[PATCH] ppc64/kdump: Fix ELF header endianess

[PATCH] ppc64/kdump: Fix ELF header endianess

[Laurent Dufour]

The ELF header created among the loading of the kdump kernel should be
flagged using the current endianess and not always as big endian.

Without this patch the data exposed in /proc/vmcore are not readable when
running in LE mode.

Signed-off-by: Laurent Dufour <ldufour@linux.vnet.ibm.com>
---
 kexec/arch/ppc64/crashdump-ppc64.c |    4 ++++
 1 file changed, 4 insertions(+)

diff --git a/kexec/arch/ppc64/crashdump-ppc64.c b/kexec/arch/ppc64/crashdump-ppc64.c
index 00a0e632f48e..6214b831fdf2 100644
--- a/kexec/arch/ppc64/crashdump-ppc64.c
+++ b/kexec/arch/ppc64/crashdump-ppc64.c
@@ -38,7 +38,11 @@
 static struct crash_elf_info elf_info64 =
 {
 	class: ELFCLASS64,
+#if BYTE_ORDER == LITTLE_ENDIAN
+	data: ELFDATA2LSB,
+#else
 	data: ELFDATA2MSB,
+#endif
 	machine: EM_PPC64,
 	page_offset: PAGE_OFFSET,
 	lowmem_limit: MAXMEM,


_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec

Re: [PATCH] ppc64/kdump: Fix ELF header endianess

[Petr Tesarik]

On Fri, 25 Jul 2014 17:07:49 +0200
Laurent Dufour <ldufour@linux.vnet.ibm.com> wrote:

> The ELF header created among the loading of the kdump kernel should be
> flagged using the current endianess and not always as big endian.
> 
> Without this patch the data exposed in /proc/vmcore are not readable when
> running in LE mode.

FWIW I can confirm that this patch has fixed kdump on my ppc64le system.

Thank you, Laurent!

Petr Tesarik

> Signed-off-by: Laurent Dufour <ldufour@linux.vnet.ibm.com>
> ---
>  kexec/arch/ppc64/crashdump-ppc64.c |    4 ++++
>  1 file changed, 4 insertions(+)
> 
> diff --git a/kexec/arch/ppc64/crashdump-ppc64.c b/kexec/arch/ppc64/crashdump-ppc64.c
> index 00a0e632f48e..6214b831fdf2 100644
> --- a/kexec/arch/ppc64/crashdump-ppc64.c
> +++ b/kexec/arch/ppc64/crashdump-ppc64.c
> @@ -38,7 +38,11 @@
>  static struct crash_elf_info elf_info64 =
>  {
>  	class: ELFCLASS64,
> +#if BYTE_ORDER == LITTLE_ENDIAN
> +	data: ELFDATA2LSB,
> +#else
>  	data: ELFDATA2MSB,
> +#endif
>  	machine: EM_PPC64,
>  	page_offset: PAGE_OFFSET,
>  	lowmem_limit: MAXMEM,
> 
> 
> _______________________________________________
> kexec mailing list
> kexec@lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/kexec


_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec

Re: [PATCH] ppc64/kdump: Fix ELF header endianess

[Laurent Dufour]

On 25/07/2014 17:07, Laurent Dufour wrote:
> The ELF header created among the loading of the kdump kernel should be
> flagged using the current endianess and not always as big endian.
> 
> Without this patch the data exposed in /proc/vmcore are not readable when
> running in LE mode.

Hi Simon,

Do you have any concerns with this patch ?
If not, would you please consider it for merge since it fixes a major
issue with dump analysis.

Thanks
Laurent.

> Signed-off-by: Laurent Dufour <ldufour@linux.vnet.ibm.com>
> ---
>  kexec/arch/ppc64/crashdump-ppc64.c |    4 ++++
>  1 file changed, 4 insertions(+)
> 
> diff --git a/kexec/arch/ppc64/crashdump-ppc64.c b/kexec/arch/ppc64/crashdump-ppc64.c
> index 00a0e632f48e..6214b831fdf2 100644
> --- a/kexec/arch/ppc64/crashdump-ppc64.c
> +++ b/kexec/arch/ppc64/crashdump-ppc64.c
> @@ -38,7 +38,11 @@
>  static struct crash_elf_info elf_info64 =
>  {
>  	class: ELFCLASS64,
> +#if BYTE_ORDER == LITTLE_ENDIAN
> +	data: ELFDATA2LSB,
> +#else
>  	data: ELFDATA2MSB,
> +#endif
>  	machine: EM_PPC64,
>  	page_offset: PAGE_OFFSET,
>  	lowmem_limit: MAXMEM,
> 
> 
> _______________________________________________
> kexec mailing list
> kexec@lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/kexec
> 


_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec

Re: [PATCH] ppc64/kdump: Fix ELF header endianess

[Simon Horman]

On Thu, Aug 28, 2014 at 03:03:11PM +0200, Laurent Dufour wrote:
> On 25/07/2014 17:07, Laurent Dufour wrote:
> > The ELF header created among the loading of the kdump kernel should be
> > flagged using the current endianess and not always as big endian.
> > 
> > Without this patch the data exposed in /proc/vmcore are not readable when
> > running in LE mode.
> 
> Hi Simon,
> 
> Do you have any concerns with this patch ?
> If not, would you please consider it for merge since it fixes a major
> issue with dump analysis.

Thanks, I have applied this patch.

_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec