From: Bjorn Helgaas <helgaas@kernel.org>
To: "Woods, Brian" <Brian.Woods@amd.com>
Cc: Thomas Gleixner <tglx@linutronix.de>,
Ingo Molnar <mingo@redhat.com>, Borislav Petkov <bp@alien8.de>,
"H. Peter Anvin" <hpa@zytor.com>,
"x86@kernel.org" <x86@kernel.org>,
Clemens Ladisch <clemens@ladisch.de>,
Jean Delvare <jdelvare@suse.com>,
Guenter Roeck <linux@roeck-us.net>, Pu Wen <puwen@hygon.cn>,
Jia Zhang <qianyue.zj@alibaba-inc.com>,
"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
"linux-hwmon@vger.kernel.org" <linux-hwmon@vger.kernel.org>,
"linux-pci@vger.kernel.org" <linux-pci@vger.kernel.org>
Subject: Re: [PATCH 2/4] x86/amd_nb: add support for newer PCI topologies
Date: Fri, 2 Nov 2018 14:59:25 -0500 [thread overview]
Message-ID: <20181102195925.GB160487@google.com> (raw)
In-Reply-To: <20181102181055.130531-3-brian.woods@amd.com>
On Fri, Nov 02, 2018 at 06:11:07PM +0000, Woods, Brian wrote:
> Add support for new processors which have multiple PCI root complexes
> per data fabric/SMN interface. The interfaces per root complex are
> redundant and should be skipped. This makes sure the DF/SMN interfaces
> get accessed via the correct root complex.
SMN?
> Ex:
> DF/SMN 0 -> 60
> 40
> 20
> 00
> DF/SMN 1 -> e0
> c0
> a0
> 80
This isn't my code, and I'm not really objecting to these changes, but
from where I sit, the fact that you need this sort of vendor-specific
topology discovery is a little bit ugly and seems like something of a
maintenance issue. You could argue that this is sort of an "AMD CPU
driver", which is entitled to be device-specific, and that does make
some sense.
But device-specific code is typically packaged as a driver that uses
driver registration interfaces like acpi_bus_register_driver(),
pci_register_driver(), etc. That gives you a consistent structure
and, more importantly, a framework for dealing with hotplug. It
doesn't look like amd_nb.c would deal well with hot-add of CPUs.
> Signed-off-by: Brian Woods <brian.woods@amd.com>
> ---
> arch/x86/kernel/amd_nb.c | 41 +++++++++++++++++++++++++++++++++++------
> 1 file changed, 35 insertions(+), 6 deletions(-)
>
> diff --git a/arch/x86/kernel/amd_nb.c b/arch/x86/kernel/amd_nb.c
> index 19d489ee2b1e..c0bf26aeb7c3 100644
> --- a/arch/x86/kernel/amd_nb.c
> +++ b/arch/x86/kernel/amd_nb.c
> @@ -213,7 +213,10 @@ int amd_cache_northbridges(void)
> const struct pci_device_id *root_ids = amd_root_ids;
> struct pci_dev *root, *misc, *link;
> struct amd_northbridge *nb;
> - u16 i = 0;
> + u16 roots_per_misc = 0;
> + u16 misc_count = 0;
> + u16 root_count = 0;
> + u16 i, j;
>
> if (amd_northbridges.num)
> return 0;
> @@ -226,26 +229,52 @@ int amd_cache_northbridges(void)
>
> misc = NULL;
> while ((misc = next_northbridge(misc, misc_ids)) != NULL)
> - i++;
> + misc_count++;
>
> - if (!i)
> + root = NULL;
> + while ((root = next_northbridge(root, root_ids)) != NULL)
> + root_count++;
> +
> + if (!misc_count)
> return -ENODEV;
>
> - nb = kcalloc(i, sizeof(struct amd_northbridge), GFP_KERNEL);
> + if (root_count) {
> + roots_per_misc = root_count / misc_count;
> +
> + /*
> + * There should be _exactly_ N roots for each DF/SMN
> + * interface.
> + */
> + if (!roots_per_misc || (root_count % roots_per_misc)) {
> + pr_info("Unsupported AMD DF/PCI configuration found\n");
> + return -ENODEV;
> + }
> + }
> +
> + nb = kcalloc(misc_count, sizeof(struct amd_northbridge), GFP_KERNEL);
> if (!nb)
> return -ENOMEM;
>
> amd_northbridges.nb = nb;
> - amd_northbridges.num = i;
> + amd_northbridges.num = misc_count;
>
> link = misc = root = NULL;
> - for (i = 0; i != amd_northbridges.num; i++) {
> + for (i = 0; i < amd_northbridges.num; i++) {
> node_to_amd_nb(i)->root = root =
> next_northbridge(root, root_ids);
> node_to_amd_nb(i)->misc = misc =
> next_northbridge(misc, misc_ids);
> node_to_amd_nb(i)->link = link =
> next_northbridge(link, link_ids);
> +
> + /*
> + * If there are more root devices than data fabric/SMN,
> + * interfaces, then the root devices per DF/SMN
> + * interface are redundant and N-1 should be skipped so
> + * they aren't mapped incorrectly.
> + */
> + for (j = 1; j < roots_per_misc; j++)
> + root = next_northbridge(root, root_ids);
> }
>
> if (amd_gart_present())
> --
> 2.11.0
>
next prev parent reply other threads:[~2018-11-02 19:59 UTC|newest]
Thread overview: 36+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-11-02 18:11 [PATCH 0/4] Update DF/SMN access and k10temp for AMD F17h M30h Woods, Brian
2018-11-02 18:11 ` [PATCH 1/4] k10temp: x86/amd_nb: consolidate shared device IDs Woods, Brian
2018-11-02 18:24 ` Guenter Roeck
2018-11-02 18:11 ` [PATCH 2/4] x86/amd_nb: add support for newer PCI topologies Woods, Brian
2018-11-02 19:59 ` Bjorn Helgaas [this message]
2018-11-02 23:29 ` Borislav Petkov
2018-11-05 21:45 ` Bjorn Helgaas
2018-11-05 21:56 ` Borislav Petkov
2018-11-06 21:42 ` Bjorn Helgaas
2018-11-06 22:00 ` Borislav Petkov
2018-11-06 23:20 ` Bjorn Helgaas
2018-11-07 9:18 ` Borislav Petkov
2018-11-07 13:38 ` Bjorn Helgaas
2018-11-07 16:07 ` Borislav Petkov
2018-11-07 17:10 ` Bjorn Helgaas
2018-11-07 17:17 ` Borislav Petkov
2018-11-07 19:50 ` Woods, Brian
2018-11-07 13:51 ` Guenter Roeck
2018-11-07 17:16 ` Bjorn Helgaas
2018-11-07 19:15 ` Srinivas Pandruvada
2018-11-07 21:31 ` Bjorn Helgaas
2018-11-07 22:42 ` Srinivas Pandruvada
2018-11-07 23:14 ` Bjorn Helgaas
2018-11-07 23:30 ` Srinivas Pandruvada
2018-11-07 23:44 ` Srinivas Pandruvada
2018-11-08 1:40 ` Guenter Roeck
2018-11-08 13:59 ` Bjorn Helgaas
2018-11-05 19:38 ` Borislav Petkov
2018-11-05 20:33 ` Woods, Brian
2018-11-05 21:42 ` Borislav Petkov
2018-11-05 23:32 ` Woods, Brian
2018-11-06 8:27 ` Borislav Petkov
2018-11-02 18:11 ` [PATCH 3/4] x86/amd_nb: add PCI device IDs for F17h M30h Woods, Brian
2018-11-02 18:11 ` [PATCH 4/4] hwmon: k10temp: add support for AMD F17h M30h CPUs Woods, Brian
2018-11-02 18:26 ` Guenter Roeck
2018-11-05 20:32 ` Borislav Petkov
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=20181102195925.GB160487@google.com \
--to=helgaas@kernel.org \
--cc=Brian.Woods@amd.com \
--cc=bp@alien8.de \
--cc=clemens@ladisch.de \
--cc=hpa@zytor.com \
--cc=jdelvare@suse.com \
--cc=linux-hwmon@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-pci@vger.kernel.org \
--cc=linux@roeck-us.net \
--cc=mingo@redhat.com \
--cc=puwen@hygon.cn \
--cc=qianyue.zj@alibaba-inc.com \
--cc=tglx@linutronix.de \
--cc=x86@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).