1. Patchew
  2. linux
  3. [PATCH v11 00/65] Fix DRM_USE_DYNAMIC_DEBUG=y
  4. View patch

[PATCH v11 01/65] dyndbg: fix NULL ptr on i386 due to section alignment

Jim Cromie posted 65 patches 3 weeks, 3 days ago
[PATCH v11 01/65] dyndbg: fix NULL ptr on i386 due to section alignment
Posted by Jim Cromie 3 weeks, 3 days ago 
When dyndbg classmaps get used (later in this series), the
__dyndbg_classes section (which has 28 byte structs on i386), causes
mis-alignment of the following __dyndbg section, resulting in a NULL
pointer deref in dynamic_debug_init().

To fix this, employ belt + suspenders:

1. move __dyndbg section above __dyndbg_classes.  This restores it to
its original position directly after the ALIGN(8), and fixes the
immediate problem.

2. add ALIGN(8) to the BOUNDED_SECTION* macros.  This aligns all
existing sections using the macro, and also fixes a future
dyndbg_class_* addition which would suffer the same misalignment on
i386.  Many of the existing macro uses already have a preceding ALIGN,
these are now redundant, but are harmless, so are left to avoid churn.

3. remove BOUNDED_SECTION* uses in ORC_UNWINDER sections.  These have
smaller alignments, *and* scripts/sorttable.c does not tolerate the
added ALIGN(8) padding.

Reported-by: kernel test robot <oliver.sang@intel.com>
Closes: https://lore.kernel.org/oe-lkp/202601211325.7e1f336-lkp@intel.com
Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
---
 include/asm-generic/vmlinux.lds.h | 16 ++++++++++++----
 1 file changed, 12 insertions(+), 4 deletions(-)

diff --git a/include/asm-generic/vmlinux.lds.h b/include/asm-generic/vmlinux.lds.h
index eeb070f330bd..a2ba7e3d9994 100644
--- a/include/asm-generic/vmlinux.lds.h
+++ b/include/asm-generic/vmlinux.lds.h
@@ -212,11 +212,13 @@
 #endif
 
 #define BOUNDED_SECTION_PRE_LABEL(_sec_, _label_, _BEGIN_, _END_)	\
+	. = ALIGN(8);							\
 	_BEGIN_##_label_ = .;						\
 	KEEP(*(_sec_))							\
 	_END_##_label_ = .;
 
 #define BOUNDED_SECTION_POST_LABEL(_sec_, _label_, _BEGIN_, _END_)	\
+	. = ALIGN(8);							\
 	_label_##_BEGIN_ = .;						\
 	KEEP(*(_sec_))							\
 	_label_##_END_ = .;
@@ -383,8 +385,8 @@
 	*(__tracepoints)						\
 	/* implement dynamic printk debug */				\
 	. = ALIGN(8);							\
-	BOUNDED_SECTION_BY(__dyndbg_classes, ___dyndbg_classes)		\
 	BOUNDED_SECTION_BY(__dyndbg, ___dyndbg)				\
+	BOUNDED_SECTION_BY(__dyndbg_classes, ___dyndbg_classes)		\
 	CODETAG_SECTIONS()						\
 	LIKELY_PROFILE()		       				\
 	BRANCH_PROFILE()						\
@@ -867,15 +869,21 @@
 #ifdef CONFIG_UNWINDER_ORC
 #define ORC_UNWIND_TABLE						\
 	.orc_header : AT(ADDR(.orc_header) - LOAD_OFFSET) {		\
-		BOUNDED_SECTION_BY(.orc_header, _orc_header)		\
+		__start_orc_header = .;					\
+		KEEP(*(.orc_header))					\
+		__stop_orc_header = .;					\
 	}								\
 	. = ALIGN(4);							\
 	.orc_unwind_ip : AT(ADDR(.orc_unwind_ip) - LOAD_OFFSET) {	\
-		BOUNDED_SECTION_BY(.orc_unwind_ip, _orc_unwind_ip)	\
+		__start_orc_unwind_ip = .;				\
+		KEEP(*(.orc_unwind_ip))					\
+		__stop_orc_unwind_ip = .;				\
 	}								\
 	. = ALIGN(2);							\
 	.orc_unwind : AT(ADDR(.orc_unwind) - LOAD_OFFSET) {		\
-		BOUNDED_SECTION_BY(.orc_unwind, _orc_unwind)		\
+		__start_orc_unwind = .;					\
+		KEEP(*(.orc_unwind))					\
+		__stop_orc_unwind = .;					\
 	}								\
 	text_size = _etext - _stext;					\
 	. = ALIGN(4);							\
-- 
2.53.0
Re: [PATCH v11 01/65] dyndbg: fix NULL ptr on i386 due to section alignment
Posted by Louis Chauvet 2 weeks, 3 days ago 
On Fri, 13 Mar 2026 07:19:26 -0600, Jim Cromie <jim.cromie@gmail.com> wrote:
> [...]
> smaller alignments, *and* scripts/sorttable.c does not tolerate the
> added ALIGN(8) padding.
> 
> Reported-by: kernel test robot <oliver.sang@intel.com>
> Closes: https://lore.kernel.org/oe-lkp/202601211325.7e1f336-lkp@intel.com
> Signed-off-by: Jim Cromie <jim.cromie@gmail.com>

comments may be wrong.

>
>
> diff --git a/include/asm-generic/vmlinux.lds.h b/include/asm-generic/vmlinux.lds.h
> index eeb070f330bd..a2ba7e3d9994 100644
> --- a/include/asm-generic/vmlinux.lds.h
> +++ b/include/asm-generic/vmlinux.lds.h
> @@ -212,11 +212,13 @@
> [ ... skip 7 lines ... ]
>  
>  #define BOUNDED_SECTION_POST_LABEL(_sec_, _label_, _BEGIN_, _END_)	\
> +	. = ALIGN(8);							\
>  	_label_##_BEGIN_ = .;						\
>  	KEEP(*(_sec_))							\
>  	_label_##_END_ = .;

This affects a lot of existing BOUNDED_SECTION_BY. I agree that it is
not a big issue (most of them already have ALIGN(8) or ALIGN(32), but
some have ALIGN(4) or just not aligned at all). I think this can increase
the size of the kernel in other places.

What do you think about a new macro or a
BOUNDED_SECTION_BY_ALIGNED(sec,label,align) with explicit aligement?

> @@ -867,15 +869,21 @@
> [ ... skip 15 lines ... ]
>  	. = ALIGN(2);							\
>  	.orc_unwind : AT(ADDR(.orc_unwind) - LOAD_OFFSET) {		\
> -		BOUNDED_SECTION_BY(.orc_unwind, _orc_unwind)		\
> +		__start_orc_unwind = .;					\
> +		KEEP(*(.orc_unwind))					\
> +		__stop_orc_unwind = .;					\

You already noticed an issue here for example, and you had to manually
expand the macro to "disable" the align. This is error-prone, I think it is
better to keep BOUNDED_SECTION_BY here.

Note: I don't understand well linker scripts and all the implications, my

Note: I don't understand well linker scripts and all the implications, my

-- 
Louis Chauvet <louis.chauvet@bootlin.com>
Re: [PATCH v11 01/65] dyndbg: fix NULL ptr on i386 due to section alignment
Posted by jim.cromie@gmail.com 2 weeks, 3 days ago 
On Fri, Mar 20, 2026 at 10:41 AM Louis Chauvet
<louis.chauvet@bootlin.com> wrote:
>
> On Fri, 13 Mar 2026 07:19:26 -0600, Jim Cromie <jim.cromie@gmail.com> wrote:
> > [...]
> > smaller alignments, *and* scripts/sorttable.c does not tolerate the
> > added ALIGN(8) padding.
> >

More specifically, it counts records in 2 sections, and insists
they're the same count,
and there's no slop/extra space.  the align(8) broke that last constraint.

> > Reported-by: kernel test robot <oliver.sang@intel.com>
> > Closes: https://lore.kernel.org/oe-lkp/202601211325.7e1f336-lkp@intel.com
> > Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
>
> comments may be wrong.


in v12, I found a way to use the makefile to conditionally include a "header"
file which invokes DYNAMIC_DEBUG_CLASSMAP_USE().

In doing this, I encountered some lost sections (lacking a KEEP in modules)
which I fixed by reusing the codetag.lds.h model, to make dyndbg.lds.h,
which reuses the (now separated out) BOUNDED_SECTION* macros,
and contains MOD_DYNDBG_SECTIONS(). which scripts/module.lds.S invokes.
So, those comments are now obsolete.


>
> >
> >
> > diff --git a/include/asm-generic/vmlinux.lds.h b/include/asm-generic/vmlinux.lds.h
> > index eeb070f330bd..a2ba7e3d9994 100644
> > --- a/include/asm-generic/vmlinux.lds.h
> > +++ b/include/asm-generic/vmlinux.lds.h
> > @@ -212,11 +212,13 @@
> > [ ... skip 7 lines ... ]
> >
> >  #define BOUNDED_SECTION_POST_LABEL(_sec_, _label_, _BEGIN_, _END_)   \
> > +     . = ALIGN(8);                                                   \
> >       _label_##_BEGIN_ = .;                                           \
> >       KEEP(*(_sec_))                                                  \
> >       _label_##_END_ = .;
>
> This affects a lot of existing BOUNDED_SECTION_BY. I agree that it is
> not a big issue (most of them already have ALIGN(8) or ALIGN(32), but
> some have ALIGN(4) or just not aligned at all). I think this can increase
> the size of the kernel in other places.
>
> What do you think about a new macro or a
> BOUNDED_SECTION_BY_ALIGNED(sec,label,align) with explicit aligement?

Thats not crazy, but Im not sure its justified by the 6 lines that the
_ALIGNED() variant would save.  It does add some modest complexity to
the macros.


>
> > @@ -867,15 +869,21 @@
> > [ ... skip 15 lines ... ]
> >       . = ALIGN(2);                                                   \
> >       .orc_unwind : AT(ADDR(.orc_unwind) - LOAD_OFFSET) {             \
> > -             BOUNDED_SECTION_BY(.orc_unwind, _orc_unwind)            \
> > +             __start_orc_unwind = .;                                 \
> > +             KEEP(*(.orc_unwind))                                    \
> > +             __stop_orc_unwind = .;                                  \
>
> You already noticed an issue here for example, and you had to manually
> expand the macro to "disable" the align. This is error-prone, I think it is
> better to keep BOUNDED_SECTION_BY here.
>

I'll look at it - there is a stack of 3-4 macros that would need adjusting.

> Note: I don't understand well linker scripts and all the implications, my
>
> Note: I don't understand well linker scripts and all the implications, my
>
> --
> Louis Chauvet <louis.chauvet@bootlin.com>

thanks Louis,
Jim

Patchew 2.1-devel-ea86937

© 2016 - 2026 Red Hat, Inc.