On Mon, 13 Dec 2010 15:22:10 +0800
Shaohua Li <shaohua.li@xxxxxxxxx> wrote:
> Add an ioctl to dump filesystem's metadata in memory in vfs. Userspace collects
> such info and uses it to do metadata readahead.
> Filesystem can hook to super_operations.metadata_incore to get metadata in
> specific approach. Next patch will give an example how to implement
> .metadata_incore in btrfs.
>
Please cc Michael Kerrisk <mtk.manpages@xxxxxxxxx> and
linux-api@xxxxxxxxxxxxxxxx I'm sure that assistance writing the
manpage would be appreciated.
>
> ...
>
> /*
> + * Copy info about metadata in memory to userspace
> + * Returns:
> + * > 0, number of metadata_incore_ent entries copied to userspace
> + * = 0, no more metadata
> + * < 0, error
> + */
> +static int ioctl_metadata_incore(struct file *filp, void __user *argp)
> +{
> + struct super_block *sb = filp->f_path.dentry->d_inode->i_sb;
> + struct metadata_incore_args args;
> + struct metadata_incore_ent ent;
> + loff_t offset, last_offset = 0;
> + ssize_t size, last_size = 0;
> + __u64 __user vec_addr;
> + int entries = 0;
> +
> + if (!sb->s_op->metadata_incore)
> + return -EOPNOTSUPP;
EOPNOTSUPP is a networking errno - it doesn't seem appropriate for an
fs ioctl.
> + if (copy_from_user(&args, (struct metadata_incore_args __user *)argp,
Unneeded typecast.
> + sizeof(args)))
> + return -EFAULT;
> +
> + /* Check the start address: needs to be page-aligned.. */
Why? The comment should tell me this.
> + if (args.offset & ~PAGE_CACHE_MASK)
> + return -EINVAL;
> +
> + if ((args.vec_size % sizeof(struct metadata_incore_ent)) != 0)
> + return -EINVAL;
> +
> + if (!access_ok(VERIFY_WRITE, args.vec_addr, args.vec_size))
Seems unneccessary - copy_to_user() checks this.
> + return -EFAULT;
> +
> + offset = args.offset;
> +
> + ent.unused = 0;
> + vec_addr = args.vec_addr;
> +
> + while (vec_addr < args.vec_addr + args.vec_size) {
> + if (signal_pending(current))
> + return -EINTR;
> + cond_resched();
> +
> + if (sb->s_op->metadata_incore(sb, &offset, &size) < 0)
> + break;
> + /* A merge or offset == 0 */
> + if (offset == last_offset + last_size) {
> + last_size += size;
> + offset = offset + size;
> + continue;
> + }
> + ent.offset = last_offset;
> + ent.size = last_size;
> + if (copy_to_user((void *)(long)vec_addr, &ent, sizeof(ent)))
> + return -EFAULT;
> + vec_addr += sizeof(ent);
> + entries++;
> +
> + last_offset = offset;
> + last_size = size;
> + ent.unused = 0;
> + offset = offset + size;
> + }
> +
> + if (last_size > 0 && vec_addr < args.vec_addr + args.vec_size) {
> + ent.offset = last_offset;
> + ent.size = last_size;
> + if (copy_to_user((void *)(long)vec_addr, &ent, sizeof(ent)))
> + return -EFAULT;
> + entries++;
> + }
> +
> + return entries;
> +}
> +
> +/*
> * When you add any new common ioctls to the switches above and below
> * please update compat_sys_ioctl() too.
> *
>
> ...
>
--
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
