Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Wed, 6 May 2020 11:43:01 +0100
From:   Marc Zyngier <maz@kernel.org>
To:     linux-pci@vger.kernel.org, linux-amlogic@lists.infradead.org,
        linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org,
        Kevin Hilman <khilman@baylibre.com>,
        Yue Wang <yue.wang@Amlogic.com>
Cc:     Bjorn Helgaas <bhelgaas@google.com>, Rob Herring <robh@kernel.org>,
        Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Subject: Re: [PATCH] PCI: amlogic: meson: Don't use FAST_LINK_MODE to set up
 link
Message-ID: <20200506114301.7288c02d@why>
In-Reply-To: <20200429164230.309922-1-maz@kernel.org>
References: <20200429164230.309922-1-maz@kernel.org>
Organization: Approximate
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Wed, 29 Apr 2020 17:42:30 +0100
Marc Zyngier <maz@kernel.org> wrote:

> My vim3l board stubbornly refuses to play ball with a bog
> standard PCIe switch (ASM1184e), spitting all kind of errors
> ranging from link never coming up to crazy things like downstream
> ports falling off the face of the planet.
> 
> Upon investigating how the PCIe RC is configured, I found the
> following nugget: the Sysnopsys DWC PCIe Reference Manual, in the
> section dedicated to the PLCR register, describes bit 7 (FAST_LINK_MODE)
> as:
> 
> "Sets all internal timers to fast mode for simulation purposes."
> 
> I completely understand the need for setting this bit from a simulation
> perspective, but what I have on my desk is actual silicon, which
> expects timers to have a nominal value (and I expect this is the
> case for most people).
> 
> Making sure the FAST_LINK_MODE bit is cleared when configuring the RC
> solves this problem.
> 
> Fixes: 9c0ef6d34fdb ("PCI: amlogic: Add the Amlogic Meson PCIe controller driver")
> Signed-off-by: Marc Zyngier <maz@kernel.org>
> ---
>  drivers/pci/controller/dwc/pci-meson.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/pci/controller/dwc/pci-meson.c b/drivers/pci/controller/dwc/pci-meson.c
> index 3715dceca1bf..ca59ba9e0ecd 100644
> --- a/drivers/pci/controller/dwc/pci-meson.c
> +++ b/drivers/pci/controller/dwc/pci-meson.c
> @@ -289,11 +289,11 @@ static void meson_pcie_init_dw(struct meson_pcie *mp)
>  	meson_cfg_writel(mp, val, PCIE_CFG0);
>  
>  	val = meson_elb_readl(mp, PCIE_PORT_LINK_CTRL_OFF);
> -	val &= ~LINK_CAPABLE_MASK;
> +	val &= ~(LINK_CAPABLE_MASK | FAST_LINK_MODE);
>  	meson_elb_writel(mp, val, PCIE_PORT_LINK_CTRL_OFF);
>  
>  	val = meson_elb_readl(mp, PCIE_PORT_LINK_CTRL_OFF);
> -	val |= LINK_CAPABLE_X1 | FAST_LINK_MODE;
> +	val |= LINK_CAPABLE_X1;
>  	meson_elb_writel(mp, val, PCIE_PORT_LINK_CTRL_OFF);
>  
>  	val = meson_elb_readl(mp, PCIE_GEN2_CTRL_OFF);

Yue, Kevin: any comment on this?

I found that the issue is reproducible even without a PCIe switch,
depending on the single device I plug in this machine (an Intel SSD
works fine, while a Marvell Ethernet adapter never shows up) as the
LTSSM times out much earlier than it really should (HW timers running
too quickly). Applying this patch makes every single device I have
lying around work fine.

Thanks,

	M.
-- 
Jazz is not dead. It just smells funny...