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>

3 files changed, 286 insertions, 0 deletions

diff --git a/usr/include/.gitignore b/usr/include/.gitignore
new file mode 100644
index 000000000..17b0ba1bd
--- /dev/null
+++ b/usr/include/.gitignore
@@ -0,0 +1,2 @@
+# SPDX-License-Identifier: GPL-2.0-only
+/*/
diff --git a/usr/include/Makefile b/usr/include/Makefile
new file mode 100644
index 000000000..ee69dd9d9
--- /dev/null
+++ b/usr/include/Makefile
@@ -0,0 +1,188 @@
+# SPDX-License-Identifier: GPL-2.0-only
+
+# Unlike the kernel space, exported headers are written in standard C.
+#  - Forbid C++ style comments
+#  - Use '__inline__', '__asm__' instead of 'inline', 'asm'
+#
+# -std=c90 (equivalent to -ansi) catches the violation of those.
+# We cannot go as far as adding -Wpedantic since it emits too many warnings.
+UAPI_CFLAGS := -std=c90 -Werror=implicit-function-declaration
+UAPI_ARCH := $(or $(HEADER_ARCH),$(SRCARCH))
+
+override c_flags = $(KBUILD_USERCFLAGS) $(UAPI_CFLAGS) -Wp,-MMD,$(depfile)
+cxx_flags = $(filter-out -Wmissing-prototypes -Wstrict-prototypes -std=%, $(KBUILD_USERCFLAGS)) -std=c++98
+
+# The following are excluded for now because they fail to build.
+#
+# Do not add a new header to the blacklist without legitimate reason.
+# Please consider to fix the header first.
+#
+# Sorted alphabetically.
+no-header-test += asm/ucontext.h
+no-header-test += drm/vmwgfx_drm.h
+no-header-test += linux/am437x-vpfe.h
+no-header-test += linux/coda.h
+no-header-test += linux/cyclades.h
+no-header-test += linux/errqueue.h
+no-header-test += linux/hdlc/ioctl.h
+no-header-test += linux/ivtv.h
+no-header-test += linux/matroxfb.h
+no-header-test += linux/omap3isp.h
+no-header-test += linux/omapfb.h
+no-header-test += linux/patchkey.h
+no-header-test += linux/phonet.h
+no-header-test += linux/sctp.h
+no-header-test += linux/sysctl.h
+no-header-test += linux/usb/audio.h
+no-header-test += linux/v4l2-mediabus.h
+no-header-test += linux/v4l2-subdev.h
+no-header-test += linux/videodev2.h
+no-header-test += linux/vm_sockets.h
+no-header-test += sound/asequencer.h
+no-header-test += sound/asoc.h
+no-header-test += sound/asound.h
+no-header-test += sound/compress_offload.h
+no-header-test += sound/emu10k1.h
+no-header-test += sound/sfnt_info.h
+no-header-test += xen/evtchn.h
+no-header-test += xen/gntdev.h
+no-header-test += xen/privcmd.h
+
+# More headers are broken in some architectures
+
+ifneq ($(filter arc openrisc xtensa nios2, $(UAPI_ARCH)),)
+no-header-test += linux/bpf_perf_event.h
+endif
+
+ifeq ($(UAPI_ARCH),sparc)
+no-header-test += asm/uctx.h
+no-header-test += asm/fbio.h
+endif
+
+# asm-generic/*.h is used by asm/*.h, and should not be included directly
+no-header-test += asm-generic/%
+
+# The following are not compatible with C++.
+#
+# Do not add a new header to the list without legitimate reason.
+# Please consider to fix the header first.
+#
+# Sorted alphabetically.
+no-header-test-cxx += linux/auto_dev-ioctl.h
+no-header-test-cxx += linux/map_to_14segment.h
+no-header-test-cxx += linux/map_to_7segment.h
+no-header-test-cxx += linux/netfilter/xt_sctp.h
+no-header-test-cxx += linux/target_core_user.h
+no-header-test-cxx += linux/vhost.h
+no-header-test-cxx += linux/vhost_types.h
+no-header-test-cxx += linux/virtio_net.h
+no-header-test-cxx += linux/virtio_ring.h
+no-header-test-cxx += scsi/fc/fc_els.h
+
+ifeq ($(UAPI_ARCH),x86)
+no-header-test-cxx += asm/elf.h
+endif
+
+# The following are using libc header and types.
+#
+# Do not add a new header to the list without legitimate reason.
+# Please consider to fix the header first.
+#
+# Sorted alphabetically.
+uses-libc += linux/a.out.h
+uses-libc += linux/atmbr2684.h
+uses-libc += linux/auto_dev-ioctl.h
+uses-libc += linux/auto_fs.h
+uses-libc += linux/auto_fs4.h
+uses-libc += linux/btrfs_tree.h
+uses-libc += linux/cec-funcs.h
+uses-libc += linux/cec.h
+uses-libc += linux/dvb/dmx.h
+uses-libc += linux/dvb/video.h
+uses-libc += linux/ethtool.h
+uses-libc += linux/ethtool_netlink.h
+uses-libc += linux/fuse.h
+uses-libc += linux/gsmmux.h
+uses-libc += linux/icmp.h
+uses-libc += linux/idxd.h
+uses-libc += linux/if.h
+uses-libc += linux/if_arp.h
+uses-libc += linux/if_bonding.h
+uses-libc += linux/if_pppox.h
+uses-libc += linux/if_tunnel.h
+uses-libc += linux/input.h
+uses-libc += linux/ip6_tunnel.h
+uses-libc += linux/joystick.h
+uses-libc += linux/llc.h
+uses-libc += linux/mctp.h
+uses-libc += linux/mdio.h
+uses-libc += linux/mii.h
+uses-libc += linux/mptcp.h
+uses-libc += linux/netdevice.h
+uses-libc += linux/netfilter/xt_RATEEST.h
+uses-libc += linux/netfilter/xt_hashlimit.h
+uses-libc += linux/netfilter/xt_physdev.h
+uses-libc += linux/netfilter/xt_rateest.h
+uses-libc += linux/netfilter_arp/arp_tables.h
+uses-libc += linux/netfilter_arp/arpt_mangle.h
+uses-libc += linux/netfilter_bridge.h
+uses-libc += linux/netfilter_bridge/ebtables.h
+uses-libc += linux/netfilter_ipv4.h
+uses-libc += linux/netfilter_ipv4/ip_tables.h
+uses-libc += linux/netfilter_ipv6.h
+uses-libc += linux/netfilter_ipv6/ip6_tables.h
+uses-libc += linux/route.h
+uses-libc += linux/shm.h
+uses-libc += linux/soundcard.h
+uses-libc += linux/string.h
+uses-libc += linux/tipc_config.h
+uses-libc += linux/uhid.h
+uses-libc += linux/uinput.h
+uses-libc += linux/vhost.h
+uses-libc += linux/vhost_types.h
+uses-libc += linux/virtio_ring.h
+uses-libc += linux/wireless.h
+uses-libc += regulator/regulator.h
+uses-libc += scsi/fc/fc_els.h
+
+ifeq ($(UAPI_ARCH),hexagon)
+uses-libc += asm/sigcontext.h
+endif
+
+ifeq ($(UAPI_ARCH),nios2)
+uses-libc += asm/ptrace.h
+endif
+
+ifeq ($(UAPI_ARCH),s390)
+uses-libc += asm/chpid.h
+uses-libc += asm/chsc.h
+endif
+
+always-y := $(patsubst $(obj)/%.h,%.hdrtest, $(shell find $(obj) -name '*.h' 2>/dev/null))
+
+# $(cc-option) forces '-x c' which breaks '-x c++' detection.
+cc-can-compile-cxx := $(call try-run,$(CC) $(CLANG_FLAGS) -c -x c++ /dev/null -o "$$TMP", 1)
+
+target-libc = $(filter $(uses-libc), $*.h)
+target-can-compile = $(filter-out $(no-header-test), $*.h)
+target-can-compile-cxx = $(and $(cc-can-compile-cxx), $(target-can-compile), $(filter-out $(no-header-test-cxx), $*.h))
+
+hdrtest-flags = -fsyntax-only -Werror \
+		-nostdinc -I $(obj) $(if $(target-libc), -I $(srctree)/usr/dummy-include)
+
+# Include the header twice to detect missing include guard.
+quiet_cmd_hdrtest = HDRTEST $<
+      cmd_hdrtest = \
+		$(CC) $(c_flags) $(hdrtest-flags) -x c /dev/null \
+			$(if $(target-can-compile), -include $< -include $<); \
+		$(if $(target-can-compile-cxx), \
+			$(CC) $(cxx_flags) $(hdrtest-flags) -x c++ /dev/null -include $<;) \
+		$(PERL) $(src)/headers_check.pl $(obj) $<; \
+		touch $@
+
+$(obj)/%.hdrtest: $(obj)/%.h $(src)/headers_check.pl FORCE
+	$(call if_changed_dep,hdrtest)
+
+# Since GNU Make 4.3, $(patsubst $(obj)/%/,%,$(wildcard $(obj)/*/)) works.
+# To support older Make versions, use a somewhat tedious way.
+clean-files += $(filter-out Makefile headers_check.pl, $(notdir $(wildcard $(obj)/*)))
diff --git a/usr/include/headers_check.pl b/usr/include/headers_check.pl
new file mode 100755
index 000000000..6cd6eb652
--- /dev/null
+++ b/usr/include/headers_check.pl
@@ -0,0 +1,96 @@
+#!/usr/bin/env perl
+# SPDX-License-Identifier: GPL-2.0
+#
+# headers_check.pl execute a number of trivial consistency checks
+#
+# Usage: headers_check.pl dir [files...]
+# dir:   dir to look for included files
+# files: list of files to check
+#
+# The script reads the supplied files line by line and:
+#
+# 1) for each include statement it checks if the
+#    included file actually exists.
+#    Only include files located in asm* and linux* are checked.
+#    The rest are assumed to be system include files.
+#
+# 2) It is checked that prototypes does not use "extern"
+#
+# 3) Check for leaked CONFIG_ symbols
+
+use warnings;
+use strict;
+use File::Basename;
+
+my ($dir, @files) = @ARGV;
+
+my $ret = 0;
+my $line;
+my $lineno = 0;
+my $filename;
+
+foreach my $file (@files) {
+	$filename = $file;
+
+	open(my $fh, '<', $filename)
+		or die "$filename: $!\n";
+	$lineno = 0;
+	while ($line = <$fh>) {
+		$lineno++;
+		&check_include();
+		&check_asm_types();
+		&check_declarations();
+	}
+	close $fh;
+}
+exit $ret;
+
+sub check_include
+{
+	if ($line =~ m/^\s*#\s*include\s+<((asm|linux).*)>/) {
+		my $inc = $1;
+		my $found;
+		$found = stat($dir . "/" . $inc);
+		if (!$found) {
+			printf STDERR "$filename:$lineno: included file '$inc' is not exported\n";
+			$ret = 1;
+		}
+	}
+}
+
+sub check_declarations
+{
+	# soundcard.h is what it is
+	if ($line =~ m/^void seqbuf_dump\(void\);/) {
+		return;
+	}
+	# drm headers are being C++ friendly
+	if ($line =~ m/^extern "C"/) {
+		return;
+	}
+	if ($line =~ m/^(\s*extern|unsigned|char|short|int|long|void)\b/) {
+		printf STDERR "$filename:$lineno: " .
+			      "userspace cannot reference function or " .
+			      "variable defined in the kernel\n";
+		$ret = 1;
+	}
+}
+
+my $linux_asm_types;
+sub check_asm_types
+{
+	if ($filename =~ /types.h|int-l64.h|int-ll64.h/o) {
+		return;
+	}
+	if ($lineno == 1) {
+		$linux_asm_types = 0;
+	} elsif ($linux_asm_types >= 1) {
+		return;
+	}
+	if ($line =~ m/^\s*#\s*include\s+<asm\/types.h>/) {
+		$linux_asm_types = 1;
+		printf STDERR "$filename:$lineno: " .
+		"include of <linux/types.h> is preferred over <asm/types.h>\n";
+		$ret = 1;
+	}
+}