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 11:02:54 -0800
From: Andy Lutomirski <>
To: Linus Torvalds <>
Cc: Andy Lutomirski <>, Al Viro <>, 
	Alan Cox <>, David Laight <>, 
	"the arch/x86 maintainers" <>, LKML <>, 
	Greg Kroah-Hartman <>, 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

On Fri, Jan 26, 2018 at 10:54 AM, Linus Torvalds
<> wrote:
> On Fri, Jan 26, 2018 at 10:23 AM, Andy Lutomirski <> wrote:
>> The issue is that doing it this way gives us, effectively:
>> long sys_foo(int a, int b)
>> {
>>   body here;
>> }
>> long SyS_foo(const struct pt_regs *regs)
>> {
>>   return sys_foo(regs->di, regs->si);
>> }
>> whereas what we want is *static* long sys_foo(...).
> How about just marking 'sys_foo()' as being always_inline (but still
> not static)? Because the case that _matters_ is that SyS_foo(), thing
> when this is enabled.
> Sure, you'll get two copies of the code (one in SyS_foo(), the other
> being the callable-from C 'sys_foo()' that is exported and almost
> never used). But that seems a fairly small price to pay. We could
> leave it for later to try to get rid of the unused copies entirely.

I could do that, but Josh Triplett will yell at me.

Anyway, I'll fiddle with it.  This isn't exactly high priority.

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.