Re: [PATCH v5 5/5] iotests: Add `vvfat` tests

[Kevin Wolf <kwolf@redhat.com>]

Am 12.06.2024 um 14:43 hat Amjad Alsharafi geschrieben:
> Added several tests to verify the implementation of the vvfat driver.
> 
> We needed a way to interact with it, so created a basic `fat16.py` driver
> that handled writing correct sectors for us.
> 
> Added `vvfat` to the non-generic formats, as its not a normal image format.
> 
> Signed-off-by: Amjad Alsharafi <amjadsharafi10@gmail.com>

> +    def truncate_file(
> +        self,
> +        entry: FatDirectoryEntry,
> +        new_size: int,
> +        allocate_non_continuous: bool = False,
> +    ):
> +        """
> +        Truncate the file at the given path to the new size.
> +        """
> +        if entry is None:
> +            return Exception("entry is None")
> +        if entry.attributes & 0x10 != 0:
> +            raise Exception(f"{entry.whole_name()} is a directory")
> +
> +        def clusters_from_size(size: int):
> +            return (
> +                size + self.boot_sector.cluster_bytes() - 1
> +            ) // self.boot_sector.cluster_bytes()
> +
> +        # First, allocate new FATs if we need to
> +        required_clusters = clusters_from_size(new_size)
> +        current_clusters = clusters_from_size(entry.size_bytes)
> +
> +        affected_clusters = set()
> +
> +        # Keep at least one cluster, easier to manage this way
> +        if required_clusters == 0:
> +            required_clusters = 1
> +        if current_clusters == 0:
> +            current_clusters = 1
> +
> +        if required_clusters > current_clusters:
> +            # Allocate new clusters
> +            cluster = entry.cluster
> +            to_add = required_clusters
> +            for _ in range(current_clusters - 1):
> +                to_add -= 1
> +                cluster = self.next_cluster(cluster)
> +            assert required_clusters > 0, "No new clusters to allocate"
> +            assert cluster is not None, "Cluster is None"
> +            assert (
> +                self.next_cluster(cluster) is None
> +            ), "Cluster is not the last cluster"
> +
> +            # Allocate new clusters
> +            for _ in range(to_add - 1):
> +                new_cluster = self.next_free_cluster()
> +                if allocate_non_continuous:
> +                    new_cluster = self.next_free_cluster_non_continuous()

The normal self.next_free_cluster() could be in an else branch. No
reason to search for a free cluster when you immediately overwrite it
anyway.

> +                self.write_fat_entry(cluster, new_cluster)
> +                self.write_fat_entry(new_cluster, 0xFFFF)
> +                cluster = new_cluster
> +
> +        elif required_clusters < current_clusters:
> +            # Truncate the file
> +            cluster = entry.cluster
> +            for _ in range(required_clusters - 1):
> +                cluster = self.next_cluster(cluster)
> +            assert cluster is not None, "Cluster is None"
> +
> +            next_cluster = self.next_cluster(cluster)
> +            # mark last as EOF
> +            self.write_fat_entry(cluster, 0xFFFF)
> +            # free the rest
> +            while next_cluster is not None:
> +                cluster = next_cluster
> +                next_cluster = self.next_cluster(next_cluster)
> +                self.write_fat_entry(cluster, 0)
> +
> +        self.flush_fats()
> +
> +        # verify number of clusters
> +        cluster = entry.cluster
> +        count = 0
> +        while cluster is not None:
> +            count += 1
> +            affected_clusters.add(cluster)
> +            cluster = self.next_cluster(cluster)
> +        assert (
> +            count == required_clusters
> +        ), f"Expected {required_clusters} clusters, got {count}"
> +
> +        # update the size
> +        entry.size_bytes = new_size
> +        self.update_direntry(entry)
> +
> +        # trigger every affected cluster
> +        for cluster in affected_clusters:
> +            first_sector = self.boot_sector.first_sector_of_cluster(cluster)
> +            first_sector_data = self.read_sectors(first_sector, 1)
> +            self.write_sectors(first_sector, first_sector_data)

Other than this, the patch looks good to me and we seem to test all the
cases that are fixed by the previous patches.

Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Tested-by: Kevin Wolf <kwolf@redhat.com>

Kevin