netdev - Re: [PATCH net-next 0/7][pull request] 1GbE Intel Wired LAN Driver Updates 2022-07-28

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 PHC
Open Source and information security mailing list archives

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

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

Date:   Mon, 1 Aug 2022 09:56:27 -0700
From:   Tony Nguyen <anthony.l.nguyen@...el.com>
To:     <davem@...emloft.net>, <kuba@...nel.org>, <pabeni@...hat.com>,
        <edumazet@...gle.com>
CC:     <netdev@...r.kernel.org>, <richardcochran@...il.com>,
        <jacob.e.keller@...el.com>
Subject: Re: [PATCH net-next 0/7][pull request] 1GbE Intel Wired LAN Driver
 Updates 2022-07-28

On 7/28/2022 11:18 AM, Tony Nguyen wrote:
> Jacob Keller says:
> 
> Convert all of the Intel drivers with PTP support to the newer .adjfine
> implementation which uses scaled parts per million.
> 
> This improves the precision of the frequency adjustments by taking advantage
> of the full scaled parts per million input coming from user space.
> 
> In addition, all implementations are converted to using the
> mul_u64_u64_div_u64 function which better handles the intermediate value.
> This function supports architecture specific instructions where possible to
> avoid loss of precision if the normal 64-bit multiplication would overflow.
> 
> Of note, the i40e implementation is now able to avoid loss of precision on
> slower link speeds by taking advantage of this to multiply by the link speed
> factor first. This results in a significantly more precise adjustment by
> allowing the calculation to impact the lower bits.
> 
> This also gets us a step closer to being able to remove the .adjfreq
> entirely by removing its use from many drivers.
> 
> I plan to follow this up with a series to update the drivers from other
> vendors and drop the .adjfreq implementation entirely.
> 
> The following are changes since commit 623cd87006983935de6c2ad8e2d50e68f1b7d6e7:
>    net: cdns,macb: use correct xlnx prefix for Xilinx
> and are available in the git repository at:
>    git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/next-queue 1GbE

Hi Dave, Jakub, Paolo, Eric,

I'm seeing this as "Accepted" on Patchworks [1], but I haven't seen a 
patchwork bot notification or am I seeing it on the tree. I think it may 
have accidentally been marked so just wanted to get it back on the radar.

Thanks,
Tony

> Jacob Keller (7):
>    ice: implement adjfine with mul_u64_u64_div_u64
>    e1000e: remove unnecessary range check in e1000e_phc_adjfreq
>    e1000e: convert .adjfreq to .adjfine
>    i40e: use mul_u64_u64_div_u64 for PTP frequency calculation
>    i40e: convert .adjfreq to .adjfine
>    ixgbe: convert .adjfreq to .adjfine
>    igb: convert .adjfreq to .adjfine
> 
>   drivers/net/ethernet/intel/e1000e/e1000.h    |  2 +-
>   drivers/net/ethernet/intel/e1000e/netdev.c   |  4 +-
>   drivers/net/ethernet/intel/e1000e/ptp.c      | 18 +++--
>   drivers/net/ethernet/intel/i40e/i40e_ptp.c   | 35 ++++------
>   drivers/net/ethernet/intel/ice/ice_ptp.c     | 16 +----
>   drivers/net/ethernet/intel/igb/igb_ptp.c     | 15 ++--
>   drivers/net/ethernet/intel/ixgbe/ixgbe_ptp.c | 73 +++++++++++---------
>   7 files changed, 75 insertions(+), 88 deletions(-)
> 

[1] 
https://patchwork.kernel.org/project/netdevbpf/list/?series=663859&state=*