Re: [Qemu-devel] [PATCH] fix entry pointer for ELF kernels loaded with -kernel option

[Henning Schild]

Find a hopefully proper patch attached. Take it or leave it.

Signed-off-by: Henning Schild <address@hidden>
---
 hw/elf_ops.h |   11 +++++++++++
 1 files changed, 11 insertions(+), 0 deletions(-)

diff --git a/hw/elf_ops.h b/hw/elf_ops.h
index fa65ce2..731a983 100644
--- a/hw/elf_ops.h
+++ b/hw/elf_ops.h
@@ -269,6 +269,17 @@ static int glue(load_elf, SZ)(const char *name, 
int fd,
                 addr = ph->p_paddr;
             }

+            /* the entry pointer in the ELF header is a virtual
+             * address, if the text segments paddr and vaddr differ
+             * we need to adjust the entry */
+            if (pentry && !translate_fn &&
+                    ph->p_vaddr != ph->p_paddr &&
+                    ehdr.e_entry >= ph->p_vaddr &&
+                    ehdr.e_entry < ph->p_vaddr + ph->p_filesz &&
+                    ph->p_flags & PF_X) {
+                *pentry = ehdr.e_entry - ph->p_vaddr + ph->p_paddr;
+            }
+
             snprintf(label, sizeof(label), "phdr #%d: %s", i, name);
             rom_add_blob_fixed(label, data, mem_size, addr);

--
1.7.8.6


Am Mittwoch, den 05.09.2012, 12:53 +0200 schrieb Kevin Wolf 
<address@hidden>:
> Am 05.09.2012 11:05, schrieb Henning Schild:
> >  This patch fixes a bug in qemu which prevents Multiboot ELF kernels
> >  from being loaded with the -kernel option. Find a full description 
> > of
> >  the problem here https://bugs.launchpad.net/qemu/+bug/1044727 .
>
> The logic looks good to me, but there are a few points about the 
> patch
> itself (see http://wiki.qemu.org/Contribute/SubmitAPatch).
>
> First thing is that the patch needs a proper Signed-off-by line. This 
> is
> absolutely crucial. The other points could be fixed manually by a
> potential patient enough maintainer, but you are the only one who can
> provide the Signed-off-by. Without it, the patch won't be applied.
>
> >  ---
> >   hw/elf_ops.h |   10 ++++++++++
> >   1 files changed, 10 insertions(+), 0 deletions(-)
> >
> >  diff --git a/hw/elf_ops.h b/hw/elf_ops.h
> >  index fa65ce2..aeddd11 100644
> >  --- a/hw/elf_ops.h
> >  +++ b/hw/elf_ops.h
> >  @@ -269,6 +269,16 @@ static int glue(load_elf, SZ)(const char 
> > *name,
> >  int fd,
>
> The patch is corrupted by line wraps. Using git 
> format-patch/send-email
> avoids this kind of problems. Alternatively, attach the patch in
> addition so that an uncorrupted version can be used for applying it.
>
> >                   addr = ph->p_paddr;
> >               }
> >
> >  +            /* the entry pointer in the ELF header is a virtual
> >  +             * address, if the text segments paddr and vaddr 
> > differ
> >  +             * we need to adjust the entry */
> >  +            if (pentry && !translate_fn &&
> >  +                ph->p_vaddr != ph->p_paddr &&
> >  +                ehdr.e_entry >= ph->p_vaddr &&
> >  +                ehdr.e_entry < ph->p_vaddr + ph->p_filesz &&
> >  +                ph->p_flags & PF_X)
> >  +                    *pentry = ehdr.e_entry - ph->p_vaddr +
> >  ph->p_paddr;
>
> The coding style problem that was already mentioned. qemu puts braces
> even for single statements.
>
> Kevin

0001-fix-entry-pointer-for-ELF-kernels-loaded-with-kernel.patch
Description: Text Data