedorenko wrote:
> On 07/07/2025 18:53, Raju Rangoju wrote:
>> @@ -1636,8 +1659,8 @@ static void xgbe_get_rx_tstamp(struct 
>> xgbe_packet_data *packet,
>>       if (XGMAC_GET_BITS_LE(rdesc->desc3, RX_CONTEXT_DESC3, TSA) &&
>>           !XGMAC_GET_BITS_LE(rdesc->desc3, RX_CONTEXT_DESC3, TSD)) {
>>           nsec = le32_to_cpu(rdesc->desc1);
>> -        nsec <<= 32;
>> -        nsec |= le32_to_cpu(rdesc->desc0);
>> +        nsec *= NSEC_PER_SEC;
>> +        nsec += le32_to_cpu(rdesc->desc0);
> 
> This change is not explained in the commit message. Looks like the
> HW timestamping has already been implemented, but in a wrong way?

Yes, the previous implementation was partial and not functional. This 
patch adds support for full implementation.

>> +    /* Calculate the addend:
>> +     *   addend = 2^32 / (PTP ref clock / (PTP clock based on SSINC))
>> +     *          = (2^32 * (PTP clock based on SSINC)) / PTP ref clock
>> +     */
>> +    if (pdata->vdata->tstamp_ptp_clock_freq)
>> +        dividend = 100000000;           // PTP clock frequency is 125MHz
>> +    else
>> +        dividend = 50000000;            // PTP clock frequency is 50MHz
> why do you use 100_000_000 value for 125MHz while 50_000_000 for 50MHz?
> it doesn't look intuitive
>

That is due to the difference in the ideal clock frequency vs
actual clock freq running on the board. I'll address the comments in v2.