linux-bcache.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Coly Li <colyli@suse.de>
To: linux-bcache@vger.kernel.org
Cc: linux-block@vger.kernel.org, linux-nvdimm@lists.linux.dev,
	axboe@kernel.dk, hare@suse.com, jack@suse.cz,
	dan.j.williams@intel.com, hch@lst.de, ying.huang@intel.com,
	Jianpeng Ma <jianpeng.ma@intel.com>,
	Qiaowei Ren <qiaowei.ren@intel.com>,
	Hannes Reinecke <hare@suse.de>
Subject: [PATCH v12 04/12] bcache: bch_nvmpg_alloc_pages() of the buddy
Date: Thu, 12 Aug 2021 01:02:16 +0800	[thread overview]
Message-ID: <20210811170224.42837-5-colyli@suse.de> (raw)
In-Reply-To: <20210811170224.42837-1-colyli@suse.de>

From: Jianpeng Ma <jianpeng.ma@intel.com>

This patch implements the bch_nvmpg_alloc_pages() of the nvm pages buddy
allocator. In terms of function, this func is like current
page-buddy-alloc. But the differences are:
a: it need owner_uuid as parameter which record owner info. And it
make those info persistence.
b: it don't need flags like GFP_*. All allocs are the equal.
c: it don't trigger other ops etc swap/recycle.

Signed-off-by: Jianpeng Ma <jianpeng.ma@intel.com>
Co-developed-by: Qiaowei Ren <qiaowei.ren@intel.com>
Signed-off-by: Qiaowei Ren <qiaowei.ren@intel.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Hannes Reinecke <hare@suse.de>
Cc: Jens Axboe <axboe@kernel.dk>
---
 drivers/md/bcache/nvm-pages.c | 210 ++++++++++++++++++++++++++++++++++
 drivers/md/bcache/nvm-pages.h |   9 ++
 2 files changed, 219 insertions(+)

diff --git a/drivers/md/bcache/nvm-pages.c b/drivers/md/bcache/nvm-pages.c
index 677fdb62f737..420b7c479057 100644
--- a/drivers/md/bcache/nvm-pages.c
+++ b/drivers/md/bcache/nvm-pages.c
@@ -50,6 +50,13 @@ unsigned long bch_nvmpg_ptr_to_offset(struct bch_nvmpg_ns *ns, void *ptr)
 	return BCH_NVMPG_OFFSET(ns_id, offset);
 }
 
+static unsigned long bch_nvmpg_ptr_to_pgoff(struct bch_nvmpg_ns *ns, void *ptr)
+{
+	unsigned long offset = (unsigned long)(ptr - ns->base_addr);
+
+	return offset >> PAGE_SHIFT;
+}
+
 static struct page *bch_nvmpg_va_to_pg(void *addr)
 {
 	return virt_to_page(addr);
@@ -268,6 +275,209 @@ static void bch_nvmpg_init_free_space(struct bch_nvmpg_ns *ns)
 	}
 }
 
+
+/* If not found, it will create if create == true */
+static struct bch_nvmpg_head *find_nvmpg_head(const char *uuid, bool create)
+{
+	struct bch_nvmpg_set_header *set_header = global_nvmpg_set->set_header;
+	struct bch_nvmpg_head *head = NULL;
+	int i;
+
+	if (set_header == NULL)
+		goto out;
+
+	for (i = 0; i < set_header->size; i++) {
+		struct bch_nvmpg_head *h = &set_header->heads[i];
+
+		if (h->state != BCH_NVMPG_HD_STAT_ALLOC)
+			continue;
+
+		if (!memcmp(uuid, h->uuid, 16)) {
+			head = h;
+			break;
+		}
+	}
+
+	if (!head && create) {
+		u32 used = set_header->used;
+
+		if (set_header->size > used) {
+			head = &set_header->heads[used];
+			memset(head, 0, sizeof(struct bch_nvmpg_head));
+			head->state = BCH_NVMPG_HD_STAT_ALLOC;
+			memcpy(head->uuid, uuid, 16);
+			global_nvmpg_set->heads_used++;
+			set_header->used++;
+		} else
+			pr_info("No free bch_nvmpg_head\n");
+	}
+
+out:
+	return head;
+}
+
+static struct bch_nvmpg_recs *find_empty_nvmpg_recs(void)
+{
+	unsigned int start;
+	struct bch_nvmpg_ns *ns = global_nvmpg_set->ns_tbl[0];
+	struct bch_nvmpg_recs *recs;
+
+	start = bitmap_find_next_zero_area(ns->recs_bitmap,
+					   BCH_MAX_PGALLOC_RECS, 0, 1, 0);
+	if (start > BCH_MAX_PGALLOC_RECS) {
+		pr_info("No free struct bch_nvmpg_recs\n");
+		return NULL;
+	}
+
+	bitmap_set(ns->recs_bitmap, start, 1);
+	recs = (struct bch_nvmpg_recs *)
+		bch_nvmpg_offset_to_ptr(BCH_NVMPG_SYSRECS_OFFSET)
+	       + start;
+
+	memset(recs, 0, sizeof(struct bch_nvmpg_recs));
+	return recs;
+}
+
+
+static struct bch_nvmpg_recs *find_nvmpg_recs(struct bch_nvmpg_ns *ns,
+					      struct bch_nvmpg_head *head,
+					      bool create)
+{
+	int ns_id = ns->sb->this_ns;
+	struct bch_nvmpg_recs *prev_recs = NULL, *recs = NULL;
+
+	recs = bch_nvmpg_offset_to_ptr(head->recs_offset[ns_id]);
+
+	/* If create=false, we return recs[nr] */
+	if (!create)
+		return recs;
+
+	/*
+	 * If create=true, it mean we need a empty struct bch_nvmpg_rec
+	 * So we should find non-empty struct bch_nvmpg_recs or alloc
+	 * new struct bch_nvmpg_recs. And return this bch_nvmpg_recs
+	 */
+	while (recs && (recs->used == recs->size)) {
+		prev_recs = recs;
+		recs = bch_nvmpg_offset_to_ptr(recs->next_offset);
+	}
+
+	/* Found empty struct bch_nvmpg_recs */
+	if (recs)
+		return recs;
+
+	/* Need alloc new struct bch_nvmpg_recs */
+	recs = find_empty_nvmpg_recs();
+	if (recs) {
+		unsigned long offset;
+
+		recs->next_offset = 0;
+		recs->head_offset = bch_nvmpg_ptr_to_offset(ns, head);
+		memcpy(recs->magic, bch_nvmpg_recs_magic, 16);
+		memcpy(recs->uuid, head->uuid, 16);
+		recs->size = BCH_NVMPG_MAX_RECS;
+		recs->used = 0;
+
+		offset = bch_nvmpg_ptr_to_offset(ns, recs);
+		if (prev_recs)
+			prev_recs->next_offset = offset;
+		else
+			head->recs_offset[ns_id] = offset;
+	}
+
+	return recs;
+}
+
+static void add_nvmpg_rec(struct bch_nvmpg_ns *ns,
+			  struct bch_nvmpg_recs *recs,
+			  void *kaddr, int order)
+{
+	int i;
+
+	for (i = 0; i < recs->size; i++) {
+		if (recs->recs[i].pgoff == 0) {
+			recs->recs[i].pgoff = bch_nvmpg_ptr_to_pgoff(ns, kaddr);
+			recs->recs[i].order = order;
+			recs->recs[i].ns_id = ns->sb->this_ns;
+			recs->used++;
+			break;
+		}
+	}
+	BUG_ON(i == recs->size);
+}
+
+
+void *bch_nvmpg_alloc_pages(int order, const char *uuid)
+{
+	void *kaddr = NULL;
+	struct bch_nvmpg_head *head;
+	int n, o;
+
+	mutex_lock(&global_nvmpg_set->lock);
+	head = find_nvmpg_head(uuid, true);
+
+	if (!head) {
+		pr_err("Cannot find bch_nvmpg_recs by uuid.\n");
+		goto unlock;
+	}
+
+	for (n = 0; n < global_nvmpg_set->total_ns; n++) {
+		struct bch_nvmpg_ns *ns = global_nvmpg_set->ns_tbl[n];
+
+		if (!ns || (ns->free < (1L << order)))
+			continue;
+
+		for (o = order; o < BCH_MAX_ORDER; o++) {
+			struct list_head *list;
+			struct page *page, *buddy_page;
+
+			if (list_empty(&ns->free_area[o]))
+				continue;
+
+			list = ns->free_area[o].next;
+			page = container_of((void *)list, struct page,
+					    zone_device_data);
+
+			list_del(list);
+
+			while (o != order) {
+				void *addr;
+				pgoff_t pgoff;
+
+				pgoff = page->index + (1L << (o - 1));
+				addr = bch_nvmpg_pgoff_to_ptr(ns, pgoff);
+				buddy_page = bch_nvmpg_va_to_pg(addr);
+				set_page_private(buddy_page, o - 1);
+				buddy_page->index = pgoff;
+				__SetPageBuddy(buddy_page);
+				list_add((struct list_head *)&buddy_page->zone_device_data,
+					 &ns->free_area[o - 1]);
+				o--;
+			}
+
+			set_page_private(page, order);
+			__ClearPageBuddy(page);
+			ns->free -= 1L << order;
+			kaddr = bch_nvmpg_pgoff_to_ptr(ns, page->index);
+			break;
+		}
+
+		if (o < BCH_MAX_ORDER) {
+			struct bch_nvmpg_recs *recs;
+
+			recs = find_nvmpg_recs(ns, head, true);
+			/* ToDo: handle pgalloc_recs==NULL */
+			add_nvmpg_rec(ns, recs, kaddr, order);
+			break;
+		}
+	}
+
+unlock:
+	mutex_unlock(&global_nvmpg_set->lock);
+	return kaddr;
+}
+EXPORT_SYMBOL_GPL(bch_nvmpg_alloc_pages);
+
 static int attach_nvmpg_set(struct bch_nvmpg_ns *ns)
 {
 	struct bch_nvmpg_sb *sb = ns->sb;
diff --git a/drivers/md/bcache/nvm-pages.h b/drivers/md/bcache/nvm-pages.h
index 2116086c4d01..1bcd7a4e1fd1 100644
--- a/drivers/md/bcache/nvm-pages.h
+++ b/drivers/md/bcache/nvm-pages.h
@@ -75,6 +75,9 @@ struct bch_nvmpg_set {
 /* Indicate which field in bch_nvmpg_sb to be updated */
 #define BCH_NVMPG_TOTAL_NS	0	/* total_ns */
 
+#define BCH_PGOFF_TO_KVADDR(pgoff)					\
+	((void *)((unsigned long)(pgoff) << PAGE_SHIFT))
+
 #define BCH_MAX_PGALLOC_RECS						\
 	(min_t(unsigned int, 64,					\
 	       (BCH_NVMPG_START - BCH_NVMPG_SYSRECS_OFFSET) /		\
@@ -88,6 +91,7 @@ unsigned long bch_nvmpg_ptr_to_offset(struct bch_nvmpg_ns *ns, void *ptr);
 struct bch_nvmpg_ns *bch_register_namespace(const char *dev_path);
 int bch_nvmpg_init(void);
 void bch_nvmpg_exit(void);
+void *bch_nvmpg_alloc_pages(int order, const char *uuid);
 
 #else
 
@@ -103,6 +107,11 @@ static inline int bch_nvmpg_init(void)
 
 static inline void bch_nvmpg_exit(void) { }
 
+static inline void *bch_nvmpg_alloc_pages(int order, const char *uuid)
+{
+	return NULL;
+}
+
 #endif /* CONFIG_BCACHE_NVM_PAGES */
 
 #endif /* _BCACHE_NVM_PAGES_H */
-- 
2.26.2


  parent reply	other threads:[~2021-08-11 17:04 UTC|newest]

Thread overview: 16+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-08-11 17:02 [PATCH v12 00/12] bcache: support NVDIMM for journaling Coly Li
2021-08-11 17:02 ` [PATCH v12 01/12] bcache: add initial data structures for nvm pages Coly Li
2021-08-11 17:02 ` [PATCH v12 02/12] bcache: initialize the nvm pages allocator Coly Li
2021-08-12  5:43   ` Dan Williams
2021-08-12  8:26     ` Coly Li
2021-08-11 17:02 ` [PATCH v12 03/12] bcache: initialization of the buddy Coly Li
2021-08-11 17:02 ` Coly Li [this message]
2021-08-11 17:02 ` [PATCH v12 05/12] bcache: bch_nvmpg_free_pages() of the buddy allocator Coly Li
2021-08-11 17:02 ` [PATCH v12 06/12] bcache: get recs list head for allocated pages by specific uuid Coly Li
2021-08-11 17:02 ` [PATCH v12 07/12] bcache: use bucket index to set GC_MARK_METADATA for journal buckets in bch_btree_gc_finish() Coly Li
2021-08-11 17:02 ` [PATCH v12 08/12] bcache: add BCH_FEATURE_INCOMPAT_NVDIMM_META into incompat feature set Coly Li
2021-08-11 17:02 ` [PATCH v12 09/12] bcache: initialize bcache journal for NVDIMM meta device Coly Li
2021-08-11 17:02 ` [PATCH v12 10/12] bcache: support storing bcache journal into " Coly Li
2021-08-11 17:02 ` [PATCH v12 11/12] bcache: read jset from NVDIMM pages for journal replay Coly Li
2021-08-11 17:02 ` [PATCH v12 12/12] bcache: add sysfs interface register_nvdimm_meta to register NVDIMM meta device Coly Li
2021-08-15 16:21 ` [PATCH v12 00/12] bcache: support NVDIMM for journaling Coly Li

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20210811170224.42837-5-colyli@suse.de \
    --to=colyli@suse.de \
    --cc=axboe@kernel.dk \
    --cc=dan.j.williams@intel.com \
    --cc=hare@suse.com \
    --cc=hare@suse.de \
    --cc=hch@lst.de \
    --cc=jack@suse.cz \
    --cc=jianpeng.ma@intel.com \
    --cc=linux-bcache@vger.kernel.org \
    --cc=linux-block@vger.kernel.org \
    --cc=linux-nvdimm@lists.linux.dev \
    --cc=qiaowei.ren@intel.com \
    --cc=ying.huang@intel.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).