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, 240 insertions, 0 deletions

diff --git a/fs/hfsplus/wrapper.c b/fs/hfsplus/wrapper.c
new file mode 100644
index 000000000..30cf4fe78
--- /dev/null
+++ b/fs/hfsplus/wrapper.c
@@ -0,0 +1,240 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ *  linux/fs/hfsplus/wrapper.c
+ *
+ * Copyright (C) 2001
+ * Brad Boyer (flar@allandria.com)
+ * (C) 2003 Ardis Technologies <roman@ardistech.com>
+ *
+ * Handling of HFS wrappers around HFS+ volumes
+ */
+
+#include <linux/fs.h>
+#include <linux/blkdev.h>
+#include <linux/cdrom.h>
+#include <linux/unaligned.h>
+
+#include "hfsplus_fs.h"
+#include "hfsplus_raw.h"
+
+struct hfsplus_wd {
+	u32 ablk_size;
+	u16 ablk_start;
+	u16 embed_start;
+	u16 embed_count;
+};
+
+/**
+ * hfsplus_submit_bio - Perform block I/O
+ * @sb: super block of volume for I/O
+ * @sector: block to read or write, for blocks of HFSPLUS_SECTOR_SIZE bytes
+ * @buf: buffer for I/O
+ * @data: output pointer for location of requested data
+ * @opf: I/O operation type and flags
+ *
+ * The unit of I/O is hfsplus_min_io_size(sb), which may be bigger than
+ * HFSPLUS_SECTOR_SIZE, and @buf must be sized accordingly. On reads
+ * @data will return a pointer to the start of the requested sector,
+ * which may not be the same location as @buf.
+ *
+ * If @sector is not aligned to the bdev logical block size it will
+ * be rounded down. For writes this means that @buf should contain data
+ * that starts at the rounded-down address. As long as the data was
+ * read using hfsplus_submit_bio() and the same buffer is used things
+ * will work correctly.
+ *
+ * Returns: %0 on success else -errno code
+ */
+int hfsplus_submit_bio(struct super_block *sb, sector_t sector,
+		       void *buf, void **data, blk_opf_t opf)
+{
+	u64 io_size = hfsplus_min_io_size(sb);
+	loff_t start = (loff_t)sector << HFSPLUS_SECTOR_SHIFT;
+	int offset = start & (io_size - 1);
+
+	if ((opf & REQ_OP_MASK) != REQ_OP_WRITE && data)
+		*data = (u8 *)buf + offset;
+
+	/*
+	 * Align sector to hardware sector size and find offset. We assume that
+	 * io_size is a power of two, which _should_ be true.
+	 */
+	sector &= ~((io_size >> HFSPLUS_SECTOR_SHIFT) - 1);
+	return bdev_rw_virt(sb->s_bdev, sector, buf, io_size, opf);
+}
+
+static int hfsplus_read_mdb(void *bufptr, struct hfsplus_wd *wd)
+{
+	u32 extent;
+	u16 attrib;
+	__be16 sig;
+
+	sig = *(__be16 *)(bufptr + HFSP_WRAPOFF_EMBEDSIG);
+	if (sig != cpu_to_be16(HFSPLUS_VOLHEAD_SIG) &&
+	    sig != cpu_to_be16(HFSPLUS_VOLHEAD_SIGX))
+		return 0;
+
+	attrib = be16_to_cpu(*(__be16 *)(bufptr + HFSP_WRAPOFF_ATTRIB));
+	if (!(attrib & HFSP_WRAP_ATTRIB_SLOCK) ||
+	   !(attrib & HFSP_WRAP_ATTRIB_SPARED))
+		return 0;
+
+	wd->ablk_size =
+		be32_to_cpu(*(__be32 *)(bufptr + HFSP_WRAPOFF_ABLKSIZE));
+	if (wd->ablk_size < HFSPLUS_SECTOR_SIZE)
+		return 0;
+	if (wd->ablk_size % HFSPLUS_SECTOR_SIZE)
+		return 0;
+	wd->ablk_start =
+		be16_to_cpu(*(__be16 *)(bufptr + HFSP_WRAPOFF_ABLKSTART));
+
+	extent = get_unaligned_be32(bufptr + HFSP_WRAPOFF_EMBEDEXT);
+	wd->embed_start = (extent >> 16) & 0xFFFF;
+	wd->embed_count = extent & 0xFFFF;
+
+	return 1;
+}
+
+static int hfsplus_get_last_session(struct super_block *sb,
+				    sector_t *start, sector_t *size)
+{
+	struct cdrom_device_info *cdi = disk_to_cdi(sb->s_bdev->bd_disk);
+
+	/* default values */
+	*start = 0;
+	*size = bdev_nr_sectors(sb->s_bdev);
+
+	if (HFSPLUS_SB(sb)->session >= 0) {
+		struct cdrom_tocentry te;
+
+		if (!cdi)
+			return -EINVAL;
+
+		te.cdte_track = HFSPLUS_SB(sb)->session;
+		te.cdte_format = CDROM_LBA;
+		if (cdrom_read_tocentry(cdi, &te) ||
+		    (te.cdte_ctrl & CDROM_DATA_TRACK) != 4) {
+			pr_err("invalid session number or type of track\n");
+			return -EINVAL;
+		}
+		*start = (sector_t)te.cdte_addr.lba << 2;
+	} else if (cdi) {
+		struct cdrom_multisession ms_info;
+
+		ms_info.addr_format = CDROM_LBA;
+		if (cdrom_multisession(cdi, &ms_info) == 0 && ms_info.xa_flag)
+			*start = (sector_t)ms_info.addr.lba << 2;
+	}
+
+	return 0;
+}
+
+/* Find the volume header and fill in some minimum bits in superblock */
+/* Takes in super block, returns true if good data read */
+int hfsplus_read_wrapper(struct super_block *sb)
+{
+	struct hfsplus_sb_info *sbi = HFSPLUS_SB(sb);
+	struct hfsplus_wd wd;
+	sector_t part_start, part_size;
+	u32 blocksize;
+	int error = 0;
+
+	error = -EINVAL;
+	blocksize = sb_min_blocksize(sb, HFSPLUS_SECTOR_SIZE);
+	if (!blocksize)
+		goto out;
+
+	sbi->min_io_size = blocksize;
+
+	if (hfsplus_get_last_session(sb, &part_start, &part_size))
+		goto out;
+
+	error = -ENOMEM;
+	sbi->s_vhdr_buf = kmalloc(hfsplus_min_io_size(sb), GFP_KERNEL);
+	if (!sbi->s_vhdr_buf)
+		goto out;
+	sbi->s_backup_vhdr_buf = kmalloc(hfsplus_min_io_size(sb), GFP_KERNEL);
+	if (!sbi->s_backup_vhdr_buf)
+		goto out_free_vhdr;
+
+reread:
+	error = hfsplus_submit_bio(sb, part_start + HFSPLUS_VOLHEAD_SECTOR,
+				   sbi->s_vhdr_buf, (void **)&sbi->s_vhdr,
+				   REQ_OP_READ);
+	if (error)
+		goto out_free_backup_vhdr;
+
+	error = -EINVAL;
+	switch (sbi->s_vhdr->signature) {
+	case cpu_to_be16(HFSPLUS_VOLHEAD_SIGX):
+		set_bit(HFSPLUS_SB_HFSX, &sbi->flags);
+		fallthrough;
+	case cpu_to_be16(HFSPLUS_VOLHEAD_SIG):
+		break;
+	case cpu_to_be16(HFSP_WRAP_MAGIC):
+		if (!hfsplus_read_mdb(sbi->s_vhdr, &wd))
+			goto out_free_backup_vhdr;
+		wd.ablk_size >>= HFSPLUS_SECTOR_SHIFT;
+		part_start += (sector_t)wd.ablk_start +
+			       (sector_t)wd.embed_start * wd.ablk_size;
+		part_size = (sector_t)wd.embed_count * wd.ablk_size;
+		goto reread;
+	default:
+		/*
+		 * Check for a partition block.
+		 *
+		 * (should do this only for cdrom/loop though)
+		 */
+		if (hfs_part_find(sb, &part_start, &part_size))
+			goto out_free_backup_vhdr;
+		goto reread;
+	}
+
+	error = hfsplus_submit_bio(sb, part_start + part_size - 2,
+				   sbi->s_backup_vhdr_buf,
+				   (void **)&sbi->s_backup_vhdr, REQ_OP_READ);
+	if (error)
+		goto out_free_backup_vhdr;
+
+	error = -EINVAL;
+	if (sbi->s_backup_vhdr->signature != sbi->s_vhdr->signature) {
+		pr_warn("invalid secondary volume header\n");
+		goto out_free_backup_vhdr;
+	}
+
+	blocksize = be32_to_cpu(sbi->s_vhdr->blocksize);
+
+	/*
+	 * Block size must be at least as large as a sector and a multiple of 2.
+	 */
+	if (blocksize < HFSPLUS_SECTOR_SIZE || ((blocksize - 1) & blocksize))
+		goto out_free_backup_vhdr;
+	sbi->alloc_blksz = blocksize;
+	sbi->alloc_blksz_shift = ilog2(blocksize);
+	blocksize = min_t(u32, sbi->alloc_blksz, PAGE_SIZE);
+
+	/*
+	 * Align block size to block offset.
+	 */
+	while (part_start & ((blocksize >> HFSPLUS_SECTOR_SHIFT) - 1))
+		blocksize >>= 1;
+
+	if (sb_set_blocksize(sb, blocksize) != blocksize) {
+		pr_err("unable to set blocksize to %u!\n", blocksize);
+		goto out_free_backup_vhdr;
+	}
+
+	sbi->blockoffset =
+		part_start >> (sb->s_blocksize_bits - HFSPLUS_SECTOR_SHIFT);
+	sbi->part_start = part_start;
+	sbi->sect_count = part_size;
+	sbi->fs_shift = sbi->alloc_blksz_shift - sb->s_blocksize_bits;
+	return 0;
+
+out_free_backup_vhdr:
+	kfree(sbi->s_backup_vhdr_buf);
+out_free_vhdr:
+	kfree(sbi->s_vhdr_buf);
+out:
+	return error;
+}