RE: [PATCH 3/3] quota: simplify the quotactl compat handling

[David Laight <David.Laight@ACULAB.COM>]

From: Christoph Hellwig
> Sent: 31 July 2020 13:22
> 
> Fold the misaligned u64 workarounds into the main quotactl flow instead
> of implementing a separate compat syscall handler.
> 
...
> +static int compat_copy_fs_quota_stat(struct compat_fs_quota_stat __user *to,
> +		struct fs_quota_stat *from)
> +{
> +	if (put_user(from->qs_version, &to->qs_version) ||
> +	    put_user(from->qs_flags, &to->qs_flags) ||
> +	    put_user(from->qs_pad, &to->qs_pad) ||
> +	    compat_copy_fs_qfilestat(&to->qs_uquota, &from->qs_uquota) ||
> +	    compat_copy_fs_qfilestat(&to->qs_gquota, &from->qs_gquota) ||
> +	    put_user(from->qs_incoredqs, &to->qs_incoredqs) ||
> +	    put_user(from->qs_btimelimit, &to->qs_btimelimit) ||
> +	    put_user(from->qs_itimelimit, &to->qs_itimelimit) ||
> +	    put_user(from->qs_rtbtimelimit, &to->qs_rtbtimelimit) ||
> +	    put_user(from->qs_bwarnlimit, &to->qs_bwarnlimit) ||
> +	    put_user(from->qs_iwarnlimit, &to->qs_iwarnlimit))
> +		return -EFAULT;
> +	return 0;
> +}

That might look better as a 'noinline' function that copied
all the fields into an on-stack struct compat_fs_quota_stat
and then did a single copy_to_user().

(I do 'like' qs_pad - I wonder what the person who added
it was smoking.)

	David

-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
Registration No: 1397386 (Wales)