libnvdimm, pmem, pfn: make pmem_rw_bytes generic and refactor pfn setup

In preparation for providing an alternative (to block device) access
mechanism to persistent memory, convert pmem_rw_bytes() to
nsio_rw_bytes().  This allows ->rw_bytes() functionality without
requiring a 'struct pmem_device' to be instantiated.

In other words, when ->rw_bytes() is in use i/o is driven through
'struct nd_namespace_io', otherwise it is driven through 'struct
pmem_device' and the block layer.  This consolidates the disjoint calls
to devm_exit_badblocks() and devm_memunmap() into a common
devm_nsio_disable() and cleans up the init path to use a unified
pmem_attach_disk() implementation.
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>

drivers/nvdimm/blk.c

@@ -324,7 +324,7 @@ static int nd_blk_probe(struct device *dev)
 	ndns->rw_bytes = nsblk_rw_bytes;
 	if (is_nd_btt(dev))
 		return nvdimm_namespace_attach_btt(ndns);
-	else if (nd_btt_probe(dev, ndns, nsblk) == 0) {
+	else if (nd_btt_probe(dev, ndns) == 0) {
 		/* we'll come back as btt-blk */
 		return -ENXIO;
 	} else

drivers/nvdimm/btt_devs.c

@@ -273,8 +273,7 @@ static int __nd_btt_probe(struct nd_btt *nd_btt,
 	return 0;
 }
 
-int nd_btt_probe(struct device *dev, struct nd_namespace_common *ndns,
-		void *drvdata)
+int nd_btt_probe(struct device *dev, struct nd_namespace_common *ndns)
 {
 	int rc;
 	struct device *btt_dev;
@@ -289,7 +288,6 @@ int nd_btt_probe(struct device *dev, struct nd_namespace_common *ndns,
 	nvdimm_bus_unlock(&ndns->dev);
 	if (!btt_dev)
 		return -ENOMEM;
-	dev_set_drvdata(btt_dev, drvdata);
 	btt_sb = devm_kzalloc(dev, sizeof(*btt_sb), GFP_KERNEL);
 	rc = __nd_btt_probe(to_nd_btt(btt_dev), ndns, btt_sb);
 	dev_dbg(dev, "%s: btt: %s\n", __func__,

drivers/nvdimm/claim.c

@@ -12,6 +12,7 @@
  */
 #include <linux/device.h>
 #include <linux/sizes.h>
+#include <linux/pmem.h>
 #include "nd-core.h"
 #include "pfn.h"
 #include "btt.h"
@@ -199,3 +200,63 @@ u64 nd_sb_checksum(struct nd_gen_sb *nd_gen_sb)
 	return sum;
 }
 EXPORT_SYMBOL(nd_sb_checksum);
+
+static int nsio_rw_bytes(struct nd_namespace_common *ndns,
+		resource_size_t offset, void *buf, size_t size, int rw)
+{
+	struct nd_namespace_io *nsio = to_nd_namespace_io(&ndns->dev);
+
+	if (unlikely(offset + size > nsio->size)) {
+		dev_WARN_ONCE(&ndns->dev, 1, "request out of range\n");
+		return -EFAULT;
+	}
+
+	if (rw == READ) {
+		unsigned int sz_align = ALIGN(size + (offset & (512 - 1)), 512);
+
+		if (unlikely(is_bad_pmem(&nsio->bb, offset / 512, sz_align)))
+			return -EIO;
+		return memcpy_from_pmem(buf, nsio->addr + offset, size);
+	} else {
+		memcpy_to_pmem(nsio->addr + offset, buf, size);
+		wmb_pmem();
+	}
+
+	return 0;
+}
+
+int devm_nsio_enable(struct device *dev, struct nd_namespace_io *nsio)
+{
+	struct resource *res = &nsio->res;
+	struct nd_namespace_common *ndns = &nsio->common;
+
+	nsio->size = resource_size(res);
+	if (!devm_request_mem_region(dev, res->start, resource_size(res),
+				dev_name(dev))) {
+		dev_warn(dev, "could not reserve region %pR\n", res);
+		return -EBUSY;
+	}
+
+	ndns->rw_bytes = nsio_rw_bytes;
+	if (devm_init_badblocks(dev, &nsio->bb))
+		return -ENOMEM;
+	nvdimm_badblocks_populate(to_nd_region(ndns->dev.parent), &nsio->bb,
+			&nsio->res);
+
+	nsio->addr = devm_memremap(dev, res->start, resource_size(res),
+			ARCH_MEMREMAP_PMEM);
+	if (IS_ERR(nsio->addr))
+		return PTR_ERR(nsio->addr);
+	return 0;
+}
+EXPORT_SYMBOL_GPL(devm_nsio_enable);
+
+void devm_nsio_disable(struct device *dev, struct nd_namespace_io *nsio)
+{
+	struct resource *res = &nsio->res;
+
+	devm_memunmap(dev, nsio->addr);
+	devm_exit_badblocks(dev, &nsio->bb);
+	devm_release_mem_region(dev, res->start, resource_size(res));
+}
+EXPORT_SYMBOL_GPL(devm_nsio_disable);

drivers/nvdimm/nd.h

@@ -13,6 +13,7 @@
 #ifndef __ND_H__
 #define __ND_H__
 #include <linux/libnvdimm.h>
+#include <linux/badblocks.h>
 #include <linux/blkdev.h>
 #include <linux/device.h>
 #include <linux/mutex.h>
@@ -197,13 +198,12 @@ struct nd_gen_sb {
 
 u64 nd_sb_checksum(struct nd_gen_sb *sb);
 #if IS_ENABLED(CONFIG_BTT)
-int nd_btt_probe(struct device *dev, struct nd_namespace_common *ndns,
-		void *drvdata);
+int nd_btt_probe(struct device *dev, struct nd_namespace_common *ndns);
 bool is_nd_btt(struct device *dev);
 struct device *nd_btt_create(struct nd_region *nd_region);
 #else
 static inline int nd_btt_probe(struct device *dev,
-		struct nd_namespace_common *ndns, void *drvdata)
+		struct nd_namespace_common *ndns)
 {
 	return -ENODEV;
 }
@@ -221,14 +221,13 @@ static inline struct device *nd_btt_create(struct nd_region *nd_region)
 
 struct nd_pfn *to_nd_pfn(struct device *dev);
 #if IS_ENABLED(CONFIG_NVDIMM_PFN)
-int nd_pfn_probe(struct device *dev, struct nd_namespace_common *ndns,
-		void *drvdata);
+int nd_pfn_probe(struct device *dev, struct nd_namespace_common *ndns);
 bool is_nd_pfn(struct device *dev);
 struct device *nd_pfn_create(struct nd_region *nd_region);
 int nd_pfn_validate(struct nd_pfn *nd_pfn);
 #else
-static inline int nd_pfn_probe(struct device *dev, struct nd_namespace_common *ndns,
-		void *drvdata)
+static inline int nd_pfn_probe(struct device *dev,
+		struct nd_namespace_common *ndns)
 {
 	return -ENODEV;
 }
@@ -272,6 +271,20 @@ const char *nvdimm_namespace_disk_name(struct nd_namespace_common *ndns,
 		char *name);
 void nvdimm_badblocks_populate(struct nd_region *nd_region,
 		struct badblocks *bb, const struct resource *res);
+#if IS_ENABLED(CONFIG_ND_CLAIM)
+int devm_nsio_enable(struct device *dev, struct nd_namespace_io *nsio);
+void devm_nsio_disable(struct device *dev, struct nd_namespace_io *nsio);
+#else
+static inline int devm_nsio_enable(struct device *dev,
+		struct nd_namespace_io *nsio)
+{
+	return -ENXIO;
+}
+static inline void devm_nsio_disable(struct device *dev,
+		struct nd_namespace_io *nsio)
+{
+}
+#endif
 int nd_blk_region_init(struct nd_region *nd_region);
 void __nd_iostat_start(struct bio *bio, unsigned long *start);
 static inline bool nd_iostat_start(struct bio *bio, unsigned long *start)
@@ -285,6 +298,19 @@ static inline bool nd_iostat_start(struct bio *bio, unsigned long *start)
 	return true;
 }
 void nd_iostat_end(struct bio *bio, unsigned long start);
+static inline bool is_bad_pmem(struct badblocks *bb, sector_t sector,
+		unsigned int len)
+{
+	if (bb->count) {
+		sector_t first_bad;
+		int num_bad;
+
+		return !!badblocks_check(bb, sector, len / 512, &first_bad,
+				&num_bad);
+	}
+
+	return false;
+}
 resource_size_t nd_namespace_blk_validate(struct nd_namespace_blk *nsblk);
 const u8 *nd_dev_to_uuid(struct device *dev);
 bool pmem_should_map_pages(struct device *dev);

drivers/nvdimm/pfn_devs.c

@@ -410,8 +410,7 @@ int nd_pfn_validate(struct nd_pfn *nd_pfn)
 }
 EXPORT_SYMBOL(nd_pfn_validate);
 
-int nd_pfn_probe(struct device *dev, struct nd_namespace_common *ndns,
-		void *drvdata)
+int nd_pfn_probe(struct device *dev, struct nd_namespace_common *ndns)
 {
 	int rc;
 	struct nd_pfn *nd_pfn;
@@ -427,7 +426,6 @@ int nd_pfn_probe(struct device *dev, struct nd_namespace_common *ndns,
 	nvdimm_bus_unlock(&ndns->dev);
 	if (!pfn_dev)
 		return -ENOMEM;
-	dev_set_drvdata(pfn_dev, drvdata);
 	pfn_sb = devm_kzalloc(dev, sizeof(*pfn_sb), GFP_KERNEL);
 	nd_pfn = to_nd_pfn(pfn_dev);
 	nd_pfn->pfn_sb = pfn_sb;

include/linux/nd.h

@@ -15,6 +15,7 @@
 #include <linux/fs.h>
 #include <linux/ndctl.h>
 #include <linux/device.h>
+#include <linux/badblocks.h>
 
 enum nvdimm_event {
 	NVDIMM_REVALIDATE_POISON,
@@ -55,13 +56,19 @@ static inline struct nd_namespace_common *to_ndns(struct device *dev)
 }
 
 /**
- * struct nd_namespace_io - infrastructure for loading an nd_pmem instance
+ * struct nd_namespace_io - device representation of a persistent memory range
  * @dev: namespace device created by the nd region driver
  * @res: struct resource conversion of a NFIT SPA table
+ * @size: cached resource_size(@res) for fast path size checks
+ * @addr: virtual address to access the namespace range
+ * @bb: badblocks list for the namespace range
  */
 struct nd_namespace_io {
 	struct nd_namespace_common common;
 	struct resource res;
+	resource_size_t size;
+	void __pmem *addr;
+	struct badblocks bb;
 };
 
 /**

tools/testing/nvdimm/Kbuild

@@ -7,6 +7,7 @@ ldflags-y += --wrap=ioremap_nocache
 ldflags-y += --wrap=iounmap
 ldflags-y += --wrap=memunmap
 ldflags-y += --wrap=__devm_request_region
+ldflags-y += --wrap=__devm_release_region
 ldflags-y += --wrap=__request_region
 ldflags-y += --wrap=__release_region
 ldflags-y += --wrap=devm_memremap_pages

tools/testing/nvdimm/test/iomap.c

@@ -239,13 +239,11 @@ struct resource *__wrap___devm_request_region(struct device *dev,
 }
 EXPORT_SYMBOL(__wrap___devm_request_region);
 
-void __wrap___release_region(struct resource *parent, resource_size_t start,
-				resource_size_t n)
+static bool nfit_test_release_region(struct resource *parent,
+		resource_size_t start, resource_size_t n)
 {
-	struct nfit_test_resource *nfit_res;
-
 	if (parent == &iomem_resource) {
-		nfit_res = get_nfit_res(start);
+		struct nfit_test_resource *nfit_res = get_nfit_res(start);
 		if (nfit_res) {
 			struct resource *res = nfit_res->res + 1;
 
@@ -254,11 +252,26 @@ void __wrap___release_region(struct resource *parent, resource_size_t start,
 						__func__, start, n, res);
 			else
 				memset(res, 0, sizeof(*res));
-			return;
+			return true;
 		}
 	}
-	__release_region(parent, start, n);
+	return false;
+}
+
+void __wrap___release_region(struct resource *parent, resource_size_t start,
+		resource_size_t n)
+{
+	if (!nfit_test_release_region(parent, start, n))
+		__release_region(parent, start, n);
 }
 EXPORT_SYMBOL(__wrap___release_region);
 
+void __wrap___devm_release_region(struct device *dev, struct resource *parent,
+		resource_size_t start, resource_size_t n)
+{
+	if (!nfit_test_release_region(parent, start, n))
+		__devm_release_region(dev, parent, start, n);
+}
+EXPORT_SYMBOL(__wrap___devm_release_region);
+
 MODULE_LICENSE("GPL v2");