Re: [Bug 215925] New: PCIe regression on Raspberry Pi Compute Module 4 (CM4) breaks booting

From: Bjorn Helgaas <helgaas@kernel.org>
To: Cyril Brulebois <kibi@debian.org>
Cc: Thorsten Leemhuis <regressions@leemhuis.info>,
	Jim Quinlan <jim2101024@gmail.com>,
	"regressions@lists.linux.dev" <regressions@lists.linux.dev>,
	Linux PCI <linux-pci@vger.kernel.org>,
	bjorn@helgaas.com
Subject: Re: [Bug 215925] New: PCIe regression on Raspberry Pi Compute Module 4 (CM4) breaks booting
Date: Tue, 10 May 2022 12:22:43 -0500	[thread overview]
Message-ID: <20220510172243.GA684299@bhelgaas> (raw)
In-Reply-To: <20220509174527.zoqhmaxfwo7udezo@mraw.org>

On Mon, May 09, 2022 at 07:45:27PM +0200, Cyril Brulebois wrote:
> Hi Bjorn,
> 
> Bjorn Helgaas <helgaas@kernel.org> (2022-05-09):
> > Cyril, 830aa6f29f07 ("PCI: brcmstb: Split brcm_pcie_setup() into two
> > funcs") reverts cleanly as of 57ae8a492116.  Does reverting it avoid
> > the regression?
> 
> I didn't even try and revert this commit before you suggested doing so
> since it was supposed to be some preliminary work. Quoting a part of
> its commit message:
> 
>     In future commits the brcm_pcie_linkup() function will be called
>     indirectly by pci_host_probe() as opposed to the host driver
>     invoking it directly.
> 
> 
> Anyway, the patch can indeed be reverted on top of v5.18-rc4 or
> v5.18-rc6 but the build fails due to the former function being removed,
> while being still called from other places:
> 
>       CC      drivers/pci/controller/pcie-brcmstb.o
>     drivers/pci/controller/pcie-brcmstb.c:199:12: warning: ‘brcm_pcie_linkup’ used but never defined
>       199 | static int brcm_pcie_linkup(struct brcm_pcie *pcie);
>           |            ^~~~~~~~~~~~~~~~
>     …
>     aarch64-linux-gnu-ld: drivers/pci/controller/pcie-brcmstb.o: in function `brcm_pcie_add_bus':
>     /home/kibi/hack/linux.git/drivers/pci/controller/pcie-brcmstb.c:527: undefined reference to `brcm_pcie_linkup'
> 
> 
> See for example:
> 
>     commit 93e41f3fca3d4a0f927b784012338c37f80a8a80
>     Author: Jim Quinlan <jim2101024@gmail.com>
>     Date:   Thu Jan 6 11:03:29 2022 -0500
>     
>         PCI: brcmstb: Add control of subdevice voltage regulators
> 
> (And that one cannot be trivially reverted.)

What if you revert 830aa6f29f07 and the subsequent brcmstb patches?

  11ed8b8624b8 ("PCI: brcmstb: Do not turn off WOL regulators on suspend")
  93e41f3fca3d ("PCI: brcmstb: Add control of subdevice voltage regulators")
  67211aadcb4b ("PCI: brcmstb: Add mechanism to turn on subdev regulators")
  830aa6f29f07 ("PCI: brcmstb: Split brcm_pcie_setup() into two funcs")

  $ git revert 11ed8b8624b8 93e41f3fca3d 67211aadcb4b 830aa6f29f07

I did that on current upstream: 9be9ed2612b5 ("Merge tag
'platform-drivers-x86-v5.18-4' of
git://git.kernel.org/pub/scm/linux/kernel/git/pdx86/platform-drivers-x86")
and it built fine on x86.