netfs: zero the tail of a short DIO/unbuffered readgrafted

The buffered read collector zero-fills the tail of a short read that
stops below the inode's i_size (netfs_clear_unread()), so a read that
races an extending write still returns the expected number of bytes.
The non-buffered collector path does no such thing: it just records
how much was transferred.

Add the same zero-fill for the non-buffered case, gated on
NETFS_SREQ_CLEAR_TAIL: a subreq's source sets that flag when a short
result from it is known to be safe to treat as a hole, as opposed to
NETFS_SREQ_HIT_EOF, which means the read genuinely ran off the end of
the file and should be reported short as-is. Only CLEAR_TAIL should
zero-fill here; a real EOF must stay a real short read.

No source currently sets CLEAR_TAIL on an unbuffered/DIO subrequest,
so this is inert on its own -- a following change teaches cifs to set
it in the one case that needs it.

Fixes: e2d46f2ec332 ("netfs: Change the read result collector to only use one work item")
Reviewed-by: David Howells <dhowells@redhat.com>
Reviewed-by: Namjae Jeon <linkinjeon@kernel.org>
Signed-off-by: Paulo Alcantara <pc@manguebit.org>
Cc: Christian Brauner <brauner@kernel.org>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Ronnie Sahlberg <ronniesahlberg@gmail.com>
Cc: Shyam Prasad N <sprasad@microsoft.com>
Cc: Tom Talpey <tom@talpey.com>
Cc: Bharath SM <bharathsm@microsoft.com>
Cc: stable@vger.kernel.org

1 files changed, 154 insertions, 0 deletions

diff --git a/drivers/base/map.c b/drivers/base/map.c
new file mode 100644
index 000000000..bde415f93
--- /dev/null
+++ b/drivers/base/map.c
@@ -0,0 +1,154 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ *  linux/drivers/base/map.c
+ *
+ * (C) Copyright Al Viro 2002,2003
+ *
+ * NOTE: data structure needs to be changed.  It works, but for large dev_t
+ * it will be too slow.  It is isolated, though, so these changes will be
+ * local to that file.
+ */
+
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <linux/mutex.h>
+#include <linux/kdev_t.h>
+#include <linux/kobject.h>
+#include <linux/kobj_map.h>
+
+struct kobj_map {
+	struct probe {
+		struct probe *next;
+		dev_t dev;
+		unsigned long range;
+		struct module *owner;
+		kobj_probe_t *get;
+		int (*lock)(dev_t, void *);
+		void *data;
+	} *probes[255];
+	struct mutex *lock;
+};
+
+int kobj_map(struct kobj_map *domain, dev_t dev, unsigned long range,
+	     struct module *module, kobj_probe_t *probe,
+	     int (*lock)(dev_t, void *), void *data)
+{
+	unsigned int n = MAJOR(dev + range - 1) - MAJOR(dev) + 1;
+	unsigned int index = MAJOR(dev);
+	unsigned int i;
+	struct probe *p;
+
+	if (n > 255)
+		n = 255;
+
+	p = kmalloc_objs(struct probe, n);
+	if (p == NULL)
+		return -ENOMEM;
+
+	for (i = 0; i < n; i++, p++) {
+		p->owner = module;
+		p->get = probe;
+		p->lock = lock;
+		p->dev = dev;
+		p->range = range;
+		p->data = data;
+	}
+	mutex_lock(domain->lock);
+	for (i = 0, p -= n; i < n; i++, p++, index++) {
+		struct probe **s = &domain->probes[index % 255];
+		while (*s && (*s)->range < range)
+			s = &(*s)->next;
+		p->next = *s;
+		*s = p;
+	}
+	mutex_unlock(domain->lock);
+	return 0;
+}
+
+void kobj_unmap(struct kobj_map *domain, dev_t dev, unsigned long range)
+{
+	unsigned int n = MAJOR(dev + range - 1) - MAJOR(dev) + 1;
+	unsigned int index = MAJOR(dev);
+	unsigned int i;
+	struct probe *found = NULL;
+
+	if (n > 255)
+		n = 255;
+
+	mutex_lock(domain->lock);
+	for (i = 0; i < n; i++, index++) {
+		struct probe **s;
+		for (s = &domain->probes[index % 255]; *s; s = &(*s)->next) {
+			struct probe *p = *s;
+			if (p->dev == dev && p->range == range) {
+				*s = p->next;
+				if (!found)
+					found = p;
+				break;
+			}
+		}
+	}
+	mutex_unlock(domain->lock);
+	kfree(found);
+}
+
+struct kobject *kobj_lookup(struct kobj_map *domain, dev_t dev, int *index)
+{
+	struct kobject *kobj;
+	struct probe *p;
+	unsigned long best = ~0UL;
+
+retry:
+	mutex_lock(domain->lock);
+	for (p = domain->probes[MAJOR(dev) % 255]; p; p = p->next) {
+		struct kobject *(*probe)(dev_t, int *, void *);
+		struct module *owner;
+		void *data;
+
+		if (p->dev > dev || p->dev + p->range - 1 < dev)
+			continue;
+		if (p->range - 1 >= best)
+			break;
+		if (!try_module_get(p->owner))
+			continue;
+		owner = p->owner;
+		data = p->data;
+		probe = p->get;
+		best = p->range - 1;
+		*index = dev - p->dev;
+		if (p->lock && p->lock(dev, data) < 0) {
+			module_put(owner);
+			continue;
+		}
+		mutex_unlock(domain->lock);
+		kobj = probe(dev, index, data);
+		/* Currently ->owner protects _only_ ->probe() itself. */
+		module_put(owner);
+		if (kobj)
+			return kobj;
+		goto retry;
+	}
+	mutex_unlock(domain->lock);
+	return NULL;
+}
+
+struct kobj_map *kobj_map_init(kobj_probe_t *base_probe, struct mutex *lock)
+{
+	struct kobj_map *p = kmalloc_obj(struct kobj_map);
+	struct probe *base = kzalloc_obj(*base);
+	int i;
+
+	if ((p == NULL) || (base == NULL)) {
+		kfree(p);
+		kfree(base);
+		return NULL;
+	}
+
+	base->dev = 1;
+	base->range = ~0;
+	base->get = base_probe;
+	for (i = 0; i < 255; i++)
+		p->probes[i] = base;
+	p->lock = lock;
+	return p;
+}