From: Jonathan Cameron <Jonathan.Cameron@Huawei.com>
To: <alison.schofield@intel.com>
Cc: Davidlohr Bueso <dave@stgolabs.net>,
Dave Jiang <dave.jiang@intel.com>,
Vishal Verma <vishal.l.verma@intel.com>,
Ira Weiny <ira.weiny@intel.com>,
Dan Williams <dan.j.williams@intel.com>,
<linux-cxl@vger.kernel.org>
Subject: Re: [PATCH v2 1/2] cxl/region: Try to add a region resource to a soft reserved parent
Date: Mon, 7 Aug 2023 15:24:07 +0100 [thread overview]
Message-ID: <20230807152407.000070e6@Huawei.com> (raw)
In-Reply-To: <6cd07222f169d951921486306187e87951853035.1691176651.git.alison.schofield@intel.com>
On Fri, 4 Aug 2023 12:31:39 -0700
alison.schofield@intel.com wrote:
> From: Alison Schofield <alison.schofield@intel.com>
>
> During region autodiscovery, the region driver always inserts the
> region resource as a child of the root decoder, a CXL WINDOW. It
> has the effect of making a soft reserved resource, with an exactly
> matching address range, a child of the region resource.
>
> It looks like this in /proc/iomem:
>
> 2080000000-29dbfffffff : CXL Window 0
> 2080000000-247fffffff : region0
> 2080000000-247fffffff : Soft Reserved
>
> Search for soft reserved resources that include the region resource,
> and add the new region resource as a child of that found resource.
> If a soft reserved resource is not found, insert to the root decoder
> as usual.
>
> With this change, it looks like this:
>
> 2080000000-29dbfffffff : CXL Window 0
> 2080000000-247fffffff : Soft Reserved
> 2080000000-247fffffff : region0
>
> This odd parenting only occurs when the resources are an exact match.
> When the region resource only uses part of a soft reserved resource,
> the parenting appears more logical like this:
>
> 2080000000-29dbfffffff : CXL Window 0
> 2080000000-287fffffff : Soft Reserved
> 2080000000-247fffffff : region0
>
> Aside from the more logical appearance, this change is in preparation
> for further cleanup in region teardown. A follow-on patch intends to
> find and free soft reserved resources upon region teardown.
>
> Signed-off-by: Alison Schofield <alison.schofield@intel.com>
Hi Alison
A comment and a question inline.
Thanks,
Jonathan
> ---
> drivers/cxl/core/region.c | 32 ++++++++++++++++++++++++++++++--
> 1 file changed, 30 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/cxl/core/region.c b/drivers/cxl/core/region.c
> index e115ba382e04..36c0c0dd5697 100644
> --- a/drivers/cxl/core/region.c
> +++ b/drivers/cxl/core/region.c
> @@ -2709,6 +2709,28 @@ static int match_region_by_range(struct device *dev, void *data)
> return rc;
> }
>
> +static int insert_resource_soft_reserved(struct resource *soft_res, void *arg)
> +{
> + struct resource *parent, *new, *res = arg;
> + bool found = false;
> +
> + parent = soft_res->parent;
> + if (!parent)
> + return 0;
> +
> + /* Caller provides a copy of soft_res. Find the actual resource. */
> + for (new = parent->child; new; new = new->sibling) {
> + if (resource_contains(new, soft_res)) {
> + found = true;
> + break;
> + }
> + }
> + if (found)
> + return insert_resource(new, res) == 0;
Why not do this in the the if above then directly return there?
Also, should we get an error here, does it make sense to carry on?
(I don't think we can get an error, but if that's the case we shouldn't
be checking for one) If it's a corner we aren't sure about, good
to add a comment on what the error case might be and what we are doing
as a result.
> +
> + return 0;
> +}
> +
> /* Establish an empty region covering the given HPA range */
> static struct cxl_region *construct_region(struct cxl_root_decoder *cxlrd,
> struct cxl_endpoint_decoder *cxled)
> @@ -2719,7 +2741,7 @@ static struct cxl_region *construct_region(struct cxl_root_decoder *cxlrd,
> struct cxl_region_params *p;
> struct cxl_region *cxlr;
> struct resource *res;
> - int rc;
> + int rc = 0;
>
> do {
> cxlr = __create_region(cxlrd, cxled->mode,
> @@ -2755,7 +2777,13 @@ static struct cxl_region *construct_region(struct cxl_root_decoder *cxlrd,
>
> *res = DEFINE_RES_MEM_NAMED(hpa->start, range_len(hpa),
> dev_name(&cxlr->dev));
> - rc = insert_resource(cxlrd->res, res);
> +
> + /* Try inserting to a Soft Reserved parent, fallback to root decoder */
> + if (walk_iomem_res_desc(IORES_DESC_SOFT_RESERVED, 0,
> + res->start, res->end, res,
> + insert_resource_soft_reserved) != 1)
> + rc = insert_resource(cxlrd->res, res);
> +
> if (rc) {
> /*
> * Platform-firmware may not have split resources like "System
next prev parent reply other threads:[~2023-08-07 14:24 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-08-04 19:31 [PATCH v2 0/2] cxl/region: Improve Soft Reserved resource handling alison.schofield
2023-08-04 19:31 ` [PATCH v2 1/2] cxl/region: Try to add a region resource to a soft reserved parent alison.schofield
2023-08-07 14:24 ` Jonathan Cameron [this message]
2023-08-15 0:10 ` Alison Schofield
2023-08-09 23:22 ` Dave Jiang
2023-08-15 0:12 ` Alison Schofield
2023-08-04 19:31 ` [PATCH v2 2/2] cxl/region: Remove a soft reserved resource at region teardown alison.schofield
2023-08-09 23:34 ` Dave Jiang
2023-08-15 0:13 ` Alison Schofield
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=20230807152407.000070e6@Huawei.com \
--to=jonathan.cameron@huawei.com \
--cc=alison.schofield@intel.com \
--cc=dan.j.williams@intel.com \
--cc=dave.jiang@intel.com \
--cc=dave@stgolabs.net \
--cc=ira.weiny@intel.com \
--cc=linux-cxl@vger.kernel.org \
--cc=vishal.l.verma@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