Merge tag 'cxl-fixes-6.12-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/cxl/cxl

Pull cxl fixes from Ira Weiny:
 "The bulk of these fixes center around an initialization order bug
  reported by Gregory Price and some additional fall out from the
  debugging effort.

  In summary, cxl_acpi and cxl_mem race and previously worked because of
  a bus_rescan_devices() while testing without modules built in.

  Unfortunately with modules built in the rescan would fail due to the
  cxl_port driver being registered late via the build order. Furthermore
  it was found bus_rescan_devices() did not guarantee a probe barrier
  which CXL was expecting. Additional fixes to cxl-test and decoder
  allocation came along as they were found in this debugging effort.

  The other fixes are pretty minor but one affects trace point data seen
  by user space.

  Summary:

   - Fix crashes when running with cxl-test code

   - Fix Trace DRAM Event Record field decodes

   - Fix module/built in initialization order errors

   - Fix use after free on decoder shutdowns

   - Fix out of order decoder allocations

   - Improve cxl-test to better reflect real world systems"

* tag 'cxl-fixes-6.12-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/cxl/cxl:
  cxl/test: Improve init-order fidelity relative to real-world systems
  cxl/port: Prevent out-of-order decoder allocation
  cxl/port: Fix use-after-free, permit out-of-order decoder shutdown
  cxl/acpi: Ensure ports ready at cxl_acpi_probe() return
  cxl/port: Fix cxl_bus_rescan() vs bus_rescan_devices()
  cxl/port: Fix CXL port initialization order when the subsystem is built-in
  cxl/events: Fix Trace DRAM Event Record
  cxl/core: Return error when cxl_endpoint_gather_bandwidth() handles a non-PCI device

Author: torvalds

drivers/base/core.c

@@ -4037,6 +4037,41 @@ int device_for_each_child_reverse(struct device *parent, void *data,
 }
 EXPORT_SYMBOL_GPL(device_for_each_child_reverse);
 
+/**
+ * device_for_each_child_reverse_from - device child iterator in reversed order.
+ * @parent: parent struct device.
+ * @from: optional starting point in child list
+ * @fn: function to be called for each device.
+ * @data: data for the callback.
+ *
+ * Iterate over @parent's child devices, starting at @from, and call @fn
+ * for each, passing it @data. This helper is identical to
+ * device_for_each_child_reverse() when @from is NULL.
+ *
+ * @fn is checked each iteration. If it returns anything other than 0,
+ * iteration stop and that value is returned to the caller of
+ * device_for_each_child_reverse_from();
+ */
+int device_for_each_child_reverse_from(struct device *parent,
+				       struct device *from, const void *data,
+				       int (*fn)(struct device *, const void *))
+{
+	struct klist_iter i;
+	struct device *child;
+	int error = 0;
+
+	if (!parent->p)
+		return 0;
+
+	klist_iter_init_node(&parent->p->klist_children, &i,
+			     (from ? &from->p->knode_parent : NULL));
+	while ((child = prev_device(&i)) && !error)
+		error = fn(child, data);
+	klist_iter_exit(&i);
+	return error;
+}
+EXPORT_SYMBOL_GPL(device_for_each_child_reverse_from);
+
 /**
  * device_find_child - device iterator for locating a particular device.
  * @parent: parent struct device

drivers/cxl/Kconfig

@@ -60,6 +60,7 @@ config CXL_ACPI
 	default CXL_BUS
 	select ACPI_TABLE_LIB
 	select ACPI_HMAT
+	select CXL_PORT
 	help
 	  Enable support for host managed device memory (HDM) resources
 	  published by a platform's ACPI CXL memory layout description.  See

drivers/cxl/Makefile

@@ -1,13 +1,21 @@
 # SPDX-License-Identifier: GPL-2.0
+
+# Order is important here for the built-in case:
+# - 'core' first for fundamental init
+# - 'port' before platform root drivers like 'acpi' so that CXL-root ports
+#   are immediately enabled
+# - 'mem' and 'pmem' before endpoint drivers so that memdevs are
+#   immediately enabled
+# - 'pci' last, also mirrors the hardware enumeration hierarchy
 obj-y += core/
-obj-$(CONFIG_CXL_PCI) += cxl_pci.o
-obj-$(CONFIG_CXL_MEM) += cxl_mem.o
+obj-$(CONFIG_CXL_PORT) += cxl_port.o
 obj-$(CONFIG_CXL_ACPI) += cxl_acpi.o
 obj-$(CONFIG_CXL_PMEM) += cxl_pmem.o
-obj-$(CONFIG_CXL_PORT) += cxl_port.o
+obj-$(CONFIG_CXL_MEM) += cxl_mem.o
+obj-$(CONFIG_CXL_PCI) += cxl_pci.o
 
-cxl_mem-y := mem.o
-cxl_pci-y := pci.o
+cxl_port-y := port.o
 cxl_acpi-y := acpi.o
 cxl_pmem-y := pmem.o security.o
-cxl_port-y := port.o
+cxl_mem-y := mem.o
+cxl_pci-y := pci.o

drivers/cxl/acpi.c

@@ -924,6 +924,13 @@ static void __exit cxl_acpi_exit(void)
 
 /* load before dax_hmem sees 'Soft Reserved' CXL ranges */
 subsys_initcall(cxl_acpi_init);
+
+/*
+ * Arrange for host-bridge ports to be active synchronous with
+ * cxl_acpi_probe() exit.
+ */
+MODULE_SOFTDEP("pre: cxl_port");
+
 module_exit(cxl_acpi_exit);
 MODULE_DESCRIPTION("CXL ACPI: Platform Support");
 MODULE_LICENSE("GPL v2");

drivers/cxl/core/cdat.c

@@ -641,6 +641,9 @@ static int cxl_endpoint_gather_bandwidth(struct cxl_region *cxlr,
 	void *ptr;
 	int rc;
 
+	if (!dev_is_pci(cxlds->dev))
+		return -ENODEV;
+
 	if (cxlds->rcd)
 		return -ENODEV;
 

drivers/cxl/core/hdm.c

@@ -712,7 +712,44 @@ static int cxl_decoder_commit(struct cxl_decoder *cxld)
 	return 0;
 }
 
-static int cxl_decoder_reset(struct cxl_decoder *cxld)
+static int commit_reap(struct device *dev, const void *data)
+{
+	struct cxl_port *port = to_cxl_port(dev->parent);
+	struct cxl_decoder *cxld;
+
+	if (!is_switch_decoder(dev) && !is_endpoint_decoder(dev))
+		return 0;
+
+	cxld = to_cxl_decoder(dev);
+	if (port->commit_end == cxld->id &&
+	    ((cxld->flags & CXL_DECODER_F_ENABLE) == 0)) {
+		port->commit_end--;
+		dev_dbg(&port->dev, "reap: %s commit_end: %d\n",
+			dev_name(&cxld->dev), port->commit_end);
+	}
+
+	return 0;
+}
+
+void cxl_port_commit_reap(struct cxl_decoder *cxld)
+{
+	struct cxl_port *port = to_cxl_port(cxld->dev.parent);
+
+	lockdep_assert_held_write(&cxl_region_rwsem);
+
+	/*
+	 * Once the highest committed decoder is disabled, free any other
+	 * decoders that were pinned allocated by out-of-order release.
+	 */
+	port->commit_end--;
+	dev_dbg(&port->dev, "reap: %s commit_end: %d\n", dev_name(&cxld->dev),
+		port->commit_end);
+	device_for_each_child_reverse_from(&port->dev, &cxld->dev, NULL,
+					   commit_reap);
+}
+EXPORT_SYMBOL_NS_GPL(cxl_port_commit_reap, CXL);
+
+static void cxl_decoder_reset(struct cxl_decoder *cxld)
 {
 	struct cxl_port *port = to_cxl_port(cxld->dev.parent);
 	struct cxl_hdm *cxlhdm = dev_get_drvdata(&port->dev);
@@ -721,14 +758,14 @@ static int cxl_decoder_reset(struct cxl_decoder *cxld)
 	u32 ctrl;
 
 	if ((cxld->flags & CXL_DECODER_F_ENABLE) == 0)
-		return 0;
+		return;
 
-	if (port->commit_end != id) {
+	if (port->commit_end == id)
+		cxl_port_commit_reap(cxld);
+	else
 		dev_dbg(&port->dev,
 			"%s: out of order reset, expected decoder%d.%d\n",
 			dev_name(&cxld->dev), port->id, port->commit_end);
-		return -EBUSY;
-	}
 
 	down_read(&cxl_dpa_rwsem);
 	ctrl = readl(hdm + CXL_HDM_DECODER0_CTRL_OFFSET(id));
@@ -741,7 +778,6 @@ static int cxl_decoder_reset(struct cxl_decoder *cxld)
 	writel(0, hdm + CXL_HDM_DECODER0_BASE_LOW_OFFSET(id));
 	up_read(&cxl_dpa_rwsem);
 
-	port->commit_end--;
 	cxld->flags &= ~CXL_DECODER_F_ENABLE;
 
 	/* Userspace is now responsible for reconfiguring this decoder */
@@ -751,8 +787,6 @@ static int cxl_decoder_reset(struct cxl_decoder *cxld)
 		cxled = to_cxl_endpoint_decoder(&cxld->dev);
 		cxled->state = CXL_DECODER_STATE_MANUAL;
 	}
-
-	return 0;
 }
 
 static int cxl_setup_hdm_decoder_from_dvsec(

drivers/cxl/core/port.c

@@ -2084,11 +2084,18 @@ static void cxl_bus_remove(struct device *dev)
 
 static struct workqueue_struct *cxl_bus_wq;
 
-static void cxl_bus_rescan_queue(struct work_struct *w)
+static int cxl_rescan_attach(struct device *dev, void *data)
 {
-	int rc = bus_rescan_devices(&cxl_bus_type);
+	int rc = device_attach(dev);
+
+	dev_vdbg(dev, "rescan: %s\n", rc ? "attach" : "detached");
 
-	pr_debug("CXL bus rescan result: %d\n", rc);
+	return 0;
+}
+
+static void cxl_bus_rescan_queue(struct work_struct *w)
+{
+	bus_for_each_dev(&cxl_bus_type, NULL, NULL, cxl_rescan_attach);
 }
 
 void cxl_bus_rescan(void)

drivers/cxl/core/region.c

@@ -232,8 +232,8 @@ static int cxl_region_invalidate_memregion(struct cxl_region *cxlr)
 				"Bypassing cpu_cache_invalidate_memregion() for testing!\n");
 			return 0;
 		} else {
-			dev_err(&cxlr->dev,
-				"Failed to synchronize CPU cache state\n");
+			dev_WARN(&cxlr->dev,
+				 "Failed to synchronize CPU cache state\n");
 			return -ENXIO;
 		}
 	}
@@ -242,19 +242,17 @@ static int cxl_region_invalidate_memregion(struct cxl_region *cxlr)
 	return 0;
 }
 
-static int cxl_region_decode_reset(struct cxl_region *cxlr, int count)
+static void cxl_region_decode_reset(struct cxl_region *cxlr, int count)
 {
 	struct cxl_region_params *p = &cxlr->params;
-	int i, rc = 0;
+	int i;
 
 	/*
-	 * Before region teardown attempt to flush, and if the flush
-	 * fails cancel the region teardown for data consistency
-	 * concerns
+	 * Before region teardown attempt to flush, evict any data cached for
+	 * this region, or scream loudly about missing arch / platform support
+	 * for CXL teardown.
 	 */
-	rc = cxl_region_invalidate_memregion(cxlr);
-	if (rc)
-		return rc;
+	cxl_region_invalidate_memregion(cxlr);
 
 	for (i = count - 1; i >= 0; i--) {
 		struct cxl_endpoint_decoder *cxled = p->targets[i];
@@ -277,23 +275,17 @@ static int cxl_region_decode_reset(struct cxl_region *cxlr, int count)
 			cxl_rr = cxl_rr_load(iter, cxlr);
 			cxld = cxl_rr->decoder;
 			if (cxld->reset)
-				rc = cxld->reset(cxld);
-			if (rc)
-				return rc;
+				cxld->reset(cxld);
 			set_bit(CXL_REGION_F_NEEDS_RESET, &cxlr->flags);
 		}
 
 endpoint_reset:
-		rc = cxled->cxld.reset(&cxled->cxld);
-		if (rc)
-			return rc;
+		cxled->cxld.reset(&cxled->cxld);
 		set_bit(CXL_REGION_F_NEEDS_RESET, &cxlr->flags);
 	}
 
 	/* all decoders associated with this region have been torn down */
 	clear_bit(CXL_REGION_F_NEEDS_RESET, &cxlr->flags);
-
-	return 0;
 }
 
 static int commit_decoder(struct cxl_decoder *cxld)
@@ -409,16 +401,8 @@ static ssize_t commit_store(struct device *dev, struct device_attribute *attr,
 		 * still pending.
 		 */
 		if (p->state == CXL_CONFIG_RESET_PENDING) {
-			rc = cxl_region_decode_reset(cxlr, p->interleave_ways);
-			/*
-			 * Revert to committed since there may still be active
-			 * decoders associated with this region, or move forward
-			 * to active to mark the reset successful
-			 */
-			if (rc)
-				p->state = CXL_CONFIG_COMMIT;
-			else
-				p->state = CXL_CONFIG_ACTIVE;
+			cxl_region_decode_reset(cxlr, p->interleave_ways);
+			p->state = CXL_CONFIG_ACTIVE;
 		}
 	}
 
@@ -794,26 +778,50 @@ static size_t show_targetN(struct cxl_region *cxlr, char *buf, int pos)
 	return rc;
 }
 
+static int check_commit_order(struct device *dev, const void *data)
+{
+	struct cxl_decoder *cxld = to_cxl_decoder(dev);
+
+	/*
+	 * if port->commit_end is not the only free decoder, then out of
+	 * order shutdown has occurred, block further allocations until
+	 * that is resolved
+	 */
+	if (((cxld->flags & CXL_DECODER_F_ENABLE) == 0))
+		return -EBUSY;
+	return 0;
+}
+
 static int match_free_decoder(struct device *dev, void *data)
 {
+	struct cxl_port *port = to_cxl_port(dev->parent);
 	struct cxl_decoder *cxld;
-	int *id = data;
+	int rc;
 
 	if (!is_switch_decoder(dev))
 		return 0;
 
 	cxld = to_cxl_decoder(dev);
 
-	/* enforce ordered allocation */
-	if (cxld->id != *id)
+	if (cxld->id != port->commit_end + 1)
 		return 0;
 
-	if (!cxld->region)
-		return 1;
-
-	(*id)++;
+	if (cxld->region) {
+		dev_dbg(dev->parent,
+			"next decoder to commit (%s) is already reserved (%s)\n",
+			dev_name(dev), dev_name(&cxld->region->dev));
+		return 0;
+	}
 
-	return 0;
+	rc = device_for_each_child_reverse_from(dev->parent, dev, NULL,
+						check_commit_order);
+	if (rc) {
+		dev_dbg(dev->parent,
+			"unable to allocate %s due to out of order shutdown\n",
+			dev_name(dev));
+		return 0;
+	}
+	return 1;
 }
 
 static int match_auto_decoder(struct device *dev, void *data)
@@ -840,7 +848,6 @@ cxl_region_find_decoder(struct cxl_port *port,
 			struct cxl_region *cxlr)
 {
 	struct device *dev;
-	int id = 0;
 
 	if (port == cxled_to_port(cxled))
 		return &cxled->cxld;
@@ -849,7 +856,7 @@ cxl_region_find_decoder(struct cxl_port *port,
 		dev = device_find_child(&port->dev, &cxlr->params,
 					match_auto_decoder);
 	else
-		dev = device_find_child(&port->dev, &id, match_free_decoder);
+		dev = device_find_child(&port->dev, NULL, match_free_decoder);
 	if (!dev)
 		return NULL;
 	/*
@@ -2054,13 +2061,7 @@ static int cxl_region_detach(struct cxl_endpoint_decoder *cxled)
 	get_device(&cxlr->dev);
 
 	if (p->state > CXL_CONFIG_ACTIVE) {
-		/*
-		 * TODO: tear down all impacted regions if a device is
-		 * removed out of order
-		 */
-		rc = cxl_region_decode_reset(cxlr, p->interleave_ways);
-		if (rc)
-			goto out;
+		cxl_region_decode_reset(cxlr, p->interleave_ways);
 		p->state = CXL_CONFIG_ACTIVE;
 	}