[V3] PCI: fu740: Drop to 2.5GT/s to fix initial device probing on some boards

[Palmer Dabbelt]

On Fri, 18 Mar 2022 08:24:30 PDT (-0700), ben.dooks at codethink.co.uk wrote:
> The fu740 PCIe core does not probe any devices on the SiFive Unmatched
> board without this fix (or having U-Boot explicitly start the PCIe via
> either boot-script or user command). The fix is to start the link at
> 2.5GT/s speeds and once the link is up then change the maximum speed back
> to the default.
>
> The U-Boot driver claims to set the link-speed to 2.5GT/s to get the probe
> to work (and U-Boot does print link up at 2.5GT/s) in the following code:
> https://source.denx.de/u-boot/u-boot/-/blob/master/drivers/pci/pcie_dw_sifive.c?id=v2022.01#L271
>
> Signed-off-by: Ben Dooks <ben.dooks at codethink.co.uk>
> --

A "--" triggers some mail handles to think the rest of this is a 
signature, git folks usually use a "---" to indicate a comment that 
shouldn't be part of what's eventually merged (like this changelog 
stuff).

> Note, this patch has had significant re-work since the previous 4
> sets, including trying to fix style, message, reliance on the U-Boot
> fix and the comments about usage of LINK_CAP and reserved fields.
> 
> v2:
> - fix issues with Gen1/2.5GTs
> - updated comment on the initial probe
> - run tests with both uninitialised and initialsed pcie from uboot
> ---
>  drivers/pci/controller/dwc/pcie-fu740.c | 52 ++++++++++++++++++++++++-
>  1 file changed, 51 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/pci/controller/dwc/pcie-fu740.c b/drivers/pci/controller/dwc/pcie-fu740.c
> index 842b7202b96e..ecac0364178a 100644
> --- a/drivers/pci/controller/dwc/pcie-fu740.c
> +++ b/drivers/pci/controller/dwc/pcie-fu740.c
> @@ -181,10 +181,60 @@ static int fu740_pcie_start_link(struct dw_pcie *pci)

Is there an errata?  IIUC this will trigger the workaround on all 
FU740s, but from the description it seems like more of a board bug than 
a chip bug.  The distinction doesn't really matter, as there's only one 
board for this chip (and I'm assuming that'll always be the case), but 
if there's an errata (or any way this is documented) it might make 
things a bit easier to sort out if we end up with another similar 
chip/board.

Either way

Acked-by: Palmer Dabbelt <palmer at rivosinc.com>

I'm assuming you, or someone else, has tested this on the board?  I'm 
pretty sure I've got one lying around somewhere, but I don't regularly 
use it.  I can dust it off if nobody else has tried this, but happy to 
avoid the need to do so.

Thanks!

>  {
>  	struct device *dev = pci->dev;
>  	struct fu740_pcie *afp = dev_get_drvdata(dev);
> +	u8 cap_exp = dw_pcie_find_capability(pci, PCI_CAP_ID_EXP);
> +	int ret;
> +	u32 orig, tmp;
> +
> +	/*
> +	 * Force 2.5GT/s when starting the link, due to some devices not
> +	 * probing at higher speeds. This happens with the PCIe switch
> +	 * on the Unmatched board when U-Boot has not initialised the PCIe.
> +	 * The fix in U-Boot is to force 2.5GT/s, which then gets cleared
> +	 * by the soft reset does by this driver.
> +	 */
> +
> +	dev_dbg(dev, "cap_exp at %x\n", cap_exp);
> +	dw_pcie_dbi_ro_wr_en(pci);
> +
> +	tmp = dw_pcie_readl_dbi(pci, cap_exp + PCI_EXP_LNKCAP);
> +	orig = tmp & PCI_EXP_LNKCAP_SLS;
> +	tmp &= ~PCI_EXP_LNKCAP_SLS;
> +	tmp |= PCI_EXP_LNKCAP_SLS_2_5GB;
> +	dw_pcie_writel_dbi(pci, cap_exp + PCI_EXP_LNKCAP, tmp);
> 
>  	/* Enable LTSSM */
>  	writel_relaxed(0x1, afp->mgmt_base + PCIEX8MGMT_APP_LTSSM_ENABLE);
> -	return 0;
> +
> +	ret = dw_pcie_wait_for_link(pci);
> +	if (ret) {
> +		dev_err(dev, "error: link did not start\n");
> +		goto err;
> +	}
> +
> +	tmp = dw_pcie_readl_dbi(pci, cap_exp + PCI_EXP_LNKCAP);
> +	if ((tmp & PCI_EXP_LNKCAP_SLS) != orig) {
> +		dev_dbg(dev, "changing speed back to original\n");
> +
> +		tmp &= ~PCI_EXP_LNKCAP_SLS;
> +		tmp |= orig;
> +		dw_pcie_writel_dbi(pci, cap_exp + PCI_EXP_LNKCAP, tmp);
> +
> +		tmp = dw_pcie_readl_dbi(pci, PCIE_LINK_WIDTH_SPEED_CONTROL);
> +		tmp |= PORT_LOGIC_SPEED_CHANGE;
> +		dw_pcie_writel_dbi(pci, PCIE_LINK_WIDTH_SPEED_CONTROL, tmp);
> +
> +		ret = dw_pcie_wait_for_link(pci);
> +		if (ret) {
> +			dev_err(dev, "error: link did not start at new speed\n");
> +			goto err;
> +		}
> +	}
> +
> +	ret = 0;
> +err:
> +	WARN_ON(ret);	/* we assume that errors will be very rare */

Maybe a message is better than a WARN_ON there?  Something users might 
be able to understand.

> +	dw_pcie_dbi_ro_wr_dis(pci);
> +	return ret;
>  }
> 
>  static int fu740_pcie_host_init(struct pcie_port *pp)
> --
> 2.35.1
> 
>