Re: [RFC PATCH 3/6] fs: provide generic compat_sys_p{read,write}64() implementations

[Linus Torvalds]

Honestly, I think the patches like this are disgusting:

On Sun, Mar 18, 2018 at 9:10 AM, Dominik Brodowski
<linux@xxxxxxxxxxxxxxxxxxxx> wrote:
> +#ifdef __ARCH_WANT_COMPAT_SYS_PREADWRITE64
> +#if defined(__ARCH_WANT_COMPAT_SYS_WITH_PADDING) && \
> +       defined(__ARCH_WANT_LE_COMPAT_SYS)
> +COMPAT_SYSCALL_DEFINE6(pread64, unsigned int, fd, char __user *, ubuf,
> +                      u32, count, u32, padding, u32, poslo, u32, poshi)
> +#elif defined(__ARCH_WANT_COMPAT_SYS_WITH_PADDING) && \
> +       !defined(__ARCH_WANT_LE_COMPAT_SYS)
> +COMPAT_SYSCALL_DEFINE6(pread64, unsigned int, fd, char __user *, ubuf,
> +                      u32, count, u32, padding, u32, poshi, u32, poslo)
> +#elif !defined(__ARCH_WANT_COMPAT_SYS_WITH_PADDING) && \
> +       defined(__ARCH_WANT_LE_COMPAT_SYS)
> +COMPAT_SYSCALL_DEFINE5(pread64, unsigned int, fd, char __user *, ubuf,
> +                      u32, count, u32, poslo, u32, poshi)
> +#else /* no padding, big endian */
> +COMPAT_SYSCALL_DEFINE5(pread64, unsigned int, fd, char __user *, ubuf,
> +                      u32, count, u32, poshi, u32, poslo)
> +#endif
> +{
> +#ifdef CONFIG_S390
> +       if ((compat_ssize_t) count < 0)
> +               return -EINVAL;
> +#endif /* CONFIG_S390 */

and we should just keep code like this entirely architecture-dependent.

It doesn't save all that many lines:

 19 files changed, 97 insertions(+), 106 deletions(-)

and the lines it adds are an unreadable mess compared to the lines it removes.

So please keep the high/low/padding stuff in the arch wrapper, and
just make them use "do_pwrite64()" and friends instead (or
"kern_pwrite64()", or whatever we ended up using as the kernel naming
for in-kernel system call wrappers).

                Linus