linux-pci.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Bjorn Helgaas <helgaas@kernel.org>
To: Lukas Wunner <lukas@wunner.de>
Cc: Jim Quinlan <jim2101024@gmail.com>,
	Nicolas Saenz Julienne <nsaenz@kernel.org>,
	Florian Fainelli <florian.fainelli@broadcom.com>,
	Lorenzo Pieralisi <lpieralisi@kernel.org>,
	Krzysztof Wilczynski <kw@linux.com>,
	Rob Herring <robh@kernel.org>,
	Bjorn Helgaas <bhelgaas@google.com>,
	bcm-kernel-feedback-list@broadcom.com, linux-pci@vger.kernel.org,
	Philipp Rosenberger <p.rosenberger@kunbus.com>,
	Cyril Brulebois <kibi@debian.org>
Subject: Re: [PATCH] PCI: brcmstb: Avoid downstream access during link training
Date: Fri, 8 Dec 2023 13:53:59 -0600	[thread overview]
Message-ID: <20231208195359.GA808176@bhelgaas> (raw)
In-Reply-To: <385baf9dbfb6b65112803998dfc0cd6f84a5e6ba.1691296765.git.lukas@wunner.de>

On Sun, Aug 06, 2023 at 06:44:50AM +0200, Lukas Wunner wrote:
> The Broadcom Set Top Box PCIe controller signals an Asynchronous SError
> Interrupt and thus causes a kernel panic when non-posted transactions
> time out.  This differs from most other PCIe controllers which return a
> fabricated "all ones" response instead.
> 
> To avoid gratuitous kernel panics, the driver reads the link status from
> a proprietary PCIE_MISC_PCIE_STATUS register and skips downstream
> accesses if the link is down.
> 
> However the bits in the proprietary register may purport that the link
> is up even though link training is still in progress (as indicated by
> the Link Training bit in the Link Status register).
> 
> This has been observed with various PCIe switches attached to a BCM2711
> (Raspberry Pi CM4):  The issue is most pronounced with the Pericom
> PI7C9X2G404SV, but has also occasionally been witnessed with the Pericom
> PI7C9X2G404SL and ASMedia ASM1184e.
> 
> Check the Link Training bit in addition to the PCIE_MISC_PCIE_STATUS
> register before performing downstream accesses.
> 
> Signed-off-by: Lukas Wunner <lukas@wunner.de>
> Cc: <stable@vger.kernel.org>

Since there was a fair bit of discussion and no obvious conclusion,
I'm dropping this for now.  Please repost if appropriate and maybe we
can get some testing/reviews/acks.

> ---
>  drivers/pci/controller/pcie-brcmstb.c | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/pci/controller/pcie-brcmstb.c b/drivers/pci/controller/pcie-brcmstb.c
> index f593a422bd63..b4abfced8e9b 100644
> --- a/drivers/pci/controller/pcie-brcmstb.c
> +++ b/drivers/pci/controller/pcie-brcmstb.c
> @@ -679,8 +679,10 @@ static bool brcm_pcie_link_up(struct brcm_pcie *pcie)
>  	u32 val = readl(pcie->base + PCIE_MISC_PCIE_STATUS);
>  	u32 dla = FIELD_GET(PCIE_MISC_PCIE_STATUS_PCIE_DL_ACTIVE_MASK, val);
>  	u32 plu = FIELD_GET(PCIE_MISC_PCIE_STATUS_PCIE_PHYLINKUP_MASK, val);
> +	u16 lnksta = readw(pcie->base + BRCM_PCIE_CAP_REGS + PCI_EXP_LNKSTA);
> +	u16 lt = FIELD_GET(PCI_EXP_LNKSTA_LT, lnksta);
>  
> -	return dla && plu;
> +	return dla && plu && !lt;
>  }
>  
>  static void __iomem *brcm_pcie_map_bus(struct pci_bus *bus,
> -- 
> 2.39.2
> 

      parent reply	other threads:[~2023-12-08 19:54 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-08-06  4:44 [PATCH] PCI: brcmstb: Avoid downstream access during link training Lukas Wunner
2023-08-06 15:15 ` Florian Fainelli
2023-08-14 19:55   ` Jim Quinlan
2023-08-06 21:43 ` Pali Rohár
2023-08-07  2:48   ` Florian Fainelli
2023-08-07  7:13     ` Pali Rohár
2023-08-07  7:28       ` Pali Rohár
2023-10-24  1:13         ` Florian Fainelli
2023-10-24  0:51 ` Bjorn Helgaas
2023-12-08 19:53 ` Bjorn Helgaas [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20231208195359.GA808176@bhelgaas \
    --to=helgaas@kernel.org \
    --cc=bcm-kernel-feedback-list@broadcom.com \
    --cc=bhelgaas@google.com \
    --cc=florian.fainelli@broadcom.com \
    --cc=jim2101024@gmail.com \
    --cc=kibi@debian.org \
    --cc=kw@linux.com \
    --cc=linux-pci@vger.kernel.org \
    --cc=lpieralisi@kernel.org \
    --cc=lukas@wunner.de \
    --cc=nsaenz@kernel.org \
    --cc=p.rosenberger@kunbus.com \
    --cc=robh@kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).