Re: [PATCH 0/2] MIPS: Malta: Fix PCI southbridge legacy resource clashes

[Bjorn Helgaas <helgaas@kernel.org>]

[+cc linux-pci]

On Tue, Oct 21, 2025 at 08:38:15PM +0100, Maciej W. Rozycki wrote:
> Hi,
> 
>  This mini patch series sorts out issues with southbridge legacy resource 
> management on the MIPS Malta platform.  Two changes turned out required, 
> because merely removing the clash would regress the PS/2 interfaces, fixed 
> by accident with the PCIBIOS_MIN_IO fix.
> 
>  This does prove nobody has used these interfaces since forever, or most 
> likely since the move to the new serio driver.  Things most likely worked 
> fine with 2.4 and I still have such old Malta kernel builds lying around 
> (though sadly no hardware to try with), although I do know I've never used 
> the PS/2 stuff with this platform, e.g. quoting an arbitrarily picked 
> 2.4.19-rc1 bootstrap log:
> 
> CPU revision is: 00018101
> Primary instruction cache 16kb, linesize 32 bytes(4 ways)
> Primary data cache 16kb, linesize 32 bytes (4 ways)
> Linux version 2.4.19-rc1 (macro@macro.ds2.pg.gda.pl) (gcc version 2.95.4 20010319 (prerelease)) #1 Fri Aug 23 02:55:02 CEST 2002
> [...]
> parport0: PC-style at 0x378 [PCSPP,EPP]
> initialize_kbd: Keyboard reset failed, no ACK
> Detected PS/2 Mouse Port.
> pty: 256 Unix98 ptys configured
> keyboard: Timeout - AT keyboard not present?(ed)
> keyboard: Timeout - AT keyboard not present?(f4)
> Serial driver version 5.05c (2001-07-08) with MANY_PORTS SHARE_IRQ DETECT_IRQ SERIAL_PCI enabled
> [...]
> 
> However to prevent the PS/2 interfaces from getting fixed and then broken 
> again with backports in a random fashion I have marked both changes for 
> backporting as appropriate.
> 
>  Bjorn, may I request that these changes be placed, with Thomas's ack of 
> course (hopefully a formality), ahead of Ilpo's commit 16fbaba2b78f 
> ("MIPS: Malta: Use pcibios_align_resource() to block io range") (or 
> whatever the latest version is, as said commit seems to be missing tags 
> updates you mentioned), and then merged via your tree?  That will prevent 
> things from breaking just to be fixed again shortly, and overall getting 
> out of sync.
> 
>  See individual commit descriptions for details.
> 
>   Maciej

Applied with Thomas's acks to pci/for-linus for v6.18, thanks!