[PATCH v2] elf: Fix GL(dl_phdr) and GL(dl_phnum) for static builds [BZ #29864]

[PATCH v2] elf: Fix GL(dl_phdr) and GL(dl_phnum) for static builds [BZ #29864]

[Adhemerval Zanella]

The 73fc4e28b9464f0e refactor did not add the GL(dl_phdr) and
GL(dl_phnum) for static build, relying on the __ehdr_start symbol,
which is always added by the static linker, to get the correct values.

This is problematic in some ways:

  - The segment may see its in-memory size differ from its in-file
    size (or the binary may have holes).  The Linux has fixed is to
    provide concise values for both AT_PHDR and AT_PHNUM (commit
    0da1d5002745c - "fs/binfmt_elf: Fix AT_PHDR for unusual ELF files")

  - Some archs (alpha for instance) the hidden weak reference is not
    correctly pulled by the static linker and  __ehdr_start address
    end up being 0, which makes GL(dl_phdr) and GL(dl_phnum) have both
    invalid values (and triggering a segfault later on libc.so while
    accessing TLS variables).

The safer fix is to just restore the previous behavior to setup
GL(dl_phdr) and GL(dl_phnum) for static based on kernel auxv.  The
__ehdr_start fallback can also be simplified by not assuming weak
linkage (as for PIE).

The libc-static.c auxv init logic is moved to dl-support.c, since
the later is build without SHARED and then GLRO macro is defined
to access the variables directly.

The _dl_phdr is also assumed to be always non NULL, since an invalid
NULL values does not trigger TLS initialization (which is used in
various libc systems).

Checked on aarch64-linux-gnu, x86_64-linux-gnu, and i686-linux-gnu.

* Changes from v1:
  - Removed auxv_values implicit initialization.
  - Removed unused header inclusion.

---
 csu/libc-start.c | 21 ---------------------
 csu/libc-tls.c   | 25 ++++++++++++-------------
 elf/dl-support.c | 46 ++++++++++++++++++++++++++++++++--------------
 3 files changed, 44 insertions(+), 48 deletions(-)

diff --git a/csu/libc-start.c b/csu/libc-start.c
index ef26a89bc4..c3bb6d09bc 100644
--- a/csu/libc-start.c
+++ b/csu/libc-start.c
@@ -262,28 +262,7 @@ LIBC_START_MAIN (int (*main) (int, char **, char ** MAIN_AUXVEC_DECL),
   }
 #  endif
   _dl_aux_init (auxvec);
-  if (GL(dl_phdr) == NULL)
 # endif
-    {
-      /* Starting from binutils-2.23, the linker will define the
-         magic symbol __ehdr_start to point to our own ELF header
-         if it is visible in a segment that also includes the phdrs.
-         So we can set up _dl_phdr and _dl_phnum even without any
-         information from auxv.  */
-
-      extern const ElfW(Ehdr) __ehdr_start
-# if BUILD_PIE_DEFAULT
-	__attribute__ ((visibility ("hidden")));
-# else
-	__attribute__ ((weak, visibility ("hidden")));
-      if (&__ehdr_start != NULL)
-# endif
-        {
-          assert (__ehdr_start.e_phentsize == sizeof *GL(dl_phdr));
-          GL(dl_phdr) = (const void *) &__ehdr_start + __ehdr_start.e_phoff;
-          GL(dl_phnum) = __ehdr_start.e_phnum;
-        }
-    }
 
   __tunables_init (__environ);
 
diff --git a/csu/libc-tls.c b/csu/libc-tls.c
index 8dabf95bcc..cdf6442c02 100644
--- a/csu/libc-tls.c
+++ b/csu/libc-tls.c
@@ -119,19 +119,18 @@ __libc_setup_tls (void)
   __tls_pre_init_tp ();
 
   /* Look through the TLS segment if there is any.  */
-  if (_dl_phdr != NULL)
-    for (phdr = _dl_phdr; phdr < &_dl_phdr[_dl_phnum]; ++phdr)
-      if (phdr->p_type == PT_TLS)
-	{
-	  /* Remember the values we need.  */
-	  memsz = phdr->p_memsz;
-	  filesz = phdr->p_filesz;
-	  initimage = (void *) phdr->p_vaddr + main_map->l_addr;
-	  align = phdr->p_align;
-	  if (phdr->p_align > max_align)
-	    max_align = phdr->p_align;
-	  break;
-	}
+  for (phdr = _dl_phdr; phdr < &_dl_phdr[_dl_phnum]; ++phdr)
+    if (phdr->p_type == PT_TLS)
+      {
+	/* Remember the values we need.  */
+	memsz = phdr->p_memsz;
+	filesz = phdr->p_filesz;
+	initimage = (void *) phdr->p_vaddr + main_map->l_addr;
+	align = phdr->p_align;
+	if (phdr->p_align > max_align)
+	  max_align = phdr->p_align;
+	break;
+      }
 
   /* Calculate the size of the static TLS surplus, with 0 auditors.  */
   _dl_tls_static_surplus_init (0);
diff --git a/elf/dl-support.c b/elf/dl-support.c
index ef0e5591e3..9714f75db0 100644
--- a/elf/dl-support.c
+++ b/elf/dl-support.c
@@ -256,6 +256,25 @@ _dl_aux_init (ElfW(auxv_t) *av)
   for (int i = 0; i < array_length (auxv_values); ++i)
     auxv_values[i] = 0;
   _dl_parse_auxv (av, auxv_values);
+
+  _dl_phdr = (void*) auxv_values[AT_PHDR];
+  _dl_phnum = auxv_values[AT_PHNUM];
+
+  if (_dl_phdr == NULL)
+    {
+      /* Starting from binutils-2.23, the linker will define the
+         magic symbol __ehdr_start to point to our own ELF header
+         if it is visible in a segment that also includes the phdrs.
+         So we can set up _dl_phdr and _dl_phnum even without any
+         information from auxv.  */
+
+      extern const ElfW(Ehdr) __ehdr_start attribute_hidden;
+      assert (__ehdr_start.e_phentsize == sizeof *GL(dl_phdr));
+      _dl_phdr = (const void *) &__ehdr_start + __ehdr_start.e_phoff;
+      _dl_phnum = __ehdr_start.e_phnum;
+    }
+
+  assert (_dl_phdr != NULL);
 }
 #endif
 
@@ -324,20 +343,19 @@ _dl_non_dynamic_init (void)
   if (_dl_platform != NULL)
     _dl_platformlen = strlen (_dl_platform);
 
-  if (_dl_phdr != NULL)
-    for (const ElfW(Phdr) *ph = _dl_phdr; ph < &_dl_phdr[_dl_phnum]; ++ph)
-      switch (ph->p_type)
-	{
-	/* Check if the stack is nonexecutable.  */
-	case PT_GNU_STACK:
-	  _dl_stack_flags = ph->p_flags;
-	  break;
-
-	case PT_GNU_RELRO:
-	  _dl_main_map.l_relro_addr = ph->p_vaddr;
-	  _dl_main_map.l_relro_size = ph->p_memsz;
-	  break;
-	}
+  for (const ElfW(Phdr) *ph = _dl_phdr; ph < &_dl_phdr[_dl_phnum]; ++ph)
+    switch (ph->p_type)
+      {
+      /* Check if the stack is nonexecutable.  */
+      case PT_GNU_STACK:
+	_dl_stack_flags = ph->p_flags;
+	break;
+
+      case PT_GNU_RELRO:
+	_dl_main_map.l_relro_addr = ph->p_vaddr;
+	_dl_main_map.l_relro_size = ph->p_memsz;
+	break;
+      }
 
   call_function_static_weak (_dl_find_object_init);
 
-- 
2.34.1

Re: [PATCH v2] elf: Fix GL(dl_phdr) and GL(dl_phnum) for static builds [BZ #29864]

[Florian Weimer]

* Adhemerval Zanella:

> The 73fc4e28b9464f0e refactor did not add the GL(dl_phdr) and
> GL(dl_phnum) for static build, relying on the __ehdr_start symbol,
> which is always added by the static linker, to get the correct values.
>
> This is problematic in some ways:
>
>   - The segment may see its in-memory size differ from its in-file
>     size (or the binary may have holes).  The Linux has fixed is to
>     provide concise values for both AT_PHDR and AT_PHNUM (commit
>     0da1d5002745c - "fs/binfmt_elf: Fix AT_PHDR for unusual ELF files")
>
>   - Some archs (alpha for instance) the hidden weak reference is not
>     correctly pulled by the static linker and  __ehdr_start address
>     end up being 0, which makes GL(dl_phdr) and GL(dl_phnum) have both
>     invalid values (and triggering a segfault later on libc.so while
>     accessing TLS variables).
>
> The safer fix is to just restore the previous behavior to setup
> GL(dl_phdr) and GL(dl_phnum) for static based on kernel auxv.  The
> __ehdr_start fallback can also be simplified by not assuming weak
> linkage (as for PIE).
>
> The libc-static.c auxv init logic is moved to dl-support.c, since
> the later is build without SHARED and then GLRO macro is defined
> to access the variables directly.
>
> The _dl_phdr is also assumed to be always non NULL, since an invalid
> NULL values does not trigger TLS initialization (which is used in
> various libc systems).
>
> Checked on aarch64-linux-gnu, x86_64-linux-gnu, and i686-linux-gnu.
>
> * Changes from v1:
>   - Removed auxv_values implicit initialization.
>   - Removed unused header inclusion.

This version looks okay to me.

Reviewed-by: Florian Weimer <fweimer@redhat.com>

Thanks,
Florian

Re: [PATCH v2] elf: Fix GL(dl_phdr) and GL(dl_phnum) for static builds [BZ #29864]

[John Paul Adrian Glaubitz]

Hi Adhemveral!

On Thu, 2023-01-12 at 11:44 -0300, Adhemerval Zanella wrote:
> The 73fc4e28b9464f0e refactor did not add the GL(dl_phdr) and
> GL(dl_phnum) for static build, relying on the __ehdr_start symbol,
> which is always added by the static linker, to get the correct values.
> 
> This is problematic in some ways:
> 
>   - The segment may see its in-memory size differ from its in-file
>     size (or the binary may have holes).  The Linux has fixed is to
>     provide concise values for both AT_PHDR and AT_PHNUM (commit
>     0da1d5002745c - "fs/binfmt_elf: Fix AT_PHDR for unusual ELF files")
> 
>   - Some archs (alpha for instance) the hidden weak reference is not
>     correctly pulled by the static linker and  __ehdr_start address
>     end up being 0, which makes GL(dl_phdr) and GL(dl_phnum) have both
>     invalid values (and triggering a segfault later on libc.so while
>     accessing TLS variables).
> 
> The safer fix is to just restore the previous behavior to setup
> GL(dl_phdr) and GL(dl_phnum) for static based on kernel auxv.  The
> __ehdr_start fallback can also be simplified by not assuming weak
> linkage (as for PIE).
> 
> The libc-static.c auxv init logic is moved to dl-support.c, since
> the later is build without SHARED and then GLRO macro is defined
> to access the variables directly.
> 
> The _dl_phdr is also assumed to be always non NULL, since an invalid
> NULL values does not trigger TLS initialization (which is used in
> various libc systems).
> 
> Checked on aarch64-linux-gnu, x86_64-linux-gnu, and i686-linux-gnu.
> 
> * Changes from v1:
>   - Removed auxv_values implicit initialization.
>   - Removed unused header inclusion.

Any chance this fix can be backported to 2.36?

Thanks,
Adrian

-- 
 .''`.  John Paul Adrian Glaubitz
: :' :  Debian Developer
`. `'   Physicist
  `-    GPG: 62FF 8A75 84E0 2956 9546  0006 7426 3B37 F5B5 F913

Re: [PATCH v2] elf: Fix GL(dl_phdr) and GL(dl_phnum) for static builds [BZ #29864]

[Adhemerval Zanella Netto]

On 02/02/23 06:36, John Paul Adrian Glaubitz wrote:
> Hi Adhemveral!
> 
> On Thu, 2023-01-12 at 11:44 -0300, Adhemerval Zanella wrote:
>> The 73fc4e28b9464f0e refactor did not add the GL(dl_phdr) and
>> GL(dl_phnum) for static build, relying on the __ehdr_start symbol,
>> which is always added by the static linker, to get the correct values.
>>
>> This is problematic in some ways:
>>
>>   - The segment may see its in-memory size differ from its in-file
>>     size (or the binary may have holes).  The Linux has fixed is to
>>     provide concise values for both AT_PHDR and AT_PHNUM (commit
>>     0da1d5002745c - "fs/binfmt_elf: Fix AT_PHDR for unusual ELF files")
>>
>>   - Some archs (alpha for instance) the hidden weak reference is not
>>     correctly pulled by the static linker and  __ehdr_start address
>>     end up being 0, which makes GL(dl_phdr) and GL(dl_phnum) have both
>>     invalid values (and triggering a segfault later on libc.so while
>>     accessing TLS variables).
>>
>> The safer fix is to just restore the previous behavior to setup
>> GL(dl_phdr) and GL(dl_phnum) for static based on kernel auxv.  The
>> __ehdr_start fallback can also be simplified by not assuming weak
>> linkage (as for PIE).
>>
>> The libc-static.c auxv init logic is moved to dl-support.c, since
>> the later is build without SHARED and then GLRO macro is defined
>> to access the variables directly.
>>
>> The _dl_phdr is also assumed to be always non NULL, since an invalid
>> NULL values does not trigger TLS initialization (which is used in
>> various libc systems).
>>
>> Checked on aarch64-linux-gnu, x86_64-linux-gnu, and i686-linux-gnu.
>>
>> * Changes from v1:
>>   - Removed auxv_values implicit initialization.
>>   - Removed unused header inclusion.
> 
> Any chance this fix can be backported to 2.36?

Done.