This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: PING^2: [PATCH 0/2] nptl: Update struct pthread_unwind_buf
- From: "H.J. Lu" <hjl dot tools at gmail dot com>
- To: "Carlos O'Donell" <carlos at redhat dot com>
- Cc: Florian Weimer <fweimer at redhat dot com>, GNU C Library <libc-alpha at sourceware dot org>
- Date: Sat, 24 Feb 2018 07:20:51 -0800
- Subject: Re: PING^2: [PATCH 0/2] nptl: Update struct pthread_unwind_buf
- Authentication-results: sourceware.org; auth=none
- References: <CAMe9rOpoaaL=2B6cjLcGdPtwtWUFp5Sd+St4i1rA+eq5c30wnw@mail.gmail.com> <dca4154b-bc86-a914-8fee-b52f03c17bdf@redhat.com>
On Fri, Feb 23, 2018 at 9:48 PM, Carlos O'Donell <carlos@redhat.com> wrote:
> On 02/21/2018 07:33 AM, H.J. Lu wrote:
>> On Thu, Feb 15, 2018 at 12:02 PM, H.J. Lu <hjl.tools@gmail.com> wrote:
>>> On Fri, Feb 9, 2018 at 7:24 AM, H.J. Lu <hjl.tools@gmail.com> wrote:
>>>> On Fri, Feb 9, 2018 at 6:33 AM, Florian Weimer <fweimer@redhat.com> wrote:
>>>>> On 02/09/2018 03:13 PM, H.J. Lu wrote:
>>>>>>
>>>>>> I built glibc master with gcc-8.0.1 -mcet -fcf-protection. Some object
>>>>>> files do get CET marker as expected. But static executable isn't:
>>>>>
>>>>>
>>>>> I said static libraries. If I compile this code (based on the example from
>>>>> the manual page) on Fedora rawhide:
>>>>>
>>>>
>>>> Glibc never provides binary compatibility with static libraries. My suggestions
>>>> are
>>>>
>>>> 1. Recompile static libraries after CET is enabled in glibc. Or
>>>> 2. Don't compile static libraries with CET.
>>>>
>>>> BTW, we don't have space to save shadow stack register with existing
>>>> cancel buf.
>>>>
>>>
>>> PING:
>>>
>>> https://sourceware.org/ml/libc-alpha/2018-02/msg00290.html
>>> https://sourceware.org/ml/libc-alpha/2018-02/msg00291.html
>>>
>>
>> PING.
>
> These patches are pending the conclusion of the discussion that the
> shadow stack pointer does not need to be restored in the unwinding
> use case, and need only be placed in such a location that if it is
> written it writes within the size of the existing truncated unwind
> buffer.
>
> Please see:
> https://www.sourceware.org/ml/libc-alpha/2018-02/msg00679.html
Please see:
https://sourceware.org/ml/libc-alpha/2018-02/msg00696.html
--
H.J.