On Wed, 16 Mar 2005 17:10:57 +1100
Paul Mackerras <[EMAIL PROTECTED]> wrote:
> Jake Moilanen writes:
>
> > It does not work w/o the sys_mprotect. It will hang in one of the first
> > few binaries.
>
> Hmmm, what distro is this with? I just tried a kernel with the patch
> below on a SLES9
On Wed, 16 Mar 2005 17:10:57 +1100
Paul Mackerras [EMAIL PROTECTED] wrote:
Jake Moilanen writes:
It does not work w/o the sys_mprotect. It will hang in one of the first
few binaries.
Hmmm, what distro is this with? I just tried a kernel with the patch
below on a SLES9 install and a
Jake Moilanen writes:
> It does not work w/o the sys_mprotect. It will hang in one of the first
> few binaries.
Hmmm, what distro is this with? I just tried a kernel with the patch
below on a SLES9 install and a Debian install and it came up and ran
just fine in both cases.
Paul.
diff -urN
On Wed, 16 Mar 2005 09:18:36 +1030
Alan Modra <[EMAIL PROTECTED]> wrote:
> On Tue, Mar 15, 2005 at 03:51:35PM -0600, Jake Moilanen wrote:
> > I believe the problem is that the last PT_LOAD entry does not have the
> > correct size, and we only mmap up to the sbss. The .sbss, .plt, and
> > .bss do
On Tue, Mar 15, 2005 at 03:51:35PM -0600, Jake Moilanen wrote:
> I believe the problem is that the last PT_LOAD entry does not have the
> correct size, and we only mmap up to the sbss. The .sbss, .plt, and
> .bss do not get mmapped with the section.
Huh? .sbss, .plt and .bss have no file
On Tue, 15 Mar 2005 09:18:04 +1100
Paul Mackerras <[EMAIL PROTECTED]> wrote:
> Jake Moilanen writes:
>
> > > I don't think I can push that upstream. What happens if you leave
> > > that out?
> >
> > The bss and the plt are in the same segment, and plt obviously needs to
> > be executable.
>
>
On Tue, 15 Mar 2005 09:18:04 +1100
Paul Mackerras [EMAIL PROTECTED] wrote:
Jake Moilanen writes:
I don't think I can push that upstream. What happens if you leave
that out?
The bss and the plt are in the same segment, and plt obviously needs to
be executable.
Yes... what I was
On Tue, Mar 15, 2005 at 03:51:35PM -0600, Jake Moilanen wrote:
I believe the problem is that the last PT_LOAD entry does not have the
correct size, and we only mmap up to the sbss. The .sbss, .plt, and
.bss do not get mmapped with the section.
Huh? .sbss, .plt and .bss have no file contents,
On Wed, 16 Mar 2005 09:18:36 +1030
Alan Modra [EMAIL PROTECTED] wrote:
On Tue, Mar 15, 2005 at 03:51:35PM -0600, Jake Moilanen wrote:
I believe the problem is that the last PT_LOAD entry does not have the
correct size, and we only mmap up to the sbss. The .sbss, .plt, and
.bss do not get
Jake Moilanen writes:
It does not work w/o the sys_mprotect. It will hang in one of the first
few binaries.
Hmmm, what distro is this with? I just tried a kernel with the patch
below on a SLES9 install and a Debian install and it came up and ran
just fine in both cases.
Paul.
diff -urN
Jake Moilanen writes:
> > I don't think I can push that upstream. What happens if you leave
> > that out?
>
> The bss and the plt are in the same segment, and plt obviously needs to
> be executable.
Yes... what I was asking was "do things actually break if you leave
that out, or does the
On Mon, 14 Mar 2005 21:13:36 +1100
Paul Mackerras <[EMAIL PROTECTED]> wrote:
> Jake Moilanen writes:
>
> > diff -puN fs/binfmt_elf.c~nx-user-ppc64 fs/binfmt_elf.c
> > --- linux-2.6-bk/fs/binfmt_elf.c~nx-user-ppc64 2005-03-08 16:08:54
> > -06:00
> > +++ linux-2.6-bk-moilanen/fs/binfmt_elf.c
Jake Moilanen writes:
> diff -puN fs/binfmt_elf.c~nx-user-ppc64 fs/binfmt_elf.c
> --- linux-2.6-bk/fs/binfmt_elf.c~nx-user-ppc642005-03-08 16:08:54
> -06:00
> +++ linux-2.6-bk-moilanen/fs/binfmt_elf.c 2005-03-08 16:08:54 -06:00
> @@ -99,6 +99,8 @@ static int set_brk(unsigned long
Jake Moilanen writes:
diff -puN fs/binfmt_elf.c~nx-user-ppc64 fs/binfmt_elf.c
--- linux-2.6-bk/fs/binfmt_elf.c~nx-user-ppc642005-03-08 16:08:54
-06:00
+++ linux-2.6-bk-moilanen/fs/binfmt_elf.c 2005-03-08 16:08:54 -06:00
@@ -99,6 +99,8 @@ static int set_brk(unsigned long start,
On Mon, 14 Mar 2005 21:13:36 +1100
Paul Mackerras [EMAIL PROTECTED] wrote:
Jake Moilanen writes:
diff -puN fs/binfmt_elf.c~nx-user-ppc64 fs/binfmt_elf.c
--- linux-2.6-bk/fs/binfmt_elf.c~nx-user-ppc64 2005-03-08 16:08:54
-06:00
+++ linux-2.6-bk-moilanen/fs/binfmt_elf.c
Jake Moilanen writes:
I don't think I can push that upstream. What happens if you leave
that out?
The bss and the plt are in the same segment, and plt obviously needs to
be executable.
Yes... what I was asking was do things actually break if you leave
that out, or does the binfmt_elf
On Wed, 9 Mar 2005 21:22:13 -0600
[EMAIL PROTECTED] (Olof Johansson) wrote:
> On Tue, Mar 08, 2005 at 05:08:26PM -0600, Jake Moilanen wrote:
> > No-exec base and user space support for PPC64.
>
> Hi, a couple of comments below.
>
Here's the revised user & base support for no-exec on ppc64
On Wed, 9 Mar 2005 21:22:13 -0600
[EMAIL PROTECTED] (Olof Johansson) wrote:
On Tue, Mar 08, 2005 at 05:08:26PM -0600, Jake Moilanen wrote:
No-exec base and user space support for PPC64.
Hi, a couple of comments below.
Here's the revised user base support for no-exec on ppc64 with Olof
On Tue, Mar 08, 2005 at 05:08:26PM -0600, Jake Moilanen wrote:
> No-exec base and user space support for PPC64.
Hi, a couple of comments below.
-Olof
> @@ -786,6 +786,7 @@ int hash_huge_page(struct mm_struct *mm,
> pte_t old_pte, new_pte;
> unsigned long hpteflags, prpn;
>
On Tue, Mar 08, 2005 at 05:08:26PM -0600, Jake Moilanen wrote:
No-exec base and user space support for PPC64.
Hi, a couple of comments below.
-Olof
@@ -786,6 +786,7 @@ int hash_huge_page(struct mm_struct *mm,
pte_t old_pte, new_pte;
unsigned long hpteflags, prpn;
long
No-exec base and user space support for PPC64.
This will prohibit user space apps that a compile w/ PT_GNU_STACK from
executing in segments that are non-executable. Non-PT_GNU_STACK
compiled apps will work as well, but will not be able to take advantage
of the no-exec feature.
Signed-off-by:
No-exec base and user space support for PPC64.
This will prohibit user space apps that a compile w/ PT_GNU_STACK from
executing in segments that are non-executable. Non-PT_GNU_STACK
compiled apps will work as well, but will not be able to take advantage
of the no-exec feature.
Signed-off-by:
22 matches
Mail list logo