svn commit: r365643 - head/bin/cp

[Rodney W. Grimes]

> cp is already fixed, people are still feeling the fallout of being
> within those revisions and needing to bootstrap their own cp. We can
> reduce the number of components these invocations rely on trivially to
> shell built-in mechanics, why not do so?

I would even go further, I would like to see the dependency on
/dev/null removed from the build, and the boot process.

>From the build because it means I would no longer have to
mount /dev in my chroots, and from the boot because I
hate to say it, but we often scribble in /dev before
devfs is mounted and if you look at root file systems
mounted on other systems you well often find a /dev/null
FILE that got created during the boot process from a >/dev/null
before devfs was mounted.

> 
> On Tue, Sep 22, 2020 at 4:40 PM Warner Losh <imp at bsdimp.com> wrote:
> >
> > So why do we need a workaround at all? cp /dev/null has been fixed, and that's way more important to get right.
> >
> > I don't want to paper-over issues with this at all, though if we use the host's (now broken) cp, I suppose that might be OK in the short term. If that's the case, then maybe this is OK.
> >
> > Otherwise, I'd strongly prefer we fix cp.
> >
> > Warner
> >
> > On Tue, Sep 22, 2020 at 3:31 PM Alan Somers <asomers at freebsd.org> wrote:
> >>
> >> +1.
> >>
> >> On Tue, Sep 22, 2020 at 3:27 PM Kyle Evans <kevans at freebsd.org> wrote:
> >>>
> >>> I'm running a build at the suggestion of mjg to confirm there aren't
> >>> any others hiding that can be converted, and I will commit after I've
> >>> verified that this is it.
> >>>
> >>> On Tue, Sep 22, 2020 at 4:02 PM Alan Somers <asomers at freebsd.org> wrote:
> >>> >
> >>> > LGTM.
> >>> >
> >>> > On Tue, Sep 22, 2020 at 2:59 PM Kyle Evans <kevans at freebsd.org> wrote:
> >>> >>
> >>> >> Perhaps:
> >>> >>
> >>> >> diff --git a/stand/i386/zfsboot/Makefile b/stand/i386/zfsboot/Makefile
> >>> >> index ff315abc0ef..7e362b43a39 100644
> >>> >> --- a/stand/i386/zfsboot/Makefile
> >>> >> +++ b/stand/i386/zfsboot/Makefile
> >>> >> @@ -81,7 +81,7 @@ zfsboot.ld: zfsboot.ldr zfsboot.bin ${BTXKERN}
> >>> >>             -o ${.TARGET} -P 1 zfsboot.bin
> >>> >>
> >>> >>  zfsboot.ldr:
> >>> >> -       cp /dev/null ${.TARGET}
> >>> >> +       :> ${.TARGET}
> >>> >>
> >>> >>  zfsboot.bin: zfsboot.out
> >>> >>         ${OBJCOPY} -S -O binary zfsboot.out ${.TARGET}
> >>> >> diff --git a/stand/libsa/Makefile b/stand/libsa/Makefile
> >>> >> index effece9e01b..63cd46a9c54 100644
> >>> >> --- a/stand/libsa/Makefile
> >>> >> +++ b/stand/libsa/Makefile
> >>> >> @@ -122,7 +122,7 @@ beforedepend:
> >>> >>         ln -sf ${SRCTOP}/include/arpa/inet.h arpa/inet.h; \
> >>> >>         ln -sf ${SRCTOP}/include/arpa/tftp.h arpa/tftp.h; \
> >>> >>         for i in _time.h _strings.h _string.h; do \
> >>> >> -               [ -f xlocale/$$i ] || cp /dev/null xlocale/$$i; \
> >>> >> +               [ -f xlocale/$$i ] || :> xlocale/$$i; \
> >>> >>         done; \
> >>> >>         for i in ${STAND_H_INC}; do \
> >>> >>                 ln -sf ${SASRC}/stand.h $$i; \
> >>> >>
> >>> >>
> >>> >> On Tue, Sep 22, 2020 at 3:58 PM Alan Somers <asomers at freebsd.org> wrote:
> >>> >> >
> >>> >> > Looks like two places in stand.  Is there any reason why Mateusz's suggestion wouldn't work?
> >>> >> >
> >>> >> > > rg -g Makefile 'cp.*/dev/null'
> >>> >> > stand/libsa/Makefile
> >>> >> > 125: [ -f xlocale/$$i ] || cp /dev/null xlocale/$$i; \
> >>> >> >
> >>> >> > stand/i386/zfsboot/Makefile
> >>> >> > 82: cp /dev/null ${.TARGET}
> >>> >> >
> >>> >> > On Tue, Sep 22, 2020 at 2:54 PM Mateusz Guzik <mjguzik at gmail.com> wrote:
> >>> >> >>
> >>> >> >> Can we instead add a workaround to the build tree?
> >>> >> >>
> >>> >> >> Where is cp /dev/null coming from anyway? Perhaps this can be patched
> >>> >> >> to touch the target file.
> >>> >> >>
> >>> >> >> On 9/22/20, Alan Somers <asomers at freebsd.org> wrote:
> >>> >> >> > On Tue, Sep 22, 2020 at 2:48 PM Kyle Evans <kevans at freebsd.org> wrote:
> >>> >> >> >
> >>> >> >> >> On Fri, Sep 11, 2020 at 3:49 PM Alan Somers <asomers at freebsd.org> wrote:
> >>> >> >> >> >
> >>> >> >> >> > Author: asomers
> >>> >> >> >> > Date: Fri Sep 11 20:49:36 2020
> >>> >> >> >> > New Revision: 365643
> >>> >> >> >> > URL: https://svnweb.freebsd.org/changeset/base/365643
> >>> >> >> >> >
> >>> >> >> >> > Log:
> >>> >> >> >> >   cp: fall back to read/write if copy_file_range fails
> >>> >> >> >> >
> >>> >> >> >> >   Even though copy_file_range has a file-system agnostic version, it
> >>> >> >> >> still
> >>> >> >> >> >   fails on devfs (perhaps because the file descriptor is non-seekable?)
> >>> >> >> >> In
> >>> >> >> >> >   that case, fallback to old-fashioned read/write. Fixes
> >>> >> >> >> >   "cp /dev/null /tmp/null"
> >>> >> >> >> >
> >>> >> >> >>
> >>> >> >> >> Hi,
> >>> >> >> >>
> >>> >> >> >> Any objection to adding a quick UPDATING entry for this? I'm seeing
> >>> >> >> >> occasional reports of this breakage as recent as today on IRC from
> >>> >> >> >> folks that were a little bit thrown off by this because it throws up
> >>> >> >> >> fairly far into the build and looks like a stand build regression
> >>> >> >> >> instead of a cp regression.
> >>> >> >> >>
> >>> >> >> >> Thanks,
> >>> >> >> >>
> >>> >> >> >> Kyle Evans
> >>> >> >> >>
> >>> >> >> >
> >>> >> >> > No objection.  Can you suggest the proper wording?
> >>> >> >> > _______________________________________________
> >>> >> >> > svn-src-all at freebsd.org mailing list
> >>> >> >> > https://lists.freebsd.org/mailman/listinfo/svn-src-all
> >>> >> >> > To unsubscribe, send any mail to "svn-src-all-unsubscribe at freebsd.org"
> >>> >> >> >
> >>> >> >>
> >>> >> >>
> >>> >> >> --
> >>> >> >> Mateusz Guzik <mjguzik gmail.com>
> 

-- 
Rod Grimes                                                 rgrimes at freebsd.org