Skip to contentSkip to conversation

resource: Use list_head to link sibling resource

PATCH Superseded Baoquan He 4 participants 19 messages Apr 19, 2018 – May 8, 2018

Subsystems (16)
DRM drivers DRM misc PCI HID +12 more subsystems

v3 of 7 Superseded Latest is v7 →

lore
Share & tools
lore

Baoquan He

author
#0 lore raw

v3 cover resource: Use list_head to link sibling resource

This patchset mainly converts strut resource's sibling list from singly linked list to doubly linked list, list_head. This is suggested by Andrew. Since I need a reversed searching on iomem_resource's IORESOURCE_SYSTEM_RAM children, the old singly linked list way makes the code in v1 post really ugly.

With this change, we only need one simple list_for_each_entry_reverse() to do reversed iteration on sibling list. The relevant codes in kernel/resource.c are more readable since those dazzling pointer operation codes for singly linked list are replaced.

With the help of list_head, in patch 0002 we can have a very simple walk_system_ram_res_rev(). And in patch 0003, will use it to search available system RAM region for kexec_buffer of kexec_file from top to down, just like we have been doing all along in kexec loading which is done in kexec-tools utility.

Note:
This patchset passed testing on my kvm guest, x86_64 arch with network enabling. The thing we need pay attetion to is that a root resource's child member need be initialized specifically with LIST_HEAD_INIT() if sttically defined or INIT_LIST_HEAD() for dynamically definition. Here Just like we do for iomem_resource/ioport_resource, or the change in get_pci_domain_busn_res().

And there are two places of change I am not very sure. One is in drivers/hv/vmbus_drv.c, the other is in drivers/pci/host/vmd.c. So will invite experts on these areas to help review.

V2 post and discussions can be found here:
https://lkml.org/lkml/2018/4/7/169

Remaining issue:
Rob raised issue that if we can make an common tree struct and helpers defined on top of list_head or a new struct, quote his saying as below. I didn't do much investigation, we may have a new struct like

struct tlist {
        void                    *parent;
        struct list_head        sibling;
        struct list_head        child;
}

Since I have some rhel dev things right now, may consider this later, put it in my TODO list. In kernel, like task_struct, task_group, they all have this kind of tree list, and have used list_head. If anyone is interested, feel free to have a try on this.

==========
The DT struct device_node also has the same tree structure with parent, child, sibling pointers and converting to list_head had been on the todo list for a while. ACPI also has some tree walking functions (drivers/acpi/acpica/pstree.c). Perhaps there should be a common tree struct and helpers defined either on top of list_head or a new struct if that saves some size.
==========

Changelog:
v2->v3:
Rename resource functions first_child() and sibling() to resource_first_chils() and resource_sibling(). Dan suggested this.

Move resource_first_chils() and resource_sibling() to linux/ioport.h and make them as inline function. Rob suggested this. Accordingly add linux/list.h including in linux/ioport.h, please help review if this bring efficiency degradation or code redundancy.

The change on struct resource {} bring two pointers of size increase, mention this in git log to make it more specifically, Rob suggested this.

v1->v2:
Use list_head instead to link resource siblings. This is suggested by Andrew.

Rewrite walk_system_ram_res_rev() after list_head is taken to link resouce siblings.

Baoquan He (3):
resource: Use list_head to link sibling resource
resource: add walk_system_ram_res_rev()
kexec_file: Load kernel at top of system RAM if required

 arch/sparc/kernel/ioport.c                  |   2 +-
 drivers/gpu/drm/drm_memory.c                |   3 +-
 drivers/gpu/drm/gma500/gtt.c                |   5 +-
 drivers/hv/vmbus_drv.c                      |  52 +++----
 drivers/input/joystick/iforce/iforce-main.c |   4 +-
 drivers/nvdimm/e820.c                       |   2 +-
 drivers/nvdimm/namespace_devs.c             |   6 +-
 drivers/nvdimm/nd.h                         |   5 +-
 drivers/of/address.c                        |   4 +-
 drivers/parisc/lba_pci.c                    |   4 +-
 drivers/pci/host/vmd.c                      |   8 +-
 drivers/pci/probe.c                         |   2 +
 drivers/pci/setup-bus.c                     |   2 +-
 include/linux/ioport.h                      |  20 ++-
 kernel/kexec_file.c                         |   2 +
 kernel/resource.c                           | 221 ++++++++++++++++------------
 16 files changed, 193 insertions(+), 149 deletions(-)
Signature · 1 line
2.13.6
author
#1 lore raw

PATCH v3 1/3 resource: Use list_head to link sibling resource

The struct resource uses singly linked list to link siblings. It's not easy to do reverse iteration on sibling list. So replace it with list_head.

And this makes codes in kernel/resource.c more readable after refactoring than pointer operation.

Besides, type of member variables of struct resource, sibling and child, are changed from 'struct resource *' to 'struct list_head'. This brings two pointers of size increase.

Cc Patrik Jakobsson <[email protected]>
Cc David Airlie <[email protected]>
Cc "K. Y. Srinivasan" <[email protected]>
Cc Haiyang Zhang <[email protected]>
Cc Stephen Hemminger <[email protected]>
Cc Dmitry Torokhov <[email protected]>
Cc Dan Williams <[email protected]>
Cc Rob Herring <[email protected]>
Cc Frank Rowand <[email protected]>
Cc Keith Busch <[email protected]>
Cc Jonathan Derrick <[email protected]>
Cc Lorenzo Pieralisi <[email protected]>
Cc Bjorn Helgaas <[email protected]>
Cc Thomas Gleixner <[email protected]>
Cc Brijesh Singh <[email protected]>
Cc "Jérôme Glisse" <[email protected]>
Cc Borislav Petkov <[email protected]>
Cc Tom Lendacky <[email protected]>
Cc Greg Kroah-Hartman <[email protected]>
Cc Yaowei Bai <[email protected]>
Cc Wei Yang <[email protected]>

---
v2->v3:
Rename resource functions first_child() and sibling() to resource_first_chils() and resource_sibling(). Dan suggested this.

Move resource_first_chils() and resource_sibling() to linux/ioport.h and make them as inline function. Rob suggested this. Accordingly add linux/list.h including in linux/ioport.h, please help review if this bring efficiency degradation or code redundancy.

The change on struct resource {} bring two pointers of size increase, mention this in git log to make it more specifically, Rob suggested this.

 arch/sparc/kernel/ioport.c                  |   2 +-
 drivers/gpu/drm/drm_memory.c                |   3 +-
 drivers/gpu/drm/gma500/gtt.c                |   5 +-
 drivers/hv/vmbus_drv.c                      |  52 ++++----
 drivers/input/joystick/iforce/iforce-main.c |   4 +-
 drivers/nvdimm/e820.c                       |   2 +-
 drivers/nvdimm/namespace_devs.c             |   6 +-
 drivers/nvdimm/nd.h                         |   5 +-
 drivers/of/address.c                        |   4 +-
 drivers/parisc/lba_pci.c                    |   4 +-
 drivers/pci/host/vmd.c                      |   8 +-
 drivers/pci/probe.c                         |   2 +
 drivers/pci/setup-bus.c                     |   2 +-
 include/linux/ioport.h                      |  17 ++-
 kernel/resource.c                           | 181 +++++++++++++---------------
 15 files changed, 148 insertions(+), 149 deletions(-)
arch/sparc/kernel/ioport.c +1 -1
index 3bcef9ce74df..4e91fbbbedcc 100644
@@ -669,7 +669,7 @@ static int sparc_io_proc_show(struct seq_file *m, void *v)
 	struct resource *root = m->private, *r;
 	const char *nm;
 
-	for (r = root->child; r != NULL; r = r->sibling) {
+	list_for_each_entry(r, &root->child, sibling) {
 		if ((nm = r->name) == NULL) nm = "???";
 		seq_printf(m, "%016llx-%016llx: %s\n",
 				(unsigned long long)r->start,
drivers/gpu/drm/drm_memory.c +1 -2
index 3c54044214db..53e300a993dc 100644
@@ -155,9 +155,8 @@ u64 drm_get_max_iomem(void)
 	struct resource *tmp;
 	resource_size_t max_iomem = 0;
 
-	for (tmp = iomem_resource.child; tmp; tmp = tmp->sibling) {
+	list_for_each_entry(tmp, &iomem_resource.child, sibling)
 		max_iomem = max(max_iomem,  tmp->end);
-	}
 
 	return max_iomem;
 }
drivers/gpu/drm/gma500/gtt.c +2 -3
index 3949b0990916..addd3bc009af 100644
@@ -565,7 +565,7 @@ int psb_gtt_init(struct drm_device *dev, int resume)
 int psb_gtt_restore(struct drm_device *dev)
 {
 	struct drm_psb_private *dev_priv = dev->dev_private;
-	struct resource *r = dev_priv->gtt_mem->child;
+	struct resource *r;
 	struct gtt_range *range;
 	unsigned int restored = 0, total = 0, size = 0;
 
@@ -573,14 +573,13 @@ int psb_gtt_restore(struct drm_device *dev)
 	mutex_lock(&dev_priv->gtt_mutex);
 	psb_gtt_init(dev, 1);
 
-	while (r != NULL) {
+	list_for_each_entry(r, &dev_priv->gtt_mem->child, sibling) {
 		range = container_of(r, struct gtt_range, resource);
 		if (range->pages) {
 			psb_gtt_insert(dev, range, 1);
 			size += range->resource.end - range->resource.start;
 			restored++;
 		}
-		r = r->sibling;
 		total++;
 	}
 	mutex_unlock(&dev_priv->gtt_mutex);
drivers/hv/vmbus_drv.c +22 -30

99 lines not on this page · raw message

drivers/input/joystick/iforce/iforce-main.c +2 -2
index daeeb4c7e3b0..5c0be27b33ff 100644
@@ -305,8 +305,8 @@ int iforce_init_device(struct iforce *iforce)
 	iforce->device_memory.end = 200;
 	iforce->device_memory.flags = IORESOURCE_MEM;
 	iforce->device_memory.parent = NULL;
-	iforce->device_memory.child = NULL;
-	iforce->device_memory.sibling = NULL;
+	INIT_LIST_HEAD(&iforce->device_memory.child);
+	INIT_LIST_HEAD(&iforce->device_memory.sibling);
 
 /*
  * Wait until device ready - until it sends its first response.
drivers/nvdimm/e820.c +1 -1
index 6f9a6ffd7cde..513e661bb0d8 100644
@@ -53,7 +53,7 @@ static int e820_pmem_probe(struct platform_device *pdev)
 		goto err;
 	platform_set_drvdata(pdev, nvdimm_bus);
 
-	for (p = iomem_resource.child; p ; p = p->sibling) {
+	list_for_each_entry(p, &iomem_resource.child, sibling) {
 		struct nd_region_desc ndr_desc;
 
 		if (p->desc != IORES_DESC_PERSISTENT_MEMORY_LEGACY)
drivers/nvdimm/namespace_devs.c +3 -3
index 28afdd668905..f53d410d9981 100644
@@ -637,7 +637,7 @@ static resource_size_t scan_allocate(struct nd_region *nd_region,
  retry:
 	first = 0;
 	for_each_dpa_resource(ndd, res) {
-		struct resource *next = res->sibling, *new_res = NULL;
+		struct resource *next = resource_sibling(res), *new_res = NULL;
 		resource_size_t allocate, available = 0;
 		enum alloc_loc loc = ALLOC_ERR;
 		const char *action;
@@ -763,7 +763,7 @@ static resource_size_t scan_allocate(struct nd_region *nd_region,
 	 * an initial "pmem-reserve pass".  Only do an initial BLK allocation
 	 * when none of the DPA space is reserved.
 	 */
-	if ((is_pmem || !ndd->dpa.child) && n == to_allocate)
+	if ((is_pmem || list_empty(&ndd->dpa.child)) && n == to_allocate)
 		return init_dpa_allocation(label_id, nd_region, nd_mapping, n);
 	return n;
 }
@@ -779,7 +779,7 @@ static int merge_dpa(struct nd_region *nd_region,
  retry:
 	for_each_dpa_resource(ndd, res) {
 		int rc;
-		struct resource *next = res->sibling;
+		struct resource *next = resource_sibling(res);
 		resource_size_t end = res->start + resource_size(res);
 
 		if (!next || strcmp(res->name, label_id->id) != 0
drivers/nvdimm/nd.h +2 -3
index 32e0364b48b9..da7da15e03e7 100644
@@ -102,11 +102,10 @@ unsigned sizeof_namespace_label(struct nvdimm_drvdata *ndd);
 		(unsigned long long) (res ? res->start : 0), ##arg)
 
 #define for_each_dpa_resource(ndd, res) \
-	for (res = (ndd)->dpa.child; res; res = res->sibling)
+	list_for_each_entry(res, &(ndd)->dpa.child, sibling)
 
 #define for_each_dpa_resource_safe(ndd, res, next) \
-	for (res = (ndd)->dpa.child, next = res ? res->sibling : NULL; \
-			res; res = next, next = next ? next->sibling : NULL)
+	list_for_each_entry_safe(res, next, &(ndd)->dpa.child, sibling)
 
 struct nd_percpu_lane {
 	int count;
drivers/of/address.c +3 -1
index 53349912ac75..e2e25719ab52 100644
@@ -330,7 +330,9 @@ int of_pci_range_to_resource(struct of_pci_range *range,
 {
 	int err;
 	res->flags = range->flags;
-	res->parent = res->child = res->sibling = NULL;
+	res->parent = NULL;
+	INIT_LIST_HEAD(&res->child);
+	INIT_LIST_HEAD(&res->sibling);
 	res->name = np->full_name;
 
 	if (res->flags & IORESOURCE_IO) {
drivers/parisc/lba_pci.c +2 -2
index 69bd98421eb1..7482bdfd1959 100644
@@ -170,8 +170,8 @@ lba_dump_res(struct resource *r, int d)
 	for (i = d; i ; --i) printk(" ");
 	printk(KERN_DEBUG "%p [%lx,%lx]/%lx\n", r,
 		(long)r->start, (long)r->end, r->flags);
-	lba_dump_res(r->child, d+2);
-	lba_dump_res(r->sibling, d);
+	lba_dump_res(resource_first_child(&r->child), d+2);
+	lba_dump_res(resource_sibling(r), d);
 }
 
 
drivers/pci/host/vmd.c +4 -4
index 930a8fa08bd6..c3000af903ea 100644
@@ -520,14 +520,14 @@ static struct pci_ops vmd_ops = {
 
 static void vmd_attach_resources(struct vmd_dev *vmd)
 {
-	vmd->dev->resource[VMD_MEMBAR1].child = &vmd->resources[1];
-	vmd->dev->resource[VMD_MEMBAR2].child = &vmd->resources[2];
+	list_add(&vmd->resources[1].sibling, &vmd->dev->resource[VMD_MEMBAR1].child);
+	list_add(&vmd->resources[2].sibling, &vmd->dev->resource[VMD_MEMBAR2].child);
 }
 
 static void vmd_detach_resources(struct vmd_dev *vmd)
 {
-	vmd->dev->resource[VMD_MEMBAR1].child = NULL;
-	vmd->dev->resource[VMD_MEMBAR2].child = NULL;
+	INIT_LIST_HEAD(&vmd->dev->resource[VMD_MEMBAR1].child);
+	INIT_LIST_HEAD(&vmd->dev->resource[VMD_MEMBAR2].child);
 }
 
 /*
drivers/pci/probe.c +2 -0
index ac91b6fd0bcd..d162c77bec29 100644
@@ -59,6 +59,8 @@ static struct resource *get_pci_domain_busn_res(int domain_nr)
 	r->res.start = 0;
 	r->res.end = 0xff;
 	r->res.flags = IORESOURCE_BUS | IORESOURCE_PCI_FIXED;
+	INIT_LIST_HEAD(&r->res.child);
+	INIT_LIST_HEAD(&r->res.sibling);
 
 	list_add_tail(&r->list, &pci_domain_busn_res_list);
 
drivers/pci/setup-bus.c +1 -1
index 072784f55ea5..0d5e30004ca6 100644
@@ -2107,7 +2107,7 @@ int pci_reassign_bridge_resources(struct pci_dev *bridge, unsigned long type)
 				continue;
 
 			/* Ignore BARs which are still in use */
-			if (res->child)
+			if (!list_empty(&res->child))
 				continue;
 
 			ret = add_to_list(&saved, bridge, res, 0, 0);
include/linux/ioport.h +15 -2
index da0ebaec25f0..225d13d3500a 100644
@@ -12,6 +12,7 @@
 #ifndef __ASSEMBLY__
 #include <linux/compiler.h>
 #include <linux/types.h>
+#include <linux/list.h>
 /*
  * Resources are tree-like, allowing
  * nesting etc..
@@ -22,7 +23,8 @@ struct resource {
 	const char *name;
 	unsigned long flags;
 	unsigned long desc;
-	struct resource *parent, *sibling, *child;
+	struct list_head child, sibling;
+	struct resource *parent;
 };
 
 /*
@@ -215,7 +217,6 @@ static inline bool resource_contains(struct resource *r1, struct resource *r2)
 	return r1->start <= r2->start && r1->end >= r2->end;
 }
 
-
 /* Convenience shorthand with allocation */
 #define request_region(start,n,name)		__request_region(&ioport_resource, (start), (n), (name), 0)
 #define request_muxed_region(start,n,name)	__request_region(&ioport_resource, (start), (n), (name), IORESOURCE_MUXED)
@@ -286,6 +287,18 @@ static inline bool resource_overlaps(struct resource *r1, struct resource *r2)
        return (r1->start <= r2->end && r1->end >= r2->start);
 }
 
+static inline struct resource *resource_sibling(struct resource *res)
+{
+	if (res->parent && !list_is_last(&res->sibling, &res->parent->child))
+		return list_next_entry(res, sibling);
+	return NULL;
+}
+
+static inline struct resource *resource_first_child(struct list_head *head)
+{
+	return list_first_entry_or_null(head, struct resource, sibling);
+}
+
 
 #endif /* __ASSEMBLY__ */
 #endif	/* _LINUX_IOPORT_H */
kernel/resource.c +87 -94

435 lines not on this page · raw message

Signature · 6 lines
2.13.6

_______________________________________________
Linux-nvdimm mailing list
[email protected]
https://lists.01.org/mailman/listinfo/linux-nvdimm
#2 lore raw
Quoted 3 lines · Baoquan He

The struct resource uses singly linked list to link siblings. It's not easy to do reverse iteration on sibling list. So replace it with list_head.

Hi, Baoquan

Besides changing the data structure, I have another proposal to do the reverse iteration. Which means it would not affect other users, if you just want a reverse iteration.

BTW, I don't think Andrew suggest to use linked-list directly. What he wants is a better solution to your first proposal in
https://patchwork.kernel.org/patch/10300819/.

Below is my proposal of resource reverse iteration without changing current design.

Quoted 1 line

From 5d7145d44fe48b98572a03884fa3a3aa82e3cef9 Mon Sep 17 00:00:00 2001

From: Wei Yang <[email protected]>
Date: Sat, 24 Mar 2018 23:25:46 +0800
Subject: [PATCH] kernel/resource: add walk_system_ram_res_rev()

As discussed on https://patchwork.kernel.org/patch/10300819/, this patch comes up with a variant implementation of walk_system_ram_res_rev(), which uses iteration instead of allocating array to store those resources.

---
 include/linux/ioport.h |   3 ++
 kernel/resource.c      | 113 +++++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 116 insertions(+)
include/linux/ioport.h +3 -0
index da0ebaec25f0..473f1d9cb97e 100644
@@ -277,6 +277,9 @@ extern int
 walk_system_ram_res(u64 start, u64 end, void *arg,
 		    int (*func)(struct resource *, void *));
 extern int
+walk_system_ram_res_rev(u64 start, u64 end, void *arg,
+		    int (*func)(struct resource *, void *));
+extern int
 walk_iomem_res_desc(unsigned long desc, unsigned long flags, u64 start, u64 end,
 		    void *arg, int (*func)(struct resource *, void *));
 
kernel/resource.c +113 -0

137 lines not on this page · raw message

Signature · 6 lines
2.15.1


-- 
Wei Yang
Help you, Help me
author
#3 lore raw

Hi Wei Yang,

Quoted 11 lines · Wei Yang
Quoted 3 lines · Baoquan He

The struct resource uses singly linked list to link siblings. It's not easy to do reverse iteration on sibling list. So replace it with list_head.

Hi, Baoquan

Besides changing the data structure, I have another proposal to do the reverse iteration. Which means it would not affect other users, if you just want a reverse iteration.

Quoted 6 lines

BTW, I don't think Andrew suggest to use linked-list directly. What he wants is a better solution to your first proposal in
https://patchwork.kernel.org/patch/10300819/.

Below is my proposal of resource reverse iteration without changing current design.

I got your mail and read it, then interrupted by other thing and forgot replying, sorry.

I am fine with your code change. As I said before, I have tried to change code per reviewers' comment, then let reviewers decide which way is better. Please feel free to post formal patches and joining discussion about this issue.

Thanks
Baoquan

Quoted 182 lines

From 5d7145d44fe48b98572a03884fa3a3aa82e3cef9 Mon Sep 17 00:00:00 2001
From: Wei Yang <[email protected]>
Date: Sat, 24 Mar 2018 23:25:46 +0800
Subject: [PATCH] kernel/resource: add walk_system_ram_res_rev()

As discussed on https://patchwork.kernel.org/patch/10300819/, this patch comes up with a variant implementation of walk_system_ram_res_rev(), which uses iteration instead of allocating array to store those resources.

---
 include/linux/ioport.h |   3 ++
 kernel/resource.c      | 113 +++++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 116 insertions(+)
include/linux/ioport.h +3 -0
index da0ebaec25f0..473f1d9cb97e 100644
@@ -277,6 +277,9 @@ extern int
 walk_system_ram_res(u64 start, u64 end, void *arg,
 		    int (*func)(struct resource *, void *));
 extern int
+walk_system_ram_res_rev(u64 start, u64 end, void *arg,
+		    int (*func)(struct resource *, void *));
+extern int
 walk_iomem_res_desc(unsigned long desc, unsigned long flags, u64 start, u64 end,
 		    void *arg, int (*func)(struct resource *, void *));
 
kernel/resource.c +113 -0
index 769109f20fb7..d4ec5fbc6875 100644
@@ -73,6 +73,38 @@ static struct resource *next_resource(struct resource *p, bool sibling_only)
 	return p->sibling;
 }
 
+static struct resource *prev_resource(struct resource *p, bool sibling_only)
+{
+	struct resource *prev;
+	if (NULL == iomem_resource.child)
+		return NULL;
+
+	if (p == NULL) {
+		prev = iomem_resource.child;
+		while (prev->sibling)
+			prev = prev->sibling;
+	} else {
+		if (p->parent->child == p) {
+			return p->parent;
+		}
+
+		for (prev = p->parent->child; prev->sibling != p;
+			prev = prev->sibling) {}
+	}
+
+	/* Caller wants to traverse through siblings only */
+	if (sibling_only)
+		return prev;
+
+	for (;prev->child;) {
+		prev = prev->child;
+
+		while (prev->sibling)
+			prev = prev->sibling;
+	}
+	return prev;
+}
+
 static void *r_next(struct seq_file *m, void *v, loff_t *pos)
 {
 	struct resource *p = v;
@@ -401,6 +433,47 @@ static int find_next_iomem_res(struct resource *res, unsigned long desc,
 	return 0;
 }
 
+/*
+ * Finds the highest iomem resource existing within [res->start.res->end).
+ * The caller must specify res->start, res->end, res->flags, and optionally
+ * desc.  If found, returns 0, res is overwritten, if not found, returns -1.
+ * This function walks the whole tree and not just first level children until
+ * and unless first_level_children_only is true.
+ */
+static int find_prev_iomem_res(struct resource *res, unsigned long desc,
+			       bool first_level_children_only)
+{
+	struct resource *p;
+
+	BUG_ON(!res);
+	BUG_ON(res->start >= res->end);
+
+	read_lock(&resource_lock);
+
+	for (p = prev_resource(NULL, first_level_children_only); p;
+		p = prev_resource(p, first_level_children_only)) {
+		if ((p->flags & res->flags) != res->flags)
+			continue;
+		if ((desc != IORES_DESC_NONE) && (desc != p->desc))
+			continue;
+		if (p->end < res->start || p->child == iomem_resource.child) {
+			p = NULL;
+			break;
+		}
+		if ((p->end >= res->start) && (p->start < res->end))
+			break;
+	}
+
+	read_unlock(&resource_lock);
+	if (!p)
+		return -1;
+	/* copy data */
+	resource_clip(res, p->start, p->end);
+	res->flags = p->flags;
+	res->desc = p->desc;
+	return 0;
+}
+
 static int __walk_iomem_res_desc(struct resource *res, unsigned long desc,
 				 bool first_level_children_only,
 				 void *arg,
@@ -422,6 +495,27 @@ static int __walk_iomem_res_desc(struct resource *res, unsigned long desc,
 	return ret;
 }
 
+static int __walk_iomem_res_rev_desc(struct resource *res, unsigned long desc,
+				 bool first_level_children_only,
+				 void *arg,
+				 int (*func)(struct resource *, void *))
+{
+	u64 orig_start = res->start;
+	int ret = -1;
+
+	while ((res->start < res->end) &&
+	       !find_prev_iomem_res(res, desc, first_level_children_only)) {
+		ret = (*func)(res, arg);
+		if (ret)
+			break;
+
+		res->end = res->start?(res->start - 1):0;
+		res->start = orig_start;
+	}
+
+	return ret;
+}
+
 /*
  * Walks through iomem resources and calls func() with matching resource
  * ranges. This walks through whole tree and not just first level children.
@@ -468,6 +562,25 @@ int walk_system_ram_res(u64 start, u64 end, void *arg,
 				     arg, func);
 }
 
+/*
+ * This function, being a variant of walk_system_ram_res(), calls the @func
+ * callback against all memory ranges of type System RAM which are marked as
+ * IORESOURCE_SYSTEM_RAM and IORESOUCE_BUSY in reversed order, i.e., from
+ * higher to lower.
+ */
+int walk_system_ram_res_rev(u64 start, u64 end, void *arg,
+				int (*func)(struct resource *, void *))
+{
+	struct resource res;
+
+	res.start = start;
+	res.end = end;
+	res.flags = IORESOURCE_SYSTEM_RAM | IORESOURCE_BUSY;
+
+	return __walk_iomem_res_rev_desc(&res, IORES_DESC_NONE, true,
+				     arg, func);
+}
+
 /*
  * This function calls the @func callback against all memory ranges, which
  * are ranges marked as IORESOURCE_MEM and IORESOUCE_BUSY.
Signature · 6 lines
2.15.1


-- 
Wei Yang
Help you, Help me
#4 lore raw
Quoted 21 lines · Baoquan He

Hi Wei Yang,

Quoted 11 lines · Wei Yang
Quoted 3 lines · Baoquan He

The struct resource uses singly linked list to link siblings. It's not easy to do reverse iteration on sibling list. So replace it with list_head.

Hi, Baoquan

Besides changing the data structure, I have another proposal to do the reverse iteration. Which means it would not affect other users, if you just want a reverse iteration.

Quoted 6 lines

BTW, I don't think Andrew suggest to use linked-list directly. What he wants is a better solution to your first proposal in
https://patchwork.kernel.org/patch/10300819/.

Below is my proposal of resource reverse iteration without changing current design.

Quoted 7 lines

I got your mail and read it, then interrupted by other thing and forgot replying, sorry.

I am fine with your code change. As I said before, I have tried to change code per reviewers' comment, then let reviewers decide which way is better. Please feel free to post formal patches and joining discussion about this issue.

Yep, while I don't have a real requirement to add the reverse version, so what is the proper way to send a patch?

A patch reply to this thread is ok?

Quoted 4 lines

Thanks
Baoquan

author
#5 lore raw
Quoted 30 lines · Wei Yang
Quoted 21 lines · Baoquan He

Hi Wei Yang,

Quoted 11 lines · Wei Yang
Quoted 3 lines · Baoquan He

The struct resource uses singly linked list to link siblings. It's not easy to do reverse iteration on sibling list. So replace it with list_head.

Hi, Baoquan

Besides changing the data structure, I have another proposal to do the reverse iteration. Which means it would not affect other users, if you just want a reverse iteration.

Quoted 6 lines

BTW, I don't think Andrew suggest to use linked-list directly. What he wants is a better solution to your first proposal in
https://patchwork.kernel.org/patch/10300819/.

Below is my proposal of resource reverse iteration without changing current design.

Quoted 7 lines

I got your mail and read it, then interrupted by other thing and forgot replying, sorry.

I am fine with your code change. As I said before, I have tried to change code per reviewers' comment, then let reviewers decide which way is better. Please feel free to post formal patches and joining discussion about this issue.

Quoted 4 lines

Yep, while I don't have a real requirement to add the reverse version, so what is the proper way to send a patch?

A patch reply to this thread is ok?

I am not sure either. Since my patches are still under reviewing. And you have pasted your patch. It depends on maintainers, mainly Andrew and other reviewers who have concerns.

#6 lore raw
Quoted 31 lines · Baoquan He
Quoted 30 lines · Wei Yang
Quoted 28 lines · Baoquan He

Hi Wei Yang,

Quoted 11 lines · Wei Yang
Quoted 3 lines · Baoquan He

The struct resource uses singly linked list to link siblings. It's not easy to do reverse iteration on sibling list. So replace it with list_head.

Hi, Baoquan

Besides changing the data structure, I have another proposal to do the reverse iteration. Which means it would not affect other users, if you just want a reverse iteration.

Quoted 6 lines

BTW, I don't think Andrew suggest to use linked-list directly. What he wants is a better solution to your first proposal in
https://patchwork.kernel.org/patch/10300819/.

Below is my proposal of resource reverse iteration without changing current design.

I got your mail and read it, then interrupted by other thing and forgot replying, sorry.

I am fine with your code change. As I said before, I have tried to change code per reviewers' comment, then let reviewers decide which way is better. Please feel free to post formal patches and joining discussion about this issue.

Quoted 8 lines
Quoted 4 lines

Yep, while I don't have a real requirement to add the reverse version, so what is the proper way to send a patch?

A patch reply to this thread is ok?

I am not sure either. Since my patches are still under reviewing. And you have pasted your patch. It depends on maintainers, mainly Andrew and other reviewers who have concerns.

Ok, thanks.

Signature · 2 lines
Wei Yang
Help you, Help me
#7 lore raw
In reply to Baoquan He
I love your patch! Yet something to improve:

Hi Baoquan,

I love your patch! Yet something to improve:

[auto build test ERROR on linus/master]
[also build test ERROR on v4.17-rc2 next-20180424]
[if your patch is applied to the wrong git tree, please drop us a note to help improve the system]

url: https://github.com/0day-ci/linux/commits/Baoquan-He/resource-Use-list_head-to-link-sibling-resource/20180419-223752
config: microblaze-mmu_defconfig (attached as .config)
compiler: microblaze-linux-gcc (GCC) 7.2.0
reproduce:
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross chmod +x ~/bin/make.cross
# save the attached .config to linux build tree
make.cross ARCH=microblaze

All errors (new ones prefixed by >>):

arch/microblaze/pci/pci-common.c: In function 'pci_process_bridge_OF_ranges':

>> arch/microblaze/pci/pci-common.c:536:44: error: incompatible types when assigning to type 'struct list_head' from type 'void *'
       res->parent = res->child = res->sibling = NULL;
                                               ^
   arch/microblaze/pci/pci-common.c: In function 'reparent_resources':
>> arch/microblaze/pci/pci-common.c:631:10: error: assignment from incompatible pointer type [-Werror=incompatible-pointer-types]
     for (pp = &parent->child; (p = *pp) != NULL; pp = &p->sibling) {
             ^
   arch/microblaze/pci/pci-common.c:631:50: error: assignment from incompatible pointer type [-Werror=incompatible-pointer-types]
     for (pp = &parent->child; (p = *pp) != NULL; pp = &p->sibling) {
                                                     ^
>> arch/microblaze/pci/pci-common.c:644:13: error: incompatible types when assigning to type 'struct list_head' from type 'struct resource *'
     res->child = *firstpp;
                ^
   arch/microblaze/pci/pci-common.c:645:15: error: incompatible types when assigning to type 'struct list_head' from type 'struct resource *'
     res->sibling = *pp;
                  ^
>> arch/microblaze/pci/pci-common.c:648:9: error: incompatible types when assigning to type 'struct resource *' from type 'struct list_head'
     for (p = res->child; p != NULL; p = p->sibling) {
            ^
   arch/microblaze/pci/pci-common.c:648:36: error: incompatible types when assigning to type 'struct resource *' from type 'struct list_head'
     for (p = res->child; p != NULL; p = p->sibling) {
                                       ^

cc1: some warnings being treated as errors

vim +536 arch/microblaze/pci/pci-common.c

d3afa58c Michal Simek        2010-01-18  387  
d3afa58c Michal Simek        2010-01-18  388  /**
d3afa58c Michal Simek        2010-01-18  389   * pci_process_bridge_OF_ranges - Parse PCI bridge resources from device tree
d3afa58c Michal Simek        2010-01-18  390   * @hose: newly allocated pci_controller to be setup
d3afa58c Michal Simek        2010-01-18  391   * @dev: device node of the host bridge
d3afa58c Michal Simek        2010-01-18  392   * @primary: set if primary bus (32 bits only, soon to be deprecated)
d3afa58c Michal Simek        2010-01-18  393   *
d3afa58c Michal Simek        2010-01-18  394   * This function will parse the "ranges" property of a PCI host bridge device
d3afa58c Michal Simek        2010-01-18  395   * node and setup the resource mapping of a pci controller based on its
d3afa58c Michal Simek        2010-01-18  396   * content.
d3afa58c Michal Simek        2010-01-18  397   *
d3afa58c Michal Simek        2010-01-18  398   * Life would be boring if it wasn't for a few issues that we have to deal
d3afa58c Michal Simek        2010-01-18  399   * with here:
d3afa58c Michal Simek        2010-01-18  400   *
d3afa58c Michal Simek        2010-01-18  401   *   - We can only cope with one IO space range and up to 3 Memory space
d3afa58c Michal Simek        2010-01-18  402   *     ranges. However, some machines (thanks Apple !) tend to split their
d3afa58c Michal Simek        2010-01-18  403   *     space into lots of small contiguous ranges. So we have to coalesce.
d3afa58c Michal Simek        2010-01-18  404   *
d3afa58c Michal Simek        2010-01-18  405   *   - We can only cope with all memory ranges having the same offset
d3afa58c Michal Simek        2010-01-18  406   *     between CPU addresses and PCI addresses. Unfortunately, some bridges
d3afa58c Michal Simek        2010-01-18  407   *     are setup for a large 1:1 mapping along with a small "window" which
d3afa58c Michal Simek        2010-01-18  408   *     maps PCI address 0 to some arbitrary high address of the CPU space in
d3afa58c Michal Simek        2010-01-18  409   *     order to give access to the ISA memory hole.
d3afa58c Michal Simek        2010-01-18  410   *     The way out of here that I've chosen for now is to always set the
d3afa58c Michal Simek        2010-01-18  411   *     offset based on the first resource found, then override it if we
d3afa58c Michal Simek        2010-01-18  412   *     have a different offset and the previous was set by an ISA hole.
d3afa58c Michal Simek        2010-01-18  413   *
d3afa58c Michal Simek        2010-01-18  414   *   - Some busses have IO space not starting at 0, which causes trouble with
d3afa58c Michal Simek        2010-01-18  415   *     the way we do our IO resource renumbering. The code somewhat deals with
d3afa58c Michal Simek        2010-01-18  416   *     it for 64 bits but I would expect problems on 32 bits.
d3afa58c Michal Simek        2010-01-18  417   *
d3afa58c Michal Simek        2010-01-18  418   *   - Some 32 bits platforms such as 4xx can have physical space larger than
d3afa58c Michal Simek        2010-01-18  419   *     32 bits so we need to use 64 bits values for the parsing
d3afa58c Michal Simek        2010-01-18  420   */
b881bc46 Greg Kroah-Hartman  2012-12-21  421  void pci_process_bridge_OF_ranges(struct pci_controller *hose,
b881bc46 Greg Kroah-Hartman  2012-12-21  422  				  struct device_node *dev, int primary)
d3afa58c Michal Simek        2010-01-18  423  {
d3afa58c Michal Simek        2010-01-18  424  	int memno = 0, isa_hole = -1;
d3afa58c Michal Simek        2010-01-18  425  	unsigned long long isa_mb = 0;
d3afa58c Michal Simek        2010-01-18  426  	struct resource *res;
4f7b6de4 Andrew Murray       2013-07-27  427  	struct of_pci_range range;
4f7b6de4 Andrew Murray       2013-07-27  428  	struct of_pci_range_parser parser;
d3afa58c Michal Simek        2010-01-18  429  
f2b8ae0e Rob Herring         2017-06-06  430  	pr_info("PCI host bridge %pOF %s ranges:\n",
f2b8ae0e Rob Herring         2017-06-06  431  	       dev, primary ? "(primary)" : "");
d3afa58c Michal Simek        2010-01-18  432  
4f7b6de4 Andrew Murray       2013-07-27  433  	/* Check for ranges property */
4f7b6de4 Andrew Murray       2013-07-27  434  	if (of_pci_range_parser_init(&parser, dev))
d3afa58c Michal Simek        2010-01-18  435  		return;
d3afa58c Michal Simek        2010-01-18  436  
d3afa58c Michal Simek        2010-01-18  437  	pr_debug("Parsing ranges property...\n");
4f7b6de4 Andrew Murray       2013-07-27  438  	for_each_of_pci_range(&parser, &range) {
d3afa58c Michal Simek        2010-01-18  439  		/* Read next ranges element */
6bd55f0b Michal Simek        2012-12-27  440  		pr_debug("pci_space: 0x%08x pci_addr:0x%016llx ",
4f7b6de4 Andrew Murray       2013-07-27  441  				range.pci_space, range.pci_addr);
6bd55f0b Michal Simek        2012-12-27  442  		pr_debug("cpu_addr:0x%016llx size:0x%016llx\n",
4f7b6de4 Andrew Murray       2013-07-27  443  					range.cpu_addr, range.size);
d3afa58c Michal Simek        2010-01-18  444  
d3afa58c Michal Simek        2010-01-18  445  		/* If we failed translation or got a zero-sized region
d3afa58c Michal Simek        2010-01-18  446  		 * (some FW try to feed us with non sensical zero sized regions
d3afa58c Michal Simek        2010-01-18  447  		 * such as power3 which look like some kind of attempt
d3afa58c Michal Simek        2010-01-18  448  		 * at exposing the VGA memory hole)
d3afa58c Michal Simek        2010-01-18  449  		 */
4f7b6de4 Andrew Murray       2013-07-27  450  		if (range.cpu_addr == OF_BAD_ADDR || range.size == 0)
d3afa58c Michal Simek        2010-01-18  451  			continue;
d3afa58c Michal Simek        2010-01-18  452  
d3afa58c Michal Simek        2010-01-18  453  		/* Act based on address space type */
d3afa58c Michal Simek        2010-01-18  454  		res = NULL;
4f7b6de4 Andrew Murray       2013-07-27  455  		switch (range.flags & IORESOURCE_TYPE_BITS) {
4f7b6de4 Andrew Murray       2013-07-27  456  		case IORESOURCE_IO:
6bd55f0b Michal Simek        2012-12-27  457  			pr_info("  IO 0x%016llx..0x%016llx -> 0x%016llx\n",
4f7b6de4 Andrew Murray       2013-07-27  458  				range.cpu_addr, range.cpu_addr + range.size - 1,
4f7b6de4 Andrew Murray       2013-07-27  459  				range.pci_addr);
d3afa58c Michal Simek        2010-01-18  460  
d3afa58c Michal Simek        2010-01-18  461  			/* We support only one IO range */
d3afa58c Michal Simek        2010-01-18  462  			if (hose->pci_io_size) {
6bd55f0b Michal Simek        2012-12-27  463  				pr_info(" \\--> Skipped (too many) !\n");
d3afa58c Michal Simek        2010-01-18  464  				continue;
d3afa58c Michal Simek        2010-01-18  465  			}
d3afa58c Michal Simek        2010-01-18  466  			/* On 32 bits, limit I/O space to 16MB */
4f7b6de4 Andrew Murray       2013-07-27  467  			if (range.size > 0x01000000)
4f7b6de4 Andrew Murray       2013-07-27  468  				range.size = 0x01000000;
d3afa58c Michal Simek        2010-01-18  469  
d3afa58c Michal Simek        2010-01-18  470  			/* 32 bits needs to map IOs here */
4f7b6de4 Andrew Murray       2013-07-27  471  			hose->io_base_virt = ioremap(range.cpu_addr,
4f7b6de4 Andrew Murray       2013-07-27  472  						range.size);
d3afa58c Michal Simek        2010-01-18  473  
d3afa58c Michal Simek        2010-01-18  474  			/* Expect trouble if pci_addr is not 0 */
d3afa58c Michal Simek        2010-01-18  475  			if (primary)
d3afa58c Michal Simek        2010-01-18  476  				isa_io_base =
d3afa58c Michal Simek        2010-01-18  477  					(unsigned long)hose->io_base_virt;
d3afa58c Michal Simek        2010-01-18  478  			/* pci_io_size and io_base_phys always represent IO
d3afa58c Michal Simek        2010-01-18  479  			 * space starting at 0 so we factor in pci_addr
d3afa58c Michal Simek        2010-01-18  480  			 */
4f7b6de4 Andrew Murray       2013-07-27  481  			hose->pci_io_size = range.pci_addr + range.size;
4f7b6de4 Andrew Murray       2013-07-27  482  			hose->io_base_phys = range.cpu_addr - range.pci_addr;
d3afa58c Michal Simek        2010-01-18  483  
d3afa58c Michal Simek        2010-01-18  484  			/* Build resource */
d3afa58c Michal Simek        2010-01-18  485  			res = &hose->io_resource;
4f7b6de4 Andrew Murray       2013-07-27  486  			range.cpu_addr = range.pci_addr;
4f7b6de4 Andrew Murray       2013-07-27  487  
d3afa58c Michal Simek        2010-01-18  488  			break;
4f7b6de4 Andrew Murray       2013-07-27  489  		case IORESOURCE_MEM:
6bd55f0b Michal Simek        2012-12-27  490  			pr_info(" MEM 0x%016llx..0x%016llx -> 0x%016llx %s\n",
4f7b6de4 Andrew Murray       2013-07-27  491  				range.cpu_addr, range.cpu_addr + range.size - 1,
4f7b6de4 Andrew Murray       2013-07-27  492  				range.pci_addr,
4f7b6de4 Andrew Murray       2013-07-27  493  				(range.pci_space & 0x40000000) ?
4f7b6de4 Andrew Murray       2013-07-27  494  				"Prefetch" : "");
d3afa58c Michal Simek        2010-01-18  495  
d3afa58c Michal Simek        2010-01-18  496  			/* We support only 3 memory ranges */
d3afa58c Michal Simek        2010-01-18  497  			if (memno >= 3) {
6bd55f0b Michal Simek        2012-12-27  498  				pr_info(" \\--> Skipped (too many) !\n");
d3afa58c Michal Simek        2010-01-18  499  				continue;
d3afa58c Michal Simek        2010-01-18  500  			}
d3afa58c Michal Simek        2010-01-18  501  			/* Handles ISA memory hole space here */
4f7b6de4 Andrew Murray       2013-07-27  502  			if (range.pci_addr == 0) {
4f7b6de4 Andrew Murray       2013-07-27  503  				isa_mb = range.cpu_addr;
d3afa58c Michal Simek        2010-01-18  504  				isa_hole = memno;
d3afa58c Michal Simek        2010-01-18  505  				if (primary || isa_mem_base == 0)
4f7b6de4 Andrew Murray       2013-07-27  506  					isa_mem_base = range.cpu_addr;
4f7b6de4 Andrew Murray       2013-07-27  507  				hose->isa_mem_phys = range.cpu_addr;
4f7b6de4 Andrew Murray       2013-07-27  508  				hose->isa_mem_size = range.size;
d3afa58c Michal Simek        2010-01-18  509  			}
d3afa58c Michal Simek        2010-01-18  510  
d3afa58c Michal Simek        2010-01-18  511  			/* We get the PCI/Mem offset from the first range or
d3afa58c Michal Simek        2010-01-18  512  			 * the, current one if the offset came from an ISA
d3afa58c Michal Simek        2010-01-18  513  			 * hole. If they don't match, bugger.
d3afa58c Michal Simek        2010-01-18  514  			 */
d3afa58c Michal Simek        2010-01-18  515  			if (memno == 0 ||
4f7b6de4 Andrew Murray       2013-07-27  516  			    (isa_hole >= 0 && range.pci_addr != 0 &&
d3afa58c Michal Simek        2010-01-18  517  			     hose->pci_mem_offset == isa_mb))
4f7b6de4 Andrew Murray       2013-07-27  518  				hose->pci_mem_offset = range.cpu_addr -
4f7b6de4 Andrew Murray       2013-07-27  519  							range.pci_addr;
4f7b6de4 Andrew Murray       2013-07-27  520  			else if (range.pci_addr != 0 &&
4f7b6de4 Andrew Murray       2013-07-27  521  				 hose->pci_mem_offset != range.cpu_addr -
4f7b6de4 Andrew Murray       2013-07-27  522  							range.pci_addr) {
6bd55f0b Michal Simek        2012-12-27  523  				pr_info(" \\--> Skipped (offset mismatch) !\n");
d3afa58c Michal Simek        2010-01-18  524  				continue;
d3afa58c Michal Simek        2010-01-18  525  			}
d3afa58c Michal Simek        2010-01-18  526  
d3afa58c Michal Simek        2010-01-18  527  			/* Build resource */
d3afa58c Michal Simek        2010-01-18  528  			res = &hose->mem_resources[memno++];
d3afa58c Michal Simek        2010-01-18  529  			break;
d3afa58c Michal Simek        2010-01-18  530  		}
70dcd942 Michal Simek        2014-10-27  531  		if (res != NULL) {
70dcd942 Michal Simek        2014-10-27  532  			res->name = dev->full_name;
70dcd942 Michal Simek        2014-10-27  533  			res->flags = range.flags;
70dcd942 Michal Simek        2014-10-27  534  			res->start = range.cpu_addr;
70dcd942 Michal Simek        2014-10-27  535  			res->end = range.cpu_addr + range.size - 1;
70dcd942 Michal Simek        2014-10-27 @536  			res->parent = res->child = res->sibling = NULL;
70dcd942 Michal Simek        2014-10-27  537  		}
d3afa58c Michal Simek        2010-01-18  538  	}
d3afa58c Michal Simek        2010-01-18  539  
d3afa58c Michal Simek        2010-01-18  540  	/* If there's an ISA hole and the pci_mem_offset is -not- matching
d3afa58c Michal Simek        2010-01-18  541  	 * the ISA hole offset, then we need to remove the ISA hole from
d3afa58c Michal Simek        2010-01-18  542  	 * the resource list for that brige
d3afa58c Michal Simek        2010-01-18  543  	 */
d3afa58c Michal Simek        2010-01-18  544  	if (isa_hole >= 0 && hose->pci_mem_offset != isa_mb) {
d3afa58c Michal Simek        2010-01-18  545  		unsigned int next = isa_hole + 1;
6bd55f0b Michal Simek        2012-12-27  546  		pr_info(" Removing ISA hole at 0x%016llx\n", isa_mb);
d3afa58c Michal Simek        2010-01-18  547  		if (next < memno)
d3afa58c Michal Simek        2010-01-18  548  			memmove(&hose->mem_resources[isa_hole],
d3afa58c Michal Simek        2010-01-18  549  				&hose->mem_resources[next],
d3afa58c Michal Simek        2010-01-18  550  				sizeof(struct resource) * (memno - next));
d3afa58c Michal Simek        2010-01-18  551  		hose->mem_resources[--memno].flags = 0;
d3afa58c Michal Simek        2010-01-18  552  	}
d3afa58c Michal Simek        2010-01-18  553  }
d3afa58c Michal Simek        2010-01-18  554  
9413d968 Bharat Kumar Gogada 2016-09-01  555  /* Display the domain number in /proc */
d3afa58c Michal Simek        2010-01-18  556  int pci_proc_domain(struct pci_bus *bus)
d3afa58c Michal Simek        2010-01-18  557  {
9413d968 Bharat Kumar Gogada 2016-09-01  558  	return pci_domain_nr(bus);
d3afa58c Michal Simek        2010-01-18  559  }
d3afa58c Michal Simek        2010-01-18  560  
d3afa58c Michal Simek        2010-01-18  561  /* This header fixup will do the resource fixup for all devices as they are
d3afa58c Michal Simek        2010-01-18  562   * probed, but not for bridge ranges
d3afa58c Michal Simek        2010-01-18  563   */
b881bc46 Greg Kroah-Hartman  2012-12-21  564  static void pcibios_fixup_resources(struct pci_dev *dev)
d3afa58c Michal Simek        2010-01-18  565  {
d3afa58c Michal Simek        2010-01-18  566  	struct pci_controller *hose = pci_bus_to_host(dev->bus);
d3afa58c Michal Simek        2010-01-18  567  	int i;
d3afa58c Michal Simek        2010-01-18  568  
d3afa58c Michal Simek        2010-01-18  569  	if (!hose) {
6bd55f0b Michal Simek        2012-12-27  570  		pr_err("No host bridge for PCI dev %s !\n",
d3afa58c Michal Simek        2010-01-18  571  		       pci_name(dev));
d3afa58c Michal Simek        2010-01-18  572  		return;
d3afa58c Michal Simek        2010-01-18  573  	}
d3afa58c Michal Simek        2010-01-18  574  	for (i = 0; i < DEVICE_COUNT_RESOURCE; i++) {
d3afa58c Michal Simek        2010-01-18  575  		struct resource *res = dev->resource + i;
d3afa58c Michal Simek        2010-01-18  576  		if (!res->flags)
d3afa58c Michal Simek        2010-01-18  577  			continue;
e5b36841 Bjorn Helgaas       2012-02-23  578  		if (res->start == 0) {
6bd55f0b Michal Simek        2012-12-27  579  			pr_debug("PCI:%s Resource %d %016llx-%016llx [%x]",
d3afa58c Michal Simek        2010-01-18  580  				 pci_name(dev), i,
d3afa58c Michal Simek        2010-01-18  581  				 (unsigned long long)res->start,
d3afa58c Michal Simek        2010-01-18  582  				 (unsigned long long)res->end,
d3afa58c Michal Simek        2010-01-18  583  				 (unsigned int)res->flags);
6bd55f0b Michal Simek        2012-12-27  584  			pr_debug("is unassigned\n");
d3afa58c Michal Simek        2010-01-18  585  			res->end -= res->start;
d3afa58c Michal Simek        2010-01-18  586  			res->start = 0;
d3afa58c Michal Simek        2010-01-18  587  			res->flags |= IORESOURCE_UNSET;
d3afa58c Michal Simek        2010-01-18  588  			continue;
d3afa58c Michal Simek        2010-01-18  589  		}
d3afa58c Michal Simek        2010-01-18  590  
aa23bdc0 Bjorn Helgaas       2012-02-23  591  		pr_debug("PCI:%s Resource %d %016llx-%016llx [%x]\n",
d3afa58c Michal Simek        2010-01-18  592  			 pci_name(dev), i,
6bd55f0b Michal Simek        2012-12-27  593  			 (unsigned long long)res->start,
d3afa58c Michal Simek        2010-01-18  594  			 (unsigned long long)res->end,
d3afa58c Michal Simek        2010-01-18  595  			 (unsigned int)res->flags);
d3afa58c Michal Simek        2010-01-18  596  	}
d3afa58c Michal Simek        2010-01-18  597  }
d3afa58c Michal Simek        2010-01-18  598  DECLARE_PCI_FIXUP_HEADER(PCI_ANY_ID, PCI_ANY_ID, pcibios_fixup_resources);
d3afa58c Michal Simek        2010-01-18  599  
d3afa58c Michal Simek        2010-01-18  600  /*
d3afa58c Michal Simek        2010-01-18  601   * We need to avoid collisions with `mirrored' VGA ports
d3afa58c Michal Simek        2010-01-18  602   * and other strange ISA hardware, so we always want the
d3afa58c Michal Simek        2010-01-18  603   * addresses to be allocated in the 0x000-0x0ff region
d3afa58c Michal Simek        2010-01-18  604   * modulo 0x400.
d3afa58c Michal Simek        2010-01-18  605   *
d3afa58c Michal Simek        2010-01-18  606   * Why? Because some silly external IO cards only decode
d3afa58c Michal Simek        2010-01-18  607   * the low 10 bits of the IO address. The 0x00-0xff region
d3afa58c Michal Simek        2010-01-18  608   * is reserved for motherboard devices that decode all 16
d3afa58c Michal Simek        2010-01-18  609   * bits, so it's ok to allocate at, say, 0x2800-0x28ff,
d3afa58c Michal Simek        2010-01-18  610   * but we want to try to avoid allocating at 0x2900-0x2bff
d3afa58c Michal Simek        2010-01-18  611   * which might have be mirrored at 0x0100-0x03ff..
d3afa58c Michal Simek        2010-01-18  612   */
01cf9d52 Bharat Kumar Gogada 2016-02-11  613  int pcibios_add_device(struct pci_dev *dev)
01cf9d52 Bharat Kumar Gogada 2016-02-11  614  {
01cf9d52 Bharat Kumar Gogada 2016-02-11  615  	dev->irq = of_irq_parse_and_map_pci(dev, 0, 0);
c86fac43 Michal Simek        2010-04-16  616  
01cf9d52 Bharat Kumar Gogada 2016-02-11  617  	return 0;
d3afa58c Michal Simek        2010-01-18  618  }
01cf9d52 Bharat Kumar Gogada 2016-02-11  619  EXPORT_SYMBOL(pcibios_add_device);
d3afa58c Michal Simek        2010-01-18  620  
d3afa58c Michal Simek        2010-01-18  621  /*
d3afa58c Michal Simek        2010-01-18  622   * Reparent resource children of pr that conflict with res
d3afa58c Michal Simek        2010-01-18  623   * under res, and make res replace those children.
d3afa58c Michal Simek        2010-01-18  624   */
d3afa58c Michal Simek        2010-01-18  625  static int __init reparent_resources(struct resource *parent,
d3afa58c Michal Simek        2010-01-18  626  				     struct resource *res)
d3afa58c Michal Simek        2010-01-18  627  {
d3afa58c Michal Simek        2010-01-18  628  	struct resource *p, **pp;
d3afa58c Michal Simek        2010-01-18  629  	struct resource **firstpp = NULL;
d3afa58c Michal Simek        2010-01-18  630  
d3afa58c Michal Simek        2010-01-18 @631  	for (pp = &parent->child; (p = *pp) != NULL; pp = &p->sibling) {
d3afa58c Michal Simek        2010-01-18  632  		if (p->end < res->start)
d3afa58c Michal Simek        2010-01-18  633  			continue;
d3afa58c Michal Simek        2010-01-18  634  		if (res->end < p->start)
d3afa58c Michal Simek        2010-01-18  635  			break;
d3afa58c Michal Simek        2010-01-18  636  		if (p->start < res->start || p->end > res->end)
d3afa58c Michal Simek        2010-01-18  637  			return -1;	/* not completely contained */
d3afa58c Michal Simek        2010-01-18  638  		if (firstpp == NULL)
d3afa58c Michal Simek        2010-01-18  639  			firstpp = pp;
d3afa58c Michal Simek        2010-01-18  640  	}
d3afa58c Michal Simek        2010-01-18  641  	if (firstpp == NULL)
d3afa58c Michal Simek        2010-01-18  642  		return -1;	/* didn't find any conflicting entries? */
d3afa58c Michal Simek        2010-01-18  643  	res->parent = parent;
d3afa58c Michal Simek        2010-01-18 @644  	res->child = *firstpp;
d3afa58c Michal Simek        2010-01-18  645  	res->sibling = *pp;
d3afa58c Michal Simek        2010-01-18  646  	*firstpp = res;
d3afa58c Michal Simek        2010-01-18  647  	*pp = NULL;
d3afa58c Michal Simek        2010-01-18 @648  	for (p = res->child; p != NULL; p = p->sibling) {
d3afa58c Michal Simek        2010-01-18  649  		p->parent = res;
d3afa58c Michal Simek        2010-01-18  650  		pr_debug("PCI: Reparented %s [%llx..%llx] under %s\n",
d3afa58c Michal Simek        2010-01-18  651  			 p->name,
d3afa58c Michal Simek        2010-01-18  652  			 (unsigned long long)p->start,
d3afa58c Michal Simek        2010-01-18  653  			 (unsigned long long)p->end, res->name);
d3afa58c Michal Simek        2010-01-18  654  	}
d3afa58c Michal Simek        2010-01-18  655  	return 0;
d3afa58c Michal Simek        2010-01-18  656  }
d3afa58c Michal Simek        2010-01-18  657  

:::::: The code at line 536 was first introduced by commit
:::::: 70dcd942dc4af3cc6c3dcc2ba499cd841c7f65a7 microblaze: Fix IO space breakage after of_pci_range_to_resource() change

:::::: TO: Michal Simek <[email protected]> :::::: CC: Michal Simek <[email protected]>

---
0-DAY kernel test infrastructure Open Source Technology Center
https://lists.01.org/pipermail/kbuild-all Intel Corporation

author
#8 lore raw
Quoted 17 lines · kbuild test robot

Hi Baoquan,

I love your patch! Yet something to improve:

[auto build test ERROR on linus/master]
[also build test ERROR on v4.17-rc2 next-20180424]
[if your patch is applied to the wrong git tree, please drop us a note to help improve the system]

url: https://github.com/0day-ci/linux/commits/Baoquan-He/resource-Use-list_head-to-link-sibling-resource/20180419-223752
config: microblaze-mmu_defconfig (attached as .config)
compiler: microblaze-linux-gcc (GCC) 7.2.0
reproduce:
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross chmod +x ~/bin/make.cross
# save the attached .config to linux build tree
make.cross ARCH=microblaze

Quoted 1 line

All errors (new ones prefixed by >>):

Thanks, below patch can fix it. Will repost including the fix.

arch/microblaze/pci/pci-common.c +18 -13
index 161f9758c631..56d189cb4be4 100644
@@ -533,7 +533,9 @@ void pci_process_bridge_OF_ranges(struct pci_controller *hose,
 			res->flags = range.flags;
 			res->start = range.cpu_addr;
 			res->end = range.cpu_addr + range.size - 1;
-			res->parent = res->child = res->sibling = NULL;
+			res->parent = NULL;
+			INIT_LIST_HEAD(&res->child);
+			INIT_LIST_HEAD(&res->sibling);
 		}
 	}
 
@@ -625,28 +627,31 @@ EXPORT_SYMBOL(pcibios_add_device);
 static int __init reparent_resources(struct resource *parent,
 				     struct resource *res)
 {
-	struct resource *p, **pp;
-	struct resource **firstpp = NULL;
+	struct resource *p, *first = NULL;
 
-	for (pp = &parent->child; (p = *pp) != NULL; pp = &p->sibling) {
+	list_for_each_entry(p, &parent->child, sibling) {
 		if (p->end < res->start)
 			continue;
 		if (res->end < p->start)
 			break;
 		if (p->start < res->start || p->end > res->end)
 			return -1;	/* not completely contained */
-		if (firstpp == NULL)
-			firstpp = pp;
+		if (first == NULL)
+			first = p;
 	}
-	if (firstpp == NULL)
+	if (first == NULL)
 		return -1;	/* didn't find any conflicting entries? */
 	res->parent = parent;
-	res->child = *firstpp;
-	res->sibling = *pp;
-	*firstpp = res;
-	*pp = NULL;
-	for (p = res->child; p != NULL; p = p->sibling) {
-		p->parent = res;
+	list_add(&res->sibling, &p->sibling.prev);
+	INIT_LIST_HEAD(&res->child);
+
+	/*
+	 * From first to p's previous sibling, they all fall into
+	 * res's region, change them as res's children.
+	 */
+	list_cut_position(&res->child, first->sibling.prev, res->sibling.prev);
+	list_for_each_entry(p, &new->child, sibling) {
+                p->parent = new;
 		pr_debug("PCI: Reparented %s [%llx..%llx] under %s\n",
 			 p->name,
 			 (unsigned long long)p->start,
Quoted 309 lines

arch/microblaze/pci/pci-common.c: In function 'pci_process_bridge_OF_ranges':

Quoted 1 line
Quoted 1 line

arch/microblaze/pci/pci-common.c:536:44: error: incompatible types when assigning to type 'struct list_head' from type 'void *'

       res->parent = res->child = res->sibling = NULL;
                                               ^

arch/microblaze/pci/pci-common.c: In function 'reparent_resources':

Quoted 1 line
Quoted 1 line

arch/microblaze/pci/pci-common.c:631:10: error: assignment from incompatible pointer type [-Werror=incompatible-pointer-types]

     for (pp = &parent->child; (p = *pp) != NULL; pp = &p->sibling) {
             ^
   arch/microblaze/pci/pci-common.c:631:50: error: assignment from incompatible pointer type [-Werror=incompatible-pointer-types]
     for (pp = &parent->child; (p = *pp) != NULL; pp = &p->sibling) {
                                                     ^
Quoted 1 line
Quoted 1 line

arch/microblaze/pci/pci-common.c:644:13: error: incompatible types when assigning to type 'struct list_head' from type 'struct resource *'

     res->child = *firstpp;
                ^
   arch/microblaze/pci/pci-common.c:645:15: error: incompatible types when assigning to type 'struct list_head' from type 'struct resource *'
     res->sibling = *pp;
                  ^
Quoted 1 line
Quoted 1 line

arch/microblaze/pci/pci-common.c:648:9: error: incompatible types when assigning to type 'struct resource *' from type 'struct list_head'

     for (p = res->child; p != NULL; p = p->sibling) {
            ^
   arch/microblaze/pci/pci-common.c:648:36: error: incompatible types when assigning to type 'struct resource *' from type 'struct list_head'
     for (p = res->child; p != NULL; p = p->sibling) {
                                       ^

cc1: some warnings being treated as errors

vim +536 arch/microblaze/pci/pci-common.c

d3afa58c Michal Simek        2010-01-18  387  
d3afa58c Michal Simek        2010-01-18  388  /**
d3afa58c Michal Simek        2010-01-18  389   * pci_process_bridge_OF_ranges - Parse PCI bridge resources from device tree
d3afa58c Michal Simek        2010-01-18  390   * @hose: newly allocated pci_controller to be setup
d3afa58c Michal Simek        2010-01-18  391   * @dev: device node of the host bridge
d3afa58c Michal Simek        2010-01-18  392   * @primary: set if primary bus (32 bits only, soon to be deprecated)
d3afa58c Michal Simek        2010-01-18  393   *
d3afa58c Michal Simek        2010-01-18  394   * This function will parse the "ranges" property of a PCI host bridge device
d3afa58c Michal Simek        2010-01-18  395   * node and setup the resource mapping of a pci controller based on its
d3afa58c Michal Simek        2010-01-18  396   * content.
d3afa58c Michal Simek        2010-01-18  397   *
d3afa58c Michal Simek        2010-01-18  398   * Life would be boring if it wasn't for a few issues that we have to deal
d3afa58c Michal Simek        2010-01-18  399   * with here:
d3afa58c Michal Simek        2010-01-18  400   *
d3afa58c Michal Simek        2010-01-18  401   *   - We can only cope with one IO space range and up to 3 Memory space
d3afa58c Michal Simek        2010-01-18  402   *     ranges. However, some machines (thanks Apple !) tend to split their
d3afa58c Michal Simek        2010-01-18  403   *     space into lots of small contiguous ranges. So we have to coalesce.
d3afa58c Michal Simek        2010-01-18  404   *
d3afa58c Michal Simek        2010-01-18  405   *   - We can only cope with all memory ranges having the same offset
d3afa58c Michal Simek        2010-01-18  406   *     between CPU addresses and PCI addresses. Unfortunately, some bridges
d3afa58c Michal Simek        2010-01-18  407   *     are setup for a large 1:1 mapping along with a small "window" which
d3afa58c Michal Simek        2010-01-18  408   *     maps PCI address 0 to some arbitrary high address of the CPU space in
d3afa58c Michal Simek        2010-01-18  409   *     order to give access to the ISA memory hole.
d3afa58c Michal Simek        2010-01-18  410   *     The way out of here that I've chosen for now is to always set the
d3afa58c Michal Simek        2010-01-18  411   *     offset based on the first resource found, then override it if we
d3afa58c Michal Simek        2010-01-18  412   *     have a different offset and the previous was set by an ISA hole.
d3afa58c Michal Simek        2010-01-18  413   *
d3afa58c Michal Simek        2010-01-18  414   *   - Some busses have IO space not starting at 0, which causes trouble with
d3afa58c Michal Simek        2010-01-18  415   *     the way we do our IO resource renumbering. The code somewhat deals with
d3afa58c Michal Simek        2010-01-18  416   *     it for 64 bits but I would expect problems on 32 bits.
d3afa58c Michal Simek        2010-01-18  417   *
d3afa58c Michal Simek        2010-01-18  418   *   - Some 32 bits platforms such as 4xx can have physical space larger than
d3afa58c Michal Simek        2010-01-18  419   *     32 bits so we need to use 64 bits values for the parsing
d3afa58c Michal Simek        2010-01-18  420   */
b881bc46 Greg Kroah-Hartman  2012-12-21  421  void pci_process_bridge_OF_ranges(struct pci_controller *hose,
b881bc46 Greg Kroah-Hartman  2012-12-21  422  				  struct device_node *dev, int primary)
d3afa58c Michal Simek        2010-01-18  423  {
d3afa58c Michal Simek        2010-01-18  424  	int memno = 0, isa_hole = -1;
d3afa58c Michal Simek        2010-01-18  425  	unsigned long long isa_mb = 0;
d3afa58c Michal Simek        2010-01-18  426  	struct resource *res;
4f7b6de4 Andrew Murray       2013-07-27  427  	struct of_pci_range range;
4f7b6de4 Andrew Murray       2013-07-27  428  	struct of_pci_range_parser parser;
d3afa58c Michal Simek        2010-01-18  429  
f2b8ae0e Rob Herring         2017-06-06  430  	pr_info("PCI host bridge %pOF %s ranges:\n",
f2b8ae0e Rob Herring         2017-06-06  431  	       dev, primary ? "(primary)" : "");
d3afa58c Michal Simek        2010-01-18  432  
4f7b6de4 Andrew Murray       2013-07-27  433  	/* Check for ranges property */
4f7b6de4 Andrew Murray       2013-07-27  434  	if (of_pci_range_parser_init(&parser, dev))
d3afa58c Michal Simek        2010-01-18  435  		return;
d3afa58c Michal Simek        2010-01-18  436  
d3afa58c Michal Simek        2010-01-18  437  	pr_debug("Parsing ranges property...\n");
4f7b6de4 Andrew Murray       2013-07-27  438  	for_each_of_pci_range(&parser, &range) {
d3afa58c Michal Simek        2010-01-18  439  		/* Read next ranges element */
6bd55f0b Michal Simek        2012-12-27  440  		pr_debug("pci_space: 0x%08x pci_addr:0x%016llx ",
4f7b6de4 Andrew Murray       2013-07-27  441  				range.pci_space, range.pci_addr);
6bd55f0b Michal Simek        2012-12-27  442  		pr_debug("cpu_addr:0x%016llx size:0x%016llx\n",
4f7b6de4 Andrew Murray       2013-07-27  443  					range.cpu_addr, range.size);
d3afa58c Michal Simek        2010-01-18  444  
d3afa58c Michal Simek        2010-01-18  445  		/* If we failed translation or got a zero-sized region
d3afa58c Michal Simek        2010-01-18  446  		 * (some FW try to feed us with non sensical zero sized regions
d3afa58c Michal Simek        2010-01-18  447  		 * such as power3 which look like some kind of attempt
d3afa58c Michal Simek        2010-01-18  448  		 * at exposing the VGA memory hole)
d3afa58c Michal Simek        2010-01-18  449  		 */
4f7b6de4 Andrew Murray       2013-07-27  450  		if (range.cpu_addr == OF_BAD_ADDR || range.size == 0)
d3afa58c Michal Simek        2010-01-18  451  			continue;
d3afa58c Michal Simek        2010-01-18  452  
d3afa58c Michal Simek        2010-01-18  453  		/* Act based on address space type */
d3afa58c Michal Simek        2010-01-18  454  		res = NULL;
4f7b6de4 Andrew Murray       2013-07-27  455  		switch (range.flags & IORESOURCE_TYPE_BITS) {
4f7b6de4 Andrew Murray       2013-07-27  456  		case IORESOURCE_IO:
6bd55f0b Michal Simek        2012-12-27  457  			pr_info("  IO 0x%016llx..0x%016llx -> 0x%016llx\n",
4f7b6de4 Andrew Murray       2013-07-27  458  				range.cpu_addr, range.cpu_addr + range.size - 1,
4f7b6de4 Andrew Murray       2013-07-27  459  				range.pci_addr);
d3afa58c Michal Simek        2010-01-18  460  
d3afa58c Michal Simek        2010-01-18  461  			/* We support only one IO range */
d3afa58c Michal Simek        2010-01-18  462  			if (hose->pci_io_size) {
6bd55f0b Michal Simek        2012-12-27  463  				pr_info(" \\--> Skipped (too many) !\n");
d3afa58c Michal Simek        2010-01-18  464  				continue;
d3afa58c Michal Simek        2010-01-18  465  			}
d3afa58c Michal Simek        2010-01-18  466  			/* On 32 bits, limit I/O space to 16MB */
4f7b6de4 Andrew Murray       2013-07-27  467  			if (range.size > 0x01000000)
4f7b6de4 Andrew Murray       2013-07-27  468  				range.size = 0x01000000;
d3afa58c Michal Simek        2010-01-18  469  
d3afa58c Michal Simek        2010-01-18  470  			/* 32 bits needs to map IOs here */
4f7b6de4 Andrew Murray       2013-07-27  471  			hose->io_base_virt = ioremap(range.cpu_addr,
4f7b6de4 Andrew Murray       2013-07-27  472  						range.size);
d3afa58c Michal Simek        2010-01-18  473  
d3afa58c Michal Simek        2010-01-18  474  			/* Expect trouble if pci_addr is not 0 */
d3afa58c Michal Simek        2010-01-18  475  			if (primary)
d3afa58c Michal Simek        2010-01-18  476  				isa_io_base =
d3afa58c Michal Simek        2010-01-18  477  					(unsigned long)hose->io_base_virt;
d3afa58c Michal Simek        2010-01-18  478  			/* pci_io_size and io_base_phys always represent IO
d3afa58c Michal Simek        2010-01-18  479  			 * space starting at 0 so we factor in pci_addr
d3afa58c Michal Simek        2010-01-18  480  			 */
4f7b6de4 Andrew Murray       2013-07-27  481  			hose->pci_io_size = range.pci_addr + range.size;
4f7b6de4 Andrew Murray       2013-07-27  482  			hose->io_base_phys = range.cpu_addr - range.pci_addr;
d3afa58c Michal Simek        2010-01-18  483  
d3afa58c Michal Simek        2010-01-18  484  			/* Build resource */
d3afa58c Michal Simek        2010-01-18  485  			res = &hose->io_resource;
4f7b6de4 Andrew Murray       2013-07-27  486  			range.cpu_addr = range.pci_addr;
4f7b6de4 Andrew Murray       2013-07-27  487  
d3afa58c Michal Simek        2010-01-18  488  			break;
4f7b6de4 Andrew Murray       2013-07-27  489  		case IORESOURCE_MEM:
6bd55f0b Michal Simek        2012-12-27  490  			pr_info(" MEM 0x%016llx..0x%016llx -> 0x%016llx %s\n",
4f7b6de4 Andrew Murray       2013-07-27  491  				range.cpu_addr, range.cpu_addr + range.size - 1,
4f7b6de4 Andrew Murray       2013-07-27  492  				range.pci_addr,
4f7b6de4 Andrew Murray       2013-07-27  493  				(range.pci_space & 0x40000000) ?
4f7b6de4 Andrew Murray       2013-07-27  494  				"Prefetch" : "");
d3afa58c Michal Simek        2010-01-18  495  
d3afa58c Michal Simek        2010-01-18  496  			/* We support only 3 memory ranges */
d3afa58c Michal Simek        2010-01-18  497  			if (memno >= 3) {
6bd55f0b Michal Simek        2012-12-27  498  				pr_info(" \\--> Skipped (too many) !\n");
d3afa58c Michal Simek        2010-01-18  499  				continue;
d3afa58c Michal Simek        2010-01-18  500  			}
d3afa58c Michal Simek        2010-01-18  501  			/* Handles ISA memory hole space here */
4f7b6de4 Andrew Murray       2013-07-27  502  			if (range.pci_addr == 0) {
4f7b6de4 Andrew Murray       2013-07-27  503  				isa_mb = range.cpu_addr;
d3afa58c Michal Simek        2010-01-18  504  				isa_hole = memno;
d3afa58c Michal Simek        2010-01-18  505  				if (primary || isa_mem_base == 0)
4f7b6de4 Andrew Murray       2013-07-27  506  					isa_mem_base = range.cpu_addr;
4f7b6de4 Andrew Murray       2013-07-27  507  				hose->isa_mem_phys = range.cpu_addr;
4f7b6de4 Andrew Murray       2013-07-27  508  				hose->isa_mem_size = range.size;
d3afa58c Michal Simek        2010-01-18  509  			}
d3afa58c Michal Simek        2010-01-18  510  
d3afa58c Michal Simek        2010-01-18  511  			/* We get the PCI/Mem offset from the first range or
d3afa58c Michal Simek        2010-01-18  512  			 * the, current one if the offset came from an ISA
d3afa58c Michal Simek        2010-01-18  513  			 * hole. If they don't match, bugger.
d3afa58c Michal Simek        2010-01-18  514  			 */
d3afa58c Michal Simek        2010-01-18  515  			if (memno == 0 ||
4f7b6de4 Andrew Murray       2013-07-27  516  			    (isa_hole >= 0 && range.pci_addr != 0 &&
d3afa58c Michal Simek        2010-01-18  517  			     hose->pci_mem_offset == isa_mb))
4f7b6de4 Andrew Murray       2013-07-27  518  				hose->pci_mem_offset = range.cpu_addr -
4f7b6de4 Andrew Murray       2013-07-27  519  							range.pci_addr;
4f7b6de4 Andrew Murray       2013-07-27  520  			else if (range.pci_addr != 0 &&
4f7b6de4 Andrew Murray       2013-07-27  521  				 hose->pci_mem_offset != range.cpu_addr -
4f7b6de4 Andrew Murray       2013-07-27  522  							range.pci_addr) {
6bd55f0b Michal Simek        2012-12-27  523  				pr_info(" \\--> Skipped (offset mismatch) !\n");
d3afa58c Michal Simek        2010-01-18  524  				continue;
d3afa58c Michal Simek        2010-01-18  525  			}
d3afa58c Michal Simek        2010-01-18  526  
d3afa58c Michal Simek        2010-01-18  527  			/* Build resource */
d3afa58c Michal Simek        2010-01-18  528  			res = &hose->mem_resources[memno++];
d3afa58c Michal Simek        2010-01-18  529  			break;
d3afa58c Michal Simek        2010-01-18  530  		}
70dcd942 Michal Simek        2014-10-27  531  		if (res != NULL) {
70dcd942 Michal Simek        2014-10-27  532  			res->name = dev->full_name;
70dcd942 Michal Simek        2014-10-27  533  			res->flags = range.flags;
70dcd942 Michal Simek        2014-10-27  534  			res->start = range.cpu_addr;
70dcd942 Michal Simek        2014-10-27  535  			res->end = range.cpu_addr + range.size - 1;
70dcd942 Michal Simek        2014-10-27 @536  			res->parent = res->child = res->sibling = NULL;
70dcd942 Michal Simek        2014-10-27  537  		}
d3afa58c Michal Simek        2010-01-18  538  	}
d3afa58c Michal Simek        2010-01-18  539  
d3afa58c Michal Simek        2010-01-18  540  	/* If there's an ISA hole and the pci_mem_offset is -not- matching
d3afa58c Michal Simek        2010-01-18  541  	 * the ISA hole offset, then we need to remove the ISA hole from
d3afa58c Michal Simek        2010-01-18  542  	 * the resource list for that brige
d3afa58c Michal Simek        2010-01-18  543  	 */
d3afa58c Michal Simek        2010-01-18  544  	if (isa_hole >= 0 && hose->pci_mem_offset != isa_mb) {
d3afa58c Michal Simek        2010-01-18  545  		unsigned int next = isa_hole + 1;
6bd55f0b Michal Simek        2012-12-27  546  		pr_info(" Removing ISA hole at 0x%016llx\n", isa_mb);
d3afa58c Michal Simek        2010-01-18  547  		if (next < memno)
d3afa58c Michal Simek        2010-01-18  548  			memmove(&hose->mem_resources[isa_hole],
d3afa58c Michal Simek        2010-01-18  549  				&hose->mem_resources[next],
d3afa58c Michal Simek        2010-01-18  550  				sizeof(struct resource) * (memno - next));
d3afa58c Michal Simek        2010-01-18  551  		hose->mem_resources[--memno].flags = 0;
d3afa58c Michal Simek        2010-01-18  552  	}
d3afa58c Michal Simek        2010-01-18  553  }
d3afa58c Michal Simek        2010-01-18  554  
9413d968 Bharat Kumar Gogada 2016-09-01  555  /* Display the domain number in /proc */
d3afa58c Michal Simek        2010-01-18  556  int pci_proc_domain(struct pci_bus *bus)
d3afa58c Michal Simek        2010-01-18  557  {
9413d968 Bharat Kumar Gogada 2016-09-01  558  	return pci_domain_nr(bus);
d3afa58c Michal Simek        2010-01-18  559  }
d3afa58c Michal Simek        2010-01-18  560  
d3afa58c Michal Simek        2010-01-18  561  /* This header fixup will do the resource fixup for all devices as they are
d3afa58c Michal Simek        2010-01-18  562   * probed, but not for bridge ranges
d3afa58c Michal Simek        2010-01-18  563   */
b881bc46 Greg Kroah-Hartman  2012-12-21  564  static void pcibios_fixup_resources(struct pci_dev *dev)
d3afa58c Michal Simek        2010-01-18  565  {
d3afa58c Michal Simek        2010-01-18  566  	struct pci_controller *hose = pci_bus_to_host(dev->bus);
d3afa58c Michal Simek        2010-01-18  567  	int i;
d3afa58c Michal Simek        2010-01-18  568  
d3afa58c Michal Simek        2010-01-18  569  	if (!hose) {
6bd55f0b Michal Simek        2012-12-27  570  		pr_err("No host bridge for PCI dev %s !\n",
d3afa58c Michal Simek        2010-01-18  571  		       pci_name(dev));
d3afa58c Michal Simek        2010-01-18  572  		return;
d3afa58c Michal Simek        2010-01-18  573  	}
d3afa58c Michal Simek        2010-01-18  574  	for (i = 0; i < DEVICE_COUNT_RESOURCE; i++) {
d3afa58c Michal Simek        2010-01-18  575  		struct resource *res = dev->resource + i;
d3afa58c Michal Simek        2010-01-18  576  		if (!res->flags)
d3afa58c Michal Simek        2010-01-18  577  			continue;
e5b36841 Bjorn Helgaas       2012-02-23  578  		if (res->start == 0) {
6bd55f0b Michal Simek        2012-12-27  579  			pr_debug("PCI:%s Resource %d %016llx-%016llx [%x]",
d3afa58c Michal Simek        2010-01-18  580  				 pci_name(dev), i,
d3afa58c Michal Simek        2010-01-18  581  				 (unsigned long long)res->start,
d3afa58c Michal Simek        2010-01-18  582  				 (unsigned long long)res->end,
d3afa58c Michal Simek        2010-01-18  583  				 (unsigned int)res->flags);
6bd55f0b Michal Simek        2012-12-27  584  			pr_debug("is unassigned\n");
d3afa58c Michal Simek        2010-01-18  585  			res->end -= res->start;
d3afa58c Michal Simek        2010-01-18  586  			res->start = 0;
d3afa58c Michal Simek        2010-01-18  587  			res->flags |= IORESOURCE_UNSET;
d3afa58c Michal Simek        2010-01-18  588  			continue;
d3afa58c Michal Simek        2010-01-18  589  		}
d3afa58c Michal Simek        2010-01-18  590  
aa23bdc0 Bjorn Helgaas       2012-02-23  591  		pr_debug("PCI:%s Resource %d %016llx-%016llx [%x]\n",
d3afa58c Michal Simek        2010-01-18  592  			 pci_name(dev), i,
6bd55f0b Michal Simek        2012-12-27  593  			 (unsigned long long)res->start,
d3afa58c Michal Simek        2010-01-18  594  			 (unsigned long long)res->end,
d3afa58c Michal Simek        2010-01-18  595  			 (unsigned int)res->flags);
d3afa58c Michal Simek        2010-01-18  596  	}
d3afa58c Michal Simek        2010-01-18  597  }
d3afa58c Michal Simek        2010-01-18  598  DECLARE_PCI_FIXUP_HEADER(PCI_ANY_ID, PCI_ANY_ID, pcibios_fixup_resources);
d3afa58c Michal Simek        2010-01-18  599  
d3afa58c Michal Simek        2010-01-18  600  /*
d3afa58c Michal Simek        2010-01-18  601   * We need to avoid collisions with `mirrored' VGA ports
d3afa58c Michal Simek        2010-01-18  602   * and other strange ISA hardware, so we always want the
d3afa58c Michal Simek        2010-01-18  603   * addresses to be allocated in the 0x000-0x0ff region
d3afa58c Michal Simek        2010-01-18  604   * modulo 0x400.
d3afa58c Michal Simek        2010-01-18  605   *
d3afa58c Michal Simek        2010-01-18  606   * Why? Because some silly external IO cards only decode
d3afa58c Michal Simek        2010-01-18  607   * the low 10 bits of the IO address. The 0x00-0xff region
d3afa58c Michal Simek        2010-01-18  608   * is reserved for motherboard devices that decode all 16
d3afa58c Michal Simek        2010-01-18  609   * bits, so it's ok to allocate at, say, 0x2800-0x28ff,
d3afa58c Michal Simek        2010-01-18  610   * but we want to try to avoid allocating at 0x2900-0x2bff
d3afa58c Michal Simek        2010-01-18  611   * which might have be mirrored at 0x0100-0x03ff..
d3afa58c Michal Simek        2010-01-18  612   */
01cf9d52 Bharat Kumar Gogada 2016-02-11  613  int pcibios_add_device(struct pci_dev *dev)
01cf9d52 Bharat Kumar Gogada 2016-02-11  614  {
01cf9d52 Bharat Kumar Gogada 2016-02-11  615  	dev->irq = of_irq_parse_and_map_pci(dev, 0, 0);
c86fac43 Michal Simek        2010-04-16  616  
01cf9d52 Bharat Kumar Gogada 2016-02-11  617  	return 0;
d3afa58c Michal Simek        2010-01-18  618  }
01cf9d52 Bharat Kumar Gogada 2016-02-11  619  EXPORT_SYMBOL(pcibios_add_device);
d3afa58c Michal Simek        2010-01-18  620  
d3afa58c Michal Simek        2010-01-18  621  /*
d3afa58c Michal Simek        2010-01-18  622   * Reparent resource children of pr that conflict with res
d3afa58c Michal Simek        2010-01-18  623   * under res, and make res replace those children.
d3afa58c Michal Simek        2010-01-18  624   */
d3afa58c Michal Simek        2010-01-18  625  static int __init reparent_resources(struct resource *parent,
d3afa58c Michal Simek        2010-01-18  626  				     struct resource *res)
d3afa58c Michal Simek        2010-01-18  627  {
d3afa58c Michal Simek        2010-01-18  628  	struct resource *p, **pp;
d3afa58c Michal Simek        2010-01-18  629  	struct resource **firstpp = NULL;
d3afa58c Michal Simek        2010-01-18  630  
d3afa58c Michal Simek        2010-01-18 @631  	for (pp = &parent->child; (p = *pp) != NULL; pp = &p->sibling) {
d3afa58c Michal Simek        2010-01-18  632  		if (p->end < res->start)
d3afa58c Michal Simek        2010-01-18  633  			continue;
d3afa58c Michal Simek        2010-01-18  634  		if (res->end < p->start)
d3afa58c Michal Simek        2010-01-18  635  			break;
d3afa58c Michal Simek        2010-01-18  636  		if (p->start < res->start || p->end > res->end)
d3afa58c Michal Simek        2010-01-18  637  			return -1;	/* not completely contained */
d3afa58c Michal Simek        2010-01-18  638  		if (firstpp == NULL)
d3afa58c Michal Simek        2010-01-18  639  			firstpp = pp;
d3afa58c Michal Simek        2010-01-18  640  	}
d3afa58c Michal Simek        2010-01-18  641  	if (firstpp == NULL)
d3afa58c Michal Simek        2010-01-18  642  		return -1;	/* didn't find any conflicting entries? */
d3afa58c Michal Simek        2010-01-18  643  	res->parent = parent;
d3afa58c Michal Simek        2010-01-18 @644  	res->child = *firstpp;
d3afa58c Michal Simek        2010-01-18  645  	res->sibling = *pp;
d3afa58c Michal Simek        2010-01-18  646  	*firstpp = res;
d3afa58c Michal Simek        2010-01-18  647  	*pp = NULL;
d3afa58c Michal Simek        2010-01-18 @648  	for (p = res->child; p != NULL; p = p->sibling) {
d3afa58c Michal Simek        2010-01-18  649  		p->parent = res;
d3afa58c Michal Simek        2010-01-18  650  		pr_debug("PCI: Reparented %s [%llx..%llx] under %s\n",
d3afa58c Michal Simek        2010-01-18  651  			 p->name,
d3afa58c Michal Simek        2010-01-18  652  			 (unsigned long long)p->start,
d3afa58c Michal Simek        2010-01-18  653  			 (unsigned long long)p->end, res->name);
d3afa58c Michal Simek        2010-01-18  654  	}
d3afa58c Michal Simek        2010-01-18  655  	return 0;
d3afa58c Michal Simek        2010-01-18  656  }
d3afa58c Michal Simek        2010-01-18  657  

:::::: The code at line 536 was first introduced by commit
:::::: 70dcd942dc4af3cc6c3dcc2ba499cd841c7f65a7 microblaze: Fix IO space breakage after of_pci_range_to_resource() change

:::::: TO: Michal Simek <[email protected]> :::::: CC: Michal Simek <[email protected]>

---
0-DAY kernel test infrastructure Open Source Technology Center
https://lists.01.org/pipermail/kbuild-all Intel Corporation

#9 lore raw
In reply to Baoquan He
I love your patch! Yet something to improve:

Hi Baoquan,

I love your patch! Yet something to improve:

[auto build test ERROR on linus/master]
[also build test ERROR on v4.17-rc2 next-20180424]
[if your patch is applied to the wrong git tree, please drop us a note to help improve the system]

url: https://github.com/0day-ci/linux/commits/Baoquan-He/resource-Use-list_head-to-link-sibling-resource/20180419-223752
config: xtensa-common_defconfig (attached as .config)
compiler: xtensa-linux-gcc (GCC) 7.2.0
reproduce:
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross chmod +x ~/bin/make.cross
# save the attached .config to linux build tree
make.cross ARCH=xtensa

All errors (new ones prefixed by >>):

In file included from arch/xtensa/lib/pci-auto.c:22:0:
arch/xtensa/include/asm/pci-bridge.h: In function 'pcibios_init_resource':

>> arch/xtensa/include/asm/pci-bridge.h:74:15: error: incompatible types when assigning to type 'struct list_head' from type 'void *'
     res->sibling = NULL;
                  ^
   arch/xtensa/include/asm/pci-bridge.h:75:13: error: incompatible types when assigning to type 'struct list_head' from type 'void *'
     res->child = NULL;
                ^

vim +74 arch/xtensa/include/asm/pci-bridge.h

9a8fd558 include/asm-xtensa/pci-bridge.h Chris Zankel 2005-06-23  65  
9a8fd558 include/asm-xtensa/pci-bridge.h Chris Zankel 2005-06-23  66  static inline void pcibios_init_resource(struct resource *res,
9a8fd558 include/asm-xtensa/pci-bridge.h Chris Zankel 2005-06-23  67  		unsigned long start, unsigned long end, int flags, char *name)
9a8fd558 include/asm-xtensa/pci-bridge.h Chris Zankel 2005-06-23  68  {
9a8fd558 include/asm-xtensa/pci-bridge.h Chris Zankel 2005-06-23  69  	res->start = start;
9a8fd558 include/asm-xtensa/pci-bridge.h Chris Zankel 2005-06-23  70  	res->end = end;
9a8fd558 include/asm-xtensa/pci-bridge.h Chris Zankel 2005-06-23  71  	res->flags = flags;
9a8fd558 include/asm-xtensa/pci-bridge.h Chris Zankel 2005-06-23  72  	res->name = name;
9a8fd558 include/asm-xtensa/pci-bridge.h Chris Zankel 2005-06-23  73  	res->parent = NULL;
9a8fd558 include/asm-xtensa/pci-bridge.h Chris Zankel 2005-06-23 @74  	res->sibling = NULL;
9a8fd558 include/asm-xtensa/pci-bridge.h Chris Zankel 2005-06-23  75  	res->child = NULL;
9a8fd558 include/asm-xtensa/pci-bridge.h Chris Zankel 2005-06-23  76  }

9a8fd558 include/asm-xtensa/pci-bridge.h Chris Zankel 2005-06-23 77

:::::: The code at line 74 was first introduced by commit
:::::: 9a8fd5589902153a134111ed7a40f9cca1f83254 [PATCH] xtensa: Architecture support for Tensilica Xtensa Part 6

:::::: TO: Chris Zankel <[email protected]> :::::: CC: Linus Torvalds <[email protected]>

---
0-DAY kernel test infrastructure Open Source Technology Center
https://lists.01.org/pipermail/kbuild-all Intel Corporation

author
#10 lore raw
Quoted 19 lines · kbuild test robot

Hi Baoquan,

I love your patch! Yet something to improve:

[auto build test ERROR on linus/master]
[also build test ERROR on v4.17-rc2 next-20180424]
[if your patch is applied to the wrong git tree, please drop us a note to help improve the system]

url: https://github.com/0day-ci/linux/commits/Baoquan-He/resource-Use-list_head-to-link-sibling-resource/20180419-223752
config: xtensa-common_defconfig (attached as .config)
compiler: xtensa-linux-gcc (GCC) 7.2.0
reproduce:
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross chmod +x ~/bin/make.cross
# save the attached .config to linux build tree
make.cross ARCH=xtensa

All errors (new ones prefixed by >>):

Quoted 7 lines

In file included from arch/xtensa/lib/pci-auto.c:22:0:
arch/xtensa/include/asm/pci-bridge.h: In function 'pcibios_init_resource':

Quoted 1 line
Quoted 1 line

arch/xtensa/include/asm/pci-bridge.h:74:15: error: incompatible types when assigning to type 'struct list_head' from type 'void *'

     res->sibling = NULL;
                  ^
   arch/xtensa/include/asm/pci-bridge.h:75:13: error: incompatible types when assigning to type 'struct list_head' from type 'void *'
     res->child = NULL;

Thanks, below fix can fix it.

arch/xtensa/include/asm/pci-bridge.h +2 -2
index 0b68c76ec1e6..f487b06817df 100644
@@ -71,8 +71,8 @@ static inline void pcibios_init_resource(struct resource *res,
 	res->flags = flags;
 	res->name = name;
 	res->parent = NULL;
-	res->sibling = NULL;
-	res->child = NULL;
+	INIT_LIST_HEAD(&res->child);
+	INIT_LIST_HEAD(&res->sibling);
 }
 
 
Quoted 27 lines
                ^

vim +74 arch/xtensa/include/asm/pci-bridge.h

9a8fd558 include/asm-xtensa/pci-bridge.h Chris Zankel 2005-06-23  65  
9a8fd558 include/asm-xtensa/pci-bridge.h Chris Zankel 2005-06-23  66  static inline void pcibios_init_resource(struct resource *res,
9a8fd558 include/asm-xtensa/pci-bridge.h Chris Zankel 2005-06-23  67  		unsigned long start, unsigned long end, int flags, char *name)
9a8fd558 include/asm-xtensa/pci-bridge.h Chris Zankel 2005-06-23  68  {
9a8fd558 include/asm-xtensa/pci-bridge.h Chris Zankel 2005-06-23  69  	res->start = start;
9a8fd558 include/asm-xtensa/pci-bridge.h Chris Zankel 2005-06-23  70  	res->end = end;
9a8fd558 include/asm-xtensa/pci-bridge.h Chris Zankel 2005-06-23  71  	res->flags = flags;
9a8fd558 include/asm-xtensa/pci-bridge.h Chris Zankel 2005-06-23  72  	res->name = name;
9a8fd558 include/asm-xtensa/pci-bridge.h Chris Zankel 2005-06-23  73  	res->parent = NULL;
9a8fd558 include/asm-xtensa/pci-bridge.h Chris Zankel 2005-06-23 @74  	res->sibling = NULL;
9a8fd558 include/asm-xtensa/pci-bridge.h Chris Zankel 2005-06-23  75  	res->child = NULL;
9a8fd558 include/asm-xtensa/pci-bridge.h Chris Zankel 2005-06-23  76  }

9a8fd558 include/asm-xtensa/pci-bridge.h Chris Zankel 2005-06-23 77

:::::: The code at line 74 was first introduced by commit
:::::: 9a8fd5589902153a134111ed7a40f9cca1f83254 [PATCH] xtensa: Architecture support for Tensilica Xtensa Part 6

:::::: TO: Chris Zankel <[email protected]> :::::: CC: Linus Torvalds <[email protected]>

---
0-DAY kernel test infrastructure Open Source Technology Center
https://lists.01.org/pipermail/kbuild-all Intel Corporation

author
#11 lore raw

PATCH v3 2/3 resource: add walk_system_ram_res_rev()

In reply to Baoquan He

This function, being a variant of walk_system_ram_res() introduced in commit 8c86e70acead ("resource: provide new functions to walk through resources"), walks through a list of all the resources of System RAM in reversed order, i.e., from higher to lower.

It will be used in kexec_file code.

Cc Andrew Morton <[email protected]>
Cc Thomas Gleixner <[email protected]>
Cc Brijesh Singh <[email protected]>
Cc "Jérôme Glisse" <[email protected]>
Cc Borislav Petkov <[email protected]>
Cc Tom Lendacky <[email protected]>
Cc Wei Yang <[email protected]>
---
 include/linux/ioport.h |  3 +++
 kernel/resource.c      | 40 ++++++++++++++++++++++++++++++++++++++++
 2 files changed, 43 insertions(+)
include/linux/ioport.h +3 -0
index 225d13d3500a..5f2cfb460a14 100644
@@ -278,6 +278,9 @@ extern int
 walk_system_ram_res(u64 start, u64 end, void *arg,
 		    int (*func)(struct resource *, void *));
 extern int
+walk_system_ram_res_rev(u64 start, u64 end, void *arg,
+			int (*func)(struct resource *, void *));
+extern int
 walk_iomem_res_desc(unsigned long desc, unsigned long flags, u64 start, u64 end,
 		    void *arg, int (*func)(struct resource *, void *));
 
kernel/resource.c +40 -0
index 4f560991c130..6c519b06e3d6 100644
@@ -23,6 +23,8 @@
 #include <linux/pfn.h>
 #include <linux/mm.h>
 #include <linux/resource_ext.h>
+#include <linux/string.h>
+#include <linux/vmalloc.h>
 #include <asm/io.h>
 
 
@@ -475,6 +477,44 @@ int walk_system_ram_res(u64 start, u64 end, void *arg,
 }
 
 /*
+ * This function, being a variant of walk_system_ram_res(), calls the @func
+ * callback against all memory ranges of type System RAM which are marked as
+ * IORESOURCE_SYSTEM_RAM and IORESOUCE_BUSY in reversed order, i.e., from
+ * higher to lower.
+ */
+int walk_system_ram_res_rev(u64 start, u64 end, void *arg,
+				int (*func)(struct resource *, void *))
+{
+	unsigned long flags;
+	struct resource *res;
+	int ret = -1;
+
+	flags = IORESOURCE_SYSTEM_RAM | IORESOURCE_BUSY;
+
+	read_lock(&resource_lock);
+	list_for_each_entry_reverse(res, &iomem_resource.child, sibling) {
+		if (start >= end)
+			break;
+		if ((res->flags & flags) != flags)
+			continue;
+		if (res->desc != IORES_DESC_NONE)
+			continue;
+		if (res->end < start)
+			break;
+
+		if ((res->end >= start) && (res->start < end)) {
+			ret = (*func)(res, arg);
+			if (ret)
+				break;
+		}
+		end = res->start - 1;
+
+	}
+	read_unlock(&resource_lock);
+	return ret;
+}
+
+/*
  * This function calls the @func callback against all memory ranges, which
  * are ranges marked as IORESOURCE_MEM and IORESOUCE_BUSY.
  */
Signature · 1 line
2.13.6
#12 lore raw

Re: [PATCH v3 2/3] resource: add walk_system_ram_res_rev()

Quoted 6 lines · Baoquan He

This function, being a variant of walk_system_ram_res() introduced in commit 8c86e70acead ("resource: provide new functions to walk through resources"), walks through a list of all the resources of System RAM in reversed order, i.e., from higher to lower.

It will be used in kexec_file code.

Of course, what is missing is the big *WHY* you need this to happen this way...

Quoted 7 lines

/*
+ * This function, being a variant of walk_system_ram_res(), calls the @func
+ * callback against all memory ranges of type System RAM which are marked as
+ * IORESOURCE_SYSTEM_RAM and IORESOUCE_BUSY in reversed order, i.e., from
+ * higher to lower.
+ */

kernel/resource.c +1 -0
+int walk_system_ram_res_rev(u64 start, u64 end, void *arg,

kernel/resource.c int walk_system_ram_res(u64 start, u64 end, void *arg, · line 495

Quoted 10 lines
kernel/resource.c +10 -0
+				int (*func)(struct resource *, void *))
+{
+	unsigned long flags;
+	struct resource *res;
+	int ret = -1;
+
+	flags = IORESOURCE_SYSTEM_RAM | IORESOURCE_BUSY;
+
+	read_lock(&resource_lock);
+	list_for_each_entry_reverse(res, &iomem_resource.child, sibling) {

... and the other thing that I'm not clear on is why are you slapping this function just like that instead of extending __walk_iomem_res_desc() to do reverse direction too?

Signature · 5 lines
Regards/Gruss,
    Boris.

SUSE Linux GmbH, GF: Felix Imendörffer, Jane Smithard, Graham Norton, HRB 21284 (AG Nürnberg)
-- 
author
#13 lore raw

Re: [PATCH v3 2/3] resource: add walk_system_ram_res_rev()

Sorry to all reviewers, I had a redhat urgent issue, this patchset discussing is delayed.

Quoted 10 lines · Borislav Petkov
Quoted 6 lines · Baoquan He

This function, being a variant of walk_system_ram_res() introduced in commit 8c86e70acead ("resource: provide new functions to walk through resources"), walks through a list of all the resources of System RAM in reversed order, i.e., from higher to lower.

It will be used in kexec_file code.

Of course, what is missing is the big *WHY* you need this to happen this way...

Sorry for that, I just ran scripts/get_maintainer.pl to get expert's name and added them into each patch. The reason this change is made is in patch 3/3. Test robot reported a code bug on the latest kernel, will repost and CC everyone in all patches.

Rob Herring asked the same question in v2, I explained to him. The discussion can be found here:
https://lkml.org/lkml/2018/4/10/484

Quoted 14 lines
Quoted 14 lines

/*
+ * This function, being a variant of walk_system_ram_res(), calls the @func
+ * callback against all memory ranges of type System RAM which are marked as
+ * IORESOURCE_SYSTEM_RAM and IORESOUCE_BUSY in reversed order, i.e., from
+ * higher to lower.
+ */

kernel/resource.c +7 -0
+int walk_system_ram_res_rev(u64 start, u64 end, void *arg,
+				int (*func)(struct resource *, void *))
+{
+	unsigned long flags;
+	struct resource *res;
+	int ret = -1;
+
Quoted 8 lines

kernel/resource.c int walk_system_ram_res(u64 start, u64 end, void *arg, · line 495

Quoted 4 lines
kernel/resource.c +4 -0
+	flags = IORESOURCE_SYSTEM_RAM | IORESOURCE_BUSY;
+
+	read_lock(&resource_lock);
+	list_for_each_entry_reverse(res, &iomem_resource.child, sibling) {

... and the other thing that I'm not clear on is why are you slapping this function just like that instead of extending __walk_iomem_res_desc() to do reverse direction too?

Yes, I planned to do as you said when started to write code. After investigation, I changed to use the way of this patch. Because __walk_iomem_res_desc() provides two interfaces by calling find_next_iomem_res(). If 'first_level_children_only' is true, it only iterates system RAM resource; if 'first_level_children_only' is false, it does depth first iteration in iomem_resource. The 1st interface is only used by walk_system_ram_res() and walk_mem_res(). While the 2nd interface need call find_next_iomem_res() which resorts to next_resource(). next_resource() will check 'sibling_only' to decide if iterate iomem_resource's children, or walk over all resources of iomem_resource with depth first. For walk_system_ram_res_rev(), we only need to iterate iomem_resource's children, and seems no one has requirement to iterate all iomem_resource's children and grand children
revresedly. For simplifying code implementation, I just did as in this patch, surely, if any suggestion or scenario given about the reversed iteration of all resources, I can change to add below functions, and based on them to implement walk_system_ram_res_rev().

walk_system_ram_res_rev

	->__walk_iomem_res_desc_rev
		->find_next_iomem_res_rev
			->next_resource_rev

Thanks
Baoquan

Quoted 3 lines

SUSE Linux GmbH, GF: Felix Imendörffer, Jane Smithard, Graham Norton, HRB 21284 (AG Nürnberg)

#14 lore raw

Re: [PATCH v3 2/3] resource: add walk_system_ram_res_rev()

Quoted 9 lines · Baoquan He

Sorry for that, I just ran scripts/get_maintainer.pl to get expert's name and added them into each patch. The reason this change is made is in patch 3/3. Test robot reported a code bug on the latest kernel, will repost and CC everyone in all patches.

Rob Herring asked the same question in v2, I explained to him. The discussion can be found here:
https://lkml.org/lkml/2018/4/10/484

... and when I open that link, the first paragraph says:

"This is the explanation I made when Andrew helped to review the v1 post:
https://lkml.org/lkml/2018/3/23/78"

Do you see the absurdity of the link chasing of your explanation?!

Instead, the explanation *WHY* should be in the commit message of the patch - not in mail replies when people ask you about it.

Also, do not use lkml.org when referencing a mail on lkml but use the Message-ID of the header. We have a proper redirector at
https://lkml.kernel.org/r/<Message-ID>

Now lemme read the reason finally...

"We need unify these two interfaces on behaviour since they are the same on essense from the users' point of view... "

That's not a good enough reason for me to cause code churn. If the only reason is: because the one does it top-down and the other bottom-up, I'm not convinced.

Signature · 5 lines
Regards/Gruss,
    Boris.

SUSE Linux GmbH, GF: Felix Imendörffer, Jane Smithard, Graham Norton, HRB 21284 (AG Nürnberg)
-- 
author
#15 lore raw

Re: [PATCH v3 2/3] resource: add walk_system_ram_res_rev()

Quoted 13 lines · Borislav Petkov
Quoted 9 lines · Baoquan He

Sorry for that, I just ran scripts/get_maintainer.pl to get expert's name and added them into each patch. The reason this change is made is in patch 3/3. Test robot reported a code bug on the latest kernel, will repost and CC everyone in all patches.

Rob Herring asked the same question in v2, I explained to him. The discussion can be found here:
https://lkml.org/lkml/2018/4/10/484

... and when I open that link, the first paragraph says:

Quoted 7 lines

"This is the explanation I made when Andrew helped to review the v1 post:
https://lkml.org/lkml/2018/3/23/78"

Do you see the absurdity of the link chasing of your explanation?!

Instead, the explanation *WHY* should be in the commit message of the patch - not in mail replies when people ask you about it.

It is in patch 3/3 and cover-letter. I often received one or several patches of a big patchset. As I said, I used ./scripts/get_maintainer.pl to each patch's reviewers, and will add all people in CC list.

Quoted 4 lines

Also, do not use lkml.org when referencing a mail on lkml but use the Message-ID of the header. We have a proper redirector at
https://lkml.kernel.org/r/<Message-ID>

I noticed maintainers merged patches with this Message-ID, could you tell how to get this Message-ID?

Quoted 9 lines

Now lemme read the reason finally...

"We need unify these two interfaces on behaviour since they are the same on essense from the users' point of view... "

That's not a good enough reason for me to cause code churn. If the only reason is: because the one does it top-down and the other bottom-up, I'm not convinced.

We have been using this top-down way for x86 64 since earlier 2013 in the KEXEC loading with command:

	'kexec -l bzImage --initrd initramfs-xx.img'

Two years later, Vivek added a new KEXEC_FILE loading, and most of job is done in kernel because we need do kernel image sinature verification. The KEXEC_FILE loading mostly is used in secure boot machine. Although more and more users like to take secure boot machine, system using KEXEC loading are still much more. I ever asked our kernel QE who takes care of kernel and kdump testing, they said in their testing systems, the proportion of legacy system vs system with secure boot is 10:1.

Before long I pinged hpa, Vivek and Yinghai who discussed and wrote the code for KEXEC loading in x86 64, asked them about the reaon they chose top-down, no reply from them. I guess they probably worried that low memory under 4G are mostly reserved by system or firmware, even though kexec/kdump have tried very hard to avoid each of them if recognized, any missed one will cause loading and booting failure. Just a guess.

In fact, there's no spec or doc to tell which one is right or not. So I finally decide to take the same top-down way for KEXEC_FILE loading because on statistics KEXEC loading is used longer and wider.

This is not a thing that one is top down, the other is bottom up. For us, they might be so different on details of code, for customers, they just think them as a same thing. They may say I just get a new machine, and still do kexec loading, why these top-down, bottom-up things come up.

And this is not causing code churn. You can see that by replacing pointer operation with list_head, code in kernel/resource.c related to child list iteration is much easier to read, e.g __request_resource(), __release_resource(). As Rob said in v2 reviewing, they have put the same replacing in DT struct device_node in todo list, and ACPI tree function in drivers/acpi/acpica/pstree.c too. I think this is why Andrew asked to try this way and see what the patches looks like.

Thanks
Baoquan

#16 lore raw

Re: [PATCH v3 2/3] resource: add walk_system_ram_res_rev()

Quoted 2 lines · Baoquan He

I noticed maintainers merged patches with this Message-ID, could you tell how to get this Message-ID?

https://en.wikipedia.org/wiki/Message-ID

Quoted 5 lines

This is not a thing that one is top down, the other is bottom up. For us, they might be so different on details of code, for customers, they just think them as a same thing. They may say I just get a new machine, and still do kexec loading, why these top-down, bottom-up things come up.

So if I read the above correctly, it doesn't matter whether top-down or bottom-up.

Quoted 3 lines

And this is not causing code churn. You can see that by replacing pointer operation with list_head, code in kernel/resource.c related to child list iteration is much easier to read,

Now *this* is starting to sound like some reason "why". If it is better readability, then say so in the commit message.

It still doesn't justify adding walk_system_ram_res_rev() though.

Signature · 5 lines
Regards/Gruss,
    Boris.

SUSE Linux GmbH, GF: Felix Imendörffer, Jane Smithard, Graham Norton, HRB 21284 (AG Nürnberg)
-- 
author
#17 lore raw

Re: [PATCH v3 2/3] resource: add walk_system_ram_res_rev()

Quoted 5 lines · Borislav Petkov
Quoted 2 lines · Baoquan He

I noticed maintainers merged patches with this Message-ID, could you tell how to get this Message-ID?

https://en.wikipedia.org/wiki/Message-ID

thx.

Quoted 9 lines
Quoted 6 lines

This is not a thing that one is top down, the other is bottom up. For us, they might be so different on details of code, for customers, they just think them as a same thing. They may say I just get a new machine, and still do kexec loading, why these top-down, bottom-up things come up.

So if I read the above correctly, it doesn't matter whether top-down or bottom-up.

Well, kexec_file loading is added because kernel signature verification has to be done in kernel space, it's implemented by porting code of user space utility kexc-tools to kernel. In essence both the kexec loading and kexec_file loading are the same, they are all used to load kernel for the furture kexec/kdump jumping. Now they have different kernel loading position, could you tell how you get the conclusion that it doesn't matter? And what's your theory basis?

And KASLR also changes kernel's position, I don't think users have to know all these details to exclude unnecessary noise.

Quoted 7 lines
Quoted 4 lines

And this is not causing code churn. You can see that by replacing pointer operation with list_head, code in kernel/resource.c related to child list iteration is much easier to read,

Now *this* is starting to sound like some reason "why". If it is better readability, then say so in the commit message.

Yes, I can add it to log in next post. While the main reason I made this is for the top down searching of system RAM in kexec_file loading.

author
#18 lore raw

PATCH v3 3/3 kexec_file: Load kernel at top of system RAM if required

In reply to Baoquan He

For kexec_file loading, if kexec_buf.top_down is 'true', the memory which is used to load kernel/initrd/purgatory is supposed to be allocated from top to down. This is what we have been doing all along in the old kexec loading interface and the kexec loading is still default setting in some distributions. However, the current kexec_file loading interface doesn't do likt this. The function arch_kexec_walk_mem() it calls ignores checking kexec_buf.top_down, but calls walk_system_ram_res() directly to go through all resources of System RAM from bottom to up, to try to find memory region which can contain the specific kexec buffer, then call locate_mem_hole_callback() to allocate memory in that found memory region from top to down. This brings confusion. These two interfaces need be unified on behaviour.

Here add checking if kexec_buf.top_down is 'true' in arch_kexec_walk_mem(), if yes, call the newly added walk_system_ram_res_rev() to find memory region from top to down to load kernel.

Cc Eric Biederman <[email protected]>
Cc Vivek Goyal <[email protected]>
Cc Dave Young <[email protected]>
Cc Andrew Morton <[email protected]>
Cc Yinghai Lu <[email protected]>
---
 kernel/kexec_file.c | 2 ++
 1 file changed, 2 insertions(+)
kernel/kexec_file.c +2 -0
index 75d8e7cf040e..7a66d9d5a534 100644
@@ -518,6 +518,8 @@ int __weak arch_kexec_walk_mem(struct kexec_buf *kbuf,
 					   IORESOURCE_SYSTEM_RAM | IORESOURCE_BUSY,
 					   crashk_res.start, crashk_res.end,
 					   kbuf, func);
+	else if (kbuf->top_down)
+		return walk_system_ram_res_rev(0, ULONG_MAX, kbuf, func);
 	else
 		return walk_system_ram_res(0, ULONG_MAX, kbuf, func);
 }
Signature · 1 line
2.13.6