From: Bjorn Helgaas <helgaas@kernel.org>
To: Nirmal Patel <nirmal.patel@linux.intel.com>
Cc: linux-pci@vger.kernel.org
Subject: Re: [PATCH] PCI: vmd: Fix domain reset operation
Date: Tue, 20 Jun 2023 12:58:52 -0500 [thread overview]
Message-ID: <20230620175852.GA51832@bhelgaas> (raw)
In-Reply-To: <20230530214706.75700-1-nirmal.patel@linux.intel.com>
On Tue, May 30, 2023 at 02:47:06PM -0700, Nirmal Patel wrote:
> During domain reset process we are accidentally enabling
> the prefetchable memory by writing 0x0 to Prefetchable Memory
> Base and Prefetchable Memory Limit registers. As a result certain
> platforms failed to boot up.
>
> Here is the quote from section 7.5.1.3.9 of PCI Express Base 6.0 spec:
>
> The Prefetchable Memory Limit register must be programmed to a smaller
> value than the Prefetchable Memory Base register if there is no
> prefetchable memory on the secondary side of the bridge.
>
> When clearing Prefetchable Memory Base, Prefetchable Memory
> Limit and Prefetchable Base Upper 32 bits, the prefetchable
> memory range becomes 0x0-0x575000fffff. As a result the
> prefetchable memory is enabled accidentally.
>
> Implementing correct operation by writing a value to Prefetchable
> Base Memory larger than the value of Prefetchable Memory Limit.
Can you make the subject line more descriptive than "Fix X"?
There are a zillion ways X might be broken, so it's useful to have a
hint about which one it is.
Actually, vmd_domain_reset() is sort of weirdly named, because I don't
see any *reset* there. It looks like it's actually disabling all
bridge windows (not just the prefetchable one, as the log suggests).
The fact that this makes a difference suggests that the bridges had
PCI_COMMAND_MEMORY set, which sounds like something you may not have
wanted.
> Signed-off-by: Nirmal Patel <nirmal.patel@linux.intel.com>
> ---
> drivers/pci/controller/vmd.c | 14 ++++++++++++--
> 1 file changed, 12 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/pci/controller/vmd.c b/drivers/pci/controller/vmd.c
> index 769eedeb8802..f3eb740e3028 100644
> --- a/drivers/pci/controller/vmd.c
> +++ b/drivers/pci/controller/vmd.c
> @@ -526,8 +526,18 @@ static void vmd_domain_reset(struct vmd_dev *vmd)
> PCI_CLASS_BRIDGE_PCI))
> continue;
>
> - memset_io(base + PCI_IO_BASE, 0,
> - PCI_ROM_ADDRESS1 - PCI_IO_BASE);
> + writel(0, base + PCI_IO_BASE);
> + writew(0xFFF0, base + PCI_MEMORY_BASE);
> + writew(0, base + PCI_MEMORY_LIMIT);
> +
> + writew(0xFFF1, base + PCI_PREF_MEMORY_BASE);
> + writew(0, base + PCI_PREF_MEMORY_LIMIT);
> +
> + writel(0xFFFFFFFF, base + PCI_PREF_BASE_UPPER32);
> + writel(0, base + PCI_PREF_LIMIT_UPPER32);
> +
> + writel(0, base + PCI_IO_BASE_UPPER16);
> + writeb(0, base + PCI_CAPABILITY_LIST);
> }
> }
> }
> --
> 2.27.0
>
next prev parent reply other threads:[~2023-06-20 17:58 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-05-30 21:47 [PATCH] PCI: vmd: Fix domain reset operation Nirmal Patel
2023-06-20 17:14 ` Patel, Nirmal
2023-06-20 17:36 ` Lorenzo Pieralisi
2023-06-20 17:58 ` Bjorn Helgaas [this message]
2023-06-20 18:07 ` Bjorn Helgaas
2023-06-21 23:02 ` Patel, Nirmal
2023-07-05 17:39 ` Patel, Nirmal
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=20230620175852.GA51832@bhelgaas \
--to=helgaas@kernel.org \
--cc=linux-pci@vger.kernel.org \
--cc=nirmal.patel@linux.intel.com \
/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