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, 215 insertions, 0 deletions

diff --git a/drivers/mtd/tests/readtest.c b/drivers/mtd/tests/readtest.c
new file mode 100644
index 000000000..99670ef91
--- /dev/null
+++ b/drivers/mtd/tests/readtest.c
@@ -0,0 +1,215 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2006-2008 Nokia Corporation
+ *
+ * Check MTD device read.
+ *
+ * Author: Adrian Hunter <ext-adrian.hunter@nokia.com>
+ */
+
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
+
+#include <linux/init.h>
+#include <linux/module.h>
+#include <linux/moduleparam.h>
+#include <linux/err.h>
+#include <linux/mtd/mtd.h>
+#include <linux/slab.h>
+#include <linux/sched.h>
+
+#include "mtd_test.h"
+
+static int dev = -EINVAL;
+module_param(dev, int, S_IRUGO);
+MODULE_PARM_DESC(dev, "MTD device number to use");
+
+static struct mtd_info *mtd;
+static unsigned char *iobuf;
+static unsigned char *iobuf1;
+static unsigned char *bbt;
+
+static int pgsize;
+static int ebcnt;
+static int pgcnt;
+
+static int read_eraseblock_by_page(int ebnum)
+{
+	int i, ret, err = 0;
+	loff_t addr = (loff_t)ebnum * mtd->erasesize;
+	void *buf = iobuf;
+	void *oobbuf = iobuf1;
+
+	for (i = 0; i < pgcnt; i++) {
+		memset(buf, 0 , pgsize);
+		ret = mtdtest_read(mtd, addr, pgsize, buf);
+		if (ret) {
+			if (!err)
+				err = ret;
+		}
+		if (mtd->oobsize) {
+			struct mtd_oob_ops ops = { };
+
+			ops.mode      = MTD_OPS_PLACE_OOB;
+			ops.len       = 0;
+			ops.retlen    = 0;
+			ops.ooblen    = mtd->oobsize;
+			ops.oobretlen = 0;
+			ops.ooboffs   = 0;
+			ops.datbuf    = NULL;
+			ops.oobbuf    = oobbuf;
+			ret = mtd_read_oob(mtd, addr, &ops);
+			if ((ret && !mtd_is_bitflip(ret)) ||
+					ops.oobretlen != mtd->oobsize) {
+				pr_err("error: read oob failed at "
+						  "%#llx\n", (long long)addr);
+				if (!err)
+					err = ret;
+				if (!err)
+					err = -EINVAL;
+			}
+			oobbuf += mtd->oobsize;
+		}
+		addr += pgsize;
+		buf += pgsize;
+	}
+
+	return err;
+}
+
+static void dump_eraseblock(int ebnum)
+{
+	int i, j, n;
+	char line[128];
+	int pg, oob;
+
+	pr_info("dumping eraseblock %d\n", ebnum);
+	n = mtd->erasesize;
+	for (i = 0; i < n;) {
+		char *p = line;
+
+		p += sprintf(p, "%05x: ", i);
+		for (j = 0; j < 32 && i < n; j++, i++)
+			p += sprintf(p, "%02x", (unsigned int)iobuf[i]);
+		printk(KERN_CRIT "%s\n", line);
+		cond_resched();
+	}
+	if (!mtd->oobsize)
+		return;
+	pr_info("dumping oob from eraseblock %d\n", ebnum);
+	n = mtd->oobsize;
+	for (pg = 0, i = 0; pg < pgcnt; pg++)
+		for (oob = 0; oob < n;) {
+			char *p = line;
+
+			p += sprintf(p, "%05x: ", i);
+			for (j = 0; j < 32 && oob < n; j++, oob++, i++)
+				p += sprintf(p, "%02x",
+					     (unsigned int)iobuf1[i]);
+			printk(KERN_CRIT "%s\n", line);
+			cond_resched();
+		}
+}
+
+static int __init mtd_readtest_init(void)
+{
+	uint64_t tmp;
+	int err, i;
+
+	printk(KERN_INFO "\n");
+	printk(KERN_INFO "=================================================\n");
+
+	if (dev < 0) {
+		pr_info("Please specify a valid mtd-device via module parameter\n");
+		return -EINVAL;
+	}
+
+	pr_info("MTD device: %d\n", dev);
+
+	mtd = get_mtd_device(NULL, dev);
+	if (IS_ERR(mtd)) {
+		err = PTR_ERR(mtd);
+		pr_err("error: Cannot get MTD device\n");
+		return err;
+	}
+
+	if (mtd->writesize == 1) {
+		pr_info("not NAND flash, assume page size is 512 "
+		       "bytes.\n");
+		pgsize = 512;
+	} else
+		pgsize = mtd->writesize;
+
+	tmp = mtd->size;
+	do_div(tmp, mtd->erasesize);
+	ebcnt = tmp;
+	pgcnt = mtd->erasesize / pgsize;
+
+	pr_info("MTD device size %llu, eraseblock size %u, "
+	       "page size %u, count of eraseblocks %u, pages per "
+	       "eraseblock %u, OOB size %u\n",
+	       (unsigned long long)mtd->size, mtd->erasesize,
+	       pgsize, ebcnt, pgcnt, mtd->oobsize);
+
+	err = -ENOMEM;
+	iobuf = kmalloc(mtd->erasesize, GFP_KERNEL);
+	if (!iobuf)
+		goto out;
+	iobuf1 = kmalloc(mtd->erasesize, GFP_KERNEL);
+	if (!iobuf1)
+		goto out;
+
+	bbt = kzalloc(ebcnt, GFP_KERNEL);
+	if (!bbt)
+		goto out;
+	err = mtdtest_scan_for_bad_eraseblocks(mtd, bbt, 0, ebcnt);
+	if (err)
+		goto out;
+
+	/* Read all eraseblocks 1 page at a time */
+	pr_info("testing page read\n");
+	for (i = 0; i < ebcnt; ++i) {
+		int ret;
+
+		if (bbt[i])
+			continue;
+		ret = read_eraseblock_by_page(i);
+		if (ret) {
+			dump_eraseblock(i);
+			if (!err)
+				err = ret;
+		}
+
+		ret = mtdtest_relax();
+		if (ret) {
+			err = ret;
+			goto out;
+		}
+	}
+
+	if (err)
+		pr_info("finished with errors\n");
+	else
+		pr_info("finished\n");
+
+out:
+
+	kfree(iobuf);
+	kfree(iobuf1);
+	kfree(bbt);
+	put_mtd_device(mtd);
+	if (err)
+		pr_info("error %d occurred\n", err);
+	printk(KERN_INFO "=================================================\n");
+	return err;
+}
+module_init(mtd_readtest_init);
+
+static void __exit mtd_readtest_exit(void)
+{
+	return;
+}
+module_exit(mtd_readtest_exit);
+
+MODULE_DESCRIPTION("Read test module");
+MODULE_AUTHOR("Adrian Hunter");
+MODULE_LICENSE("GPL");