When KCOV is enabled all functions get instrumented, unless
the __no_sanitize_coverage attribute is used. To prepare for
__no_sanitize_coverage being applied to __init functions, we have to
handle differences in how GCC's inline optimizations get resolved. For
s390 this requires forcing a couple functions to be inline with
__always_inline.
Signed-off-by: Kees Cook <kees@kernel.org>
---
Cc: Madhavan Srinivasan <maddy@linux.ibm.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Nicholas Piggin <npiggin@gmail.com>
Cc: Christophe Leroy <christophe.leroy@csgroup.eu>
Cc: Naveen N Rao <naveen@kernel.org>
Cc: "Ritesh Harjani (IBM)" <ritesh.list@gmail.com>
Cc: "Aneesh Kumar K.V" <aneesh.kumar@linux.ibm.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: <linuxppc-dev@lists.ozlabs.org>
---
arch/powerpc/mm/book3s64/hash_utils.c | 2 +-
arch/powerpc/mm/book3s64/radix_pgtable.c | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/arch/powerpc/mm/book3s64/hash_utils.c b/arch/powerpc/mm/book3s64/hash_utils.c
index 5158aefe4873..93f1e1eb5ea6 100644
--- a/arch/powerpc/mm/book3s64/hash_utils.c
+++ b/arch/powerpc/mm/book3s64/hash_utils.c
@@ -409,7 +409,7 @@ static DEFINE_RAW_SPINLOCK(linear_map_kf_hash_lock);
static phys_addr_t kfence_pool;
-static inline void hash_kfence_alloc_pool(void)
+static __always_inline void hash_kfence_alloc_pool(void)
{
if (!kfence_early_init_enabled())
goto err;
diff --git a/arch/powerpc/mm/book3s64/radix_pgtable.c b/arch/powerpc/mm/book3s64/radix_pgtable.c
index 9f764bc42b8c..3238e9ed46b5 100644
--- a/arch/powerpc/mm/book3s64/radix_pgtable.c
+++ b/arch/powerpc/mm/book3s64/radix_pgtable.c
@@ -363,7 +363,7 @@ static int __meminit create_physical_mapping(unsigned long start,
}
#ifdef CONFIG_KFENCE
-static inline phys_addr_t alloc_kfence_pool(void)
+static __always_inline phys_addr_t alloc_kfence_pool(void)
{
phys_addr_t kfence_pool;
--
2.34.1Kees Cook <kees@kernel.org> writes:
> When KCOV is enabled all functions get instrumented, unless
> the __no_sanitize_coverage attribute is used. To prepare for
> __no_sanitize_coverage being applied to __init functions, we have to
> handle differences in how GCC's inline optimizations get resolved. For
> s390 this requires forcing a couple functions to be inline with
> __always_inline.
>
> Signed-off-by: Kees Cook <kees@kernel.org>
> ---
> Cc: Madhavan Srinivasan <maddy@linux.ibm.com>
> Cc: Michael Ellerman <mpe@ellerman.id.au>
> Cc: Nicholas Piggin <npiggin@gmail.com>
> Cc: Christophe Leroy <christophe.leroy@csgroup.eu>
> Cc: Naveen N Rao <naveen@kernel.org>
> Cc: "Ritesh Harjani (IBM)" <ritesh.list@gmail.com>
> Cc: "Aneesh Kumar K.V" <aneesh.kumar@linux.ibm.com>
> Cc: Andrew Morton <akpm@linux-foundation.org>
> Cc: <linuxppc-dev@lists.ozlabs.org>
> ---
> arch/powerpc/mm/book3s64/hash_utils.c | 2 +-
> arch/powerpc/mm/book3s64/radix_pgtable.c | 2 +-
> 2 files changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/arch/powerpc/mm/book3s64/hash_utils.c b/arch/powerpc/mm/book3s64/hash_utils.c
> index 5158aefe4873..93f1e1eb5ea6 100644
> --- a/arch/powerpc/mm/book3s64/hash_utils.c
> +++ b/arch/powerpc/mm/book3s64/hash_utils.c
> @@ -409,7 +409,7 @@ static DEFINE_RAW_SPINLOCK(linear_map_kf_hash_lock);
>
> static phys_addr_t kfence_pool;
>
> -static inline void hash_kfence_alloc_pool(void)
> +static __always_inline void hash_kfence_alloc_pool(void)
> {
> if (!kfence_early_init_enabled())
> goto err;
> diff --git a/arch/powerpc/mm/book3s64/radix_pgtable.c b/arch/powerpc/mm/book3s64/radix_pgtable.c
> index 9f764bc42b8c..3238e9ed46b5 100644
> --- a/arch/powerpc/mm/book3s64/radix_pgtable.c
> +++ b/arch/powerpc/mm/book3s64/radix_pgtable.c
> @@ -363,7 +363,7 @@ static int __meminit create_physical_mapping(unsigned long start,
> }
>
> #ifdef CONFIG_KFENCE
> -static inline phys_addr_t alloc_kfence_pool(void)
> +static __always_inline phys_addr_t alloc_kfence_pool(void)
> {
> phys_addr_t kfence_pool;
>
I remember seeing a warning msg around .init.text section. Let me dig
that...
... Here it is: https://lore.kernel.org/oe-kbuild-all/202504190552.mnFGs5sj-lkp@intel.com/
I am not sure why it only complains for hash_debug_pagealloc_alloc_slots().
I believe there should me more functions to mark with __init here.
Anyways, here is the patch of what I had in mind.. I am not a compiler expert,
so please let me know your thoughts on this.
-ritesh
From 59d64dc0014ccb4ae13ed08ab596738628ee23b1 Mon Sep 17 00:00:00 2001
Message-Id: <59d64dc0014ccb4ae13ed08ab596738628ee23b1.1748084756.git.ritesh.list@gmail.com>
From: "Ritesh Harjani (IBM)" <ritesh.list@gmail.com>
Date: Sat, 24 May 2025 16:14:08 +0530
Subject: [RFC] powerpc/mm/book3s64: Move few kfence & debug_pagealloc
related calls to __init section
Move few kfence and debug_pagealloc related functions in hash_utils.c
and radix_pgtable.c to __init sections since these are only invoked once
by an __init function during system initialization.
i.e.
- hash_debug_pagealloc_alloc_slots()
- hash_kfence_alloc_pool()
- hash_kfence_map_pool()
The above 3 functions only gets called by __init htab_initialize().
- alloc_kfence_pool()
- map_kfence_pool()
The above 2 functions only gets called by __init radix_init_pgtable()
This should also help fix warning msgs like:
>> WARNING: modpost: vmlinux: section mismatch in reference:
hash_debug_pagealloc_alloc_slots+0xb0 (section: .text) ->
memblock_alloc_try_nid (section: .init.text)
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202504190552.mnFGs5sj-lkp@intel.com/
Signed-off-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com>
---
arch/powerpc/mm/book3s64/hash_utils.c | 6 +++---
arch/powerpc/mm/book3s64/radix_pgtable.c | 4 ++--
2 files changed, 5 insertions(+), 5 deletions(-)
diff --git a/arch/powerpc/mm/book3s64/hash_utils.c b/arch/powerpc/mm/book3s64/hash_utils.c
index 5158aefe4873..4693c464fc5a 100644
--- a/arch/powerpc/mm/book3s64/hash_utils.c
+++ b/arch/powerpc/mm/book3s64/hash_utils.c
@@ -343,7 +343,7 @@ static inline bool hash_supports_debug_pagealloc(void)
static u8 *linear_map_hash_slots;
static unsigned long linear_map_hash_count;
static DEFINE_RAW_SPINLOCK(linear_map_hash_lock);
-static void hash_debug_pagealloc_alloc_slots(void)
+static __init void hash_debug_pagealloc_alloc_slots(void)
{
if (!hash_supports_debug_pagealloc())
return;
@@ -409,7 +409,7 @@ static DEFINE_RAW_SPINLOCK(linear_map_kf_hash_lock);
static phys_addr_t kfence_pool;
-static inline void hash_kfence_alloc_pool(void)
+static __init void hash_kfence_alloc_pool(void)
{
if (!kfence_early_init_enabled())
goto err;
@@ -445,7 +445,7 @@ static inline void hash_kfence_alloc_pool(void)
disable_kfence();
}
-static inline void hash_kfence_map_pool(void)
+static __init void hash_kfence_map_pool(void)
{
unsigned long kfence_pool_start, kfence_pool_end;
unsigned long prot = pgprot_val(PAGE_KERNEL);
diff --git a/arch/powerpc/mm/book3s64/radix_pgtable.c b/arch/powerpc/mm/book3s64/radix_pgtable.c
index 311e2112d782..ed226ee1569a 100644
--- a/arch/powerpc/mm/book3s64/radix_pgtable.c
+++ b/arch/powerpc/mm/book3s64/radix_pgtable.c
@@ -363,7 +363,7 @@ static int __meminit create_physical_mapping(unsigned long start,
}
#ifdef CONFIG_KFENCE
-static inline phys_addr_t alloc_kfence_pool(void)
+static __init phys_addr_t alloc_kfence_pool(void)
{
phys_addr_t kfence_pool;
@@ -393,7 +393,7 @@ static inline phys_addr_t alloc_kfence_pool(void)
return 0;
}
-static inline void map_kfence_pool(phys_addr_t kfence_pool)
+static __init void map_kfence_pool(phys_addr_t kfence_pool)
{
if (!kfence_pool)
return;
--
2.39.5
On Sat, May 24, 2025 at 04:13:02PM +0530, Ritesh Harjani wrote:
> Kees Cook <kees@kernel.org> writes:
>
> > When KCOV is enabled all functions get instrumented, unless
> > the __no_sanitize_coverage attribute is used. To prepare for
> > __no_sanitize_coverage being applied to __init functions, we have to
> > handle differences in how GCC's inline optimizations get resolved. For
> > s390 this requires forcing a couple functions to be inline with
> > __always_inline.
> >
> > Signed-off-by: Kees Cook <kees@kernel.org>
> > ---
> > Cc: Madhavan Srinivasan <maddy@linux.ibm.com>
> > Cc: Michael Ellerman <mpe@ellerman.id.au>
> > Cc: Nicholas Piggin <npiggin@gmail.com>
> > Cc: Christophe Leroy <christophe.leroy@csgroup.eu>
> > Cc: Naveen N Rao <naveen@kernel.org>
> > Cc: "Ritesh Harjani (IBM)" <ritesh.list@gmail.com>
> > Cc: "Aneesh Kumar K.V" <aneesh.kumar@linux.ibm.com>
> > Cc: Andrew Morton <akpm@linux-foundation.org>
> > Cc: <linuxppc-dev@lists.ozlabs.org>
> > ---
> > arch/powerpc/mm/book3s64/hash_utils.c | 2 +-
> > arch/powerpc/mm/book3s64/radix_pgtable.c | 2 +-
> > 2 files changed, 2 insertions(+), 2 deletions(-)
> >
> > diff --git a/arch/powerpc/mm/book3s64/hash_utils.c b/arch/powerpc/mm/book3s64/hash_utils.c
> > index 5158aefe4873..93f1e1eb5ea6 100644
> > --- a/arch/powerpc/mm/book3s64/hash_utils.c
> > +++ b/arch/powerpc/mm/book3s64/hash_utils.c
> > @@ -409,7 +409,7 @@ static DEFINE_RAW_SPINLOCK(linear_map_kf_hash_lock);
> >
> > static phys_addr_t kfence_pool;
> >
> > -static inline void hash_kfence_alloc_pool(void)
> > +static __always_inline void hash_kfence_alloc_pool(void)
> > {
> > if (!kfence_early_init_enabled())
> > goto err;
> > diff --git a/arch/powerpc/mm/book3s64/radix_pgtable.c b/arch/powerpc/mm/book3s64/radix_pgtable.c
> > index 9f764bc42b8c..3238e9ed46b5 100644
> > --- a/arch/powerpc/mm/book3s64/radix_pgtable.c
> > +++ b/arch/powerpc/mm/book3s64/radix_pgtable.c
> > @@ -363,7 +363,7 @@ static int __meminit create_physical_mapping(unsigned long start,
> > }
> >
> > #ifdef CONFIG_KFENCE
> > -static inline phys_addr_t alloc_kfence_pool(void)
> > +static __always_inline phys_addr_t alloc_kfence_pool(void)
> > {
> > phys_addr_t kfence_pool;
> >
>
> I remember seeing a warning msg around .init.text section. Let me dig
> that...
>
> ... Here it is: https://lore.kernel.org/oe-kbuild-all/202504190552.mnFGs5sj-lkp@intel.com/
>
> I am not sure why it only complains for hash_debug_pagealloc_alloc_slots().
> I believe there should me more functions to mark with __init here.
> Anyways, here is the patch of what I had in mind.. I am not a compiler expert,
> so please let me know your thoughts on this.
Yeah, this looks good. I'll snag your patch and drop mine. :)
-Kees
--
Kees Cook
On Thu, 2025-05-22 at 21:39 -0700, Kees Cook wrote: > When KCOV is enabled all functions get instrumented, unless > the __no_sanitize_coverage attribute is used. To prepare for > __no_sanitize_coverage being applied to __init functions, we have to > handle differences in how GCC's inline optimizations get resolved. > For > s390 this requires forcing a couple functions to be inline with I assume you mean powerpc here, though I'm sure my employer is happy that you're at least confusing us with IBM's other architecture :) -- Andrew Donnellan OzLabs, ADL Canberra ajd@linux.ibm.com IBM Australia Limited
On May 22, 2025 10:24:30 PM PDT, Andrew Donnellan <ajd@linux.ibm.com> wrote: >On Thu, 2025-05-22 at 21:39 -0700, Kees Cook wrote: >> When KCOV is enabled all functions get instrumented, unless >> the __no_sanitize_coverage attribute is used. To prepare for >> __no_sanitize_coverage being applied to __init functions, we have to >> handle differences in how GCC's inline optimizations get resolved. >> For >> s390 this requires forcing a couple functions to be inline with > >I assume you mean powerpc here, though I'm sure my employer is happy >that you're at least confusing us with IBM's other architecture :) Whoops! Yes. Paste-o on my part. The rest of the sentence was updated correctly though. :) -- Kees Cook
© 2016 - 2025 Red Hat, Inc.