[PATCH] kbuild: modversions: Fix relative CRC byte order interpretion

[PATCH] kbuild: modversions: Fix relative CRC byte order interpretion

[Fredrik Noring]

Fix commit 56067812d5b0 ("kbuild: modversions: add infrastructure for
emitting relative CRCs") where CRCs are interpreted in host byte order
rather than proper kernel byte order.

For example, when loading BE modules compiled with a LE system, the
error "disagrees about version of symbol module_layout" is produced.
A message such as "Found checksum D7FA6856 vs module 5668FAD7" will be
given with debug enabled, which indicates an obvious endian problem.

The general solution is to use the macro TO_NATIVE, as is done in
similar cases throughout modpost.c. With this correction it has been
verified that BE modules compiled with a LE system load properly into
a BE kernel.

Signed-off-by: Fredrik Noring <noring@nocrew.org>
Fixes: 56067812d5b0 ("kbuild: modversions: add infrastructure for emitting relative CRCs")
Cc: stable@vger.kernel.org
---
diff --git a/scripts/mod/modpost.c b/scripts/mod/modpost.c
index 0b0d1080b1c5..f277e116e0eb 100644
--- a/scripts/mod/modpost.c
+++ b/scripts/mod/modpost.c
@@ -639,7 +639,7 @@ static void handle_modversions(struct module *mod, struct elf_info *info,
 			       info->sechdrs[sym->st_shndx].sh_offset -
 			       (info->hdr->e_type != ET_REL ?
 				info->sechdrs[sym->st_shndx].sh_addr : 0);
-			crc = *crcp;
+			crc = TO_NATIVE(*crcp);
 		}
 		sym_update_crc(symname + strlen("__crc_"), mod, crc,
 				export);

Re: [PATCH] kbuild: modversions: Fix relative CRC byte order interpretion

[Fredrik Noring]

Masahiro, Michal --

> Fix commit 56067812d5b0 ("kbuild: modversions: add infrastructure for
> emitting relative CRCs") where CRCs are interpreted in host byte order
> rather than proper kernel byte order.
> 
> For example, when loading BE modules compiled with a LE system, the
> error "disagrees about version of symbol module_layout" is produced.
> A message such as "Found checksum D7FA6856 vs module 5668FAD7" will be
> given with debug enabled, which indicates an obvious endian problem.

I realise that the description above is perhaps a bit too brief; although
the error is triggered by loading modules, the precise location of the
error is within __kcrctab within the kernel image.

> The general solution is to use the macro TO_NATIVE, as is done in
> similar cases throughout modpost.c. With this correction it has been
> verified that BE modules compiled with a LE system load properly into
> a BE kernel.

One could note that it has also been tested with a LE kernel compiled
with a LE system, in which case TO_NATIVE returns its value unmodified
since the byte orders match. This is by far the common case.

Would you like me to send an updated description?

Fredrik

> Signed-off-by: Fredrik Noring <noring@nocrew.org>
> Fixes: 56067812d5b0 ("kbuild: modversions: add infrastructure for emitting relative CRCs")
> Cc: stable@vger.kernel.org
> ---
> diff --git a/scripts/mod/modpost.c b/scripts/mod/modpost.c
> index 0b0d1080b1c5..f277e116e0eb 100644
> --- a/scripts/mod/modpost.c
> +++ b/scripts/mod/modpost.c
> @@ -639,7 +639,7 @@ static void handle_modversions(struct module *mod, struct elf_info *info,
>  			       info->sechdrs[sym->st_shndx].sh_offset -
>  			       (info->hdr->e_type != ET_REL ?
>  				info->sechdrs[sym->st_shndx].sh_addr : 0);
> -			crc = *crcp;
> +			crc = TO_NATIVE(*crcp);
>  		}
>  		sym_update_crc(symname + strlen("__crc_"), mod, crc,
>  				export);

Re: [PATCH] kbuild: modversions: Fix relative CRC byte order interpretion

[Masahiro Yamada]

Hi Fredrik,



On Sun, Mar 24, 2019 at 3:37 PM Fredrik Noring <noring@nocrew.org> wrote:
>
> Masahiro, Michal --
>
> > Fix commit 56067812d5b0 ("kbuild: modversions: add infrastructure for
> > emitting relative CRCs") where CRCs are interpreted in host byte order
> > rather than proper kernel byte order.
> >
> > For example, when loading BE modules compiled with a LE system, the
> > error "disagrees about version of symbol module_layout" is produced.
> > A message such as "Found checksum D7FA6856 vs module 5668FAD7" will be
> > given with debug enabled, which indicates an obvious endian problem.
>
> I realise that the description above is perhaps a bit too brief; although
> the error is triggered by loading modules, the precise location of the
> error is within __kcrctab within the kernel image.
>
> > The general solution is to use the macro TO_NATIVE, as is done in
> > similar cases throughout modpost.c. With this correction it has been
> > verified that BE modules compiled with a LE system load properly into
> > a BE kernel.
>
> One could note that it has also been tested with a LE kernel compiled
> with a LE system, in which case TO_NATIVE returns its value unmodified
> since the byte orders match. This is by far the common case.
>
> Would you like me to send an updated description?


Either will do.

If you send v2, could you please check the possible
typo in the subject?

interpretion -> interpretation ?



Masahiro



> Fredrik
>
> > Signed-off-by: Fredrik Noring <noring@nocrew.org>
> > Fixes: 56067812d5b0 ("kbuild: modversions: add infrastructure for emitting relative CRCs")
> > Cc: stable@vger.kernel.org
> > ---
> > diff --git a/scripts/mod/modpost.c b/scripts/mod/modpost.c
> > index 0b0d1080b1c5..f277e116e0eb 100644
> > --- a/scripts/mod/modpost.c
> > +++ b/scripts/mod/modpost.c
> > @@ -639,7 +639,7 @@ static void handle_modversions(struct module *mod, struct elf_info *info,
> >                              info->sechdrs[sym->st_shndx].sh_offset -
> >                              (info->hdr->e_type != ET_REL ?
> >                               info->sechdrs[sym->st_shndx].sh_addr : 0);
> > -                     crc = *crcp;
> > +                     crc = TO_NATIVE(*crcp);
> >               }
> >               sym_update_crc(symname + strlen("__crc_"), mod, crc,
> >                               export);



-- 
Best Regards
Masahiro Yamada

[PATCH v2] kbuild: modversions: Fix relative CRC byte order interpretation

[Fredrik Noring]

Fix commit 56067812d5b0 ("kbuild: modversions: add infrastructure for
emitting relative CRCs") where CRCs are interpreted in host byte order
rather than proper kernel byte order. The bug is conditional on
CONFIG_MODULE_REL_CRCS.

For example, when loading a BE module into a BE kernel compiled with a LE
system, the error "disagrees about version of symbol module_layout" is
produced. A message such as "Found checksum D7FA6856 vs module 5668FAD7"
will be given with debug enabled, which indicates an obvious endian
problem within __kcrctab within the kernel image.

The general solution is to use the macro TO_NATIVE, as is done in
similar cases throughout modpost.c. With this correction it has been
verified that a BE kernel compiled with a LE system accepts BE modules.

This change has also been verified with a LE kernel compiled with a LE
system, in which case TO_NATIVE returns its value unmodified since the
byte orders match. This is by far the common case.

Fixes: 56067812d5b0 ("kbuild: modversions: add infrastructure for emitting relative CRCs")
Signed-off-by: Fredrik Noring <noring@nocrew.org>
Cc: stable@vger.kernel.org
---
Thank you for your review, Masahiro,

Changes in v2:
- Fix subject spelling mistake.
- Note that the bug is conditional on CONFIG_MODULE_REL_CRCS.
- Clarify kernel image and __kcrctab compilation details.
- Mention verification of matching LE byte order common case.

Fredrik
---
diff --git a/scripts/mod/modpost.c b/scripts/mod/modpost.c
index 0b0d1080b1c5..f277e116e0eb 100644
--- a/scripts/mod/modpost.c
+++ b/scripts/mod/modpost.c
@@ -639,7 +639,7 @@ static void handle_modversions(struct module *mod, struct elf_info *info,
 			       info->sechdrs[sym->st_shndx].sh_offset -
 			       (info->hdr->e_type != ET_REL ?
 				info->sechdrs[sym->st_shndx].sh_addr : 0);
-			crc = *crcp;
+			crc = TO_NATIVE(*crcp);
 		}
 		sym_update_crc(symname + strlen("__crc_"), mod, crc,
 				export);

Re: [PATCH v2] kbuild: modversions: Fix relative CRC byte order interpretation

[Masahiro Yamada]

On Thu, Mar 28, 2019 at 3:13 AM Fredrik Noring <noring@nocrew.org> wrote:
>
> Fix commit 56067812d5b0 ("kbuild: modversions: add infrastructure for
> emitting relative CRCs") where CRCs are interpreted in host byte order
> rather than proper kernel byte order. The bug is conditional on
> CONFIG_MODULE_REL_CRCS.
>
> For example, when loading a BE module into a BE kernel compiled with a LE
> system, the error "disagrees about version of symbol module_layout" is
> produced. A message such as "Found checksum D7FA6856 vs module 5668FAD7"
> will be given with debug enabled, which indicates an obvious endian
> problem within __kcrctab within the kernel image.
>
> The general solution is to use the macro TO_NATIVE, as is done in
> similar cases throughout modpost.c. With this correction it has been
> verified that a BE kernel compiled with a LE system accepts BE modules.
>
> This change has also been verified with a LE kernel compiled with a LE
> system, in which case TO_NATIVE returns its value unmodified since the
> byte orders match. This is by far the common case.
>
> Fixes: 56067812d5b0 ("kbuild: modversions: add infrastructure for emitting relative CRCs")
> Signed-off-by: Fredrik Noring <noring@nocrew.org>
> Cc: stable@vger.kernel.org


Applied to linux-kbuild/fixes. Thanks.



> ---
> Thank you for your review, Masahiro,
>
> Changes in v2:
> - Fix subject spelling mistake.
> - Note that the bug is conditional on CONFIG_MODULE_REL_CRCS.
> - Clarify kernel image and __kcrctab compilation details.
> - Mention verification of matching LE byte order common case.
>
> Fredrik
> ---
> diff --git a/scripts/mod/modpost.c b/scripts/mod/modpost.c
> index 0b0d1080b1c5..f277e116e0eb 100644
> --- a/scripts/mod/modpost.c
> +++ b/scripts/mod/modpost.c
> @@ -639,7 +639,7 @@ static void handle_modversions(struct module *mod, struct elf_info *info,
>                                info->sechdrs[sym->st_shndx].sh_offset -
>                                (info->hdr->e_type != ET_REL ?
>                                 info->sechdrs[sym->st_shndx].sh_addr : 0);
> -                       crc = *crcp;
> +                       crc = TO_NATIVE(*crcp);
>                 }
>                 sym_update_crc(symname + strlen("__crc_"), mod, crc,
>                                 export);
>


-- 
Best Regards
Masahiro Yamada