Mike Rapoport <[email protected]> writes:

> Hi Christophe,
>
> On Fri, Nov 13, 2026 at 03:36:01PM +0100, Christophe Leroy (CS GROUP) wrote:
>> Hi Mike,
>> 
>> Looks like I missed this change ...
>> 

I didn't initially get what Christophe meant here. Then I came across this
reply of from him[1]. So looks like somehow between v3 to v8, something got
changed? 

    On powerpc book3s/32, vmalloc space is not executable. Only modules 
    space is executable. X/NX cannot be set on a per page basis, it can only 
    be set on a 256 Mbytes segment basis.

    See commit c49643319715 ("powerpc/32s: Only leave NX unset on segments 
    used for modules") and 6ca055322da8 ("powerpc/32s: Use dedicated segment 
    for modules with STRICT_KERNEL_RWX") and 7bee31ad8e2f ("powerpc/32s: Fix 
    is_module_segment() when MODULES_VADDR is defined").

    So if your intention is still to have an executable kprobes, then you 
    can't use vmalloc address space.

[1]: 
https://lore.kernel.org/all/[email protected]/

>> Le 05/05/2024 à 18:06, Mike Rapoport a écrit :
>> > From: "Mike Rapoport (IBM)" <[email protected]>
>> > 
>> > powerpc overrides kprobes::alloc_insn_page() to remove writable
>> > permissions when STRICT_MODULE_RWX is on.
>> > 
>> > Add definition of EXECMEM_KRPOBES to execmem_params to allow using the
>> > generic kprobes::alloc_insn_page() with the desired permissions.
>> > 
>> > As powerpc uses breakpoint instructions to inject kprobes, it does not
>> > need to constrain kprobe allocations to the modules area and can use the
>> > entire vmalloc address space.
>> 
>> Using breakpoint means the segment must still be executable.
>> On book3s/32 VMALLOC segment is not executable, while module segment is.
>> 
>> See commit 6ca055322da8 ("powerpc/32s: Use dedicated segment for modules
>> with STRICT_KERNEL_RWX")
>> 
>> See report and fix at 
>> https://lore.kernel.org/all/ed03d290d440d579a81ae0b7de5243721291d903.1794543909.git.ritesh.l...@gmail.com/
>
> I'll take a look, thanks!
>  

Thanks Mike for taking a look. Sorry, I should have ccd'd you there, I
guess I missed it. However, the changes are these one here [2][3]

[2]: 
https://lore.kernel.org/linuxppc-dev/82c4973addebd82923c460cb44221f5b76146afe.1794572365.git.ritesh.l...@gmail.com/
[3]: 
https://lore.kernel.org/linuxppc-dev/f75f2cf97f44375fd2c8820e0caeab703cdb371c.1794572365.git.ritesh.l...@gmail.com/

-ritesh


Reply via email to