Re: [PATCH RFC net-next 10/10] net: dsa: allow use of phylink managed EEE support
From: Vladimir Oltean <olteanv@gmail.com>
Date: 2025-01-14 20:33:37
Also in:
linux-arm-kernel
On Tue, Jan 14, 2025 at 08:02:55PM +0000, Russell King (Oracle) wrote:
quoted
What is the reason for including this patch with this set, where it is of no use until at least one DSA driver provides the new API implementations?No criticism against you, but I guess you didn't read the cover message? I tend not to read cover messages. I've been wondering for a while now whether anyone actually does and thus whether they are worth the effort of writing anything beyond providing a message ID to tie a series together and a diffstat for the series. Here's the relevant bit: "The remainder of the patches convert mvneta, lan743x and stmmac, add support for mvneta, and add the basics that will be necessary into the DSA code for DSA drivers to make use of this. "I would like to get patches 1 through 9 into net-next before the merge window, but we're running out of time for that." So, it's included in this RFC series not because I'm intending it to be merged, but so that people can see what DSA requires to make it functional there, and provide review comments if they see fit - which you have done, thanks. I'm sure if I'd said "I have patches for DSA" you'd have responded asking to see them! Of course, I do have changes that will require this - mt753x - but that patch is not quite ready because this series I've posted has seen a few changes recently to remove stuff that was never settled (the LPI timer questions, whether it should be validated, clamped, should phylink provide a software timer if the LPI timer is out of range, etc.) That's more proof that text in cover messages is utterly useless!
Ok, for your information, I had read the cover letter in its entirety, including the paragraph which you are highlighting here again. I still find it of limited usefulness for me to see and not be able to leave meaningful feedback on a DSA conversion broken down into sets which offer an incomplete image, especially when the API itself is still subject to change. I wanted to leave an Acked-by but I didn't fully understand where the conversion is going, so I asked the question instead. I wonder what you would have done in my position. I wish you a beautiful evening.