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, 225 insertions, 0 deletions

diff --git a/drivers/clocksource/timer-integrator-ap.c b/drivers/clocksource/timer-integrator-ap.c
new file mode 100644
index 000000000..a4c700b11
--- /dev/null
+++ b/drivers/clocksource/timer-integrator-ap.c
@@ -0,0 +1,225 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Integrator/AP timer driver
+ * Copyright (C) 2000-2003 Deep Blue Solutions Ltd
+ * Copyright (c) 2014, Linaro Limited
+ */
+
+#include <linux/clk.h>
+#include <linux/clocksource.h>
+#include <linux/of.h>
+#include <linux/of_irq.h>
+#include <linux/of_address.h>
+#include <linux/clockchips.h>
+#include <linux/interrupt.h>
+#include <linux/sched_clock.h>
+
+#include "timer-sp.h"
+
+static void __iomem * sched_clk_base;
+
+static u64 notrace integrator_read_sched_clock(void)
+{
+	return -readl(sched_clk_base + TIMER_VALUE);
+}
+
+static int __init integrator_clocksource_init(unsigned long inrate,
+					      void __iomem *base)
+{
+	u32 ctrl = TIMER_CTRL_ENABLE | TIMER_CTRL_PERIODIC;
+	unsigned long rate = inrate;
+	int ret;
+
+	if (rate >= 1500000) {
+		rate /= 16;
+		ctrl |= TIMER_CTRL_DIV16;
+	}
+
+	writel(0xffff, base + TIMER_LOAD);
+	writel(ctrl, base + TIMER_CTRL);
+
+	ret = clocksource_mmio_init(base + TIMER_VALUE, "timer2",
+				    rate, 200, 16, clocksource_mmio_readl_down);
+	if (ret)
+		return ret;
+
+	sched_clk_base = base;
+	sched_clock_register(integrator_read_sched_clock, 16, rate);
+
+	return 0;
+}
+
+static unsigned long timer_reload;
+static void __iomem * clkevt_base;
+
+/*
+ * IRQ handler for the timer
+ */
+static irqreturn_t integrator_timer_interrupt(int irq, void *dev_id)
+{
+	struct clock_event_device *evt = dev_id;
+
+	/* clear the interrupt */
+	writel(1, clkevt_base + TIMER_INTCLR);
+
+	evt->event_handler(evt);
+
+	return IRQ_HANDLED;
+}
+
+static int clkevt_shutdown(struct clock_event_device *evt)
+{
+	u32 ctrl = readl(clkevt_base + TIMER_CTRL) & ~TIMER_CTRL_ENABLE;
+
+	/* Disable timer */
+	writel(ctrl, clkevt_base + TIMER_CTRL);
+	return 0;
+}
+
+static int clkevt_set_oneshot(struct clock_event_device *evt)
+{
+	u32 ctrl = readl(clkevt_base + TIMER_CTRL) &
+		   ~(TIMER_CTRL_ENABLE | TIMER_CTRL_PERIODIC);
+
+	/* Leave the timer disabled, .set_next_event will enable it */
+	writel(ctrl, clkevt_base + TIMER_CTRL);
+	return 0;
+}
+
+static int clkevt_set_periodic(struct clock_event_device *evt)
+{
+	u32 ctrl = readl(clkevt_base + TIMER_CTRL) & ~TIMER_CTRL_ENABLE;
+
+	/* Disable timer */
+	writel(ctrl, clkevt_base + TIMER_CTRL);
+
+	/* Enable the timer and start the periodic tick */
+	writel(timer_reload, clkevt_base + TIMER_LOAD);
+	ctrl |= TIMER_CTRL_PERIODIC | TIMER_CTRL_ENABLE;
+	writel(ctrl, clkevt_base + TIMER_CTRL);
+	return 0;
+}
+
+static int clkevt_set_next_event(unsigned long next, struct clock_event_device *evt)
+{
+	unsigned long ctrl = readl(clkevt_base + TIMER_CTRL);
+
+	writel(ctrl & ~TIMER_CTRL_ENABLE, clkevt_base + TIMER_CTRL);
+	writel(next, clkevt_base + TIMER_LOAD);
+	writel(ctrl | TIMER_CTRL_ENABLE, clkevt_base + TIMER_CTRL);
+
+	return 0;
+}
+
+static struct clock_event_device integrator_clockevent = {
+	.name			= "timer1",
+	.features		= CLOCK_EVT_FEAT_PERIODIC |
+				  CLOCK_EVT_FEAT_ONESHOT,
+	.set_state_shutdown	= clkevt_shutdown,
+	.set_state_periodic	= clkevt_set_periodic,
+	.set_state_oneshot	= clkevt_set_oneshot,
+	.tick_resume		= clkevt_shutdown,
+	.set_next_event		= clkevt_set_next_event,
+	.rating			= 300,
+};
+
+static int integrator_clockevent_init(unsigned long inrate,
+				      void __iomem *base, int irq)
+{
+	unsigned long rate = inrate;
+	unsigned int ctrl = 0;
+	int ret;
+
+	clkevt_base = base;
+	/* Calculate and program a divisor */
+	if (rate > 0x100000 * HZ) {
+		rate /= 256;
+		ctrl |= TIMER_CTRL_DIV256;
+	} else if (rate > 0x10000 * HZ) {
+		rate /= 16;
+		ctrl |= TIMER_CTRL_DIV16;
+	}
+	timer_reload = rate / HZ;
+	writel(ctrl, clkevt_base + TIMER_CTRL);
+
+	ret = request_irq(irq, integrator_timer_interrupt,
+			  IRQF_TIMER | IRQF_IRQPOLL, "timer",
+			  &integrator_clockevent);
+	if (ret)
+		return ret;
+
+	clockevents_config_and_register(&integrator_clockevent,
+					rate,
+					1,
+					0xffffU);
+	return 0;
+}
+
+static int __init integrator_ap_timer_init_of(struct device_node *node)
+{
+	const char *path;
+	void __iomem *base;
+	int err;
+	int irq;
+	struct clk *clk;
+	unsigned long rate;
+	struct device_node *alias_node;
+
+	base = of_io_request_and_map(node, 0, "integrator-timer");
+	if (IS_ERR(base))
+		return PTR_ERR(base);
+
+	clk = of_clk_get(node, 0);
+	if (IS_ERR(clk)) {
+		pr_err("No clock for %pOFn\n", node);
+		return PTR_ERR(clk);
+	}
+	clk_prepare_enable(clk);
+	rate = clk_get_rate(clk);
+	writel(0, base + TIMER_CTRL);
+
+	err = of_property_read_string(of_aliases,
+				"arm,timer-primary", &path);
+	if (err) {
+		pr_warn("Failed to read property\n");
+		return err;
+	}
+
+	alias_node = of_find_node_by_path(path);
+
+	/*
+	 * The pointer is used as an identifier not as a pointer, we
+	 * can drop the refcount on the of__node immediately after
+	 * getting it.
+	 */
+	of_node_put(alias_node);
+
+	if (node == alias_node)
+		/* The primary timer lacks IRQ, use as clocksource */
+		return integrator_clocksource_init(rate, base);
+
+	err = of_property_read_string(of_aliases,
+				"arm,timer-secondary", &path);
+	if (err) {
+		pr_warn("Failed to read property\n");
+		return err;
+	}
+
+	alias_node = of_find_node_by_path(path);
+
+	of_node_put(alias_node);
+
+	if (node == alias_node) {
+		/* The secondary timer will drive the clock event */
+		irq = irq_of_parse_and_map(node, 0);
+		return integrator_clockevent_init(rate, base, irq);
+	}
+
+	pr_info("Timer @%p unused\n", base);
+	clk_disable_unprepare(clk);
+
+	return 0;
+}
+
+TIMER_OF_DECLARE(integrator_ap_timer, "arm,integrator-timer",
+		       integrator_ap_timer_init_of);