[PATCH] Document M_PERTURB

[PATCH] Document M_PERTURB

[Andreas Schwab]

2010-07-26  Andreas Schwab  <schwab@redhat.com>

	* manual/memory.texi (Malloc Tunable Parameters): Document
	M_PERTURB.

diff --git a/manual/memory.texi b/manual/memory.texi
index 59ea1ee..6ae408d 100644
--- a/manual/memory.texi
+++ b/manual/memory.texi
@@ -702,6 +702,11 @@ be allocated via @code{mmap}.
 @item M_MMAP_MAX
 The maximum number of chunks to allocate with @code{mmap}.  Setting this
 to zero disables all use of @code{mmap}.
+@item M_PERTURB
+If non-zero, memory blocks are filled with a value depending on
+this parameter when they are allocated (except when allocated by
+@code{calloc}) and freed.  This can be used to debug the use of
+uninitialized or freed heap memory.
 @end table
 
 @end deftypefun
-- 
1.7.1.1


-- 
Andreas Schwab, schwab@redhat.com
GPG Key fingerprint = D4E8 DBE3 3813 BB5D FA84  5EC7 45C6 250E 6F00 984E
"And now for something completely different."

Re: [PATCH] Document M_PERTURB

[Roland McGrath]

I would s/a value/values/ since there is one value for alloc and one value
for free, both based on the perturb byte.  I think it should also say
"... depending on the low-order byte of this value," so nobody can think
that passing in more than 8 bits has any effect.  With those tweaks, it's
fine to commit it.

Thanks,
Roland

Re: [PATCH] Document M_PERTURB

[Jakub Jelinek]

On Mon, Jul 26, 2010 at 01:01:11PM -0700, Roland McGrath wrote:
> I would s/a value/values/ since there is one value for alloc and one value
> for free, both based on the perturb byte.  I think it should also say
> "... depending on the low-order byte of this value," so nobody can think
> that passing in more than 8 bits has any effect.

It actually has.  MALLOC_PERTURB_=0 doesn't do anything, while
MALLOC_PERTURB_=256 does.

	Jakub

Re: [PATCH] Document M_PERTURB

[Andreas Schwab]

Roland McGrath <roland@redhat.com> writes:

> I think it should also say "... depending on the low-order byte of
> this value," so nobody can think that passing in more than 8 bits has
> any effect.

Depends on how much of it is considered an implementation detail.

Andreas.

-- 
Andreas Schwab, schwab@redhat.com
GPG Key fingerprint = D4E8 DBE3 3813 BB5D FA84  5EC7 45C6 250E 6F00 984E
"And now for something completely different."

Re: [PATCH] Document M_PERTURB

[Roland McGrath]

> Roland McGrath <roland@redhat.com> writes:
> 
> > I think it should also say "... depending on the low-order byte of
> > this value," so nobody can think that passing in more than 8 bits has
> > any effect.
> 
> Depends on how much of it is considered an implementation detail.

True.  Perhaps just say "some low-order bits of this parameter".