[PATCH v6 6/6] pmem: implement pmem_recovery_write()

[Jane Chu <jane.chu@oracle.com>]

The recovery write thread started out as a normal pwrite thread and
when the filesystem was told about potential media error in the
range, filesystem turns the normal pwrite to a dax_recovery_write.

The recovery write consists of clearing media poison, clearing page
HWPoison bit, reenable page-wide read-write permission, flush the
caches and finally write.  A competing pread thread will be held
off during the recovery process since data read back might not be
valid, and this is achieved by clearing the badblock records after
the recovery write is complete. Competing recovery write threads
are serialized by pmem device level .recovery_lock.

Signed-off-by: Jane Chu <jane.chu@oracle.com>
---
 drivers/nvdimm/pmem.c | 49 ++++++++++++++++++++++++++++++++++++++++---
 drivers/nvdimm/pmem.h |  1 +
 2 files changed, 47 insertions(+), 3 deletions(-)

diff --git a/drivers/nvdimm/pmem.c b/drivers/nvdimm/pmem.c
index 18f357fbef69..c4fbd9426075 100644
--- a/drivers/nvdimm/pmem.c
+++ b/drivers/nvdimm/pmem.c
@@ -321,15 +321,57 @@ static const struct dax_operations pmem_dax_ops = {
 	.zero_page_range = pmem_dax_zero_page_range,
 };
 
+/*
+ * The recovery write thread started out as a normal pwrite thread and
+ * when the filesystem was told about potential media error in the
+ * range, filesystem turns the normal pwrite to a dax_recovery_write.
+ *
+ * The recovery write consists of clearing media poison, clearing page
+ * HWPoison bit, reenable page-wide read-write permission, flush the
+ * caches and finally write.  A competing pread thread will be held
+ * off during the recovery process since data read back might not be
+ * valid, and this is achieved by clearing the badblock records after
+ * the recovery write is complete. Competing recovery write threads
+ * are serialized by pmem device level .recovery_lock.
+ */
 static size_t pmem_recovery_write(struct dev_pagemap *pgmap, pgoff_t pgoff,
 		void *addr, size_t bytes, void *iter)
 {
 	struct pmem_device *pmem = pgmap->owner;
+	size_t olen, len, off;
+	phys_addr_t pmem_off;
+	struct device *dev = pmem->bb.dev;
+	unsigned int blks;
 
-	dev_warn(pmem->bb.dev, "%s: not yet implemented\n", __func__);
+	off = (unsigned long)addr & ~PAGE_MASK;
+	len = PFN_PHYS(PFN_UP(off + bytes));
+	if (!is_bad_pmem(&pmem->bb, PFN_PHYS(pgoff) / 512, len))
+		return _copy_from_iter_flushcache(addr, bytes, iter);
 
-	/* XXX more later */
-	return 0;
+	/*
+	 * Not page-aligned range cannot be recovered. This should not
+	 * happen unless something else went wrong.
+	 */
+	if (off || !(PAGE_ALIGNED(bytes))) {
+		dev_warn(dev, "Found poison, but addr(%p) or bytes(%#lx) not page aligned\n",
+			addr, bytes);
+		return (size_t) -EIO;
+	}
+
+	mutex_lock(&pmem->recovery_lock);
+	pmem_off = PFN_PHYS(pgoff) + pmem->data_offset;
+	if (__pmem_clear_poison(pmem, pmem_off, len, &blks) != BLK_STS_OK) {
+		dev_warn(dev, "pmem dimm poison clearing failed\n");
+		mutex_unlock(&pmem->recovery_lock);
+		return (size_t) -EIO;
+	}
+
+	olen = _copy_from_iter_flushcache(addr, bytes, iter);
+	if (blks > 0)
+		clear_bb(pmem, to_sect(pmem, pmem_off), blks);
+
+	mutex_unlock(&pmem->recovery_lock);
+	return olen;
 }
 
 static ssize_t write_cache_show(struct device *dev,
@@ -520,6 +562,7 @@ static int pmem_attach_disk(struct device *dev,
 	if (rc)
 		goto out_cleanup_dax;
 	dax_write_cache(dax_dev, nvdimm_has_cache(nd_region));
+	mutex_init(&pmem->recovery_lock);
 	pmem->dax_dev = dax_dev;
 
 	rc = device_add_disk(dev, disk, pmem_attribute_groups);
diff --git a/drivers/nvdimm/pmem.h b/drivers/nvdimm/pmem.h
index 823eeffa7298..bad7d1b05691 100644
--- a/drivers/nvdimm/pmem.h
+++ b/drivers/nvdimm/pmem.h
@@ -24,6 +24,7 @@ struct pmem_device {
 	struct dax_device	*dax_dev;
 	struct gendisk		*disk;
 	struct dev_pagemap	pgmap;
+	struct mutex		recovery_lock;
 };
 
 long __pmem_direct_access(struct pmem_device *pmem, pgoff_t pgoff,
-- 
2.18.4