virtio_blk: set the zone write granularitygrafted

virtblk_read_zoned_limits() reads the write granularity that the device
reports in virtio_blk_zoned_characteristics and assigns it to the
physical block size and to io_min, but never to the limit that is named
after it. queue_limits.zone_write_granularity is left at zero, so
blk_validate_zoned_limits() raises it to the logical block size:

	if (lim->zone_write_granularity < lim->logical_block_size)
		lim->zone_write_granularity = lim->logical_block_size;

A device that reports a granularity coarser than its logical block size,
which is what the field exists to express, therefore has it silently
reduced. A 512e host managed disk passed through to a guest reports a
logical block size of 512 and a write granularity of 4096, and the guest
ends up with a zone write granularity of 512.

bio_split_alignment() returns lim->zone_write_granularity if it is non-zero
and bio_split_io_at() may split a bio with as per bio_split_alignment().
This can real to the write getting rejected by the host drive, as the write
is not aligned to the physical block size.

zonefs also takes its block size from bdev_zone_write_granularity(), so it
would incorrectly use 512 on a disk that requires 4096.

sd_zbc_read_zones() sets the limit from the physical block size for the
same reason. NVMe ZNS and null_blk leave it unset, but the fallback
gives the right answer for them, as their write granularity is the
logical block size. virtio carries a separate value that may exceed it.

Set the zone write granularity from the value that the device reports.

Fixes: 95bfec41bd3d ("virtio-blk: add support for zoned block devices")
Signed-off-by: Niklas Cassel <cassel@kernel.org>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Link: https://patch.msgid.link/20260918140641.2031075-2-cassel@kernel.org
Signed-off-by: Jens Axboe <axboe@kernel.dk>

1 files changed, 212 insertions, 0 deletions

diff --git a/samples/livepatch/livepatch-shadow-mod.c b/samples/livepatch/livepatch-shadow-mod.c
new file mode 100644
index 000000000..5d83ad5a8
--- /dev/null
+++ b/samples/livepatch/livepatch-shadow-mod.c
@@ -0,0 +1,212 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (C) 2017 Joe Lawrence <joe.lawrence@redhat.com>
+ */
+
+/*
+ * livepatch-shadow-mod.c - Shadow variables, buggy module demo
+ *
+ * Purpose
+ * -------
+ *
+ * As a demonstration of livepatch shadow variable API, this module
+ * introduces memory leak behavior that livepatch modules
+ * livepatch-shadow-fix1.ko and livepatch-shadow-fix2.ko correct and
+ * enhance.
+ *
+ * WARNING - even though the livepatch-shadow-fix modules patch the
+ * memory leak, please load these modules at your own risk -- some
+ * amount of memory may leaked before the bug is patched.
+ *
+ *
+ * Usage
+ * -----
+ *
+ * Step 1 - Load the buggy demonstration module:
+ *
+ *   insmod samples/livepatch/livepatch-shadow-mod.ko
+ *
+ * Watch dmesg output for a few moments to see new dummy being allocated
+ * and a periodic cleanup check.  (Note: a small amount of memory is
+ * being leaked.)
+ *
+ *
+ * Step 2 - Load livepatch fix1:
+ *
+ *   insmod samples/livepatch/livepatch-shadow-fix1.ko
+ *
+ * Continue watching dmesg and note that now livepatch_fix1_dummy_free()
+ * and livepatch_fix1_dummy_alloc() are logging messages about leaked
+ * memory and eventually leaks prevented.
+ *
+ *
+ * Step 3 - Load livepatch fix2 (on top of fix1):
+ *
+ *   insmod samples/livepatch/livepatch-shadow-fix2.ko
+ *
+ * This module extends functionality through shadow variables, as a new
+ * "check" counter is added to the dummy structure.  Periodic dmesg
+ * messages will log these as dummies are cleaned up.
+ *
+ *
+ * Step 4 - Cleanup
+ *
+ * Unwind the demonstration by disabling the livepatch fix modules, then
+ * removing them and the demo module:
+ *
+ *   echo 0 > /sys/kernel/livepatch/livepatch_shadow_fix2/enabled
+ *   echo 0 > /sys/kernel/livepatch/livepatch_shadow_fix1/enabled
+ *   rmmod livepatch-shadow-fix2
+ *   rmmod livepatch-shadow-fix1
+ *   rmmod livepatch-shadow-mod
+ */
+
+
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/sched.h>
+#include <linux/slab.h>
+#include <linux/stat.h>
+#include <linux/workqueue.h>
+
+MODULE_LICENSE("GPL");
+MODULE_AUTHOR("Joe Lawrence <joe.lawrence@redhat.com>");
+MODULE_DESCRIPTION("Buggy module for shadow variable demo");
+
+/* Allocate new dummies every second */
+#define ALLOC_PERIOD	1
+/* Check for expired dummies after a few new ones have been allocated */
+#define CLEANUP_PERIOD	(3 * ALLOC_PERIOD)
+/* Dummies expire after a few cleanup instances */
+#define EXPIRE_PERIOD	(4 * CLEANUP_PERIOD)
+
+/*
+ * Keep a list of all the dummies so we can clean up any residual ones
+ * on module exit
+ */
+static LIST_HEAD(dummy_list);
+static DEFINE_MUTEX(dummy_list_mutex);
+
+struct dummy {
+	struct list_head list;
+	unsigned long jiffies_expire;
+};
+
+static __used noinline struct dummy *dummy_alloc(void)
+{
+	struct dummy *d;
+	int *leak;
+
+	d = kzalloc(sizeof(*d), GFP_KERNEL);
+	if (!d)
+		return NULL;
+
+	d->jiffies_expire = jiffies + secs_to_jiffies(EXPIRE_PERIOD);
+
+	/* Oops, forgot to save leak! */
+	leak = kzalloc(sizeof(*leak), GFP_KERNEL);
+	if (!leak) {
+		kfree(d);
+		return NULL;
+	}
+
+	pr_info("%s: dummy @ %p, expires @ %lx\n",
+		__func__, d, d->jiffies_expire);
+
+	return d;
+}
+
+static __used noinline void dummy_free(struct dummy *d)
+{
+	pr_info("%s: dummy @ %p, expired = %lx\n",
+		__func__, d, d->jiffies_expire);
+
+	kfree(d);
+}
+
+static __used noinline bool dummy_check(struct dummy *d,
+					   unsigned long jiffies)
+{
+	return time_after(jiffies, d->jiffies_expire);
+}
+
+/*
+ * alloc_work_func: allocates new dummy structures, allocates additional
+ *                  memory, aptly named "leak", but doesn't keep
+ *                  permanent record of it.
+ */
+
+static void alloc_work_func(struct work_struct *work);
+static DECLARE_DELAYED_WORK(alloc_dwork, alloc_work_func);
+
+static void alloc_work_func(struct work_struct *work)
+{
+	struct dummy *d;
+
+	d = dummy_alloc();
+	if (!d)
+		return;
+
+	mutex_lock(&dummy_list_mutex);
+	list_add(&d->list, &dummy_list);
+	mutex_unlock(&dummy_list_mutex);
+
+	schedule_delayed_work(&alloc_dwork, secs_to_jiffies(ALLOC_PERIOD));
+}
+
+/*
+ * cleanup_work_func: frees dummy structures.  Without knownledge of
+ *                    "leak", it leaks the additional memory that
+ *                    alloc_work_func created.
+ */
+
+static void cleanup_work_func(struct work_struct *work);
+static DECLARE_DELAYED_WORK(cleanup_dwork, cleanup_work_func);
+
+static void cleanup_work_func(struct work_struct *work)
+{
+	struct dummy *d, *tmp;
+	unsigned long j;
+
+	j = jiffies;
+	pr_info("%s: jiffies = %lx\n", __func__, j);
+
+	mutex_lock(&dummy_list_mutex);
+	list_for_each_entry_safe(d, tmp, &dummy_list, list) {
+
+		/* Kick out and free any expired dummies */
+		if (dummy_check(d, j)) {
+			list_del(&d->list);
+			dummy_free(d);
+		}
+	}
+	mutex_unlock(&dummy_list_mutex);
+
+	schedule_delayed_work(&cleanup_dwork, secs_to_jiffies(CLEANUP_PERIOD));
+}
+
+static int livepatch_shadow_mod_init(void)
+{
+	schedule_delayed_work(&alloc_dwork, secs_to_jiffies(ALLOC_PERIOD));
+	schedule_delayed_work(&cleanup_dwork, secs_to_jiffies(CLEANUP_PERIOD));
+
+	return 0;
+}
+
+static void livepatch_shadow_mod_exit(void)
+{
+	struct dummy *d, *tmp;
+
+	/* Wait for any dummies at work */
+	cancel_delayed_work_sync(&alloc_dwork);
+	cancel_delayed_work_sync(&cleanup_dwork);
+
+	/* Cleanup residual dummies */
+	list_for_each_entry_safe(d, tmp, &dummy_list, list) {
+		list_del(&d->list);
+		dummy_free(d);
+	}
+}
+
+module_init(livepatch_shadow_mod_init);
+module_exit(livepatch_shadow_mod_exit);