[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20200728125359.1e9f9b92@kicinski-fedora-pc1c0hjn.dhcp.thefacebook.com>
Date: Tue, 28 Jul 2020 12:53:59 -0700
From: Jakub Kicinski <kuba@...nel.org>
To: Ioana Ciornei <ioana.ciornei@....com>
Cc: "davem@...emloft.net" <davem@...emloft.net>,
"netdev@...r.kernel.org" <netdev@...r.kernel.org>
Subject: Re: [PATCH net-next 0/2] dpaa2-eth: add reset control for debugfs
stats
On Tue, 28 Jul 2020 19:33:46 +0000 Ioana Ciornei wrote:
> > Subject: Re: [PATCH net-next 0/2] dpaa2-eth: add reset control for debugfs stats
> > No, come on, you know what we're going to say to a debugfs patch like this...
>
> Eh, I figured it was worth a try since I saw that i40e also supports clearing
> the stats through debugfs.
The stuff that got snuck into i40e is the reason I pay very close
attention to Intel patches these days.
> > Is there anything dpaa2-specific here? We should be able to add a common API
> > for this.
>
> No, there is nothing dpaa2-specific. The common API would be in the
> 'ethtool --reset' area or do you have anything other in mind?
We have this huge ongoing discussion about devlink reset/reload:
https://lore.kernel.org/netdev/1590908625-10952-1-git-send-email-vasundhara-v.volam@broadcom.com/
https://lore.kernel.org/netdev/1595847753-2234-1-git-send-email-moshe@mellanox.com/
Perhaps it could fit in there? I presume your reset is on the device
level? If it's per netdev, I'd personally not be opposed to extending
the ethtool API.
Powered by blists - more mailing lists