[net-next,v2] sections: global data can be in .bss

Commit Message

Antoine Tenart Nov. 22, 2021, 2:24 p.m. UTC

When checking an address is located in a global data section also check
for the .bss section as global variables initialized to 0 can be in
there (-fzero-initialized-in-bss).

This was found when looking at ensure_safe_net_sysctl which was failing
to detect non-init sysctl pointing to a global data section when the
data was in the .bss section.

Signed-off-by: Antoine Tenart <[e-mail]>
Acked-by: Steven Rostedt (VMware) <[e-mail]>
---

A few remarks:

- This still targets net-next but I added Arnd if he prefers to take it
  through the 'asm-generic' tree, now that is_kernel_core_data is in
  include/asm-generic/.

- I kept the Acked-by tag as the change is the same really, the
  difference is the core_kernel_data function was renamed to
  is_kernel_core_data and moved since then.

- @Jonathon: with your analysis and suggestion I think you should be
  listed as a co-developer. If that's fine please say so, and reply
  with both a Co-developed-by and a Signed-off-by tags.

Since v1:
  - Grouped the .data and .bss checks in the same function.

v1 was https://lore.kernel.org/all/[e-mail]/T/

Thanks!
Antoine

 include/asm-generic/sections.h | 14 ++++++++++----
 1 file changed, 10 insertions(+), 4 deletions(-)

Comments

Arnd Bergmann Nov. 22, 2021, 2:39 p.m. UTC | #1

On Mon, Nov 22, 2021 at 3:24 PM Antoine Tenart <[e-mail]> wrote:
>
> When checking an address is located in a global data section also check
> for the .bss section as global variables initialized to 0 can be in
> there (-fzero-initialized-in-bss).
>
> This was found when looking at ensure_safe_net_sysctl which was failing
> to detect non-init sysctl pointing to a global data section when the
> data was in the .bss section.
>
> Signed-off-by: Antoine Tenart <[e-mail]>
> Acked-by: Steven Rostedt (VMware) <[e-mail]>
> ---
>
> A few remarks:
>
> - This still targets net-next but I added Arnd if he prefers to take it
>   through the 'asm-generic' tree, now that is_kernel_core_data is in
>   include/asm-generic/.

I have nothing else for asm-generic at the moment, please take
this through net-next.

Acked-by: Arnd Bergmann <[e-mail]>

[e-mail] Nov. 22, 2021, 3 p.m. UTC | #2

Hello:

This patch was applied to netdev/net-next.git (master)
by David S. Miller <[e-mail]>:

On Mon, 22 Nov 2021 15:24:56 +0100 you wrote:
> When checking an address is located in a global data section also check
> for the .bss section as global variables initialized to 0 can be in
> there (-fzero-initialized-in-bss).
> 
> This was found when looking at ensure_safe_net_sysctl which was failing
> to detect non-init sysctl pointing to a global data section when the
> data was in the .bss section.
> 
> [...]

Here is the summary with links:
  - [net-next,v2] sections: global data can be in .bss
    https://git.kernel.org/netdev/net-next/c/cb902b332f95

You are awesome, thank you!

Jonathon Reinhart Nov. 22, 2021, 4:56 p.m. UTC | #3

On Mon, Nov 22, 2021 at 9:24 AM Antoine Tenart <[e-mail]> wrote:
>
> When checking an address is located in a global data section also check
> for the .bss section as global variables initialized to 0 can be in
> there (-fzero-initialized-in-bss).
>
> This was found when looking at ensure_safe_net_sysctl which was failing
> to detect non-init sysctl pointing to a global data section when the
> data was in the .bss section.
>
> Signed-off-by: Antoine Tenart <[e-mail]>
> Acked-by: Steven Rostedt (VMware) <[e-mail]>

Co-Developed-by: Jonathon Reinhart <[e-mail]>
Signed-off-by: Jonathon Reinhart <[e-mail]>

> ---
>
> A few remarks:
>
> - This still targets net-next but I added Arnd if he prefers to take it
>   through the 'asm-generic' tree, now that is_kernel_core_data is in
>   include/asm-generic/.
>
> - I kept the Acked-by tag as the change is the same really, the
>   difference is the core_kernel_data function was renamed to
>   is_kernel_core_data and moved since then.
>
> - @Jonathon: with your analysis and suggestion I think you should be
>   listed as a co-developer. If that's fine please say so, and reply
>   with both a Co-developed-by and a Signed-off-by tags.

Added, thanks. Although it appears I may have missed the boat.

Antoine Tenart Nov. 22, 2021, 4:59 p.m. UTC | #4

Quoting Jonathon Reinhart (2021-11-22 17:56:55)
> On Mon, Nov 22, 2021 at 9:24 AM Antoine Tenart <[e-mail]> wrote:
> >
> > - @Jonathon: with your analysis and suggestion I think you should be
> >   listed as a co-developer. If that's fine please say so, and reply
> >   with both a Co-developed-by and a Signed-off-by tags.
> 
> Added, thanks. Although it appears I may have missed the boat.

Yes, the patch was applied quickly. Anyway, thanks for the investigation!

Antoine

Patch

diff --git a/include/asm-generic/sections.h b/include/asm-generic/sections.h
index 1dfadb2e878d..76a0f16e56cf 100644
--- a/include/asm-generic/sections.h
+++ b/include/asm-generic/sections.h
@@ -130,18 +130,24 @@  static inline bool init_section_intersects(void *virt, size_t size)
 
 /**
  * is_kernel_core_data - checks if the pointer address is located in the
- *			 .data section
+ *			 .data or .bss section
  *
  * @addr: address to check
  *
- * Returns: true if the address is located in .data, false otherwise.
+ * Returns: true if the address is located in .data or .bss, false otherwise.
  * Note: On some archs it may return true for core RODATA, and false
  *       for others. But will always be true for core RW data.
  */
 static inline bool is_kernel_core_data(unsigned long addr)
 {
-	return addr >= (unsigned long)_sdata &&
-	       addr < (unsigned long)_edata;
+	if (addr >= (unsigned long)_sdata && addr < (unsigned long)_edata)
+		return true;
+
+	if (addr >= (unsigned long)__bss_start &&
+	    addr < (unsigned long)__bss_stop)
+		return true;
+
+	return false;
 }
 
 /**