[PATCH v2 0/6] nvme: few fixups

[PATCH v2 0/6] nvme: few fixups

[dmukhin]

The series adds few fixes for NVMe subsystem.

Patch 1 adds logline for tracing misbehaving NVMe drives.
Patch 2 adds plumbing for blk_flush()/blk_dflush().
Patch 3 introduces NVMe flush command.
Patch 4 enables `flush` for the currently selected block device.
Patch 5 makes nvme_shutdown() symbol public to enable use of it from
        board_quiesce_devices()
Patch 6 tiny fixup for QEMU command for manual NVMe tests.

Denis Mukhin (6):
  drivers: nvme: Log I/O timeouts
  drivers: block: Introduce blk_flush()/blk_dflush()
  drivers: nvme: Implement flush command
  cmd: Add flush support for all blk devices
  drivers: nvme: Export nvme_shutdown()
  docs: nvme: Update QEMU command for testing

 cmd/blk_common.c                  | 14 ++++++++++++++
 cmd/ide.c                         |  1 +
 cmd/nvme.c                        |  1 +
 cmd/pvblock.c                     |  1 +
 cmd/sata.c                        |  1 +
 cmd/scsi.c                        |  1 +
 cmd/usb.c                         |  1 +
 cmd/virtio.c                      |  1 +
 disk/disk-uclass.c                |  6 ++++++
 doc/develop/driver-model/nvme.rst |  2 +-
 drivers/block/blk-uclass.c        | 18 ++++++++++++++++++
 drivers/nvme/nvme.c               | 24 +++++++++++++++++++++++-
 drivers/nvme/nvme.h               |  7 -------
 include/blk.h                     | 28 ++++++++++++++++++++++++++++
 include/nvme.h                    | 11 +++++++++++
 include/part.h                    |  8 ++++++++
 16 files changed, 116 insertions(+), 9 deletions(-)

-- 
2.54.0

[PATCH v2 1/6] drivers: nvme: Log I/O timeouts

[dmukhin]

From: Denis Mukhin <dmukhin@ford.com> 

Current code silently swallows any timed-out commands scheduled
to NVMe. Log those to be able to debug any potential problems with
the NVMe hardware/firmware.

Signed-off-by: Denis Mukhin <dmukhin@ford.com>
Reviewed-by: Neil Armstrong <neil.armstrong@linaro.org>
---
Changes since v1:
- added Neil's R-b
---
 drivers/nvme/nvme.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/nvme/nvme.c b/drivers/nvme/nvme.c
index 0631b190b978..c3c44e50f19a 100644
--- a/drivers/nvme/nvme.c
+++ b/drivers/nvme/nvme.c
@@ -182,8 +182,10 @@ static int nvme_submit_sync_cmd(struct nvme_queue *nvmeq,
 		if ((status & 0x01) == phase)
 			break;
 		if (timeout_us > 0 && (timer_get_us() - start_time)
-		    >= timeout_us)
+		    >= timeout_us) {
+			pr_warn("nvme: cmd %#x timed out\n", cmd->common.command_id);
 			return -ETIMEDOUT;
+		}
 	}
 
 	ops = (struct nvme_ops *)nvmeq->dev->udev->driver->ops;
-- 
2.54.0

Re: [PATCH v2 1/6] drivers: nvme: Log I/O timeouts

[Simon Glass]

Hi Denis,

On 2026-05-29T03:44:35, None <dmukhin@ford.com> wrote:
> drivers: nvme: Log I/O timeouts
>
> Current code silently swallows any timed-out commands scheduled
> to NVMe. Log those to be able to debug any potential problems with
> the NVMe hardware/firmware.
>
> Signed-off-by: Denis Mukhin <dmukhin@ford.com>
> Reviewed-by: Neil Armstrong <neil.armstrong@linaro.org>
>
> drivers/nvme/nvme.c | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)

> diff --git a/drivers/nvme/nvme.c b/drivers/nvme/nvme.c
> @@ -179,8 +179,10 @@ static int nvme_submit_sync_cmd(struct nvme_queue *nvmeq,
>               if ((status & 0x01) == phase)
>                       break;
>               if (timeout_us > 0 && (timer_get_us() - start_time)
> -                 >= timeout_us)
> +                 >= timeout_us) {
> +                     pr_warn("nvme: cmd %#x timed out\n", cmd->common.command_id);
> +                     return -ETIMEDOUT;
> +             }

Reviewed-by: Simon Glass <sjg@chromium.org>

A few suggestions:

- command_id is just an incrementing sequence (see nvme_get_cmd_id()),
so on its own it tells you very little. How about logging
cmd->common.opcode as well - that identifies the command type and is
the useful bit for debugging.

- With multiple NVMe controllers this message won't say which one
stalled. Since nvmeq->dev->udev is to hand, dev_warn() would be more
informative than pr_warn(), and would drop the need for the 'nvme: '
prefix. Or you could use log_warning() and print out what you want

- The new line is ~85 columns; please wrap it.

Regards,
Simon

[PATCH v2 2/6] drivers: block: Introduce blk_flush()/blk_dflush()

[dmukhin]

From: Denis Mukhin <dmukhin@ford.com> 

Add generic flush operations for committing dirty data to the storage
device.

This provides a common block-layer interface for flushing pending
writes, allowing callers to ensure that data buffered by the block
device or its backing implementation is written out to the underlying
storage.

Signed-off-by: Denis Mukhin <dmukhin@ford.com>
---
Changes since v1:
- new patch
---
 disk/disk-uclass.c         |  6 ++++++
 drivers/block/blk-uclass.c | 18 ++++++++++++++++++
 include/blk.h              | 28 ++++++++++++++++++++++++++++
 include/part.h             |  8 ++++++++
 4 files changed, 60 insertions(+)

diff --git a/disk/disk-uclass.c b/disk/disk-uclass.c
index ee3cc4407d76..1a87f7dffa7c 100644
--- a/disk/disk-uclass.c
+++ b/disk/disk-uclass.c
@@ -122,6 +122,11 @@ unsigned long disk_blk_erase(struct udevice *dev, lbaint_t start,
 			 blkcnt);
 }
 
+unsigned long disk_blk_flush(struct udevice *dev)
+{
+	return blk_flush(dev_get_parent(dev));
+}
+
 UCLASS_DRIVER(partition) = {
 	.id		= UCLASS_PARTITION,
 	.per_device_plat_auto	= sizeof(struct disk_part),
@@ -132,6 +137,7 @@ static const struct blk_ops blk_part_ops = {
 	.read	= disk_blk_read,
 	.write	= disk_blk_write,
 	.erase	= disk_blk_erase,
+	.flush	= disk_blk_flush,
 };
 
 U_BOOT_DRIVER(blk_partition) = {
diff --git a/drivers/block/blk-uclass.c b/drivers/block/blk-uclass.c
index 73c24fd91763..0be1fdab1ba5 100644
--- a/drivers/block/blk-uclass.c
+++ b/drivers/block/blk-uclass.c
@@ -514,6 +514,19 @@ long blk_erase(struct udevice *dev, lbaint_t start, lbaint_t blkcnt)
 	return ops->erase(dev, start, blkcnt);
 }
 
+long blk_flush(struct udevice *dev)
+{
+	struct blk_desc *desc = dev_get_uclass_plat(dev);
+	const struct blk_ops *ops = blk_get_ops(dev);
+
+	if (!ops->flush)
+		return -ENOSYS;
+
+	blkcache_invalidate(desc->uclass_id, desc->devnum);
+
+	return ops->flush(dev);
+}
+
 ulong blk_dread(struct blk_desc *desc, lbaint_t start, lbaint_t blkcnt,
 		void *buffer)
 {
@@ -531,6 +544,11 @@ ulong blk_derase(struct blk_desc *desc, lbaint_t start, lbaint_t blkcnt)
 	return blk_erase(desc->bdev, start, blkcnt);
 }
 
+ulong blk_dflush(struct blk_desc *desc)
+{
+	return blk_flush(desc->bdev);
+}
+
 int blk_find_from_parent(struct udevice *parent, struct udevice **devp)
 {
 	struct udevice *dev;
diff --git a/include/blk.h b/include/blk.h
index 8d1b70cabd31..98e1db7a58f1 100644
--- a/include/blk.h
+++ b/include/blk.h
@@ -99,6 +99,7 @@ struct blk_desc {
 	unsigned long	(*block_erase)(struct blk_desc *block_dev,
 				       lbaint_t start,
 				       lbaint_t blkcnt);
+	unsigned long	(*block_flush)(struct blk_desc *block_dev);
 	void		*priv;		/* driver private struct pointer */
 #endif
 };
@@ -275,6 +276,14 @@ struct blk_ops {
 	 */
 	int (*buffer_aligned)(struct udevice *dev, struct bounce_buffer *state);
 #endif	/* CONFIG_BOUNCE_BUFFER */
+
+	/**
+	 * flush() -  commit all dirty data to storage
+	 *
+	 * @dev:	Device to flush
+	 * @return 0 if OK, -ve on error
+	 */
+	unsigned long (*flush)(struct udevice *dev);
 };
 
 #if CONFIG_IS_ENABLED(BLK)
@@ -291,6 +300,7 @@ unsigned long blk_dwrite(struct blk_desc *block_dev, lbaint_t start,
 			 lbaint_t blkcnt, const void *buffer);
 unsigned long blk_derase(struct blk_desc *block_dev, lbaint_t start,
 			 lbaint_t blkcnt);
+unsigned long blk_dflush(struct blk_desc *block_dev);
 
 #endif /* BLK */
 
@@ -331,6 +341,14 @@ long blk_write(struct udevice *dev, lbaint_t start, lbaint_t blkcnt,
  */
 long blk_erase(struct udevice *dev, lbaint_t start, lbaint_t blkcnt);
 
+/**
+ * blk_flush() - Commit data to a block device
+ *
+ * @dev: Device to flush
+ * @return 0 if operation succeeded, or -ve on error.
+ */
+long blk_flush(struct udevice *dev);
+
 /**
  * blk_find_device() - Find a block device
  *
@@ -559,6 +577,16 @@ static inline ulong blk_derase(struct blk_desc *block_dev, lbaint_t start,
 	return block_dev->block_erase(block_dev, start, blkcnt);
 }
 
+static inline ulong blk_dflush(struct blk_desc *block_dev)
+{
+	if (!block_dev->block_flush)
+		return -ENOSYS;
+
+	blkcache_invalidate(block_dev->uclass_id, block_dev->devnum);
+
+	return block_dev->block_flush(block_dev);
+}
+
 /**
  * struct blk_driver - Driver for block interface types
  *
diff --git a/include/part.h b/include/part.h
index 15daacd7faaa..63982d7b9370 100644
--- a/include/part.h
+++ b/include/part.h
@@ -454,6 +454,14 @@ ulong disk_blk_write(struct udevice *dev, lbaint_t start, lbaint_t blkcnt,
  */
 ulong disk_blk_erase(struct udevice *dev, lbaint_t start, lbaint_t blkcnt);
 
+/**
+ * disk_blk_flush() - commit data to a disk
+ *
+ * @dev:	Device to flush
+ * Return:	0 success, or -ve error number (see the IS_ERR_VALUE() macro
+ */
+ulong disk_blk_flush(struct udevice *dev);
+
 /*
  * We don't support printing partition information in SPL and only support
  * getting partition information in a few cases.
-- 
2.54.0

Re: [PATCH v2 2/6] drivers: block: Introduce blk_flush()/blk_dflush()

[Simon Glass]

uHi Denis,

On 2026-05-29T03:44:35, None <dmukhin@ford.com> wrote:
> drivers: block: Introduce blk_flush()/blk_dflush()
>
> Add generic flush operations for committing dirty data to the storage
> device.
>
> This provides a common block-layer interface for flushing pending
> writes, allowing callers to ensure that data buffered by the block
> device or its backing implementation is written out to the underlying
> storage.
>
> Signed-off-by: Denis Mukhin <dmukhin@ford.com>
>
> disk/disk-uclass.c         |  6 ++++++
>  drivers/block/blk-uclass.c | 18 ++++++++++++++++++
>  include/blk.h              | 28 ++++++++++++++++++++++++++++
>  include/part.h             |  8 ++++++++
>  4 files changed, 60 insertions(+)

> diff --git a/drivers/block/blk-uclass.c b/drivers/block/blk-uclass.c
> @@ -514,6 +514,19 @@ long blk_erase(struct udevice *dev, lbaint_t start, lbaint_t blkcnt)
> +long blk_flush(struct udevice *dev)
> +{
> +     struct blk_desc *desc = dev_get_uclass_plat(dev);
> +     const struct blk_ops *ops = blk_get_ops(dev);
> +
> +     if (!ops->flush)
> +             return -ENOSYS;
> +
> +     blkcache_invalidate(desc->uclass_id, desc->devnum);
> +
> +     return ops->flush(dev);
> +}

Why invalidate the read cache here? Flush pushes dirty data down to
the media; it doesn't change what a subsequent read returns, so the
cached contents are still valid. Erase/write mutate device state, but
flush does not. Please drop the blkcache_invalidate() call (and the
matching one in the non-DM inline below).

> diff --git a/include/blk.h b/include/blk.h
> @@ -99,6 +99,7 @@ struct blk_desc {
>       unsigned long   (*block_erase)(struct blk_desc *block_dev,
>                                      lbaint_t start,
>                                      lbaint_t blkcnt);
> +     unsigned long   (*block_flush)(struct blk_desc *block_dev);

This field lives in the !CONFIG_BLK (legacy) branch, but no legacy
driver in the series wires it up. The only consumer (NVMe) is DM-only.
Are you going to need the legacy plumbing?

> diff --git a/include/blk.h b/include/blk.h
> @@ -275,6 +276,14 @@ struct blk_ops {
> +     /**
> +      * flush() -  commit all dirty data to storage
> +      *
> +      * @dev:        Device to flush
> +      * @return 0 if OK, -ve on error
> +      */
> +     unsigned long (*flush)(struct udevice *dev);
>  };

Two small things: stray double space after the dash (flush() -
commit), and please place this next to .erase so the
read/write/erase/flush group is together.

> diff --git a/include/part.h b/include/part.h
> @@ -454,6 +454,14 @@ ulong disk_blk_write(struct udevice *dev, lbaint_t start, lbaint_t blkcnt,
> +/**
> + * disk_blk_flush() - commit data to a disk
> + *
> + * @dev:     Device to flush
> + * Return:   0 success, or -ve error number (see the IS_ERR_VALUE() macro
> + */
> +ulong disk_blk_flush(struct udevice *dev);

Missing closing parenthesis after IS_ERR_VALUE() macro.

> diff --git a/disk/disk-uclass.c b/disk/disk-uclass.c
> @@ -122,6 +122,11 @@ unsigned long disk_blk_erase(struct udevice *dev, lbaint_t start,
> +unsigned long disk_blk_flush(struct udevice *dev)
> +{
> +     return blk_flush(dev_get_parent(dev));
> +}

Please note in the commit message that flushing a partition flushes
the entire underlying device (the API has no start/blkcnt), so callers
don't expect partition-scoped semantics.

Regards,
Simon

[PATCH v2 3/6] drivers: nvme: Implement flush command

[dmukhin]

From: Denis Mukhin <dmukhin@ford.com> 

Implement the NVM flush command (opcode 0x00) to allow callers to
synchronize any buffered data with the backing NVMe storage.

Signed-off-by: Denis Mukhin <dmukhin@ford.com>
---
Changes since v1:
- moved generic disk_blk_flush() to a separate patch
---
 drivers/nvme/nvme.c | 20 ++++++++++++++++++++
 1 file changed, 20 insertions(+)

diff --git a/drivers/nvme/nvme.c b/drivers/nvme/nvme.c
index c3c44e50f19a..d9f099a11593 100644
--- a/drivers/nvme/nvme.c
+++ b/drivers/nvme/nvme.c
@@ -819,9 +819,29 @@ static ulong nvme_blk_write(struct udevice *udev, lbaint_t blknr,
 	return nvme_blk_rw(udev, blknr, blkcnt, (void *)buffer, false);
 }
 
+/*
+ * NVM Flush command (opcode 0x00).
+ *
+ * Applies to a single namespace; the controller must commit all dirty
+ * data for that namespace to storage before completing the command.
+ */
+static ulong nvme_blk_flush(struct udevice *udev)
+{
+	struct nvme_ns *ns = dev_get_priv(udev);
+	struct nvme_dev *dev = ns->dev;
+	struct nvme_command c;
+
+	memset(&c, 0, sizeof(c));
+	c.common.opcode = nvme_cmd_flush;
+	c.common.nsid = cpu_to_le32(ns->ns_id);
+
+	return nvme_submit_sync_cmd(dev->queues[NVME_IO_Q], &c, NULL, IO_TIMEOUT);
+}
+
 static const struct blk_ops nvme_blk_ops = {
 	.read	= nvme_blk_read,
 	.write	= nvme_blk_write,
+	.flush	= nvme_blk_flush,
 };
 
 U_BOOT_DRIVER(nvme_blk) = {
-- 
2.54.0

Re: [PATCH v2 3/6] drivers: nvme: Implement flush command

[Simon Glass]

kOn 2026-05-29T03:44:35, None <dmukhin@ford.com> wrote:
> drivers: nvme: Implement flush command
>
> Implement the NVM flush command (opcode 0x00) to allow callers to
> synchronize any buffered data with the backing NVMe storage.
>
> Signed-off-by: Denis Mukhin <dmukhin@ford.com>
>
> drivers/nvme/nvme.c | 20 ++++++++++++++++++++
>  1 file changed, 20 insertions(+)

Reviewed-by: Simon Glass <sjg@chromium.org>

[PATCH v2 4/6] cmd: Add flush support for all blk devices

[dmukhin]

From: Denis Mukhin <dmukhin@ford.com> 

Introduce `flush` subcommand for all blk devices to allow committing
dirty data explicitly to the given block device.

Signed-off-by: Denis Mukhin <dmukhin@ford.com>
---
Changes since v1:
- updated commit message
- updated blk commands to enable flush
---
 cmd/blk_common.c | 14 ++++++++++++++
 cmd/ide.c        |  1 +
 cmd/nvme.c       |  1 +
 cmd/pvblock.c    |  1 +
 cmd/sata.c       |  1 +
 cmd/scsi.c       |  1 +
 cmd/usb.c        |  1 +
 cmd/virtio.c     |  1 +
 8 files changed, 21 insertions(+)

diff --git a/cmd/blk_common.c b/cmd/blk_common.c
index 56529702a470..d3b00e10f17c 100644
--- a/cmd/blk_common.c
+++ b/cmd/blk_common.c
@@ -37,6 +37,20 @@ int blk_common_cmd(int argc, char *const argv[], enum uclass_id uclass_id,
 				printf("\nno %s partition table available\n",
 				       if_name);
 			return CMD_RET_SUCCESS;
+		} else if (strncmp(argv[1], "flush", 5) == 0) {
+			struct blk_desc *desc;
+			int ret;
+
+			ret = blk_get_desc(uclass_id, *cur_devnump, &desc);
+			if (ret)
+				return CMD_RET_FAILURE;
+
+			if (blk_dflush(desc)) {
+				printf("\nfailed to flush device %s\n", if_name);
+				return CMD_RET_FAILURE;
+			}
+
+			return CMD_RET_SUCCESS;
 		}
 		return CMD_RET_USAGE;
 	case 3:
diff --git a/cmd/ide.c b/cmd/ide.c
index f99fb6f58246..16833e30319c 100644
--- a/cmd/ide.c
+++ b/cmd/ide.c
@@ -63,6 +63,7 @@ U_BOOT_CMD(ide, 5, 1, do_ide,
 	   "IDE sub-system",
 	   "reset - reset IDE controller\n"
 	   "ide info  - show available IDE devices\n"
+	   "ide flush - commit all dirty data to the current IDE device\n"
 	   "ide device [dev] - show or set current device\n"
 	   "ide part [dev] - print partition table of one or all IDE devices\n"
 	   "ide read  addr blk# cnt\n"
diff --git a/cmd/nvme.c b/cmd/nvme.c
index f2c9acba5c32..dbccb69042bc 100644
--- a/cmd/nvme.c
+++ b/cmd/nvme.c
@@ -47,6 +47,7 @@ U_BOOT_CMD(
 	"scan - scan NVMe devices\n"
 	"nvme detail - show details of current NVMe device\n"
 	"nvme info - show all available NVMe devices\n"
+	"nvme flush - commit all dirty data to the current NVMe device\n"
 	"nvme device [dev] - show or set current NVMe device\n"
 	"nvme part [dev] - print partition table of one or all NVMe devices\n"
 	"nvme read addr blk# cnt - read `cnt' blocks starting at block\n"
diff --git a/cmd/pvblock.c b/cmd/pvblock.c
index 3a83ac9cd92c..5e4a98a6b13b 100644
--- a/cmd/pvblock.c
+++ b/cmd/pvblock.c
@@ -20,6 +20,7 @@ int do_pvblock(struct cmd_tbl *cmdtp, int flag, int argc, char *const argv[])
 U_BOOT_CMD(pvblock, 5, 1, do_pvblock,
 	   "Xen para-virtualized block device",
 	   "info  - show available block devices\n"
+	   "pvblock flush - commit all dirty data to the current device\n"
 	   "pvblock device [dev] - show or set current device\n"
 	   "pvblock part [dev] - print partition table of one or all devices\n"
 	   "pvblock read  addr blk# cnt\n"
diff --git a/cmd/sata.c b/cmd/sata.c
index 8b923f9378b2..417e027cba4f 100644
--- a/cmd/sata.c
+++ b/cmd/sata.c
@@ -119,6 +119,7 @@ U_BOOT_CMD(
 	"init - init SATA sub system\n"
 	"sata stop [dev] - disable SATA sub system or device\n"
 	"sata info - show available SATA devices\n"
+	"sata flush - commit all dirty data to the current SATA device\n"
 	"sata device [dev] - show or set current device\n"
 	"sata part [dev] - print partition table\n"
 	"sata read addr blk# cnt\n"
diff --git a/cmd/scsi.c b/cmd/scsi.c
index ad7d8a4b6644..39be208f337a 100644
--- a/cmd/scsi.c
+++ b/cmd/scsi.c
@@ -47,6 +47,7 @@ U_BOOT_CMD(
 	"SCSI sub-system",
 	"scsi info  - show available SCSI devices\n"
 	"scsi scan  - (re-)scan SCSI bus\n"
+	"scsi flush - commit all dirty data to the current SCSI device\n"
 	"scsi device [dev] - show or set current device\n"
 	"scsi part [dev] - print partition table of one or all SCSI devices\n"
 	"scsi read addr blk# cnt - read `cnt' blocks starting at block `blk#'\n"
diff --git a/cmd/usb.c b/cmd/usb.c
index 13a2996c1f00..3e156f54c50b 100644
--- a/cmd/usb.c
+++ b/cmd/usb.c
@@ -706,6 +706,7 @@ U_BOOT_CMD(
 	"    Available modes: J, K, S[E0_NAK], P[acket], F[orce_Enable]\n"
 #ifdef CONFIG_USB_STORAGE
 	"usb storage - show details of USB storage devices\n"
+	"usb flush - commit all dirty data to the current USB storage\n"
 	"usb dev [dev] - show or set current USB storage device\n"
 	"usb part [dev] - print partition table of one or all USB storage"
 	"    devices\n"
diff --git a/cmd/virtio.c b/cmd/virtio.c
index a42a563ab727..94f1e82cc902 100644
--- a/cmd/virtio.c
+++ b/cmd/virtio.c
@@ -44,6 +44,7 @@ U_BOOT_CMD(
 	"virtio block devices sub-system",
 	"scan - initialize virtio bus\n"
 	"virtio info - show all available virtio block devices\n"
+	"virtio flush - commit all dirty data to the current virtio block device\n"
 	"virtio device [dev] - show or set current virtio block device\n"
 	"virtio part [dev] - print partition table of one or all virtio block devices\n"
 	"virtio read addr blk# cnt - read `cnt' blocks starting at block\n"
-- 
2.54.0

Re: [PATCH v2 4/6] cmd: Add flush support for all blk devices

[Simon Glass]

Hi Denis,

On 2026-05-29T03:44:35, None <dmukhin@ford.com> wrote:
> cmd: Add flush support for all blk devices
>
> Introduce flush subcommand for all blk devices to allow committing
> dirty data explicitly to the given block device.
>
> Signed-off-by: Denis Mukhin <dmukhin@ford.com>
>
> cmd/blk_common.c | 14 ++++++++++++++
>  cmd/ide.c        |  1 +
>  cmd/nvme.c       |  1 +
>  cmd/pvblock.c    |  1 +
>  cmd/sata.c       |  1 +
>  cmd/scsi.c       |  1 +
>  cmd/usb.c        |  1 +
>  cmd/virtio.c     |  1 +
>  8 files changed, 21 insertions(+)

> diff --git a/cmd/blk_common.c b/cmd/blk_common.c
> @@ -37,6 +37,20 @@ int blk_common_cmd(int argc, char *const argv[], enum uclass_id uclass_id,
> +             } else if (strncmp(argv[1], 'flush', 5) == 0) {
> +                     struct blk_desc *desc;
> +                     int ret;
> +
> +                     ret = blk_get_desc(uclass_id, *cur_devnump, &desc);
> +                     if (ret)
> +                             return CMD_RET_FAILURE;
> +
> +                     if (blk_dflush(desc)) {
> +                             printf("\nfailed to flush device %s\n", if_name);
> +                             return CMD_RET_FAILURE;
> +                     }
> +
> +                     return CMD_RET_SUCCESS;
> +             }

Please print the return value from blk_dflush() and the device number
in the failure message - otherwise a user can't tell -ENOSYS from a
genuine I/O error. Compare with the erase arm just below which reports
the block count. You can use %dE to get an error string if the board
enables it.

Do you want to print something on success?

It is a bit misleading to suggest that all the commands support
'flush'. But I suppose we can worry about that later.

> diff --git a/cmd/blk_common.c b/cmd/blk_common.c
> @@ -37,6 +37,20 @@ int blk_common_cmd(int argc, char *const argv[], enum uclass_id uclass_id,
> +             } else if (strncmp(argv[1], 'flush', 5) == 0) {

cmd/blkmap.c also routes through blk_common_cmd(), but dispatches via
U_BOOT_SUBCMD_MKENT() so flush never reaches this function unless you
add a matching subcmd entry. Either add it for consistency or call out
in the commit message that blkmap is intentionally excluded.

Regards,
Simon

Re: [PATCH v2 4/6] cmd: Add flush support for all blk devices

[dmukhin]

On Fri, May 29, 2026 at 04:57:48AM -0500, Simon Glass wrote:
> Hi Denis,
> 
> On 2026-05-29T03:44:35, None <dmukhin@ford.com> wrote:
> > cmd: Add flush support for all blk devices
> >
> > Introduce flush subcommand for all blk devices to allow committing
> > dirty data explicitly to the given block device.
> >
> > Signed-off-by: Denis Mukhin <dmukhin@ford.com>
> >
> > cmd/blk_common.c | 14 ++++++++++++++
> >  cmd/ide.c        |  1 +
> >  cmd/nvme.c       |  1 +
> >  cmd/pvblock.c    |  1 +
> >  cmd/sata.c       |  1 +
> >  cmd/scsi.c       |  1 +
> >  cmd/usb.c        |  1 +
> >  cmd/virtio.c     |  1 +
> >  8 files changed, 21 insertions(+)
> 
> > diff --git a/cmd/blk_common.c b/cmd/blk_common.c
> > @@ -37,6 +37,20 @@ int blk_common_cmd(int argc, char *const argv[], enum uclass_id uclass_id,
> > +             } else if (strncmp(argv[1], 'flush', 5) == 0) {
> > +                     struct blk_desc *desc;
> > +                     int ret;
> > +
> > +                     ret = blk_get_desc(uclass_id, *cur_devnump, &desc);
> > +                     if (ret)
> > +                             return CMD_RET_FAILURE;
> > +
> > +                     if (blk_dflush(desc)) {
> > +                             printf("\nfailed to flush device %s\n", if_name);
> > +                             return CMD_RET_FAILURE;
> > +                     }
> > +
> > +                     return CMD_RET_SUCCESS;
> > +             }
> 
> Please print the return value from blk_dflush() and the device number
> in the failure message - otherwise a user can't tell -ENOSYS from a
> genuine I/O error. Compare with the erase arm just below which reports
> the block count. You can use %dE to get an error string if the board
> enables it.
> 
> Do you want to print something on success?
> 
> It is a bit misleading to suggest that all the commands support
> 'flush'. But I suppose we can worry about that later.
> 
> > diff --git a/cmd/blk_common.c b/cmd/blk_common.c
> > @@ -37,6 +37,20 @@ int blk_common_cmd(int argc, char *const argv[], enum uclass_id uclass_id,
> > +             } else if (strncmp(argv[1], 'flush', 5) == 0) {
> 
> cmd/blkmap.c also routes through blk_common_cmd(), but dispatches via
> U_BOOT_SUBCMD_MKENT() so flush never reaches this function unless you
> add a matching subcmd entry. Either add it for consistency or call out
> in the commit message that blkmap is intentionally excluded.

Thanks for the feedback!
Will update the patch.

> 
> Regards,
> Simon

Re: [PATCH v2 4/6] cmd: Add flush support for all blk devices

[Sean Anderson]

On 5/28/26 23:44, dmukhin@ford.com wrote:
> From: Denis Mukhin <dmukhin@ford.com>
> 
> Introduce `flush` subcommand for all blk devices to allow committing
> dirty data explicitly to the given block device.

Shouldn't this be done automatically as part of blk_write? Similar to
how all FS operations in U-Boot look like

- Mount FS
- Perform operation
- Unmount FS

so the FS is never in an inconsistent state once an operation completes.

Is the performance impact too much? I'm just concerned that a lot of
existing code assumes that blk_write is persistent and that it's OK
to e.g. immediately reset after blk_write returns.

--Sean

Re: [PATCH v2 4/6] cmd: Add flush support for all blk devices

[Simon Glass]

Hi Sean,

On Wed, 3 Jun 2026 at 11:43, Sean Anderson <seanga2@gmail.com> wrote:
>
> On 5/28/26 23:44, dmukhin@ford.com wrote:
> > From: Denis Mukhin <dmukhin@ford.com>
> >
> > Introduce `flush` subcommand for all blk devices to allow committing
> > dirty data explicitly to the given block device.
>
> Shouldn't this be done automatically as part of blk_write? Similar to
> how all FS operations in U-Boot look like
>
> - Mount FS
> - Perform operation
> - Unmount FS
>
> so the FS is never in an inconsistent state once an operation completes.
>
> Is the performance impact too much? I'm just concerned that a lot of
> existing code assumes that blk_write is persistent and that it's OK
> to e.g. immediately reset after blk_write returns.

This is actually for a device-level flush, so more like a Linux 'sync'
than just a write. If the hardware supports / requires it, then it
seems valuable to me.

Regards,
Simon

Re: [PATCH v2 4/6] cmd: Add flush support for all blk devices

[dmukhin]

Hi Sean,

Thanks for the feedback!

On Wed, Jun 03, 2026 at 01:43:10PM -0400, Sean Anderson wrote:
> On 5/28/26 23:44, dmukhin@ford.com wrote:
> > From: Denis Mukhin <dmukhin@ford.com>
> > 
> > Introduce `flush` subcommand for all blk devices to allow committing
> > dirty data explicitly to the given block device.
> 
> Shouldn't this be done automatically as part of blk_write? Similar to
> how all FS operations in U-Boot look like
> 
> - Mount FS
> - Perform operation
> - Unmount FS
> 
> so the FS is never in an inconsistent state once an operation completes.
> 
> Is the performance impact too much? I'm just concerned that a lot of
> existing code assumes that blk_write is persistent and that it's OK
> to e.g. immediately reset after blk_write returns.

In my case I need to update the GPT-backed boot counter; there are two
GPT headers, primary and backup, so I used blk_dflush() to guarantee that
both headers are updated before jumping into the OS.

So in my scenario I need to ensure that those 2 writes will land to the
NVMe.

I think FUA suggested by Neil in v1 feedback should be a robust solution
for NVMe, but I went with a 'flush' solution first.

> 
> --Sean

[PATCH v2 5/6] drivers: nvme: Export nvme_shutdown()

[dmukhin]

From: Denis Mukhin <dmukhin@ford.com> 

Make nvme_shutdown() public so that it can be re-used in
board-specific board_quiesce_devices() before OS hand-off.

Signed-off-by: Denis Mukhin <dmukhin@ford.com>
---
Changes since v1:
- n/a
---
 drivers/nvme/nvme.h |  7 -------
 include/nvme.h      | 11 +++++++++++
 2 files changed, 11 insertions(+), 7 deletions(-)

diff --git a/drivers/nvme/nvme.h b/drivers/nvme/nvme.h
index bc1d612dde40..25a8b0b17772 100644
--- a/drivers/nvme/nvme.h
+++ b/drivers/nvme/nvme.h
@@ -698,11 +698,4 @@ struct nvme_ops {
  */
 int nvme_init(struct udevice *udev);
 
-/**
- * nvme_shutdown() - Shutdown NVM Express device
- * @udev:	The NVM Express device
- * Return: 0 if OK, -ve on error
- */
-int nvme_shutdown(struct udevice *udev);
-
 #endif /* __DRIVER_NVME_H__ */
diff --git a/include/nvme.h b/include/nvme.h
index 2cdf8ce320c1..c7b301300097 100644
--- a/include/nvme.h
+++ b/include/nvme.h
@@ -90,4 +90,15 @@ int nvme_print_info(struct udevice *udev);
  */
 int nvme_get_namespace_id(struct udevice *udev, u32 *ns_id, u8 *eui64);
 
+/**
+ * nvme_shutdown() - Shutdown NVM Express device
+ *
+ * Must be called before booting an OS to ensure cache is flushed
+ * and the controller is in a clean state for OS to re-initialize.
+ *
+ * @udev:	The NVM Express device
+ * Return: 0 if OK, -ve on error
+ */
+int nvme_shutdown(struct udevice *udev);
+
 #endif /* __NVME_H__ */
-- 
2.54.0

Re: [PATCH v2 5/6] drivers: nvme: Export nvme_shutdown()

[Simon Glass]

On 2026-05-29T03:44:35, None <dmukhin@ford.com> wrote:
> drivers: nvme: Export nvme_shutdown()
>
> Make nvme_shutdown() public so that it can be re-used in
> board-specific board_quiesce_devices() before OS hand-off.
>
> Signed-off-by: Denis Mukhin <dmukhin@ford.com>
>
> drivers/nvme/nvme.h |  7 -------
>  include/nvme.h      | 11 +++++++++++
>  2 files changed, 11 insertions(+), 7 deletions(-)

Reviewed-by: Simon Glass <sjg@chromium.org>

[PATCH v2 6/6] docs: nvme: Update QEMU command for testing

[dmukhin]

From: Denis Mukhin <dmukhin@ford.com> 

Switch to x86-64 QEMU machine for NVMe testing in the docs.
Also, add `format=raw` for empty NVMe disks QEMU command line.

Signed-off-by: Denis Mukhin <dmukhin@ford.com>
---
Changes since v1:
- n/a
---
 doc/develop/driver-model/nvme.rst | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/doc/develop/driver-model/nvme.rst b/doc/develop/driver-model/nvme.rst
index 755181331213..0e0624925e0d 100644
--- a/doc/develop/driver-model/nvme.rst
+++ b/doc/develop/driver-model/nvme.rst
@@ -95,4 +95,4 @@ Example command line to call QEMU x86 below with emulated NVMe device:
 
 .. code-block:: bash
 
-  $ ./qemu-system-i386 -drive file=nvme.img,if=none,id=drv0 -device nvme,drive=drv0,serial=QEMUNVME0001 -bios u-boot.rom
+  $ qemu-system-x86_64 -nographic -drive file=nvme.img,format=raw,if=none,id=drv0 -device nvme,drive=drv0,serial=QEMUNVME0001 -bios u-boot.rom
-- 
2.54.0

Re: [PATCH v2 6/6] docs: nvme: Update QEMU command for testing

[Simon Glass]

Hi Denis,

On 2026-05-29T03:44:35, None <dmukhin@ford.com> wrote:
> docs: nvme: Update QEMU command for testing
>
> Switch to x86-64 QEMU machine for NVMe testing in the docs.
> Also, add format=raw for empty NVMe disks QEMU command line.
>
> Signed-off-by: Denis Mukhin <dmukhin@ford.com>
>
> doc/develop/driver-model/nvme.rst | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)

> diff --git a/doc/develop/driver-model/nvme.rst b/doc/develop/driver-model/nvme.rst
> @@ -95,4 +95,4 @@ Example command line to call QEMU x86 below with emulated NVMe device:
>
>  .. code-block:: bash
>
> -  $ ./qemu-system-i386 -drive file=nvme.img,if=none,id=drv0 -device nvme,drive=drv0,serial=QEMUNVME0001 -bios u-boot.rom
> +  $ qemu-system-x86_64 -nographic -drive file=nvme.img,format=raw,if=none,id=drv0 -device nvme,drive=drv0,serial=QEMUNVME0001 -bios u-boot.rom

The commit message lists two changes but the diff also adds
-nographic. Please mention it, or drop it if not required.

Also, format=raw is needed for any raw image to silence QEMU's probing
warning, not specifically for empty disks. Something like "specify
format=raw explicitly to avoid QEMU's raw-format auto-detection
warning" would read better.

Regards,
Simon

Re: [v2,0/6] nvme: few fixups

[Simon Glass]

Hi Denis,

On 2026-05-29T03:44:35, None <dmukhin@ford.com> wrote:

> Patch 4 enables flush for the currently selected block device.

It adds the flush subcommand to all the block-uclass front-end
commands (ide, nvme, pvblock, sata, scsi, usb, virtio) in one go.
Please reword.

> The series adds few fixes for NVMe subsystem.

This is more than NVMe fixes - patch 2 adds a new block-uclass op and
patch 4 touches seven command files. Since you are adding a new uclass
operation, please add a test for blk_flush()/blk_dflush() under
test/dm/blk.c covering success and -ENOSYS when .flush is absent.

> Patch 5 makes nvme_shutdown() symbol public to enable use of it from
>         board_quiesce_devices()

Really we should mark the driver as DM_FLAG_ACTIVE_DMA and then this
happens automatically. The board_quiesce_devices() function should go
away at some point.

Regards,
Simon

Re: (subset) [PATCH v2 0/6] nvme: few fixups

[Neil Armstrong]

Hi,

On Thu, 28 May 2026 20:44:35 -0700, dmukhin@ford.com wrote:
> The series adds few fixes for NVMe subsystem.
> 
> Patch 1 adds logline for tracing misbehaving NVMe drives.
> Patch 2 adds plumbing for blk_flush()/blk_dflush().
> Patch 3 introduces NVMe flush command.
> Patch 4 enables `flush` for the currently selected block device.
> Patch 5 makes nvme_shutdown() symbol public to enable use of it from
>         board_quiesce_devices()
> Patch 6 tiny fixup for QEMU command for manual NVMe tests.
> 
> [...]

Thanks, Applied to https://source.denx.de/u-boot/custodians/u-boot-ufs (u-boot-nvme-fixes)

[1/6] drivers: nvme: Log I/O timeouts
      https://source.denx.de/u-boot/custodians/u-boot-ufs/-/commit/389363d287585d8135a554c4419a4f580346b9f8

-- 
Neil

Re: [PATCH v2 0/6] nvme: few fixups

[neil.armstrong]

Hi,

On 5/29/26 05:44, dmukhin@ford.com wrote:
> The series adds few fixes for NVMe subsystem.
> 
> Patch 1 adds logline for tracing misbehaving NVMe drives.
> Patch 2 adds plumbing for blk_flush()/blk_dflush().
> Patch 3 introduces NVMe flush command.
> Patch 4 enables `flush` for the currently selected block device.

As my thread on v1, please enable FUA or disable WC to be sure we don't use the WC.

Neil

> Patch 5 makes nvme_shutdown() symbol public to enable use of it from
>          board_quiesce_devices()
> Patch 6 tiny fixup for QEMU command for manual NVMe tests.
> 
> Denis Mukhin (6):
>    drivers: nvme: Log I/O timeouts
>    drivers: block: Introduce blk_flush()/blk_dflush()
>    drivers: nvme: Implement flush command
>    cmd: Add flush support for all blk devices
>    drivers: nvme: Export nvme_shutdown()
>    docs: nvme: Update QEMU command for testing
> 
>   cmd/blk_common.c                  | 14 ++++++++++++++
>   cmd/ide.c                         |  1 +
>   cmd/nvme.c                        |  1 +
>   cmd/pvblock.c                     |  1 +
>   cmd/sata.c                        |  1 +
>   cmd/scsi.c                        |  1 +
>   cmd/usb.c                         |  1 +
>   cmd/virtio.c                      |  1 +
>   disk/disk-uclass.c                |  6 ++++++
>   doc/develop/driver-model/nvme.rst |  2 +-
>   drivers/block/blk-uclass.c        | 18 ++++++++++++++++++
>   drivers/nvme/nvme.c               | 24 +++++++++++++++++++++++-
>   drivers/nvme/nvme.h               |  7 -------
>   include/blk.h                     | 28 ++++++++++++++++++++++++++++
>   include/nvme.h                    | 11 +++++++++++
>   include/part.h                    |  8 ++++++++
>   16 files changed, 116 insertions(+), 9 deletions(-)
> 

Re: [PATCH v2 0/6] nvme: few fixups

[dmukhin]

Hi Neil,

Thanks for the feedback!

On Thu, Jun 04, 2026 at 08:32:50AM +0200, neil.armstrong@linaro.org wrote:
> Hi,
> 
> On 5/29/26 05:44, dmukhin@ford.com wrote:
> > The series adds few fixes for NVMe subsystem.
> > 
> > Patch 1 adds logline for tracing misbehaving NVMe drives.
> > Patch 2 adds plumbing for blk_flush()/blk_dflush().
> > Patch 3 introduces NVMe flush command.
> > Patch 4 enables `flush` for the currently selected block device.
> 
> As my thread on v1, please enable FUA or disable WC to be sure we don't use the WC.

Yes, sorry, I did not have time to respond to v1 feedback (just sent out
an email):

   https://lore.kernel.org/u-boot/aiJFzMdcCu5sAktz@kraken/

I will look into FUA for NVMe.

But I also think it makes sense to add flush to the code base as
an alternative solution for ensuring data is stored.

--
Denis