[PATCH] Zero supplied stat buffers in functions that pretend to stat

[PATCH] Zero supplied stat buffers in functions that pretend to stat

[Gary Benson]

Hi all,

GDB has five places where it pretends to stat for bfd_openr_iovec.
Four of these only set the incoming buffer's st_size, leaving the
other fields unchanged, which is to say very likely populated with
random values from the stack.  remote_bfd_iovec_stat was fixed in
0a93529c56714b1da3d7106d3e0300764f8bb81c; this commit fixes the
other four.

Built and and regtested on RHEL6.6 x86_64.

Ok to commit?

Cheers,
Gary

gdb/ChangeLog:

	* jit.c (mem_bfd_iovec_stat): Zero supplied buffer.
	* minidebug.c (lzma_stat): Likewise.
	* solib-spu.c (spu_bfd_iovec_stat): Likewise.
	* spu-linux-nat.c (spu_bfd_iovec_stat): Likewise.
---
 gdb/ChangeLog       |    7 +++++++
 gdb/jit.c           |    1 +
 gdb/minidebug.c     |    1 +
 gdb/solib-spu.c     |    1 +
 gdb/spu-linux-nat.c |    1 +
 5 files changed, 11 insertions(+), 0 deletions(-)

diff --git a/gdb/jit.c b/gdb/jit.c
index e872c8f..f977ea6 100644
--- a/gdb/jit.c
+++ b/gdb/jit.c
@@ -126,6 +126,7 @@ mem_bfd_iovec_stat (struct bfd *abfd, void *stream, struct stat *sb)
 {
   struct target_buffer *buffer = (struct target_buffer*) stream;
 
+  memset (sb, 0, sizeof (struct stat));
   sb->st_size = buffer->size;
   return 0;
 }
diff --git a/gdb/minidebug.c b/gdb/minidebug.c
index cc20914..98c2187 100644
--- a/gdb/minidebug.c
+++ b/gdb/minidebug.c
@@ -241,6 +241,7 @@ lzma_stat (struct bfd *abfd,
 {
   struct gdb_lzma_stream *lstream = stream;
 
+  memset (sb, 0, sizeof (struct stat));
   sb->st_size = lzma_index_uncompressed_size (lstream->index);
   return 0;
 }
diff --git a/gdb/solib-spu.c b/gdb/solib-spu.c
index 250cf21..44fbf91 100644
--- a/gdb/solib-spu.c
+++ b/gdb/solib-spu.c
@@ -313,6 +313,7 @@ spu_bfd_iovec_stat (bfd *abfd, void *stream, struct stat *sb)
      table to find the extent of the last section but that seems
      pointless when the size is needed only for checks of other
      parsed values in dbxread.c.  */
+  memset (sb, 0, sizeof (struct stat));
   sb->st_size = INT_MAX;
   return 0;
 }
diff --git a/gdb/spu-linux-nat.c b/gdb/spu-linux-nat.c
index b0942a9..a043f53 100644
--- a/gdb/spu-linux-nat.c
+++ b/gdb/spu-linux-nat.c
@@ -313,6 +313,7 @@ spu_bfd_iovec_stat (struct bfd *abfd, void *stream, struct stat *sb)
      table to find the extent of the last section but that seems
      pointless when the size is needed only for checks of other
      parsed values in dbxread.c.  */
+  memset (sb, 0, sizeof (struct stat));
   sb->st_size = INT_MAX;
   return 0;
 }
-- 
1.7.1

Re: [PATCH] Zero supplied stat buffers in functions that pretend to stat

[Pedro Alves]

On 04/13/2015 10:40 PM, Gary Benson wrote:
> Hi all,
> 
> GDB has five places where it pretends to stat for bfd_openr_iovec.
> Four of these only set the incoming buffer's st_size, leaving the
> other fields unchanged, which is to say very likely populated with
> random values from the stack.  remote_bfd_iovec_stat was fixed in
> 0a93529c56714b1da3d7106d3e0300764f8bb81c; this commit fixes the
> other four.
> 
> Built and and regtested on RHEL6.6 x86_64.
> 
> Ok to commit?

Eh, how apropos for the bfd cache discussion.

OK, thanks.

-- 
Pedro Alves

Re: [PATCH] Zero supplied stat buffers in functions that pretend to stat

[Gary Benson]

Pedro Alves wrote:
> On 04/13/2015 10:40 PM, Gary Benson wrote:
> > GDB has five places where it pretends to stat for bfd_openr_iovec.
> > Four of these only set the incoming buffer's st_size, leaving the
> > other fields unchanged, which is to say very likely populated with
> > random values from the stack.  remote_bfd_iovec_stat was fixed in
> > 0a93529c56714b1da3d7106d3e0300764f8bb81c; this commit fixes the
> > other four.
> > 
> > Built and and regtested on RHEL6.6 x86_64.
> > 
> > Ok to commit?
> 
> Eh, how apropos for the bfd cache discussion.

Yeah, I've been meaning to reply to that :)

> OK, thanks.

Will push it later.

Cheers,
Gary

-- 
http://gbenson.net/

Re: [PATCH] Zero supplied stat buffers in functions that pretend to stat

[Gary Benson]

Gary Benson wrote:
> Pedro Alves wrote:
> > On 04/13/2015 10:40 PM, Gary Benson wrote:
> > > GDB has five places where it pretends to stat for bfd_openr_iovec.
> > > Four of these only set the incoming buffer's st_size, leaving the
> > > other fields unchanged, which is to say very likely populated with
> > > random values from the stack.  remote_bfd_iovec_stat was fixed in
> > > 0a93529c56714b1da3d7106d3e0300764f8bb81c; this commit fixes the
> > > other four.
> > > 
> > > Built and and regtested on RHEL6.6 x86_64.
> > > 
> > > Ok to commit?
> > 
> > Eh, how apropos for the bfd cache discussion.
> 
> Yeah, I've been meaning to reply to that :)
> 
> > OK, thanks.
> 
> Will push it later.

Pushed.

Cheers,
Gary

-- 
http://gbenson.net/