On 08/03/2017 08:01, Fam Zheng wrote:
> Linux exposes a separate limit, /sys/block/.../queue/max_segments, which
> in the worst case can be more restrictive than BLKSECTGET (as they are
> two different things). Similar to the BLKSECTGET story, guests don't see
> this limit and send big requests will get -EINVAL error on SG_IO.
>
> Lean on the safer side to clamp max_transfer according to max_segments
> and page size, because in the end what host HBA gets is the mapped host
> pages rather than a guest buffer.
>
> Signed-off-by: Fam Zheng <f...@redhat.com>
>
> ---
>
> v2: Use /sys/dev/block/MAJOR:MINOR/queue/max_segments. [Paolo]
> ---
> block/file-posix.c | 47 +++++++++++++++++++++++++++++++++++++++++++++++
> 1 file changed, 47 insertions(+)
>
> diff --git a/block/file-posix.c b/block/file-posix.c
> index 4de1abd..c4c0663 100644
> --- a/block/file-posix.c
> +++ b/block/file-posix.c
> @@ -668,6 +668,48 @@ static int hdev_get_max_transfer_length(BlockDriverState
> *bs, int fd)
> #endif
> }
>
> +static int hdev_get_max_segments(const struct stat *st)
> +{
> +#ifdef CONFIG_LINUX
> + char buf[32];
> + const char *end;
> + char *sysfspath;
> + int ret;
> + int fd = -1;
> + long max_segments;
> +
> + sysfspath = g_strdup_printf("/sys/dev/block/%u:%u/queue/max_segments",
> + major(st->st_rdev), minor(st->st_rdev));
> + fd = open(sysfspath, O_RDONLY);
> + if (fd == -1) {
> + ret = -errno;
> + goto out;
> + }
> + do {
> + ret = read(fd, buf, sizeof(buf));
> + } while (ret == -1 && errno == EINTR);
> + if (ret < 0) {
> + ret = -errno;
> + goto out;
> + } else if (ret == 0) {
> + ret = -EIO;
> + goto out;
> + }
> + buf[ret] = 0;
> + /* The file is ended with '\n', pass 'end' to accept that. */
> + ret = qemu_strtol(buf, &end, 10, &max_segments);
> + if (ret == 0 && end && *end == '\n') {
> + ret = max_segments;
> + }
> +
> +out:
> + g_free(sysfspath);
> + return ret;
> +#else
> + return -ENOTSUP;
> +#endif
> +}
> +
> static void raw_refresh_limits(BlockDriverState *bs, Error **errp)
> {
> BDRVRawState *s = bs->opaque;
> @@ -679,6 +721,11 @@ static void raw_refresh_limits(BlockDriverState *bs,
> Error **errp)
> if (ret > 0 && ret <= BDRV_REQUEST_MAX_BYTES) {
> bs->bl.max_transfer = pow2floor(ret);
> }
> + ret = hdev_get_max_segments(&st);
> + if (ret > 0) {
> + bs->bl.max_transfer = MIN(bs->bl.max_transfer,
> + ret * getpagesize());
> + }
> }
> }
>
>
Reviewed-by: Paolo Bonzini <pbonz...@redhat.com>
Thanks!
