Re: [RFC PATCH 00/10] Use __kernel_[u]long_t for x32 user space compatibility
From: David Daney
Date: Thu May 17 2012 - 19:26:08 EST
On 05/17/2012 04:11 PM, H. Peter Anvin wrote:
On 05/17/2012 04:07 PM, David Daney wrote:
Has anybody checked how this affects MIPS n32 userspace?
I think it totally breaks it.
Do you have any basis whatsoever for that statement?
You should have asked for a 'solid basis'.
My basis is that the name '__kernel_ulong_t' implies, in my mind, that
it would have the width of a kernel unsigned long.
Really it should be called something like __abi_alternate_ulong_t.
This should have
zero effect on any non-x32 platforms.
After further reflection on this, you are probably right.
Sorry for raising the alarm (or would that be crying Wolf?).
In addition, 109a1f32 (sysinfo: Use explicit types in<linux/sysinfo.h>)
is probably bad. I think it may need to be reverted, or somebody should
fix all the __kernel_{,u}long_t definitions for the ABI that may have
been broken by the change.
You realize __kernel_[u]long_t didn't even exist until the 3.4 kernel,
right?
Yes.
David Daney
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/