[PATCH 4/4] elf_fdpic_core_dump(): fix build break in 2.6.33-rc1

[PATCH 4/4] elf_fdpic_core_dump(): fix build break in 2.6.33-rc1

[David Howells]

From: Daisuke HATAYAMA <d.hatayama@jp.fujitsu.com>

Commit f6151dfea21496d43dbaba32cfcd9c9f404769bc causes ELF-FDPIC to break
during building, so this patch fixes it together with a correction for the
printk format.

Signed-off-by: Daisuke HATAYAMA <d.hatayama@jp.fujitsu.com>
Reviewed-by: Masami Hiramatsu <mhiramat@redhat.com>
Acked-by: Paul Mundt <lethal@linux-sh.org>
Signed-off-by: David Howells <dhowells@redhat.com>
---

 fs/binfmt_elf_fdpic.c |    8 ++++----
 1 files changed, 4 insertions(+), 4 deletions(-)


diff --git a/fs/binfmt_elf_fdpic.c b/fs/binfmt_elf_fdpic.c
index 2917ab1..384a19f 100644
--- a/fs/binfmt_elf_fdpic.c
+++ b/fs/binfmt_elf_fdpic.c
@@ -1807,11 +1807,11 @@ static int elf_fdpic_core_dump(struct coredump_params *cprm)
 	ELF_CORE_WRITE_EXTRA_DATA;
 #endif
 
-	if (file->f_pos != offset) {
+	if (cprm->file->f_pos != offset) {
 		/* Sanity check */
-		printk(KERN_WARNING
-		       "elf_core_dump: file->f_pos (%lld) != offset (%lld)\n",
-		       file->f_pos, offset);
+		printk(KERN_WARNING "elf_fdpic_core_dump:"
+		       " cprm->file->f_pos (%lld) != offset (%lld)\n",
+		       cprm->file->f_pos, offset);
 	}
 
 end_coredump:

[PATCH 1/4] FDPIC: Respect PT_GNU_STACK exec protection markings when creating NOMMU stack

[David Howells]

From: Mike Frysinger <vapier@gentoo.org>

The current code will load the stack size and protection markings, but then
only use the markings in the MMU code path.  The NOMMU code path always passes
PROT_EXEC to the mmap() call.  While this doesn't matter to most people whilst
the code is running, it will cause a pointless icache flush when starting every
FDPIC application.  Typically this icache flush will be of a region on the
order of 128KB in size, or may be the entire icache, depending on the
facilities available on the CPU.

In the case where the arch default behaviour seems to be desired
(EXSTACK_DEFAULT), we probe VM_STACK_FLAGS for VM_EXEC to determine whether we
should be setting PROT_EXEC or not.

For arches that support an MPU (Memory Protection Unit - an MMU without the
virtual mapping capability), setting PROT_EXEC or not will make an important
difference.

It should be noted that this change also affects the executability of the brk
region, since ELF-FDPIC has that share with the stack.  However, this is
probably irrelevant as NOMMU programs aren't likely to use the brk region,
preferring instead allocation via mmap().

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: David Howells <dhowells@redhat.com>
---

 arch/blackfin/include/asm/page.h |    5 +++++
 arch/frv/include/asm/page.h      |    2 --
 fs/binfmt_elf_fdpic.c            |   13 +++++++++++--
 3 files changed, 16 insertions(+), 4 deletions(-)


diff --git a/arch/blackfin/include/asm/page.h b/arch/blackfin/include/asm/page.h
index 944a07c..1d04e40 100644
--- a/arch/blackfin/include/asm/page.h
+++ b/arch/blackfin/include/asm/page.h
@@ -10,4 +10,9 @@
 #include <asm-generic/page.h>
 #define MAP_NR(addr) (((unsigned long)(addr)-PAGE_OFFSET) >> PAGE_SHIFT)
 
+#define VM_DATA_DEFAULT_FLAGS \
+	(VM_READ | VM_WRITE | \
+	((current->personality & READ_IMPLIES_EXEC) ? VM_EXEC : 0 ) | \
+		 VM_MAYREAD | VM_MAYWRITE | VM_MAYEXEC)
+
 #endif
diff --git a/arch/frv/include/asm/page.h b/arch/frv/include/asm/page.h
index 25c6a50..8c97068 100644
--- a/arch/frv/include/asm/page.h
+++ b/arch/frv/include/asm/page.h
@@ -63,12 +63,10 @@ extern unsigned long max_pfn;
 #define virt_addr_valid(kaddr)	pfn_valid(__pa(kaddr) >> PAGE_SHIFT)
 
 
-#ifdef CONFIG_MMU
 #define VM_DATA_DEFAULT_FLAGS \
 	(VM_READ | VM_WRITE | \
 	((current->personality & READ_IMPLIES_EXEC) ? VM_EXEC : 0 ) | \
 		 VM_MAYREAD | VM_MAYWRITE | VM_MAYEXEC)
-#endif
 
 #endif /* __ASSEMBLY__ */
 
diff --git a/fs/binfmt_elf_fdpic.c b/fs/binfmt_elf_fdpic.c
index 7dc8599..c57d9ce 100644
--- a/fs/binfmt_elf_fdpic.c
+++ b/fs/binfmt_elf_fdpic.c
@@ -171,6 +171,9 @@ static int load_elf_fdpic_binary(struct linux_binprm *bprm,
 #ifdef ELF_FDPIC_PLAT_INIT
 	unsigned long dynaddr;
 #endif
+#ifndef CONFIG_MMU
+	unsigned long stack_prot;
+#endif
 	struct file *interpreter = NULL; /* to shut gcc up */
 	char *interpreter_name = NULL;
 	int executable_stack;
@@ -316,6 +319,8 @@ static int load_elf_fdpic_binary(struct linux_binprm *bprm,
 	 * defunct, deceased, etc. after this point we have to exit via
 	 * error_kill */
 	set_personality(PER_LINUX_FDPIC);
+	if (elf_read_implies_exec(&exec_params.hdr, executable_stack))
+		current->personality |= READ_IMPLIES_EXEC;
 	set_binfmt(&elf_fdpic_format);
 
 	current->mm->start_code = 0;
@@ -377,9 +382,13 @@ static int load_elf_fdpic_binary(struct linux_binprm *bprm,
 	if (stack_size < PAGE_SIZE * 2)
 		stack_size = PAGE_SIZE * 2;
 
+	stack_prot = PROT_READ | PROT_WRITE;
+	if (executable_stack == EXSTACK_ENABLE_X ||
+	    (executable_stack == EXSTACK_DEFAULT && VM_STACK_FLAGS & VM_EXEC))
+		stack_prot |= PROT_EXEC;
+
 	down_write(&current->mm->mmap_sem);
-	current->mm->start_brk = do_mmap(NULL, 0, stack_size,
-					 PROT_READ | PROT_WRITE | PROT_EXEC,
+	current->mm->start_brk = do_mmap(NULL, 0, stack_size, stack_prot,
 					 MAP_PRIVATE | MAP_ANONYMOUS |
 					 MAP_UNINITIALIZED | MAP_GROWSDOWN,
 					 0);

[PATCH 2/4] NOMMU: Avoiding duplicate icache flushes of shared maps

[David Howells]

From: Mike Frysinger <vapier.adi@gmail.com>

When working with FDPIC, there are many shared mappings of read-only code
regions between applications (the C library, applet packages like busybox,
etc.), but the current do_mmap_pgoff() function will issue an icache flush
whenever a VMA is added to an MM instead of only doing it when the map is
initially created.

The flush can instead be done when a region is first mmapped PROT_EXEC.  Note
that we may not rely on the first mapping of a region being executable - it's
possible for it to be PROT_READ only, so we have to remember whether we've
flushed the region or not, and then flush the entire region when a bit of it is
made executable.

However, this also affects the brk area.  That will no longer be executable.
We can mprotect() it to PROT_EXEC on MPU-mode kernels, but for NOMMU mode
kernels, when it increases the brk allocation, making sys_brk() flush the extra
from the icache should suffice.  The brk area probably isn't used by NOMMU
programs since the brk area can only use up the leavings from the stack
allocation, where the stack allocation is larger than requested.

Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
---

 include/linux/mm_types.h |    2 ++
 mm/nommu.c               |   11 ++++++++---
 2 files changed, 10 insertions(+), 3 deletions(-)


diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h
index 84a524a..84d020b 100644
--- a/include/linux/mm_types.h
+++ b/include/linux/mm_types.h
@@ -123,6 +123,8 @@ struct vm_region {
 	struct file	*vm_file;	/* the backing file or NULL */
 
 	atomic_t	vm_usage;	/* region usage count */
+	bool		vm_icache_flushed : 1; /* true if the icache has been flushed for
+						* this region */
 };
 
 /*
diff --git a/mm/nommu.c b/mm/nommu.c
index 6f9248f..a8d1752 100644
--- a/mm/nommu.c
+++ b/mm/nommu.c
@@ -432,6 +432,7 @@ SYSCALL_DEFINE1(brk, unsigned long, brk)
 	/*
 	 * Ok, looks good - let it rip.
 	 */
+	flush_icache_range(mm->brk, brk);
 	return mm->brk = brk;
 }
 
@@ -1353,10 +1354,14 @@ unsigned long do_mmap_pgoff(struct file *file,
 share:
 	add_vma_to_mm(current->mm, vma);
 
-	up_write(&nommu_region_sem);
+	/* we flush the region from the icache only when the first executable
+	 * mapping of it is made  */
+	if (vma->vm_flags & VM_EXEC && !region->vm_icache_flushed) {
+		flush_icache_range(region->vm_start, region->vm_end);
+		region->vm_icache_flushed = true;
+	}
 
-	if (prot & PROT_EXEC)
-		flush_icache_range(result, result + len);
+	up_write(&nommu_region_sem);
 
 	kleave(" = %lx", result);
 	return result;

[PATCH 3/4] NOMMU: Use copy_*_user_page() in access_process_vm()

[David Howells]

From: Jie Zhang <jie.zhang@analog.com>

The MMU code uses the copy_*_user_page() variants in access_process_vm()
rather than copy_*_user() as the former includes an icache flush.  This is
important when doing things like setting software breakpoints with gdb.
So switch the NOMMU code over to do the same.

This patch makes the reasonable assumption that copy_from_user_page() won't
fail - which is probably fine, as we've checked the VMA from which we're
copying is usable, and the copy is not allowed to cross VMAs.  The one case
where it might go wrong is if the VMA is a device rather than RAM, and that
device returns an error which - in which case rubbish will be returned rather
than EIO.

Signed-off-by: Jie Zhang <jie.zhang@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: David Howells <dhowells@redhat.com>
Acked-by: David McCullough <david_mccullough@mcafee.com>
Acked-by: Paul Mundt <lethal@linux-sh.org>
Acked-by: Greg Ungerer <gerg@uclinux.org>
---

 mm/nommu.c |    6 ++++--
 1 files changed, 4 insertions(+), 2 deletions(-)


diff --git a/mm/nommu.c b/mm/nommu.c
index a8d1752..1777386 100644
--- a/mm/nommu.c
+++ b/mm/nommu.c
@@ -1921,9 +1921,11 @@ int access_process_vm(struct task_struct *tsk, unsigned long addr, void *buf, in
 
 		/* only read or write mappings where it is permitted */
 		if (write && vma->vm_flags & VM_MAYWRITE)
-			len -= copy_to_user((void *) addr, buf, len);
+			copy_to_user_page(vma, NULL, addr,
+					 (void *) addr, buf, len);
 		else if (!write && vma->vm_flags & VM_MAYREAD)
-			len -= copy_from_user(buf, (void *) addr, len);
+			copy_from_user_page(vma, NULL, addr,
+					    buf, (void *) addr, len);
 		else
 			len = 0;
 	} else {

[PATCH 4/4] elf_fdpic_core_dump(): fix build break in 2.6.33-rc1

[David Howells]

From: Daisuke HATAYAMA <d.hatayama@jp.fujitsu.com>

Commit f6151dfea21496d43dbaba32cfcd9c9f404769bc causes ELF-FDPIC to break
during building, so this patch fixes it together with a correction for the
printk format.

Signed-off-by: Daisuke HATAYAMA <d.hatayama@jp.fujitsu.com>
Reviewed-by: Masami Hiramatsu <mhiramat@redhat.com>
Acked-by: Paul Mundt <lethal@linux-sh.org>
Signed-off-by: David Howells <dhowells@redhat.com>
---

 fs/binfmt_elf_fdpic.c |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)


diff --git a/fs/binfmt_elf_fdpic.c b/fs/binfmt_elf_fdpic.c
index c57d9ce..384a19f 100644
--- a/fs/binfmt_elf_fdpic.c
+++ b/fs/binfmt_elf_fdpic.c
@@ -1809,8 +1809,8 @@ static int elf_fdpic_core_dump(struct coredump_params *cprm)
 
 	if (cprm->file->f_pos != offset) {
 		/* Sanity check */
-		printk(KERN_WARNING
-		       "elf_core_dump: file->f_pos (%lld) != offset (%lld)\n",
+		printk(KERN_WARNING "elf_fdpic_core_dump:"
+		       " cprm->file->f_pos (%lld) != offset (%lld)\n",
 		       cprm->file->f_pos, offset);
 	}
 

Re: [PATCH 4/4] elf_fdpic_core_dump(): fix build break in 2.6.33-rc1

[Paul Mundt]

On Wed, Jan 06, 2010 at 05:23:33PM +0000, David Howells wrote:
> From: Daisuke HATAYAMA <d.hatayama@jp.fujitsu.com>
> 
> Commit f6151dfea21496d43dbaba32cfcd9c9f404769bc causes ELF-FDPIC to break
> during building, so this patch fixes it together with a correction for the
> printk format.
> 
> Signed-off-by: Daisuke HATAYAMA <d.hatayama@jp.fujitsu.com>
> Reviewed-by: Masami Hiramatsu <mhiramat@redhat.com>
> Acked-by: Paul Mundt <lethal@linux-sh.org>
> Signed-off-by: David Howells <dhowells@redhat.com>

I already pushed this through my tree, so you can just ignore this one.

Re: [PATCH 4/4] elf_fdpic_core_dump(): fix build break in 2.6.33-rc1

[Linus Torvalds]

On Wed, 6 Jan 2010, David Howells wrote:
>
> From: Daisuke HATAYAMA <d.hatayama@jp.fujitsu.com>
> 
> Commit f6151dfea21496d43dbaba32cfcd9c9f404769bc causes ELF-FDPIC to break
> during building, so this patch fixes it together with a correction for the
> printk format.

The comment doesn't match the patch. There's no build breakage fix (the 
actual fix was already merged, and your patch reflects that), there's only 
a printk format "fix" that is actually a de-provement because it splits 
the string across multiple lines.

So a strong NAK to this patch.

		Linus

Re: [PATCH 4/4] elf_fdpic_core_dump(): fix build break in 2.6.33-rc1

[David Howells]

Linus Torvalds <torvalds@linux-foundation.org> wrote:

> So a strong NAK to this patch.

Okay, forget that one then.  Can you take the others?

David