[v2] elf: Check invalid hole in PT_LOAD segments [BZ #28838]

  Changes in v2:

1. Update commit log.

commit 163f625cf9becbb82dfec63a29e566324129c0cd
Author: H.J. Lu <hjl.tools@gmail.com>
Date:   Tue Dec 21 12:35:47 2021 -0800

    elf: Remove excessive p_align check on PT_LOAD segments [BZ #28688]

removed the p_align check against the page size.  It caused the loader
error or crash on elf/tst-p_align3 when loading elf/tst-p_alignmod3.so,
which has the invalid p_align in PT_LOAD segments, added by

commit d8d94863ef125a392b929732b37e07dc927fbcd1
Author: H.J. Lu <hjl.tools@gmail.com>
Date:   Tue Dec 21 13:42:28 2021 -0800

The loader failure caused by a negative length passed to __mprotect is
random, depending on architecture and toolchain.  Update _dl_map_segments
to detect invalid holes.  This fixes BZ #28838.
---
 elf/dl-map-segments.h | 3 +++
 1 file changed, 3 insertions(+)
  

* H. J. Lu:

> Changes in v2:
>
> 1. Update commit log.
>
> commit 163f625cf9becbb82dfec63a29e566324129c0cd
> Author: H.J. Lu <hjl.tools@gmail.com>
> Date:   Tue Dec 21 12:35:47 2021 -0800
>
>     elf: Remove excessive p_align check on PT_LOAD segments [BZ #28688]
>
> removed the p_align check against the page size.  It caused the loader
> error or crash on elf/tst-p_align3 when loading elf/tst-p_alignmod3.so,
> which has the invalid p_align in PT_LOAD segments, added by
>
> commit d8d94863ef125a392b929732b37e07dc927fbcd1
> Author: H.J. Lu <hjl.tools@gmail.com>
> Date:   Tue Dec 21 13:42:28 2021 -0800
>
> The loader failure caused by a negative length passed to __mprotect is
> random, depending on architecture and toolchain.  Update _dl_map_segments
> to detect invalid holes.  This fixes BZ #28838.
> ---
>  elf/dl-map-segments.h | 3 +++
>  1 file changed, 3 insertions(+)
>
> diff --git a/elf/dl-map-segments.h b/elf/dl-map-segments.h
> index b3513e7909..024175b2d5 100644
> --- a/elf/dl-map-segments.h
> +++ b/elf/dl-map-segments.h
> @@ -112,6 +112,9 @@ _dl_map_segments (struct link_map *l, int fd,
>               unallocated.  Then jump into the normal segment-mapping loop to
>               handle the portion of the segment past the end of the file
>               mapping.  */
> +	  if (__glibc_unlikely (loadcmds[nloadcmds - 1].mapstart <
> +				c->mapend))
> +	    return N_("ELF load command address/offset not page-aligned");
>            if (__glibc_unlikely
>                (__mprotect ((caddr_t) (l->l_addr + c->mapend),
>                             loadcmds[nloadcmds - 1].mapstart - c->mapend,

This looks okay to me.  Avoiding the invalid mprotect call is a good
idea.

Thanks,
Florian
  

On Mon, Feb 21, 2022 at 6:32 AM Florian Weimer <fweimer@redhat.com> wrote:
>
> * H. J. Lu:
>
> > Changes in v2:
> >
> > 1. Update commit log.
> >
> > commit 163f625cf9becbb82dfec63a29e566324129c0cd
> > Author: H.J. Lu <hjl.tools@gmail.com>
> > Date:   Tue Dec 21 12:35:47 2021 -0800
> >
> >     elf: Remove excessive p_align check on PT_LOAD segments [BZ #28688]
> >
> > removed the p_align check against the page size.  It caused the loader
> > error or crash on elf/tst-p_align3 when loading elf/tst-p_alignmod3.so,
> > which has the invalid p_align in PT_LOAD segments, added by
> >
> > commit d8d94863ef125a392b929732b37e07dc927fbcd1
> > Author: H.J. Lu <hjl.tools@gmail.com>
> > Date:   Tue Dec 21 13:42:28 2021 -0800
> >
> > The loader failure caused by a negative length passed to __mprotect is
> > random, depending on architecture and toolchain.  Update _dl_map_segments
> > to detect invalid holes.  This fixes BZ #28838.
> > ---
> >  elf/dl-map-segments.h | 3 +++
> >  1 file changed, 3 insertions(+)
> >
> > diff --git a/elf/dl-map-segments.h b/elf/dl-map-segments.h
> > index b3513e7909..024175b2d5 100644
> > --- a/elf/dl-map-segments.h
> > +++ b/elf/dl-map-segments.h
> > @@ -112,6 +112,9 @@ _dl_map_segments (struct link_map *l, int fd,
> >               unallocated.  Then jump into the normal segment-mapping loop to
> >               handle the portion of the segment past the end of the file
> >               mapping.  */
> > +       if (__glibc_unlikely (loadcmds[nloadcmds - 1].mapstart <
> > +                             c->mapend))
> > +         return N_("ELF load command address/offset not page-aligned");
> >            if (__glibc_unlikely
> >                (__mprotect ((caddr_t) (l->l_addr + c->mapend),
> >                             loadcmds[nloadcmds - 1].mapstart - c->mapend,
>
> This looks okay to me.  Avoiding the invalid mprotect call is a good
> idea.
>
>

I am checking it in and backporting it to 2.35 branch.

Thanks.