From mboxrd@z Thu Jan  1 00:00:00 1970
From: szabolcs.nagy@arm.com (Szabolcs Nagy)
Date: Thu, 26 May 2016 15:50:01 +0100
Subject: [PATCH 01/23] all: syscall wrappers: add documentation
In-Reply-To: <20160526142057.GA7456@e104818-lin.cambridge.arm.com>
References: <6293194.tGy03QJ9ME@wuerfel>
 <20160525.135039.244098606649448826.davem@davemloft.net>
 <6407614.fdv5XFSBue@wuerfel>
 <20160525.142821.1719403997976778673.davem@davemloft.net>
 <20160526142057.GA7456@e104818-lin.cambridge.arm.com>
Message-ID: <57470D19.2000501@arm.com>
To: linux-arm-kernel@lists.infradead.org
List-Id: linux-arm-kernel.lists.infradead.org

On 26/05/16 15:20, Catalin Marinas wrote:
> While writing the above, I realised the current ILP32 patches still miss
> on converting pointers passed from user space (unless I got myself
> confused in macros). The new __SC_WRAP() and COMPAT_SYSCALL_WRAPx()
> macros take care of zero or sign extension via __SC_COMPAT_CAST().
> However, we have two more existing cases which I don't see covered:
> 
> a) Native syscalls taking a pointer argument and invoked directly from
>    ILP32. For example, sys_read() takes a pointer but I don't see any
>    __SC_WRAP added by patch 5
> 
> b) Current compat syscalls taking a pointer argument. For example,
>    compat_sys_vmsplice() gets the iov32 pointer and the compiler assumes
>    it is a 64-bit variable. I don't see where the upper half is zeroed
> 

on x32 sign/zero extension is currently left to userspace,
which is difficult to deal with, (long long)arg does the
wrong thing for pointer args.

> We can solve (a) by adding more __SC_WRAP annotations in the generic
> unistd.h. For (b), we would need an __SC_DELOUSE with a bit of penalty
> on AArch32/compat support where it isn't needed. So maybe davem has a
> point on the overall impact of always zeroing the upper half of the
> arguments ;) (both from a performance and maintainability perspective).
> I guess this part of the ABI is still up for discussion.
>