Hi Tom, On Mon, 26 May 2025 at 15:42, Tom Rini <tr...@konsulko.com> wrote: > > On Sun, May 25, 2025 at 09:43:14AM +0100, Simon Glass wrote: > > > Hi Tom, > > > > Bloblist is IMO a bit of a mess, with 2-3 maintainers who don't agree > > on the direction. > > > > You have requested that I: > > - implement a handoff protocol for x86 in the spec > > - do the same in the code > > - convert x86 boards to use OF_BLOBLIST > > - remove BLOBLIST_FIXED > > > > I also want to enable VBE, which would require further work. I'm > > sending a v4 today. > > > > I am willing to do the above, but to do so, I want to make bloblist a > > first-class citizen, available on any arch, built into global_data, > > with no reliance on weak functions. > > > > Please can you allow me to maintain bloblist? > > First, OK, yes, fine.
Thank you, I appreciate your flexibility on this. I'll prepare a v5 for that series. > > Second, for the record, I am compromising and "Saying Yes" to what you > want to do, over what I think is a bad idea. We have had a mechanism to > allow for saving off registers early so we can make use of them a little > bit later as a "first class citizen" for 14 years come this July. > Bloblist is not unique or new. What you want to do instead will come > back to be a problem for some other valid use case later on and require > some further architecture work at that time (or be an "ugly workaround" > that we have to accept if we want that case to work). I'm not sure if it would be helpful for me to respond to this now, since I have already stated how I see things differently. But perhaps we can take up this discussion again once my series is applied and we can see the result? Regards, SImon
