Re: [PATCH 1/1] wifi: nl80211: Add support for plumbing SAE groups to driver

[Johannes Berg <johannes@sipsolutions.net>]

On Mon, 2024-02-12 at 09:25 +0200, Kalle Valo wrote:

> What driver is going to use these new crypto settings? Or is this for an
> out-of-tree driver?
> 

I'm sure it's for an out-of-tree driver.

This is the _entirety_ of "@broadcom"'s wireless contributions with
"--since=2020" (somewhat arbitrarily chosen, though going a bit further
back has some "real" work in brcmfmac), as far as I can tell:

Arend Van Spriel (1):
      cfg80211: adapt to new channelization of the 6GHz band

Arend van Spriel (23):
      cfg80211: add VHT rate entries for MCS-10 and MCS-11
      brcmfmac: use different error value for invalid ram base address
      brcmfmac: increase core revision column aligning core list
      brcmfmac: add xtlv support to firmware interface layer
      brcmfmac: support chipsets with different core enumeration space
      wifi: cfg80211: fix memory leak in query_regdb_file()
      wifi: brcmfmac: add function to unbind device to bus layer api
      wifi: brcmfmac: add firmware vendor info in driver info
      wifi: brcmfmac: add support for vendor-specific firmware api
      wifi: brcmfmac: add support for Cypress firmware api
      wifi: brcmfmac: add support Broadcom BCA firmware api
      wifi: brcmfmac: add vendor name in revinfo debugfs file
      wifi: brcmfmac: introduce BRCMFMAC exported symbols namespace
      wifi: brcmfmac: avoid handling disabled channels for survey dump
      wifi: brcmfmac: avoid NULL-deref in survey dump for 2G only device
      wifi: brcmfmac: fix regression for Broadcom PCIe wifi devices
      wifi: brcmfmac: change cfg80211_set_channel() name and signature
      wifi: brcmfmac: export firmware interface functions
      wifi: brcmfmac: add per-vendor feature detection callback
      wifi: brcmfmac: move feature overrides before feature_disable
      wifi: brcmfmac: avoid invalid list operation when vendor attach fails
      wifi: brcmfmac: allow per-vendor event handling
      wifi: brcmfmac: add linefeed at end of file

Vinayak Yadawad (4):
      wifi: cfg80211: Allow P2P client interface to indicate port authorization
      cfg80211: Update Transition Disable policy during port authorization
      wifi: cfg80211: Allow AP/P2PGO to indicate port authorization to peer STA/P2PClient
      wifi: nl80211: Extend del pmksa support for SAE and OWE security


So looks to me like Broadcom doesn't want its (real) drivers to work in
upstream, so I guess we really ought to just stop accommodating for them
in the wireless stack... This only works if we collaborate, and I've
said this before: I can't maintain something well that I cannot see (and
possibly change) the user(s) of.

I guess if Broadcom's plans change they can start by submitting drivers
that actually use the relevant infrastructure.

And note that I've said this to Qualcomm before: I don't really want to
and can't (well) maintain a lot of stuff in the tree that exists there
solely to make out-of-tree drivers happy.

And @Broadcom: we really _want_ you to contribute upstream. But that
shouldn't be dumping APIs over the wall when you need them and letting
us sort out everything else ...

johannes