From: Andy Lutomirski <luto@amacapital.net>
To: Stas Sergeev <stsp@list.ru>
Cc: Linus Torvalds <torvalds@linux-foundation.org>,
Raymond Jennings <shentino@gmail.com>,
Cyrill Gorcunov <gorcunov@gmail.com>,
Pavel Emelyanov <xemul@parallels.com>,
Linux kernel <linux-kernel@vger.kernel.org>
Subject: Re: [regression] x86/signal/64: Fix SS handling for signals delivered to 64-bit programs breaks dosemu
Date: Wed, 2 Sep 2015 15:25:15 -0700 [thread overview]
Message-ID: <CALCETrV4TLiuy6jHE7wRFz1jiwXH7Tro6meBcBZ24eRAZwfS4g@mail.gmail.com> (raw)
In-Reply-To: <55E7774F.6060907@list.ru>
On Wed, Sep 2, 2015 at 3:25 PM, Stas Sergeev <stsp@list.ru> wrote:
> 03.09.2015 00:39, Andy Lutomirski пишет:
>
>> On Wed, Sep 2, 2015 at 2:01 PM, Stas Sergeev <stsp@list.ru> wrote:
>>>
>>> 02.09.2015 22:06, Andy Lutomirski пишет:
>>>
>>>> On Wed, Sep 2, 2015 at 11:23 AM, Stas Sergeev <stsp@list.ru> wrote:
>>>>>
>>>>> 02.09.2015 21:17, Andy Lutomirski пишет:
>>>>>>
>>>>>> On Wed, Sep 2, 2015 at 10:46 AM, Stas Sergeev <stsp@list.ru> wrote:
>>>>>>>
>>>>>>> 02.09.2015 17:21, Andy Lutomirski пишет:
>>>>>>>>>>
>>>>>>>>>> This should work for old DOSEMU. It's a bit gross, but it has the
>>>>>>>>>> nice benefit that everyone (even things that aren't DOSEMU) gain
>>>>>>>>>> the
>>>>>>>>>> ability to catch signals thrown from bogus SS contexts, which
>>>>>>>>>> probably
>>>>>>>>>> improves debugability. It's also nice to not have the SA flag.
>>>>>>>>>
>>>>>>>>> Pros:
>>>>>>>>> - No new SA flag
>>>>>>>>> - May improve debugability in some unknown scenario where people
>>>>>>>>> do not want to just use the new flag to get their things improved
>>>>>>>>>
>>>>>>>>> Cons:
>>>>>>>>> - Does not allow to cleanly use siglongjmp(), as then there is a
>>>>>>>>> risk
>>>>>>>>> to jump to 64bit code with bad SS
>>>>>>>>
>>>>>>>> What's the issue here? I don't understand.
>>>>>>>>
>>>>>>>> On musl, (sig)longjmp just restores rsp, rbx, rbp, and r12-r15, so
>>>>>>>> it
>>>>>>>> won't be affected. AFAIK all implementations of siglongjmp are
>>>>>>>> likely
>>>>>>>> to call sigprocmask or similar, and that will clobber SS. I'm not
>>>>>>>> aware of an implementation of siglongjmp that uses sigreturn.
>>>>>>>
>>>>>>> I am not saying siglongjmp() will be affected.
>>>>>>> Quite the opposite: it won't, which is bad. :)
>>>>>>> If you have always correct SS, you can use siglongjmp(). If you have
>>>>>>> broken SS at times, siglongjmp() will be an asking for troubles, as
>>>>>>> it exactly does not restore SS.
>>>>>>> dosemu could do a good use of siglongjmp() to get back to 64bit code
>>>>>>> from its sighandler.
>>>>>>
>>>>>> This seems like it would be relying unpleasantly heavily on libc
>>>>>> internals.
>>>>>
>>>>> Could you please clarify?
>>>>> If kernel always passes the right SS to the sighandler, then what's
>>>>> the problem?
>>>>
>>>> What's the exact siglongjmp usage you have in mind? Signal context
>>>> isn't normally involved AFAIK.
>>>
>>> dosemu needs 2 return pathes:
>>> 1. to DOS code
>>> 2. to 64bit code (dosemu is not all in a sighandler, right?)
>>>
>>> How it is currently achieved:
>>> dosemu1:
>>> 1. sigreturn() + iret (to DOS)
>>> 2. modify sigcontext -> sigreturn() (to 64bit asm helper)
>>>
>>> dosemu2:
>>> 1. sigreturn() + iret (to DOS)
>>> 2. modify sigcontext -> sigreturn() -> longjmp() (to 64bit C-coded)
>>
>> So you're modifying sigcontext such that it returns to a C function
>> that calls longjmp?
>
> Yes.
>
>>> How dosemu2 is supposed to do this:
>>> 1. sigreturn() (to DOS)
>>> 2. siglongjmp() (to 64bit C-coded)
>>
>> This should work fine on any kernel, right?
>
> 1 - not.
> 2 - maybe.
> If, as you say, siglongjmp() restores SS, I need to try it out.
> (there is also a problem that most siglongjmp() implementations
> are incompatible with sigaltstack(), but this is not what you can fix).
>
1 - definitely needs kernel changes. I was referring to #2.
2 - siglongjmp probably varies in its behavior across different libc
implementations. My point is that siglongjmp isn't a kernel-provided
thing.
>> For backwards compat, we either need the default behavior to be
>> unchanged, or we need the default behavior to be something that works
>> with existing dosemu. For existing dosemu, the only interesting cases
>> (I think) are signal delivery from *valid* 16-bit context, in which
>> case we need to preserve SS so that the signal handler can read it out
>> with mov ..., %ss, and sigreturn to 64-bit mode for the IRET
>> trampoline. For sigreturn, IIUC old dosemu will replace the saved CS
>> with a 64-bit code segment selector and won't touch the saved SS
>> because it doesn't know about the saved SS. Those dosemu versions
>> don't care what SS actually contains after sigreturn, because they're
>> immediately going to change it again using IRET. So we just need to
>> make sure we return without faulting.
>>
>> New dosemu2 would like to sigreturn directly back to 16-bit mode, so
>> it needs the kernel to honor the saved ss value and restore it,
>> possibly changed by dosemu.
>>
>> We obviously can't require old dosemu to set an SA flag to keep
>> working. But, if we can get away with it, I think it's somewhat
>> preferable not to require new DOSEMU to set an SA flag either.
>>
>> This has one major benefit at least: if new dosemu loads some random
>> library that installs some async signal handler (SIGALRM for example),
>> everything will work with regard to CS and SS.
>
> This case is covered if we do both things together: use
> your heuristic when SA_hyz is not set, and don't use it
> when its set. In this case dosemu2 will be able to request
> the proper SS delivery for its sighandlers, but the 3rd-party
> sighandlers will work too.
> I think we have never discussed the possibility of doing
> both things together, even though I have proposed it many
> times.
> After discussing this full-blown solution, we can think about
> reducing it, either by removing the heuristic or by removing
> SA_hyz, but discussing the full one would be nice too.
> Your opinion is likely that no one will use this SA_hyz in
> presence of the heuristic that "seems to work anyway".
> But in the light of extending it for TLS (with a new flag),
> I wouldn't be so sure. You can also document it as a
> needed flag when user code touches SS, and then it will
> be used. dosemu1 code that doesn't use it, will eventually
> be forgotten. So IMHO whether it will be used, is fully up
> to how will you market it. :)
I'll think about it. I'll think about FS and GS, too, although that's
still a longer-term thing.
--Andy
--
Andy Lutomirski
AMA Capital Management, LLC
next prev parent reply other threads:[~2015-09-02 22:25 UTC|newest]
Thread overview: 121+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-08-12 0:17 [regression] x86/signal/64: Fix SS handling for signals delivered to 64-bit programs breaks dosemu Stas Sergeev
2015-08-12 0:38 ` Andy Lutomirski
2015-08-12 8:02 ` Stas Sergeev
2015-08-12 16:19 ` Andy Lutomirski
2015-08-12 17:00 ` Stas Sergeev
2015-08-12 18:25 ` Andy Lutomirski
2015-08-12 18:55 ` Stas Sergeev
2015-08-12 19:20 ` Andy Lutomirski
2015-08-12 19:55 ` Stas Sergeev
2015-08-12 20:01 ` Andy Lutomirski
2015-08-12 20:14 ` Stas Sergeev
2015-08-12 20:28 ` Andy Lutomirski
2015-08-12 20:45 ` Stas Sergeev
2015-08-12 20:47 ` Andy Lutomirski
2015-08-12 20:55 ` Stas Sergeev
2015-08-12 21:37 ` Andy Lutomirski
2015-08-12 21:50 ` Stas Sergeev
2015-08-12 22:00 ` Andy Lutomirski
2015-08-13 8:39 ` Ingo Molnar
2015-08-13 10:14 ` Stas Sergeev
2015-08-13 12:44 ` Stas Sergeev
2015-08-13 14:58 ` Andy Lutomirski
2015-08-13 15:22 ` Stas Sergeev
2015-08-13 15:38 ` Andy Lutomirski
2015-08-13 16:03 ` Stas Sergeev
2015-08-13 16:09 ` Andy Lutomirski
2015-08-13 16:20 ` Stas Sergeev
2015-08-13 16:24 ` Andy Lutomirski
2015-08-13 16:38 ` Stas Sergeev
2015-08-13 16:42 ` Andy Lutomirski
2015-08-13 16:48 ` Stas Sergeev
2015-08-13 16:59 ` Andy Lutomirski
2015-08-13 17:13 ` Stas Sergeev
2015-08-13 17:17 ` Andy Lutomirski
2015-08-13 18:00 ` Stas Sergeev
2015-08-13 18:05 ` Andy Lutomirski
2015-08-13 18:19 ` Stas Sergeev
2015-08-13 18:25 ` Andy Lutomirski
2015-08-13 18:35 ` Stas Sergeev
2015-08-22 12:38 ` Ingo Molnar
2015-08-22 14:19 ` Stas Sergeev
2015-08-23 6:25 ` Ingo Molnar
2015-08-13 11:08 ` Stas Sergeev
2015-08-13 15:37 ` Linus Torvalds
2015-08-13 15:43 ` Andy Lutomirski
2015-08-13 16:19 ` Linus Torvalds
2015-08-13 16:23 ` Andy Lutomirski
2015-08-13 16:34 ` Linus Torvalds
2015-08-13 16:43 ` Linus Torvalds
2015-08-13 16:44 ` Andy Lutomirski
2015-08-13 17:00 ` Brian Gerst
2015-08-18 6:29 ` Stas Sergeev
2015-08-18 22:42 ` Andy Lutomirski
2015-08-18 22:47 ` Andy Lutomirski
2015-08-19 9:35 ` Stas Sergeev
2015-08-19 15:46 ` Andy Lutomirski
2015-08-19 16:30 ` Stas Sergeev
2015-09-02 5:12 ` Andy Lutomirski
2015-09-02 9:17 ` Stas Sergeev
2015-09-02 14:21 ` Andy Lutomirski
2015-09-02 15:02 ` Andy Lutomirski
2015-09-02 17:46 ` Stas Sergeev
2015-09-02 18:17 ` Andy Lutomirski
2015-09-02 18:23 ` Stas Sergeev
2015-09-02 19:06 ` Andy Lutomirski
2015-09-02 21:01 ` Stas Sergeev
2015-09-02 21:39 ` Andy Lutomirski
2015-09-02 22:25 ` Stas Sergeev
2015-09-02 22:25 ` Andy Lutomirski [this message]
2015-09-02 23:01 ` Stas Sergeev
2015-08-19 10:10 ` Stas Sergeev
2015-08-19 15:35 ` Andy Lutomirski
2015-08-14 8:10 ` Cyrill Gorcunov
2015-08-13 17:51 ` Stas Sergeev
2015-08-13 18:35 ` Linus Torvalds
2015-08-13 18:41 ` Andy Lutomirski
2015-08-13 19:05 ` Stas Sergeev
2015-08-13 19:49 ` Andy Lutomirski
2015-08-13 20:09 ` Stas Sergeev
2015-08-13 19:53 ` Linus Torvalds
2015-08-13 20:08 ` Cyrill Gorcunov
2015-08-13 20:09 ` Linus Torvalds
2015-08-13 21:42 ` Raymond Jennings
2015-08-13 21:46 ` Linus Torvalds
2015-08-13 22:01 ` Raymond Jennings
2015-08-13 22:05 ` Stas Sergeev
2015-08-13 23:05 ` Linus Torvalds
2015-08-13 23:18 ` Linus Torvalds
2015-08-13 23:35 ` Raymond Jennings
2015-08-13 23:43 ` Stas Sergeev
2015-08-14 0:02 ` Linus Torvalds
2015-08-13 22:02 ` Stas Sergeev
2015-08-13 22:11 ` Andy Lutomirski
2015-08-13 22:25 ` Stas Sergeev
2015-08-13 22:29 ` Andy Lutomirski
2015-08-13 22:51 ` Stas Sergeev
2015-08-13 23:00 ` Andy Lutomirski
2015-08-13 23:17 ` Stas Sergeev
2015-08-14 0:00 ` Stas Sergeev
2015-08-14 0:05 ` Andy Lutomirski
2015-08-14 0:17 ` Stas Sergeev
2015-08-14 0:27 ` Linus Torvalds
2015-08-14 0:50 ` Stas Sergeev
2015-08-14 1:21 ` Andy Lutomirski
2015-08-14 1:32 ` Stas Sergeev
2015-08-14 1:37 ` Andy Lutomirski
2015-08-14 2:03 ` Stas Sergeev
2015-08-18 6:19 ` Stas Sergeev
2015-08-14 0:08 ` Linus Torvalds
2015-08-14 0:24 ` Andy Lutomirski
2015-08-14 0:40 ` Linus Torvalds
2015-08-14 7:22 ` Cyrill Gorcunov
2015-08-14 10:02 ` Pavel Emelyanov
2015-08-14 10:53 ` Cyrill Gorcunov
2015-08-13 18:57 ` Stas Sergeev
2015-08-13 19:01 ` Andy Lutomirski
2015-08-13 19:13 ` Stas Sergeev
2015-08-13 19:37 ` Linus Torvalds
2015-08-13 19:59 ` Stas Sergeev
2015-08-13 20:07 ` Linus Torvalds
2015-08-18 6:40 ` Stas Sergeev
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=CALCETrV4TLiuy6jHE7wRFz1jiwXH7Tro6meBcBZ24eRAZwfS4g@mail.gmail.com \
--to=luto@amacapital.net \
--cc=gorcunov@gmail.com \
--cc=linux-kernel@vger.kernel.org \
--cc=shentino@gmail.com \
--cc=stsp@list.ru \
--cc=torvalds@linux-foundation.org \
--cc=xemul@parallels.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).