[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Thu, 3 Sep 2020 20:08:19 +0200
From: Arnd Bergmann <arnd@...db.de>
To: musl@...ts.openwall.com
Subject: Re: [PATCH 07/14] Emulate wait4 using waitid
On Thu, Sep 3, 2020 at 5:41 PM Stefan O'Rear <sorear@...tmail.com> wrote:
>
> On Thu, Sep 3, 2020, at 11:36 AM, Arnd Bergmann wrote:
> > On Thu, Sep 3, 2020 at 4:56 PM Stefan O'Rear <sorear@...tmail.com> wrote:
> > >
> > > On Thu, Sep 3, 2020, at 7:23 AM, Stefan O'Rear wrote:
> > > > + case CLD_STOPPED:
> > > > + case CLD_TRAPPED:
> > > > + sw = ((info.si_status&0xff) << 8) + 0x7f;
> > > > + break;
> > >
> > > This is trying to be defensive but it is the cause of the strace issue
> > > in the cover letter since the ptrace interface generates si_status
> > > greater than 8 bits which must be visible in WSTOPSIG; the v2 will not
> > > mask here.
> >
> > Ah, I was trying to find out what exactly the masking was for since
> > I did not have that in my original version of the same function for Arm:
>
> Where is this coming from? I didn't want to blindly copy code from glibc
> and I couldn't find any documentation of how the mapping is supposed
> to work other than the POSIX descriptions of the wait and waitid functions
> (which is why I missed the non-POSIX ptrace cases).
>
> If there's an authoritative description of the mapping I would like to match
> it exactly.
I did this prototype implementation to test the syscall changes before
I submitted them into the mainline kernel:
https://git.linaro.org/people/arnd.bergmann/musl-y2038.git/
This was never meant to get merged, but I made sure that I could
get LTP to pass just using the new minimum set of syscalls.
Arnd
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.
