Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Fri, 26 Jan 2018 11:17:29 +0000
From: David Laight <David.Laight@...LAB.COM>
To: 'Andy Lutomirski' <>, Linus Torvalds
CC: the arch/x86 maintainers <>, LKML
	<>, Greg Kroah-Hartman
	<>, Alan Cox <>, Jann Horn
	<>, Samuel Neves <>, Dan Williams
	<>, Kernel Hardening
	<>, Borislav Petkov <>
Subject: RE: [PATCH] x86/retpoline/entry: Disable the entire SYSCALL64 fast
 path with retpolines on

From: Andy Lutomirski
> Sent: 25 January 2018 21:31
> Another potentially useful if rather minor optimization would be to
> rejigger the SYSCALL_DEFINE macros a bit.  Currently we treat all
> syscalls like this:
> long func(long arg0, long arg1, long arg2, long arg3, long arg4, long arg5);
> I wonder if we'd be better off doing:
> long func(const struct pt_regs *regs);
> and autogenerating:
> static long SyS_read(const struct pt_regs *regs)
> {
>    return sys_reg(regs->di, ...);
> }

NetBSD (and the other BSD?) defines a structure for the arguments to each syscall.
On systems where all function arguments are put on stack the user stack that
contains the arguments is copied into a kernel buffer.
For amd64 I changed the register save area layout so that the arguments were in
the right order [1]. Then added an extra area for the extra arguments that had to be
read from the user stack.
Just passing a pointer into the save area has to be better than reading
all the values again.

[1] There was some horrid fallout from that :-(


Powered by blists - more mailing lists

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.