Linux CXL
 help / color / mirror / Atom feed
From: Robert Richter <rrichter@amd.com>
To: Dave Jiang <dave.jiang@intel.com>
Cc: linux-cxl@vger.kernel.org, dave@stgolabs.net,
	jonathan.cameron@huawei.com, alison.schofield@intel.com,
	vishal.l.verma@intel.com, ira.weiny@intel.com,
	dan.j.williams@intel.com
Subject: [REWORK][PATCH] cxl: Remove core/acpi.c and cxl core dependency on ACPI
Date: Fri, 11 Jul 2025 13:37:20 +0200	[thread overview]
Message-ID: <aHD3cDOZ4R3Vm7YA@rric.localdomain> (raw)
In-Reply-To: <5282f0d4-a25b-4fc5-9890-9f50052bd4df@intel.com>

Hi Dave,

On 03.07.25 16:49:11, Dave Jiang wrote:
> 
> 
> On 6/10/25 10:29 AM, Dave Jiang wrote:
> > It was a mistake to introduce core/acpi.c and putting ACPI dependency on
> > cxl_core when adding the extended linear cache support. Add a callback
> > in the cxl_root_decoder to retrieve the extended linear cache size from
> > ACPI via the cxl_acpi driver.
> > 
> > Signed-off-by: Dave Jiang <dave.jiang@intel.com>
> 
> Applied to cxl/next
> 1a45e1cd694fdf2d2a22e3c70a6917ff39fe77ab

I am sorry to come back to this patch that late in the process. While
rebasing my current region rework and address translation patches I
came across your patch and realized that I have a much more simpler
rework of that in my patch queue. I extracted that patch below from my
series I am working on. Could you take a look at it? The changes are
less intrusive and part of my rework required for address translation.
If it looks good to you and during review, I rather prefer to replace
your patch with this one. Note that I have follow on patches that
reduce more code and remove the cxl_extended_linear_cache_resize()
function entirely.

Patch below bases on:

 ac0fe6a57317 cxl: make cxl_bus_type constant

Many thanks,

-Robert




From fd3f1f5b33313bb10a2e51861031015ac48a8cd1 Mon Sep 17 00:00:00 2001
From: Robert Richter <rrichter@amd.com>
Date: Fri, 11 Jul 2025 11:56:56 +0200
Subject: [PATCH] cxl: Remove core/acpi.c and cxl core dependency on ACPI

From Dave [1]:

"""
It was a mistake to introduce core/acpi.c and putting ACPI dependency on
cxl_core when adding the extended linear cache support.
"""

Current implementation calls hmat_get_extended_linear_cache_size() of
the ACPI subsystem. That external reference causes issue running
cxl_test as there is no way to "mock" that function and ignore it when
using cxl test.

Instead of working around that using cxlrd ops and extensively
expanding cxl_test code [1], just move HMAT calls out of the core
module to cxl_acpi. Implement this by adding a @cache_size member to
struct cxl_root_decoder. During initialization the cache size is
determined and added to the root decoder object in cxl_acpi. Later on
in cxl_core the cache_size parameter is used to setup extended linear
caching.

[1] https://patch.msgid.link/20250610172938.139428-1-dave.jiang@intel.com

Cc: Dave Jiang <dave.jiang@intel.com>
Signed-off-by: Robert Richter <rrichter@amd.com>
---
 drivers/cxl/acpi.c        | 59 +++++++++++++++++++++++++++++++++++++++
 drivers/cxl/core/Makefile |  1 -
 drivers/cxl/core/acpi.c   | 11 --------
 drivers/cxl/core/core.h   |  2 --
 drivers/cxl/core/region.c |  7 +----
 drivers/cxl/cxl.h         |  1 +
 6 files changed, 61 insertions(+), 20 deletions(-)
 delete mode 100644 drivers/cxl/core/acpi.c

diff --git a/drivers/cxl/acpi.c b/drivers/cxl/acpi.c
index a1a99ec3f12c..712624cba2b6 100644
--- a/drivers/cxl/acpi.c
+++ b/drivers/cxl/acpi.c
@@ -335,6 +335,63 @@ static int add_or_reset_cxl_resource(struct resource *parent, struct resource *r
 	return rc;
 }
 
+static int cxl_acpi_set_cache_size(struct cxl_root_decoder *cxlrd)
+{
+	struct cxl_decoder *cxld = &cxlrd->cxlsd.cxld;
+	struct range *hpa = &cxld->hpa_range;
+	resource_size_t size = range_len(hpa);
+	resource_size_t start = hpa->start;
+	resource_size_t cache_size;
+	struct resource res;
+	int nid, rc;
+
+	res = DEFINE_RES(start, size, 0);
+	nid = phys_to_target_node(start);
+
+	rc = hmat_get_extended_linear_cache_size(&res, nid, &cache_size);
+	if (rc)
+		return rc;
+
+	/*
+	 * The cache range is expected to be within the CFMWS.
+	 * Currently there is only support cache_size == cxl_size. CXL
+	 * size is then half of the total CFMWS window size.
+	 */
+	size = size >> 1;
+	if (cache_size && size != cache_size) {
+		dev_warn(&cxld->dev,
+			 "Extended Linear Cache size %pa != CXL size %pa. No Support!",
+			 &cache_size, &size);
+		return -ENXIO;
+	}
+
+	cxlrd->cache_size = cache_size;
+
+	return 0;
+}
+
+static void cxl_setup_extended_linear_cache(struct cxl_root_decoder *cxlrd)
+{
+	int rc;
+
+	rc = cxl_acpi_set_cache_size(cxlrd);
+	if (!rc)
+		return;
+
+	if (rc != -EOPNOTSUPP) {
+		/*
+		 * Failing to support extended linear cache region resize does not
+		 * prevent the region from functioning. Only causes cxl list showing
+		 * incorrect region size.
+		 */
+		dev_warn(cxlrd->cxlsd.cxld.dev.parent,
+			 "Extended linear cache calculation failed rc:%d\n", rc);
+	}
+
+	/* Ignoring return code */
+	cxlrd->cache_size = 0;
+}
+
 DEFINE_FREE(put_cxlrd, struct cxl_root_decoder *,
 	    if (!IS_ERR_OR_NULL(_T)) put_device(&_T->cxlsd.cxld.dev))
 DEFINE_FREE(del_cxl_resource, struct resource *, if (_T) del_cxl_resource(_T))
@@ -394,6 +451,8 @@ static int __cxl_parse_cfmws(struct acpi_cedt_cfmws *cfmws,
 		ig = CXL_DECODER_MIN_GRANULARITY;
 	cxld->interleave_granularity = ig;
 
+	cxl_setup_extended_linear_cache(cxlrd);
+
 	if (cfmws->interleave_arithmetic == ACPI_CEDT_CFMWS_ARITHMETIC_XOR) {
 		if (ways != 1 && ways != 3) {
 			cxims_ctx = (struct cxl_cxims_context) {
diff --git a/drivers/cxl/core/Makefile b/drivers/cxl/core/Makefile
index 79e2ef81fde8..5ad8fef210b5 100644
--- a/drivers/cxl/core/Makefile
+++ b/drivers/cxl/core/Makefile
@@ -15,7 +15,6 @@ cxl_core-y += hdm.o
 cxl_core-y += pmu.o
 cxl_core-y += cdat.o
 cxl_core-y += ras.o
-cxl_core-y += acpi.o
 cxl_core-$(CONFIG_TRACING) += trace.o
 cxl_core-$(CONFIG_CXL_REGION) += region.o
 cxl_core-$(CONFIG_CXL_MCE) += mce.o
diff --git a/drivers/cxl/core/acpi.c b/drivers/cxl/core/acpi.c
deleted file mode 100644
index f13b4dae6ac5..000000000000
--- a/drivers/cxl/core/acpi.c
+++ /dev/null
@@ -1,11 +0,0 @@
-// SPDX-License-Identifier: GPL-2.0-only
-/* Copyright(c) 2024 Intel Corporation. All rights reserved. */
-#include <linux/acpi.h>
-#include "cxl.h"
-#include "core.h"
-
-int cxl_acpi_get_extended_linear_cache_size(struct resource *backing_res,
-					    int nid, resource_size_t *size)
-{
-	return hmat_get_extended_linear_cache_size(backing_res, nid, size);
-}
diff --git a/drivers/cxl/core/core.h b/drivers/cxl/core/core.h
index 29b61828a847..9462aea9ce9d 100644
--- a/drivers/cxl/core/core.h
+++ b/drivers/cxl/core/core.h
@@ -120,8 +120,6 @@ int cxl_port_get_switch_dport_bandwidth(struct cxl_port *port,
 int cxl_ras_init(void);
 void cxl_ras_exit(void);
 int cxl_gpf_port_setup(struct cxl_dport *dport);
-int cxl_acpi_get_extended_linear_cache_size(struct resource *backing_res,
-					    int nid, resource_size_t *size);
 
 #ifdef CONFIG_CXL_FEATURES
 struct cxl_feat_entry *
diff --git a/drivers/cxl/core/region.c b/drivers/cxl/core/region.c
index 6e5e1460068d..bc542a7142c0 100644
--- a/drivers/cxl/core/region.c
+++ b/drivers/cxl/core/region.c
@@ -3282,15 +3282,10 @@ static int cxl_extended_linear_cache_resize(struct cxl_region *cxlr,
 {
 	struct cxl_root_decoder *cxlrd = to_cxl_root_decoder(cxlr->dev.parent);
 	struct cxl_region_params *p = &cxlr->params;
-	int nid = phys_to_target_node(res->start);
 	resource_size_t size = resource_size(res);
 	resource_size_t cache_size, start;
-	int rc;
-
-	rc = cxl_acpi_get_extended_linear_cache_size(res, nid, &cache_size);
-	if (rc)
-		return rc;
 
+	cache_size = cxlrd->cache_size;
 	if (!cache_size)
 		return 0;
 
diff --git a/drivers/cxl/cxl.h b/drivers/cxl/cxl.h
index e7b66ca1d423..4643a95ca111 100644
--- a/drivers/cxl/cxl.h
+++ b/drivers/cxl/cxl.h
@@ -432,6 +432,7 @@ typedef u64 (*cxl_hpa_to_spa_fn)(struct cxl_root_decoder *cxlrd, u64 hpa);
  */
 struct cxl_root_decoder {
 	struct resource *res;
+	resource_size_t cache_size;
 	atomic_t region_id;
 	cxl_hpa_to_spa_fn hpa_to_spa;
 	void *platform_data;
-- 
2.39.5



  reply	other threads:[~2025-07-11 11:37 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-06-10 17:29 [PATCH v3] cxl: Remove core/acpi.c and cxl core dependency on ACPI Dave Jiang
2025-06-11 13:30 ` Jonathan Cameron
2025-07-02 20:53 ` Alison Schofield
2025-07-03 23:48   ` Dave Jiang
2025-07-03 23:49 ` Dave Jiang
2025-07-11 11:37   ` Robert Richter [this message]
2025-07-11 14:58     ` [REWORK][PATCH] " Dave Jiang
2025-07-11 15:17       ` Robert Richter

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=aHD3cDOZ4R3Vm7YA@rric.localdomain \
    --to=rrichter@amd.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=jonathan.cameron@huawei.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