[Toybox] [PATCH] macOS: numerous fixes.

[enh]

On Fri, Jun 7, 2019 at 8:16 PM Rob Landley <rob at landley.net> wrote:
>
>
>
> On 6/7/19 6:29 PM, enh via Toybox wrote:
> > This patch adds a BSD version of xgetmountlist (for the path ==
> > NULL case only), tested on macOS. It also papers over the differences
> > between macOS' and Linux's xattr APIs. For once I think the macOS
> > one is better. The imitation of mknodat I've had to write swings
> > things back in Linux's favor though.
> >
> > BSD calls f_frsize by the name f_iosize instead. (FWIW, it looks
> > like this is meaningless on Linux and actually meaningful on macOS.)
> >
> > I've added one #if to toys/ --- I'm calling pathconf in stat.c to
> > work around the absence of f_namelen, and have left a TODO with an
> > explanation. I'm not sure what the best fix is here, so punting.
> >
> > No-one can agree what f_fsid is, even if they're all basically the
> > same, so work around the `val` versus `__val` issue between macOS
> > and Linux.
> >
> > With this patch, it's now possible to build cp/mv/install and stat
> > for macOS too. (Which completes the set of "toybox commands currently
> > used on Linux as part of the AOSP build" if you ignore stuff that
> > deals with processes, which I doubt we'll ever be able to support
> > for lack of any API.)
>
> I have a tab open with
> https://svnweb.freebsd.org/base/head/lib/libprocstat/libprocstat.c?revision=341448&view=markup
> which is the BSD library they use instead of reading /proc, and under the covers
> it seems to be a sysctl() and then a whole lot of parsing of the binary
> structure(s) it returns?

macOS doesn't seem to have the sysctl or the kvm alternative.

*but*

nm says macOS ps calls proc_pidinfo. grep says that's in <libproc.h>.
so that seems to be the approved macOS API for this? no man pages
though...

> 5               if (procstat->type == PROCSTAT_SYSCTL) {
> 1786                    name[0] = CTL_KERN;
> 1787                    name[1] = KERN_PROC;
> 1788                    name[2] = env ? KERN_PROC_ENV : KERN_PROC_ARGS;
> 1789                    name[3] = kp->ki_pid;
> 1790                    len = nchr;
> 1791                    error = sysctl(name, nitems(name), av->buf, &len, NULL, 0);
>
> It's doable, I just need to get a development environment set up and clear a
> couple weeks of focus.
>
> The way toys/*/ps.c is set up, the only function that would need to be replaced
> is get_ps(). Unfortunately, that's about 300 lines of code (starting at line
> 700), and the new one can't cheat and go "the slot layout started life as the
> order fields occur in /proc/$PID/stat so we get 2/3 of them for free with a for
> (;;) loop reading long long and the units are what the filter/search/display
> code expects..."
>
> On the bright side, we don't have to populate _all_ the fields, at least not to
> start. The code won't care if most of them stay zero.
>
> (I need to break up ps.c so there aren't 5 commands and 1900 lines of code in a
> single file. I should make a lib/ps.c, I just need to turn the TT.stuff into a
> structure in lib/lib.h and do the surgery...
>
> Lemme get toysh to a good checkin point first...
>
> Rob