public inbox for binutils@sourceware.org
 help / color / mirror / Atom feed
* Re: Update objcopy's --section-alignment option
@ 2024-04-11  7:36 Alan Modra
  0 siblings, 0 replies; 2+ messages in thread
From: Alan Modra @ 2024-04-11  7:36 UTC (permalink / raw)
  To: binutils

ubsan: shift exponent 255 is too large for 64-bit type

I should have known oss-fuzz wouldn't be satisfied so easily.  The pef
format allows quite silly section alignments in object files.

	* objcopy.c (setup_section): Limit shift exponent when checking
	vma and lma for alignment.

diff --git a/binutils/objcopy.c b/binutils/objcopy.c
index d9abfdfbb39..d91ba123c01 100644
--- a/binutils/objcopy.c
+++ b/binutils/objcopy.c
@@ -4340,7 +4340,9 @@ setup_section (bfd *ibfd, sec_ptr isection, void *obfdarg)
      and the VMA was not set by the user
      and the section does not have relocations associated with it
      then warn the user.  */
-  if ((osection->vma & (((bfd_vma) 1 << alignment) - 1)) != 0
+  if (osection->vma != 0
+      && (alignment >= sizeof (bfd_vma) * CHAR_BIT
+	  || (osection->vma & (((bfd_vma) 1 << alignment) - 1)) != 0)
       && alignment != bfd_section_alignment (isection)
       && change_section_address == 0
       && ! vma_set_by_user
@@ -4352,7 +4354,9 @@ setup_section (bfd *ibfd, sec_ptr isection, void *obfdarg)
   /* Similar check for a non-aligned LMA.
      FIXME: Since this is only an LMA, maybe it does not matter if
      it is not aligned ?  */
-  if ((osection->lma & (((bfd_vma) 1 << alignment) - 1)) != 0
+  if (osection->lma != 0
+      && (alignment >= sizeof (bfd_vma) * CHAR_BIT
+	  || (osection->lma & (((bfd_vma) 1 << alignment) - 1)) != 0)
       && alignment != bfd_section_alignment (isection)
       && change_section_address == 0
       && ! lma_set_by_user

-- 
Alan Modra
Australia Development Lab, IBM

^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: Update objcopy's --section-alignment option
@ 2024-04-04 10:32 Alan Modra
  0 siblings, 0 replies; 2+ messages in thread
From: Alan Modra @ 2024-04-04 10:32 UTC (permalink / raw)
  To: binutils

ubsan: left shift of 1 by 31 places cannot be represented in type 'int'

	* objcopy.c (setup_section): Avoid undefined behaviour when
	checking vma and lma for alignment.

diff --git a/binutils/objcopy.c b/binutils/objcopy.c
index 77ab9080946..d9abfdfbb39 100644
--- a/binutils/objcopy.c
+++ b/binutils/objcopy.c
@@ -4340,7 +4340,7 @@ setup_section (bfd *ibfd, sec_ptr isection, void *obfdarg)
      and the VMA was not set by the user
      and the section does not have relocations associated with it
      then warn the user.  */
-  if (osection->vma & ((1 << alignment) - 1)
+  if ((osection->vma & (((bfd_vma) 1 << alignment) - 1)) != 0
       && alignment != bfd_section_alignment (isection)
       && change_section_address == 0
       && ! vma_set_by_user
@@ -4352,7 +4352,7 @@ setup_section (bfd *ibfd, sec_ptr isection, void *obfdarg)
   /* Similar check for a non-aligned LMA.
      FIXME: Since this is only an LMA, maybe it does not matter if
      it is not aligned ?  */
-  if (osection->lma & ((1 << alignment) - 1)
+  if ((osection->lma & (((bfd_vma) 1 << alignment) - 1)) != 0
       && alignment != bfd_section_alignment (isection)
       && change_section_address == 0
       && ! lma_set_by_user

-- 
Alan Modra
Australia Development Lab, IBM

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2024-04-11  7:36 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2024-04-11  7:36 Update objcopy's --section-alignment option Alan Modra
  -- strict thread matches above, loose matches on Subject: below --
2024-04-04 10:32 Alan Modra

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).