From: Kees Cook <keescook@chromium.org>
To: Thomas Gleixner <tglx@linutronix.de>
Cc: LKML <linux-kernel@vger.kernel.org>,
x86@kernel.org, linux-arch@vger.kernel.org,
Will Deacon <will@kernel.org>, Arnd Bergmann <arnd@arndb.de>,
Mark Rutland <mark.rutland@arm.com>,
Keno Fischer <keno@juliacomputing.com>,
Paolo Bonzini <pbonzini@redhat.com>,
kvm@vger.kernel.org,
Gabriel Krisman Bertazi <krisman@collabora.com>
Subject: Re: [patch V3 01/13] entry: Provide generic syscall entry functionality
Date: Fri, 17 Jul 2020 10:49:13 -0700 [thread overview]
Message-ID: <202007171045.FB4A586F1D@keescook> (raw)
In-Reply-To: <87d04vt98w.fsf@nanos.tec.linutronix.de>
On Thu, Jul 16, 2020 at 11:55:59PM +0200, Thomas Gleixner wrote:
> Kees Cook <keescook@chromium.org> writes:
> > On Thu, Jul 16, 2020 at 08:22:09PM +0200, Thomas Gleixner wrote:
> >> This code is needlessly duplicated and different in all
> >> architectures.
> >>
> >> Provide a generic version based on the x86 implementation which has all the
> >> RCU and instrumentation bits right.
> >
> > Ahh! You're reading my mind!
>
> I told you about that plan at the last conference over a beer :)
Thank you for incepting it in my head, then! ;)
> > [1] https://lore.kernel.org/lkml/20200716193141.4068476-2-krisman@collabora.com/
>
> Saw that fly by. *shudder*
Aw, it's nice. Better emulation! :)
>
> >> +/*
> >> + * Define dummy _TIF work flags if not defined by the architecture or for
> >> + * disabled functionality.
> >> + */
> >
> > When I was thinking about this last week I was pondering having a split
> > between the arch-agnositc TIF flags and the arch-specific TIF flags, and
> > that each arch could have a single "there is agnostic work to be done"
> > TIF in their thread_info, and the agnostic flags could live in
> > task_struct or something. Anyway, I'll keep reading...
>
> That's going to be nasty. We rather go and expand the TIF storage to
> 64bit. And then do the following in a generic header:
I though the point was to make the TIF_WORK check as fast as possible,
even on the 32-bit word systems. I mean it's not a huge performance hit,
but *shrug*
>
> #ifndef TIF_ARCH_SPECIFIC
> # define TIF_ARCH_SPECIFIC
> #endif
>
> enum tif_bits {
> TIF_NEED_RESCHED = 0,
> TIF_...,
> TIF_LAST_GENERIC,
> TIF_ARCH_SPECIFIC,
> };
>
> and in the arch specific one:
>
> #define TIF_ARCH_SPECIFIC \
> TIF_ARCH_1, \
> TIF_ARCH_2,
>
> or something like that.
Okay, yeah, that can work.
> > There's been some recent confusion over "has the syscall changed,
> > or did seccomp request it be skipped?" that was explored in arm64[2]
> > (though I see Will and Keno in CC already). There might need to be a
> > clearer way to distinguish between "wild userspace issued a -1 syscall"
> > and "seccomp or ptrace asked for the syscall to be skipped". The
> > difference is mostly about when ENOSYS gets set, with respect to calls
> > to syscall_set_return_value(), but if the syscall gets changed, the arch
> > may need to recheck the value and consider ENOSYS, etc. IIUC, what Will
> > ended up with[3] was having syscall_trace_enter() return the syscall return
> > value instead of the new syscall.
>
> I was chatting with Will about that yesterday. IIRC he plans to fix the
> immediate issue on arm64 first and then move arm64 over to the generic
> variant. That's the reason why I reshuffled the patch series so the
> generic parts are first which allows me to provide will a branch with
> just those. If there are any changes needed we can just feed them back
> into that branch and fixup the affected architecture trees.
>
> IOW, that should not block progress on this stuff.
Ok, great! I just wanted to make sure that didn't surprise anyone. :)
--
Kees Cook
next prev parent reply other threads:[~2020-07-17 17:49 UTC|newest]
Thread overview: 31+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-07-16 18:22 [patch V3 00/13] entry, x86, kvm: Generic entry/exit functionality for host and guest Thomas Gleixner
2020-07-16 18:22 ` [patch V3 01/13] entry: Provide generic syscall entry functionality Thomas Gleixner
2020-07-16 20:52 ` Kees Cook
2020-07-16 21:55 ` Thomas Gleixner
2020-07-17 17:49 ` Kees Cook [this message]
2020-07-17 19:29 ` Thomas Gleixner
2020-07-17 21:56 ` Andy Lutomirski
2020-07-18 14:16 ` Thomas Gleixner
2020-07-18 14:41 ` Andy Lutomirski
2020-07-19 10:17 ` Thomas Gleixner
2020-07-19 15:25 ` Andy Lutomirski
2020-07-20 6:50 ` Thomas Gleixner
2020-07-27 22:28 ` Andy Lutomirski
2020-07-16 18:22 ` [patch V3 02/13] entry: Provide generic syscall exit function Thomas Gleixner
2020-07-16 20:55 ` Kees Cook
2020-07-16 21:28 ` Thomas Gleixner
2020-07-16 18:22 ` [patch V3 03/13] entry: Provide generic interrupt entry/exit code Thomas Gleixner
2020-07-16 18:22 ` [patch V3 04/13] entry: Provide infrastructure for work before exiting to guest mode Thomas Gleixner
2020-07-16 18:22 ` [patch V3 05/13] x86/entry: Consolidate check_user_regs() Thomas Gleixner
2020-07-16 20:56 ` Kees Cook
2020-07-16 18:22 ` [patch V3 06/13] x86/entry: Consolidate 32/64 bit syscall entry Thomas Gleixner
2020-07-16 18:22 ` [patch V3 07/13] x86/ptrace: Provide pt_regs helpers for entry/exit Thomas Gleixner
2020-07-16 20:57 ` Kees Cook
2020-07-16 18:22 ` [patch V3 08/13] x86/entry: Use generic syscall entry function Thomas Gleixner
2020-07-16 21:13 ` Kees Cook
2020-07-16 21:33 ` Thomas Gleixner
2020-07-16 18:22 ` [patch V3 09/13] x86/entry: Use generic syscall exit functionality Thomas Gleixner
2020-07-16 18:22 ` [patch V3 10/13] x86/entry: Cleanup idtentry_entry/exit_user Thomas Gleixner
2020-07-16 18:22 ` [patch V3 11/13] x86/entry: Use generic interrupt entry/exit code Thomas Gleixner
2020-07-16 18:22 ` [patch V3 12/13] x86/entry: Cleanup idtentry_enter/exit Thomas Gleixner
2020-07-16 18:22 ` [patch V3 13/13] x86/kvm: Use generic exit to guest work function Thomas Gleixner
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=202007171045.FB4A586F1D@keescook \
--to=keescook@chromium.org \
--cc=arnd@arndb.de \
--cc=keno@juliacomputing.com \
--cc=krisman@collabora.com \
--cc=kvm@vger.kernel.org \
--cc=linux-arch@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mark.rutland@arm.com \
--cc=pbonzini@redhat.com \
--cc=tglx@linutronix.de \
--cc=will@kernel.org \
--cc=x86@kernel.org \
/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).