Re: [PATCH v8 1/4] file-posix: add tracking of the zone write pointers

[Stefan Hajnoczi <stefanha@redhat.com>]

[-- Attachment #1: Type: text/plain, Size: 6749 bytes --]

On Tue, Apr 04, 2023 at 11:32:36PM +0800, Sam Li wrote:
> Since Linux doesn't have a user API to issue zone append operations to
> zoned devices from user space, the file-posix driver is modified to add
> zone append emulation using regular writes. To do this, the file-posix
> driver tracks the wp location of all zones of the device. It uses an
> array of uint64_t. The most significant bit of each wp location indicates
> if the zone type is conventional zones.
> 
> The zones wp can be changed due to the following operations issued:
> - zone reset: change the wp to the start offset of that zone
> - zone finish: change to the end location of that zone
> - write to a zone
> - zone append
> 
> Signed-off-by: Sam Li <faithilikerun@gmail.com>
> ---
>  block/file-posix.c               | 168 ++++++++++++++++++++++++++++++-
>  include/block/block-common.h     |  14 +++
>  include/block/block_int-common.h |   5 +
>  3 files changed, 184 insertions(+), 3 deletions(-)
> 
> diff --git a/block/file-posix.c b/block/file-posix.c
> index 65efe5147e..bc58f7193b 100644
> --- a/block/file-posix.c
> +++ b/block/file-posix.c
> @@ -1324,6 +1324,88 @@ static int hdev_get_max_segments(int fd, struct stat *st)
>  #endif
>  }
>  
> +#if defined(CONFIG_BLKZONED)
> +/*
> + * If the reset_all flag is true, then the wps of zone whose state is
> + * not readonly or offline should be all reset to the start sector.
> + * Else, take the real wp of the device.
> + */
> +static int get_zones_wp(int fd, BlockZoneWps *wps, int64_t offset,
> +                        unsigned int nrz, bool reset_all)
> +{
> +    struct blk_zone *blkz;
> +    size_t rep_size;
> +    uint64_t sector = offset >> BDRV_SECTOR_BITS;
> +    int ret, n = 0, i = 0;
> +    rep_size = sizeof(struct blk_zone_report) + nrz * sizeof(struct blk_zone);
> +    g_autofree struct blk_zone_report *rep = NULL;
> +
> +    rep = g_malloc(rep_size);
> +    blkz = (struct blk_zone *)(rep + 1);
> +    while (n < nrz) {
> +        memset(rep, 0, rep_size);
> +        rep->sector = sector;
> +        rep->nr_zones = nrz - n;
> +
> +        do {
> +            ret = ioctl(fd, BLKREPORTZONE, rep);
> +        } while (ret != 0 && errno == EINTR);
> +        if (ret != 0) {
> +            error_report("%d: ioctl BLKREPORTZONE at %" PRId64 " failed %d",
> +                    fd, offset, errno);
> +            return -errno;
> +        }
> +
> +        if (!rep->nr_zones) {
> +            break;
> +        }
> +
> +        for (i = 0; i < rep->nr_zones; i++, n++) {
> +            /*
> +             * The wp tracking cares only about sequential writes required and
> +             * sequential write preferred zones so that the wp can advance to
> +             * the right location.
> +             * Use the most significant bit of the wp location to indicate the
> +             * zone type: 0 for SWR/SWP zones and 1 for conventional zones.
> +             */
> +            if (blkz[i].type == BLK_ZONE_TYPE_CONVENTIONAL) {
> +                wps->wp[i] &= 1ULL << 63;

Should this be |= instead of &=? I think the intention is to set the
bit.

> +            } else {
> +                switch(blkz[i].cond) {
> +                case BLK_ZONE_COND_FULL:
> +                case BLK_ZONE_COND_READONLY:
> +                    /* Zone not writable */
> +                    wps->wp[i] = (blkz[i].start + blkz[i].len) << BDRV_SECTOR_BITS;

wps->wp[i] looks wrong in two cases:
1. After the first iteration of the while (n < nrz) loop.
2. When offset > 0.

I think there should be a j variable that tracks the index into wp[]. It
should be initialized outside the while loop based on offset and
incremented inside the for loop.

> +                    break;
> +                case BLK_ZONE_COND_OFFLINE:
> +                    /* Zone not writable nor readable */
> +                    wps->wp[i] = (blkz[i].start) << BDRV_SECTOR_BITS;
> +                    break;
> +                default:
> +                    if (reset_all) {
> +                        wps->wp[i] = blkz[i].start << BDRV_SECTOR_BITS;
> +                    } else {
> +                        wps->wp[i] = blkz[i].wp << BDRV_SECTOR_BITS;
> +                    }
> +                    break;
> +                }
> +            }
> +        }
> +        sector = blkz[i - 1].start + blkz[i - 1].len;
> +    }
> +
> +    return 0;
> +}
> +
> +static void update_zones_wp(int fd, BlockZoneWps *wps, int64_t offset,
> +                            unsigned int nrz)
> +{
> +    if (get_zones_wp(fd, wps, offset, nrz, 0) < 0) {
> +        error_report("update zone wp failed");
> +    }
> +}
> +#endif
> +
>  static void raw_refresh_limits(BlockDriverState *bs, Error **errp)
>  {
>      BDRVRawState *s = bs->opaque;
> @@ -1413,6 +1495,23 @@ static void raw_refresh_limits(BlockDriverState *bs, Error **errp)
>          if (ret >= 0) {
>              bs->bl.max_active_zones = ret;
>          }
> +
> +        ret = get_sysfs_long_val(&st, "physical_block_size");
> +        if (ret >= 0) {
> +            bs->bl.write_granularity = ret;
> +        }
> +
> +        /* The refresh_limits() function can be called multiple times. */
> +        bs->wps = NULL;

This needs to be g_free(bs->wps) to avoid leaking the odl bs->wps
memory.

(There are more complex solutions that reuse bs->wps when nr_zones has
not gotten larger, but freeing and allocating a new one is the simple
solution for now. This code isn't performance-critical.)

> +        bs->wps = g_malloc(sizeof(BlockZoneWps) +
> +                sizeof(int64_t) * bs->bl.nr_zones);
> +        ret = get_zones_wp(s->fd, bs->wps, 0, bs->bl.nr_zones, 0);
> +        if (ret < 0) {
> +            error_setg_errno(errp, -ret, "report wps failed");
> +            bs->wps = NULL;
> +            return;
> +        }
> +        qemu_co_mutex_init(&bs->wps->colock);
>          return;
>      }
>  out:
> @@ -2338,9 +2437,15 @@ static int coroutine_fn raw_co_prw(BlockDriverState *bs, uint64_t offset,
>  {
>      BDRVRawState *s = bs->opaque;
>      RawPosixAIOData acb;
> +    int ret;
>  
>      if (fd_open(bs) < 0)
>          return -EIO;
> +#if defined(CONFIG_BLKZONED)
> +    if (type & QEMU_AIO_WRITE && bs->wps) {
> +        QEMU_LOCK_GUARD(&bs->wps->colock);
> +    }

QEMU_LOCK_GUARD() has lexical scope so its lifetime ends at the end of
the {} block. Therefore it doesn't lock for the remainder of the
function.

This is a case where it's necessary to use qemu_co_mutex_lock() directly
and remember to qemu_co_mutex_unlock() at the exit point of this
function.

Stefan

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]