* [PATCH 1/2] riscv: Fix asan-stack clang build
@ 2021-10-27 4:58 Alexandre Ghiti
2021-10-27 4:58 ` [PATCH 2/2] riscv: Fix CONFIG_KASAN_STACK build Alexandre Ghiti
2021-10-27 23:06 ` [PATCH 1/2] riscv: Fix asan-stack clang build Palmer Dabbelt
0 siblings, 2 replies; 10+ messages in thread
From: Alexandre Ghiti @ 2021-10-27 4:58 UTC (permalink / raw)
To: Paul Walmsley, Palmer Dabbelt, Albert Ou, Andrey Ryabinin,
Alexander Potapenko, Andrey Konovalov, Dmitry Vyukov,
linux-riscv, linux-kernel, kasan-dev
Cc: Alexandre Ghiti, Nathan Chancellor
Nathan reported that because KASAN_SHADOW_OFFSET was not defined in
Kconfig, it prevents asan-stack from getting disabled with clang even
when CONFIG_KASAN_STACK is disabled: fix this by defining the
corresponding config.
Reported-by: Nathan Chancellor <nathan@kernel.org>
Signed-off-by: Alexandre Ghiti <alexandre.ghiti@canonical.com>
---
arch/riscv/Kconfig | 6 ++++++
arch/riscv/include/asm/kasan.h | 3 +--
arch/riscv/mm/kasan_init.c | 3 +++
3 files changed, 10 insertions(+), 2 deletions(-)
diff --git a/arch/riscv/Kconfig b/arch/riscv/Kconfig
index c1abbc876e5b..79250b1ed54e 100644
--- a/arch/riscv/Kconfig
+++ b/arch/riscv/Kconfig
@@ -162,6 +162,12 @@ config PAGE_OFFSET
default 0xffffffff80000000 if 64BIT && MAXPHYSMEM_2GB
default 0xffffffe000000000 if 64BIT && MAXPHYSMEM_128GB
+config KASAN_SHADOW_OFFSET
+ hex
+ depends on KASAN_GENERIC
+ default 0xdfffffc800000000 if 64BIT
+ default 0xffffffff if 32BIT
+
config ARCH_FLATMEM_ENABLE
def_bool !NUMA
diff --git a/arch/riscv/include/asm/kasan.h b/arch/riscv/include/asm/kasan.h
index a2b3d9cdbc86..b00f503ec124 100644
--- a/arch/riscv/include/asm/kasan.h
+++ b/arch/riscv/include/asm/kasan.h
@@ -30,8 +30,7 @@
#define KASAN_SHADOW_SIZE (UL(1) << ((CONFIG_VA_BITS - 1) - KASAN_SHADOW_SCALE_SHIFT))
#define KASAN_SHADOW_START KERN_VIRT_START
#define KASAN_SHADOW_END (KASAN_SHADOW_START + KASAN_SHADOW_SIZE)
-#define KASAN_SHADOW_OFFSET (KASAN_SHADOW_END - (1ULL << \
- (64 - KASAN_SHADOW_SCALE_SHIFT)))
+#define KASAN_SHADOW_OFFSET _AC(CONFIG_KASAN_SHADOW_OFFSET, UL)
void kasan_init(void);
asmlinkage void kasan_early_init(void);
diff --git a/arch/riscv/mm/kasan_init.c b/arch/riscv/mm/kasan_init.c
index d7189c8714a9..8175e98b9073 100644
--- a/arch/riscv/mm/kasan_init.c
+++ b/arch/riscv/mm/kasan_init.c
@@ -17,6 +17,9 @@ asmlinkage void __init kasan_early_init(void)
uintptr_t i;
pgd_t *pgd = early_pg_dir + pgd_index(KASAN_SHADOW_START);
+ BUILD_BUG_ON(KASAN_SHADOW_OFFSET !=
+ KASAN_SHADOW_END - (1UL << (64 - KASAN_SHADOW_SCALE_SHIFT)));
+
for (i = 0; i < PTRS_PER_PTE; ++i)
set_pte(kasan_early_shadow_pte + i,
mk_pte(virt_to_page(kasan_early_shadow_page),
--
2.30.2
^ permalink raw reply related [flat|nested] 10+ messages in thread
* [PATCH 2/2] riscv: Fix CONFIG_KASAN_STACK build
2021-10-27 4:58 [PATCH 1/2] riscv: Fix asan-stack clang build Alexandre Ghiti
@ 2021-10-27 4:58 ` Alexandre Ghiti
2021-10-27 23:06 ` [PATCH 1/2] riscv: Fix asan-stack clang build Palmer Dabbelt
1 sibling, 0 replies; 10+ messages in thread
From: Alexandre Ghiti @ 2021-10-27 4:58 UTC (permalink / raw)
To: Paul Walmsley, Palmer Dabbelt, Albert Ou, Andrey Ryabinin,
Alexander Potapenko, Andrey Konovalov, Dmitry Vyukov,
linux-riscv, linux-kernel, kasan-dev
Cc: Alexandre Ghiti
Now that CONFIG_KASAN_SHADOW_OFFSET is correctly defined, the boot
hung while populating the shadow memory right after the call to
kasan_populate_early_shadow: when calling this function, all the shadow
memory is already populated with kasan_early_shadow_pte which has
PAGE_KERNEL protection.
kasan_populate_early_shadow write-protects the mapping of the range
of addresses passed in argument in zero_pte_populate, which actually
write-protects all the shadow memory mapping since
kasan_early_shadow_pte is used for all the shadow memory at this point.
And then when using memblock API to populate the shadow memory, the first
write access to the kernel stack triggers a trap.
We already manually populate all the shadow memory in kasan_early_init
and we write-protect kasan_early_shadow_pte at the end of kasan_init
which makes the call to kasan_populate_early_shadow superfluous so
we can remove it.
Signed-off-by: Alexandre Ghiti <alexandre.ghiti@canonical.com>
---
arch/riscv/mm/kasan_init.c | 7 -------
1 file changed, 7 deletions(-)
diff --git a/arch/riscv/mm/kasan_init.c b/arch/riscv/mm/kasan_init.c
index 8175e98b9073..8df937902630 100644
--- a/arch/riscv/mm/kasan_init.c
+++ b/arch/riscv/mm/kasan_init.c
@@ -175,13 +175,6 @@ void __init kasan_init(void)
phys_addr_t p_start, p_end;
u64 i;
- /*
- * Populate all kernel virtual address space with kasan_early_shadow_page
- * except for the linear mapping and the modules/kernel/BPF mapping.
- */
- kasan_populate_early_shadow((void *)KASAN_SHADOW_START,
- (void *)kasan_mem_to_shadow((void *)
- VMEMMAP_END));
if (IS_ENABLED(CONFIG_KASAN_VMALLOC))
kasan_shallow_populate(
(void *)kasan_mem_to_shadow((void *)VMALLOC_START),
--
2.30.2
^ permalink raw reply related [flat|nested] 10+ messages in thread
* Re: [PATCH 1/2] riscv: Fix asan-stack clang build
2021-10-27 4:58 [PATCH 1/2] riscv: Fix asan-stack clang build Alexandre Ghiti
2021-10-27 4:58 ` [PATCH 2/2] riscv: Fix CONFIG_KASAN_STACK build Alexandre Ghiti
@ 2021-10-27 23:06 ` Palmer Dabbelt
2021-10-28 4:15 ` Alexandre Ghiti
1 sibling, 1 reply; 10+ messages in thread
From: Palmer Dabbelt @ 2021-10-27 23:06 UTC (permalink / raw)
To: alexandre.ghiti
Cc: Paul Walmsley, aou, ryabinin.a.a, glider, andreyknvl, dvyukov,
linux-riscv, linux-kernel, kasan-dev, alexandre.ghiti, nathan
On Tue, 26 Oct 2021 21:58:42 PDT (-0700), alexandre.ghiti@canonical.com wrote:
> Nathan reported that because KASAN_SHADOW_OFFSET was not defined in
> Kconfig, it prevents asan-stack from getting disabled with clang even
> when CONFIG_KASAN_STACK is disabled: fix this by defining the
> corresponding config.
>
> Reported-by: Nathan Chancellor <nathan@kernel.org>
> Signed-off-by: Alexandre Ghiti <alexandre.ghiti@canonical.com>
> ---
> arch/riscv/Kconfig | 6 ++++++
> arch/riscv/include/asm/kasan.h | 3 +--
> arch/riscv/mm/kasan_init.c | 3 +++
> 3 files changed, 10 insertions(+), 2 deletions(-)
>
> diff --git a/arch/riscv/Kconfig b/arch/riscv/Kconfig
> index c1abbc876e5b..79250b1ed54e 100644
> --- a/arch/riscv/Kconfig
> +++ b/arch/riscv/Kconfig
> @@ -162,6 +162,12 @@ config PAGE_OFFSET
> default 0xffffffff80000000 if 64BIT && MAXPHYSMEM_2GB
> default 0xffffffe000000000 if 64BIT && MAXPHYSMEM_128GB
>
> +config KASAN_SHADOW_OFFSET
> + hex
> + depends on KASAN_GENERIC
> + default 0xdfffffc800000000 if 64BIT
> + default 0xffffffff if 32BIT
I thought I posted this somewhere, but this is exactly what my first
guess was. The problem is that it's hanging on boot for me. I don't
really have anything exotic going on, it's just a defconfig with
CONFIG_KASAN=y running in QEMU.
Does this boot for you?
> +
> config ARCH_FLATMEM_ENABLE
> def_bool !NUMA
>
> diff --git a/arch/riscv/include/asm/kasan.h b/arch/riscv/include/asm/kasan.h
> index a2b3d9cdbc86..b00f503ec124 100644
> --- a/arch/riscv/include/asm/kasan.h
> +++ b/arch/riscv/include/asm/kasan.h
> @@ -30,8 +30,7 @@
> #define KASAN_SHADOW_SIZE (UL(1) << ((CONFIG_VA_BITS - 1) - KASAN_SHADOW_SCALE_SHIFT))
> #define KASAN_SHADOW_START KERN_VIRT_START
> #define KASAN_SHADOW_END (KASAN_SHADOW_START + KASAN_SHADOW_SIZE)
> -#define KASAN_SHADOW_OFFSET (KASAN_SHADOW_END - (1ULL << \
> - (64 - KASAN_SHADOW_SCALE_SHIFT)))
> +#define KASAN_SHADOW_OFFSET _AC(CONFIG_KASAN_SHADOW_OFFSET, UL)
>
> void kasan_init(void);
> asmlinkage void kasan_early_init(void);
> diff --git a/arch/riscv/mm/kasan_init.c b/arch/riscv/mm/kasan_init.c
> index d7189c8714a9..8175e98b9073 100644
> --- a/arch/riscv/mm/kasan_init.c
> +++ b/arch/riscv/mm/kasan_init.c
> @@ -17,6 +17,9 @@ asmlinkage void __init kasan_early_init(void)
> uintptr_t i;
> pgd_t *pgd = early_pg_dir + pgd_index(KASAN_SHADOW_START);
>
> + BUILD_BUG_ON(KASAN_SHADOW_OFFSET !=
> + KASAN_SHADOW_END - (1UL << (64 - KASAN_SHADOW_SCALE_SHIFT)));
> +
> for (i = 0; i < PTRS_PER_PTE; ++i)
> set_pte(kasan_early_shadow_pte + i,
> mk_pte(virt_to_page(kasan_early_shadow_page),
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH 1/2] riscv: Fix asan-stack clang build
2021-10-27 23:06 ` [PATCH 1/2] riscv: Fix asan-stack clang build Palmer Dabbelt
@ 2021-10-28 4:15 ` Alexandre Ghiti
2021-10-28 5:02 ` Palmer Dabbelt
0 siblings, 1 reply; 10+ messages in thread
From: Alexandre Ghiti @ 2021-10-28 4:15 UTC (permalink / raw)
To: Palmer Dabbelt
Cc: Paul Walmsley, aou, ryabinin.a.a, glider, andreyknvl, dvyukov,
linux-riscv, linux-kernel, kasan-dev, nathan
On Thu, Oct 28, 2021 at 1:06 AM Palmer Dabbelt <palmer@dabbelt.com> wrote:
>
> On Tue, 26 Oct 2021 21:58:42 PDT (-0700), alexandre.ghiti@canonical.com wrote:
> > Nathan reported that because KASAN_SHADOW_OFFSET was not defined in
> > Kconfig, it prevents asan-stack from getting disabled with clang even
> > when CONFIG_KASAN_STACK is disabled: fix this by defining the
> > corresponding config.
> >
> > Reported-by: Nathan Chancellor <nathan@kernel.org>
> > Signed-off-by: Alexandre Ghiti <alexandre.ghiti@canonical.com>
> > ---
> > arch/riscv/Kconfig | 6 ++++++
> > arch/riscv/include/asm/kasan.h | 3 +--
> > arch/riscv/mm/kasan_init.c | 3 +++
> > 3 files changed, 10 insertions(+), 2 deletions(-)
> >
> > diff --git a/arch/riscv/Kconfig b/arch/riscv/Kconfig
> > index c1abbc876e5b..79250b1ed54e 100644
> > --- a/arch/riscv/Kconfig
> > +++ b/arch/riscv/Kconfig
> > @@ -162,6 +162,12 @@ config PAGE_OFFSET
> > default 0xffffffff80000000 if 64BIT && MAXPHYSMEM_2GB
> > default 0xffffffe000000000 if 64BIT && MAXPHYSMEM_128GB
> >
> > +config KASAN_SHADOW_OFFSET
> > + hex
> > + depends on KASAN_GENERIC
> > + default 0xdfffffc800000000 if 64BIT
> > + default 0xffffffff if 32BIT
>
> I thought I posted this somewhere, but this is exactly what my first
> guess was. The problem is that it's hanging on boot for me. I don't
> really have anything exotic going on, it's just a defconfig with
> CONFIG_KASAN=y running in QEMU.
>
> Does this boot for you?
Yes with the 2nd patch of this series which fixes the issue
encountered here. And that's true I copied/pasted this part of your
patch which was better than what I had initially done, sorry I should
have mentioned you did that, please add a Codeveloped-by or something
like that.
Thanks,
Alex
>
> > +
> > config ARCH_FLATMEM_ENABLE
> > def_bool !NUMA
> >
> > diff --git a/arch/riscv/include/asm/kasan.h b/arch/riscv/include/asm/kasan.h
> > index a2b3d9cdbc86..b00f503ec124 100644
> > --- a/arch/riscv/include/asm/kasan.h
> > +++ b/arch/riscv/include/asm/kasan.h
> > @@ -30,8 +30,7 @@
> > #define KASAN_SHADOW_SIZE (UL(1) << ((CONFIG_VA_BITS - 1) - KASAN_SHADOW_SCALE_SHIFT))
> > #define KASAN_SHADOW_START KERN_VIRT_START
> > #define KASAN_SHADOW_END (KASAN_SHADOW_START + KASAN_SHADOW_SIZE)
> > -#define KASAN_SHADOW_OFFSET (KASAN_SHADOW_END - (1ULL << \
> > - (64 - KASAN_SHADOW_SCALE_SHIFT)))
> > +#define KASAN_SHADOW_OFFSET _AC(CONFIG_KASAN_SHADOW_OFFSET, UL)
> >
> > void kasan_init(void);
> > asmlinkage void kasan_early_init(void);
> > diff --git a/arch/riscv/mm/kasan_init.c b/arch/riscv/mm/kasan_init.c
> > index d7189c8714a9..8175e98b9073 100644
> > --- a/arch/riscv/mm/kasan_init.c
> > +++ b/arch/riscv/mm/kasan_init.c
> > @@ -17,6 +17,9 @@ asmlinkage void __init kasan_early_init(void)
> > uintptr_t i;
> > pgd_t *pgd = early_pg_dir + pgd_index(KASAN_SHADOW_START);
> >
> > + BUILD_BUG_ON(KASAN_SHADOW_OFFSET !=
> > + KASAN_SHADOW_END - (1UL << (64 - KASAN_SHADOW_SCALE_SHIFT)));
> > +
> > for (i = 0; i < PTRS_PER_PTE; ++i)
> > set_pte(kasan_early_shadow_pte + i,
> > mk_pte(virt_to_page(kasan_early_shadow_page),
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH 1/2] riscv: Fix asan-stack clang build
2021-10-28 4:15 ` Alexandre Ghiti
@ 2021-10-28 5:02 ` Palmer Dabbelt
2021-10-28 5:30 ` Alexandre Ghiti
0 siblings, 1 reply; 10+ messages in thread
From: Palmer Dabbelt @ 2021-10-28 5:02 UTC (permalink / raw)
To: alexandre.ghiti
Cc: Paul Walmsley, aou, ryabinin.a.a, glider, andreyknvl, dvyukov,
linux-riscv, linux-kernel, kasan-dev, nathan
On Wed, 27 Oct 2021 21:15:28 PDT (-0700), alexandre.ghiti@canonical.com wrote:
> On Thu, Oct 28, 2021 at 1:06 AM Palmer Dabbelt <palmer@dabbelt.com> wrote:
>>
>> On Tue, 26 Oct 2021 21:58:42 PDT (-0700), alexandre.ghiti@canonical.com wrote:
>> > Nathan reported that because KASAN_SHADOW_OFFSET was not defined in
>> > Kconfig, it prevents asan-stack from getting disabled with clang even
>> > when CONFIG_KASAN_STACK is disabled: fix this by defining the
>> > corresponding config.
>> >
>> > Reported-by: Nathan Chancellor <nathan@kernel.org>
>> > Signed-off-by: Alexandre Ghiti <alexandre.ghiti@canonical.com>
>> > ---
>> > arch/riscv/Kconfig | 6 ++++++
>> > arch/riscv/include/asm/kasan.h | 3 +--
>> > arch/riscv/mm/kasan_init.c | 3 +++
>> > 3 files changed, 10 insertions(+), 2 deletions(-)
>> >
>> > diff --git a/arch/riscv/Kconfig b/arch/riscv/Kconfig
>> > index c1abbc876e5b..79250b1ed54e 100644
>> > --- a/arch/riscv/Kconfig
>> > +++ b/arch/riscv/Kconfig
>> > @@ -162,6 +162,12 @@ config PAGE_OFFSET
>> > default 0xffffffff80000000 if 64BIT && MAXPHYSMEM_2GB
>> > default 0xffffffe000000000 if 64BIT && MAXPHYSMEM_128GB
>> >
>> > +config KASAN_SHADOW_OFFSET
>> > + hex
>> > + depends on KASAN_GENERIC
>> > + default 0xdfffffc800000000 if 64BIT
>> > + default 0xffffffff if 32BIT
>>
>> I thought I posted this somewhere, but this is exactly what my first
>> guess was. The problem is that it's hanging on boot for me. I don't
>> really have anything exotic going on, it's just a defconfig with
>> CONFIG_KASAN=y running in QEMU.
>>
>> Does this boot for you?
>
> Yes with the 2nd patch of this series which fixes the issue
> encountered here. And that's true I copied/pasted this part of your
> patch which was better than what I had initially done, sorry I should
> have mentioned you did that, please add a Codeveloped-by or something
> like that.
Not sure if I'm missing something, but it's still not booting for me.
I've put what I'm testing on palmer/to-test, it's these two on top of
fixes and merged into Linus' tree
* 6d7d351902ff - (HEAD -> to-test, palmer/to-test) Merge remote-tracking branch 'palmer/fixes' into to-test (7 minutes ago) <Palmer Dabbelt>
|\
| * 782551edf8f8 - (palmer/fixes) riscv: Fix CONFIG_KASAN_STACK build (6 hours ago) <Alexandre Ghiti>
| * 47383e5b3c4f - riscv: Fix asan-stack clang build (6 hours ago) <Alexandre Ghiti>
| * 64a19591a293 - (riscv/fixes) riscv: fix misalgned trap vector base address (9 hours ago) <Chen Lu>
* | 1fc596a56b33 - (palmer/master, linus/master, linus/HEAD, master) Merge tag 'trace-v5.15-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace (11 hours ago) <Linus Torvalds>
Am I missing something else?
>
> Thanks,
>
> Alex
>
>>
>> > +
>> > config ARCH_FLATMEM_ENABLE
>> > def_bool !NUMA
>> >
>> > diff --git a/arch/riscv/include/asm/kasan.h b/arch/riscv/include/asm/kasan.h
>> > index a2b3d9cdbc86..b00f503ec124 100644
>> > --- a/arch/riscv/include/asm/kasan.h
>> > +++ b/arch/riscv/include/asm/kasan.h
>> > @@ -30,8 +30,7 @@
>> > #define KASAN_SHADOW_SIZE (UL(1) << ((CONFIG_VA_BITS - 1) - KASAN_SHADOW_SCALE_SHIFT))
>> > #define KASAN_SHADOW_START KERN_VIRT_START
>> > #define KASAN_SHADOW_END (KASAN_SHADOW_START + KASAN_SHADOW_SIZE)
>> > -#define KASAN_SHADOW_OFFSET (KASAN_SHADOW_END - (1ULL << \
>> > - (64 - KASAN_SHADOW_SCALE_SHIFT)))
>> > +#define KASAN_SHADOW_OFFSET _AC(CONFIG_KASAN_SHADOW_OFFSET, UL)
>> >
>> > void kasan_init(void);
>> > asmlinkage void kasan_early_init(void);
>> > diff --git a/arch/riscv/mm/kasan_init.c b/arch/riscv/mm/kasan_init.c
>> > index d7189c8714a9..8175e98b9073 100644
>> > --- a/arch/riscv/mm/kasan_init.c
>> > +++ b/arch/riscv/mm/kasan_init.c
>> > @@ -17,6 +17,9 @@ asmlinkage void __init kasan_early_init(void)
>> > uintptr_t i;
>> > pgd_t *pgd = early_pg_dir + pgd_index(KASAN_SHADOW_START);
>> >
>> > + BUILD_BUG_ON(KASAN_SHADOW_OFFSET !=
>> > + KASAN_SHADOW_END - (1UL << (64 - KASAN_SHADOW_SCALE_SHIFT)));
>> > +
>> > for (i = 0; i < PTRS_PER_PTE; ++i)
>> > set_pte(kasan_early_shadow_pte + i,
>> > mk_pte(virt_to_page(kasan_early_shadow_page),
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH 1/2] riscv: Fix asan-stack clang build
2021-10-28 5:02 ` Palmer Dabbelt
@ 2021-10-28 5:30 ` Alexandre Ghiti
2021-10-28 5:34 ` Alexandre Ghiti
0 siblings, 1 reply; 10+ messages in thread
From: Alexandre Ghiti @ 2021-10-28 5:30 UTC (permalink / raw)
To: Palmer Dabbelt
Cc: Paul Walmsley, aou, ryabinin.a.a, glider, andreyknvl, dvyukov,
linux-riscv, linux-kernel, kasan-dev, nathan
On Thu, Oct 28, 2021 at 7:02 AM Palmer Dabbelt <palmer@dabbelt.com> wrote:
>
> On Wed, 27 Oct 2021 21:15:28 PDT (-0700), alexandre.ghiti@canonical.com wrote:
> > On Thu, Oct 28, 2021 at 1:06 AM Palmer Dabbelt <palmer@dabbelt.com> wrote:
> >>
> >> On Tue, 26 Oct 2021 21:58:42 PDT (-0700), alexandre.ghiti@canonical.com wrote:
> >> > Nathan reported that because KASAN_SHADOW_OFFSET was not defined in
> >> > Kconfig, it prevents asan-stack from getting disabled with clang even
> >> > when CONFIG_KASAN_STACK is disabled: fix this by defining the
> >> > corresponding config.
> >> >
> >> > Reported-by: Nathan Chancellor <nathan@kernel.org>
> >> > Signed-off-by: Alexandre Ghiti <alexandre.ghiti@canonical.com>
> >> > ---
> >> > arch/riscv/Kconfig | 6 ++++++
> >> > arch/riscv/include/asm/kasan.h | 3 +--
> >> > arch/riscv/mm/kasan_init.c | 3 +++
> >> > 3 files changed, 10 insertions(+), 2 deletions(-)
> >> >
> >> > diff --git a/arch/riscv/Kconfig b/arch/riscv/Kconfig
> >> > index c1abbc876e5b..79250b1ed54e 100644
> >> > --- a/arch/riscv/Kconfig
> >> > +++ b/arch/riscv/Kconfig
> >> > @@ -162,6 +162,12 @@ config PAGE_OFFSET
> >> > default 0xffffffff80000000 if 64BIT && MAXPHYSMEM_2GB
> >> > default 0xffffffe000000000 if 64BIT && MAXPHYSMEM_128GB
> >> >
> >> > +config KASAN_SHADOW_OFFSET
> >> > + hex
> >> > + depends on KASAN_GENERIC
> >> > + default 0xdfffffc800000000 if 64BIT
> >> > + default 0xffffffff if 32BIT
> >>
> >> I thought I posted this somewhere, but this is exactly what my first
> >> guess was. The problem is that it's hanging on boot for me. I don't
> >> really have anything exotic going on, it's just a defconfig with
> >> CONFIG_KASAN=y running in QEMU.
> >>
> >> Does this boot for you?
> >
> > Yes with the 2nd patch of this series which fixes the issue
> > encountered here. And that's true I copied/pasted this part of your
> > patch which was better than what I had initially done, sorry I should
> > have mentioned you did that, please add a Codeveloped-by or something
> > like that.
>
> Not sure if I'm missing something, but it's still not booting for me.
> I've put what I'm testing on palmer/to-test, it's these two on top of
> fixes and merged into Linus' tree
>
> * 6d7d351902ff - (HEAD -> to-test, palmer/to-test) Merge remote-tracking branch 'palmer/fixes' into to-test (7 minutes ago) <Palmer Dabbelt>
> |\
> | * 782551edf8f8 - (palmer/fixes) riscv: Fix CONFIG_KASAN_STACK build (6 hours ago) <Alexandre Ghiti>
> | * 47383e5b3c4f - riscv: Fix asan-stack clang build (6 hours ago) <Alexandre Ghiti>
> | * 64a19591a293 - (riscv/fixes) riscv: fix misalgned trap vector base address (9 hours ago) <Chen Lu>
> * | 1fc596a56b33 - (palmer/master, linus/master, linus/HEAD, master) Merge tag 'trace-v5.15-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace (11 hours ago) <Linus Torvalds>
>
> Am I missing something else?
Hmm, that's weird, I have just done the same: cherry-picked both my
commits on top of fixes (64a19591a293) and it boots fine with KASAN
enabled. Maybe a config thing? I pushed my branch here:
https://github.com/AlexGhiti/riscv-linux/tree/int/alex/kasan_stack_fixes_rebase
>
> >
> > Thanks,
> >
> > Alex
> >
> >>
> >> > +
> >> > config ARCH_FLATMEM_ENABLE
> >> > def_bool !NUMA
> >> >
> >> > diff --git a/arch/riscv/include/asm/kasan.h b/arch/riscv/include/asm/kasan.h
> >> > index a2b3d9cdbc86..b00f503ec124 100644
> >> > --- a/arch/riscv/include/asm/kasan.h
> >> > +++ b/arch/riscv/include/asm/kasan.h
> >> > @@ -30,8 +30,7 @@
> >> > #define KASAN_SHADOW_SIZE (UL(1) << ((CONFIG_VA_BITS - 1) - KASAN_SHADOW_SCALE_SHIFT))
> >> > #define KASAN_SHADOW_START KERN_VIRT_START
> >> > #define KASAN_SHADOW_END (KASAN_SHADOW_START + KASAN_SHADOW_SIZE)
> >> > -#define KASAN_SHADOW_OFFSET (KASAN_SHADOW_END - (1ULL << \
> >> > - (64 - KASAN_SHADOW_SCALE_SHIFT)))
> >> > +#define KASAN_SHADOW_OFFSET _AC(CONFIG_KASAN_SHADOW_OFFSET, UL)
> >> >
> >> > void kasan_init(void);
> >> > asmlinkage void kasan_early_init(void);
> >> > diff --git a/arch/riscv/mm/kasan_init.c b/arch/riscv/mm/kasan_init.c
> >> > index d7189c8714a9..8175e98b9073 100644
> >> > --- a/arch/riscv/mm/kasan_init.c
> >> > +++ b/arch/riscv/mm/kasan_init.c
> >> > @@ -17,6 +17,9 @@ asmlinkage void __init kasan_early_init(void)
> >> > uintptr_t i;
> >> > pgd_t *pgd = early_pg_dir + pgd_index(KASAN_SHADOW_START);
> >> >
> >> > + BUILD_BUG_ON(KASAN_SHADOW_OFFSET !=
> >> > + KASAN_SHADOW_END - (1UL << (64 - KASAN_SHADOW_SCALE_SHIFT)));
> >> > +
> >> > for (i = 0; i < PTRS_PER_PTE; ++i)
> >> > set_pte(kasan_early_shadow_pte + i,
> >> > mk_pte(virt_to_page(kasan_early_shadow_page),
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH 1/2] riscv: Fix asan-stack clang build
2021-10-28 5:30 ` Alexandre Ghiti
@ 2021-10-28 5:34 ` Alexandre Ghiti
2021-10-28 6:45 ` Palmer Dabbelt
0 siblings, 1 reply; 10+ messages in thread
From: Alexandre Ghiti @ 2021-10-28 5:34 UTC (permalink / raw)
To: Palmer Dabbelt
Cc: Paul Walmsley, aou, ryabinin.a.a, glider, andreyknvl, dvyukov,
linux-riscv, linux-kernel, kasan-dev, nathan
On Thu, Oct 28, 2021 at 7:30 AM Alexandre Ghiti
<alexandre.ghiti@canonical.com> wrote:
>
> On Thu, Oct 28, 2021 at 7:02 AM Palmer Dabbelt <palmer@dabbelt.com> wrote:
> >
> > On Wed, 27 Oct 2021 21:15:28 PDT (-0700), alexandre.ghiti@canonical.com wrote:
> > > On Thu, Oct 28, 2021 at 1:06 AM Palmer Dabbelt <palmer@dabbelt.com> wrote:
> > >>
> > >> On Tue, 26 Oct 2021 21:58:42 PDT (-0700), alexandre.ghiti@canonical.com wrote:
> > >> > Nathan reported that because KASAN_SHADOW_OFFSET was not defined in
> > >> > Kconfig, it prevents asan-stack from getting disabled with clang even
> > >> > when CONFIG_KASAN_STACK is disabled: fix this by defining the
> > >> > corresponding config.
> > >> >
> > >> > Reported-by: Nathan Chancellor <nathan@kernel.org>
> > >> > Signed-off-by: Alexandre Ghiti <alexandre.ghiti@canonical.com>
> > >> > ---
> > >> > arch/riscv/Kconfig | 6 ++++++
> > >> > arch/riscv/include/asm/kasan.h | 3 +--
> > >> > arch/riscv/mm/kasan_init.c | 3 +++
> > >> > 3 files changed, 10 insertions(+), 2 deletions(-)
> > >> >
> > >> > diff --git a/arch/riscv/Kconfig b/arch/riscv/Kconfig
> > >> > index c1abbc876e5b..79250b1ed54e 100644
> > >> > --- a/arch/riscv/Kconfig
> > >> > +++ b/arch/riscv/Kconfig
> > >> > @@ -162,6 +162,12 @@ config PAGE_OFFSET
> > >> > default 0xffffffff80000000 if 64BIT && MAXPHYSMEM_2GB
> > >> > default 0xffffffe000000000 if 64BIT && MAXPHYSMEM_128GB
> > >> >
> > >> > +config KASAN_SHADOW_OFFSET
> > >> > + hex
> > >> > + depends on KASAN_GENERIC
> > >> > + default 0xdfffffc800000000 if 64BIT
> > >> > + default 0xffffffff if 32BIT
> > >>
> > >> I thought I posted this somewhere, but this is exactly what my first
> > >> guess was. The problem is that it's hanging on boot for me. I don't
> > >> really have anything exotic going on, it's just a defconfig with
> > >> CONFIG_KASAN=y running in QEMU.
> > >>
> > >> Does this boot for you?
> > >
> > > Yes with the 2nd patch of this series which fixes the issue
> > > encountered here. And that's true I copied/pasted this part of your
> > > patch which was better than what I had initially done, sorry I should
> > > have mentioned you did that, please add a Codeveloped-by or something
> > > like that.
> >
> > Not sure if I'm missing something, but it's still not booting for me.
> > I've put what I'm testing on palmer/to-test, it's these two on top of
> > fixes and merged into Linus' tree
> >
> > * 6d7d351902ff - (HEAD -> to-test, palmer/to-test) Merge remote-tracking branch 'palmer/fixes' into to-test (7 minutes ago) <Palmer Dabbelt>
> > |\
> > | * 782551edf8f8 - (palmer/fixes) riscv: Fix CONFIG_KASAN_STACK build (6 hours ago) <Alexandre Ghiti>
> > | * 47383e5b3c4f - riscv: Fix asan-stack clang build (6 hours ago) <Alexandre Ghiti>
> > | * 64a19591a293 - (riscv/fixes) riscv: fix misalgned trap vector base address (9 hours ago) <Chen Lu>
> > * | 1fc596a56b33 - (palmer/master, linus/master, linus/HEAD, master) Merge tag 'trace-v5.15-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace (11 hours ago) <Linus Torvalds>
> >
> > Am I missing something else?
>
> Hmm, that's weird, I have just done the same: cherry-picked both my
> commits on top of fixes (64a19591a293) and it boots fine with KASAN
> enabled. Maybe a config thing? I pushed my branch here:
> https://github.com/AlexGhiti/riscv-linux/tree/int/alex/kasan_stack_fixes_rebase
I pushed the config I use and that boots in that branch, maybe there's
another issue somewhere.
>
> >
> > >
> > > Thanks,
> > >
> > > Alex
> > >
> > >>
> > >> > +
> > >> > config ARCH_FLATMEM_ENABLE
> > >> > def_bool !NUMA
> > >> >
> > >> > diff --git a/arch/riscv/include/asm/kasan.h b/arch/riscv/include/asm/kasan.h
> > >> > index a2b3d9cdbc86..b00f503ec124 100644
> > >> > --- a/arch/riscv/include/asm/kasan.h
> > >> > +++ b/arch/riscv/include/asm/kasan.h
> > >> > @@ -30,8 +30,7 @@
> > >> > #define KASAN_SHADOW_SIZE (UL(1) << ((CONFIG_VA_BITS - 1) - KASAN_SHADOW_SCALE_SHIFT))
> > >> > #define KASAN_SHADOW_START KERN_VIRT_START
> > >> > #define KASAN_SHADOW_END (KASAN_SHADOW_START + KASAN_SHADOW_SIZE)
> > >> > -#define KASAN_SHADOW_OFFSET (KASAN_SHADOW_END - (1ULL << \
> > >> > - (64 - KASAN_SHADOW_SCALE_SHIFT)))
> > >> > +#define KASAN_SHADOW_OFFSET _AC(CONFIG_KASAN_SHADOW_OFFSET, UL)
> > >> >
> > >> > void kasan_init(void);
> > >> > asmlinkage void kasan_early_init(void);
> > >> > diff --git a/arch/riscv/mm/kasan_init.c b/arch/riscv/mm/kasan_init.c
> > >> > index d7189c8714a9..8175e98b9073 100644
> > >> > --- a/arch/riscv/mm/kasan_init.c
> > >> > +++ b/arch/riscv/mm/kasan_init.c
> > >> > @@ -17,6 +17,9 @@ asmlinkage void __init kasan_early_init(void)
> > >> > uintptr_t i;
> > >> > pgd_t *pgd = early_pg_dir + pgd_index(KASAN_SHADOW_START);
> > >> >
> > >> > + BUILD_BUG_ON(KASAN_SHADOW_OFFSET !=
> > >> > + KASAN_SHADOW_END - (1UL << (64 - KASAN_SHADOW_SCALE_SHIFT)));
> > >> > +
> > >> > for (i = 0; i < PTRS_PER_PTE; ++i)
> > >> > set_pte(kasan_early_shadow_pte + i,
> > >> > mk_pte(virt_to_page(kasan_early_shadow_page),
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH 1/2] riscv: Fix asan-stack clang build
2021-10-28 5:34 ` Alexandre Ghiti
@ 2021-10-28 6:45 ` Palmer Dabbelt
2021-10-28 7:13 ` Alexandre Ghiti
0 siblings, 1 reply; 10+ messages in thread
From: Palmer Dabbelt @ 2021-10-28 6:45 UTC (permalink / raw)
To: alexandre.ghiti
Cc: Paul Walmsley, aou, ryabinin.a.a, glider, andreyknvl, dvyukov,
linux-riscv, linux-kernel, kasan-dev, nathan
On Wed, 27 Oct 2021 22:34:32 PDT (-0700), alexandre.ghiti@canonical.com wrote:
> On Thu, Oct 28, 2021 at 7:30 AM Alexandre Ghiti
> <alexandre.ghiti@canonical.com> wrote:
>>
>> On Thu, Oct 28, 2021 at 7:02 AM Palmer Dabbelt <palmer@dabbelt.com> wrote:
>> >
>> > On Wed, 27 Oct 2021 21:15:28 PDT (-0700), alexandre.ghiti@canonical.com wrote:
>> > > On Thu, Oct 28, 2021 at 1:06 AM Palmer Dabbelt <palmer@dabbelt.com> wrote:
>> > >>
>> > >> On Tue, 26 Oct 2021 21:58:42 PDT (-0700), alexandre.ghiti@canonical.com wrote:
>> > >> > Nathan reported that because KASAN_SHADOW_OFFSET was not defined in
>> > >> > Kconfig, it prevents asan-stack from getting disabled with clang even
>> > >> > when CONFIG_KASAN_STACK is disabled: fix this by defining the
>> > >> > corresponding config.
>> > >> >
>> > >> > Reported-by: Nathan Chancellor <nathan@kernel.org>
>> > >> > Signed-off-by: Alexandre Ghiti <alexandre.ghiti@canonical.com>
>> > >> > ---
>> > >> > arch/riscv/Kconfig | 6 ++++++
>> > >> > arch/riscv/include/asm/kasan.h | 3 +--
>> > >> > arch/riscv/mm/kasan_init.c | 3 +++
>> > >> > 3 files changed, 10 insertions(+), 2 deletions(-)
>> > >> >
>> > >> > diff --git a/arch/riscv/Kconfig b/arch/riscv/Kconfig
>> > >> > index c1abbc876e5b..79250b1ed54e 100644
>> > >> > --- a/arch/riscv/Kconfig
>> > >> > +++ b/arch/riscv/Kconfig
>> > >> > @@ -162,6 +162,12 @@ config PAGE_OFFSET
>> > >> > default 0xffffffff80000000 if 64BIT && MAXPHYSMEM_2GB
>> > >> > default 0xffffffe000000000 if 64BIT && MAXPHYSMEM_128GB
>> > >> >
>> > >> > +config KASAN_SHADOW_OFFSET
>> > >> > + hex
>> > >> > + depends on KASAN_GENERIC
>> > >> > + default 0xdfffffc800000000 if 64BIT
>> > >> > + default 0xffffffff if 32BIT
>> > >>
>> > >> I thought I posted this somewhere, but this is exactly what my first
>> > >> guess was. The problem is that it's hanging on boot for me. I don't
>> > >> really have anything exotic going on, it's just a defconfig with
>> > >> CONFIG_KASAN=y running in QEMU.
>> > >>
>> > >> Does this boot for you?
>> > >
>> > > Yes with the 2nd patch of this series which fixes the issue
>> > > encountered here. And that's true I copied/pasted this part of your
>> > > patch which was better than what I had initially done, sorry I should
>> > > have mentioned you did that, please add a Codeveloped-by or something
>> > > like that.
OK, those should probably be in the opposite order (though it looks like
they're inter-dependent, which makes things a bit trickier).
>> >
>> > Not sure if I'm missing something, but it's still not booting for me.
>> > I've put what I'm testing on palmer/to-test, it's these two on top of
>> > fixes and merged into Linus' tree
>> >
>> > * 6d7d351902ff - (HEAD -> to-test, palmer/to-test) Merge remote-tracking branch 'palmer/fixes' into to-test (7 minutes ago) <Palmer Dabbelt>
>> > |\
>> > | * 782551edf8f8 - (palmer/fixes) riscv: Fix CONFIG_KASAN_STACK build (6 hours ago) <Alexandre Ghiti>
>> > | * 47383e5b3c4f - riscv: Fix asan-stack clang build (6 hours ago) <Alexandre Ghiti>
>> > | * 64a19591a293 - (riscv/fixes) riscv: fix misalgned trap vector base address (9 hours ago) <Chen Lu>
>> > * | 1fc596a56b33 - (palmer/master, linus/master, linus/HEAD, master) Merge tag 'trace-v5.15-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace (11 hours ago) <Linus Torvalds>
>> >
>> > Am I missing something else?
>>
>> Hmm, that's weird, I have just done the same: cherry-picked both my
>> commits on top of fixes (64a19591a293) and it boots fine with KASAN
>> enabled. Maybe a config thing? I pushed my branch here:
>> https://github.com/AlexGhiti/riscv-linux/tree/int/alex/kasan_stack_fixes_rebase
>
> I pushed the config I use and that boots in that branch, maybe there's
> another issue somewhere.
CONFIG_KASAN_VMALLOC=n is what's causing the failure. I'm testing both
polarities of that, looks like your config has =y. I haven't looked any
further as I'm pretty much cooked for tonight, but if you don't have
time then I'll try to find some time tomorrow.
>
>>
>> >
>> > >
>> > > Thanks,
>> > >
>> > > Alex
>> > >
>> > >>
>> > >> > +
>> > >> > config ARCH_FLATMEM_ENABLE
>> > >> > def_bool !NUMA
>> > >> >
>> > >> > diff --git a/arch/riscv/include/asm/kasan.h b/arch/riscv/include/asm/kasan.h
>> > >> > index a2b3d9cdbc86..b00f503ec124 100644
>> > >> > --- a/arch/riscv/include/asm/kasan.h
>> > >> > +++ b/arch/riscv/include/asm/kasan.h
>> > >> > @@ -30,8 +30,7 @@
>> > >> > #define KASAN_SHADOW_SIZE (UL(1) << ((CONFIG_VA_BITS - 1) - KASAN_SHADOW_SCALE_SHIFT))
>> > >> > #define KASAN_SHADOW_START KERN_VIRT_START
>> > >> > #define KASAN_SHADOW_END (KASAN_SHADOW_START + KASAN_SHADOW_SIZE)
>> > >> > -#define KASAN_SHADOW_OFFSET (KASAN_SHADOW_END - (1ULL << \
>> > >> > - (64 - KASAN_SHADOW_SCALE_SHIFT)))
>> > >> > +#define KASAN_SHADOW_OFFSET _AC(CONFIG_KASAN_SHADOW_OFFSET, UL)
>> > >> >
>> > >> > void kasan_init(void);
>> > >> > asmlinkage void kasan_early_init(void);
>> > >> > diff --git a/arch/riscv/mm/kasan_init.c b/arch/riscv/mm/kasan_init.c
>> > >> > index d7189c8714a9..8175e98b9073 100644
>> > >> > --- a/arch/riscv/mm/kasan_init.c
>> > >> > +++ b/arch/riscv/mm/kasan_init.c
>> > >> > @@ -17,6 +17,9 @@ asmlinkage void __init kasan_early_init(void)
>> > >> > uintptr_t i;
>> > >> > pgd_t *pgd = early_pg_dir + pgd_index(KASAN_SHADOW_START);
>> > >> >
>> > >> > + BUILD_BUG_ON(KASAN_SHADOW_OFFSET !=
>> > >> > + KASAN_SHADOW_END - (1UL << (64 - KASAN_SHADOW_SCALE_SHIFT)));
>> > >> > +
>> > >> > for (i = 0; i < PTRS_PER_PTE; ++i)
>> > >> > set_pte(kasan_early_shadow_pte + i,
>> > >> > mk_pte(virt_to_page(kasan_early_shadow_page),
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH 1/2] riscv: Fix asan-stack clang build
2021-10-28 6:45 ` Palmer Dabbelt
@ 2021-10-28 7:13 ` Alexandre Ghiti
2021-10-28 15:11 ` Palmer Dabbelt
0 siblings, 1 reply; 10+ messages in thread
From: Alexandre Ghiti @ 2021-10-28 7:13 UTC (permalink / raw)
To: Palmer Dabbelt
Cc: Paul Walmsley, aou, ryabinin.a.a, glider, andreyknvl, dvyukov,
linux-riscv, linux-kernel, kasan-dev, nathan
On Thu, Oct 28, 2021 at 8:45 AM Palmer Dabbelt <palmer@dabbelt.com> wrote:
>
> On Wed, 27 Oct 2021 22:34:32 PDT (-0700), alexandre.ghiti@canonical.com wrote:
> > On Thu, Oct 28, 2021 at 7:30 AM Alexandre Ghiti
> > <alexandre.ghiti@canonical.com> wrote:
> >>
> >> On Thu, Oct 28, 2021 at 7:02 AM Palmer Dabbelt <palmer@dabbelt.com> wrote:
> >> >
> >> > On Wed, 27 Oct 2021 21:15:28 PDT (-0700), alexandre.ghiti@canonical.com wrote:
> >> > > On Thu, Oct 28, 2021 at 1:06 AM Palmer Dabbelt <palmer@dabbelt.com> wrote:
> >> > >>
> >> > >> On Tue, 26 Oct 2021 21:58:42 PDT (-0700), alexandre.ghiti@canonical.com wrote:
> >> > >> > Nathan reported that because KASAN_SHADOW_OFFSET was not defined in
> >> > >> > Kconfig, it prevents asan-stack from getting disabled with clang even
> >> > >> > when CONFIG_KASAN_STACK is disabled: fix this by defining the
> >> > >> > corresponding config.
> >> > >> >
> >> > >> > Reported-by: Nathan Chancellor <nathan@kernel.org>
> >> > >> > Signed-off-by: Alexandre Ghiti <alexandre.ghiti@canonical.com>
> >> > >> > ---
> >> > >> > arch/riscv/Kconfig | 6 ++++++
> >> > >> > arch/riscv/include/asm/kasan.h | 3 +--
> >> > >> > arch/riscv/mm/kasan_init.c | 3 +++
> >> > >> > 3 files changed, 10 insertions(+), 2 deletions(-)
> >> > >> >
> >> > >> > diff --git a/arch/riscv/Kconfig b/arch/riscv/Kconfig
> >> > >> > index c1abbc876e5b..79250b1ed54e 100644
> >> > >> > --- a/arch/riscv/Kconfig
> >> > >> > +++ b/arch/riscv/Kconfig
> >> > >> > @@ -162,6 +162,12 @@ config PAGE_OFFSET
> >> > >> > default 0xffffffff80000000 if 64BIT && MAXPHYSMEM_2GB
> >> > >> > default 0xffffffe000000000 if 64BIT && MAXPHYSMEM_128GB
> >> > >> >
> >> > >> > +config KASAN_SHADOW_OFFSET
> >> > >> > + hex
> >> > >> > + depends on KASAN_GENERIC
> >> > >> > + default 0xdfffffc800000000 if 64BIT
> >> > >> > + default 0xffffffff if 32BIT
> >> > >>
> >> > >> I thought I posted this somewhere, but this is exactly what my first
> >> > >> guess was. The problem is that it's hanging on boot for me. I don't
> >> > >> really have anything exotic going on, it's just a defconfig with
> >> > >> CONFIG_KASAN=y running in QEMU.
> >> > >>
> >> > >> Does this boot for you?
> >> > >
> >> > > Yes with the 2nd patch of this series which fixes the issue
> >> > > encountered here. And that's true I copied/pasted this part of your
> >> > > patch which was better than what I had initially done, sorry I should
> >> > > have mentioned you did that, please add a Codeveloped-by or something
> >> > > like that.
>
> OK, those should probably be in the opposite order (though it looks like
> they're inter-dependent, which makes things a bit trickier).
>
> >> >
> >> > Not sure if I'm missing something, but it's still not booting for me.
> >> > I've put what I'm testing on palmer/to-test, it's these two on top of
> >> > fixes and merged into Linus' tree
> >> >
> >> > * 6d7d351902ff - (HEAD -> to-test, palmer/to-test) Merge remote-tracking branch 'palmer/fixes' into to-test (7 minutes ago) <Palmer Dabbelt>
> >> > |\
> >> > | * 782551edf8f8 - (palmer/fixes) riscv: Fix CONFIG_KASAN_STACK build (6 hours ago) <Alexandre Ghiti>
> >> > | * 47383e5b3c4f - riscv: Fix asan-stack clang build (6 hours ago) <Alexandre Ghiti>
> >> > | * 64a19591a293 - (riscv/fixes) riscv: fix misalgned trap vector base address (9 hours ago) <Chen Lu>
> >> > * | 1fc596a56b33 - (palmer/master, linus/master, linus/HEAD, master) Merge tag 'trace-v5.15-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace (11 hours ago) <Linus Torvalds>
> >> >
> >> > Am I missing something else?
> >>
> >> Hmm, that's weird, I have just done the same: cherry-picked both my
> >> commits on top of fixes (64a19591a293) and it boots fine with KASAN
> >> enabled. Maybe a config thing? I pushed my branch here:
> >> https://github.com/AlexGhiti/riscv-linux/tree/int/alex/kasan_stack_fixes_rebase
> >
> > I pushed the config I use and that boots in that branch, maybe there's
> > another issue somewhere.
>
> CONFIG_KASAN_VMALLOC=n is what's causing the failure. I'm testing both
> polarities of that, looks like your config has =y. I haven't looked any
> further as I'm pretty much cooked for tonight, but if you don't have
> time then I'll try to find some time tomorrow.
>
Arf, that was obvious and just under my nose: without KASAN_VMALLOC,
kasan_populate_early_shadow is called and creates the same issue that
the second patch fixes.
I'll send a v2 today and try to swap both patches to avoid having a
non-bootable kernel commit.
Alex
> >
> >>
> >> >
> >> > >
> >> > > Thanks,
> >> > >
> >> > > Alex
> >> > >
> >> > >>
> >> > >> > +
> >> > >> > config ARCH_FLATMEM_ENABLE
> >> > >> > def_bool !NUMA
> >> > >> >
> >> > >> > diff --git a/arch/riscv/include/asm/kasan.h b/arch/riscv/include/asm/kasan.h
> >> > >> > index a2b3d9cdbc86..b00f503ec124 100644
> >> > >> > --- a/arch/riscv/include/asm/kasan.h
> >> > >> > +++ b/arch/riscv/include/asm/kasan.h
> >> > >> > @@ -30,8 +30,7 @@
> >> > >> > #define KASAN_SHADOW_SIZE (UL(1) << ((CONFIG_VA_BITS - 1) - KASAN_SHADOW_SCALE_SHIFT))
> >> > >> > #define KASAN_SHADOW_START KERN_VIRT_START
> >> > >> > #define KASAN_SHADOW_END (KASAN_SHADOW_START + KASAN_SHADOW_SIZE)
> >> > >> > -#define KASAN_SHADOW_OFFSET (KASAN_SHADOW_END - (1ULL << \
> >> > >> > - (64 - KASAN_SHADOW_SCALE_SHIFT)))
> >> > >> > +#define KASAN_SHADOW_OFFSET _AC(CONFIG_KASAN_SHADOW_OFFSET, UL)
> >> > >> >
> >> > >> > void kasan_init(void);
> >> > >> > asmlinkage void kasan_early_init(void);
> >> > >> > diff --git a/arch/riscv/mm/kasan_init.c b/arch/riscv/mm/kasan_init.c
> >> > >> > index d7189c8714a9..8175e98b9073 100644
> >> > >> > --- a/arch/riscv/mm/kasan_init.c
> >> > >> > +++ b/arch/riscv/mm/kasan_init.c
> >> > >> > @@ -17,6 +17,9 @@ asmlinkage void __init kasan_early_init(void)
> >> > >> > uintptr_t i;
> >> > >> > pgd_t *pgd = early_pg_dir + pgd_index(KASAN_SHADOW_START);
> >> > >> >
> >> > >> > + BUILD_BUG_ON(KASAN_SHADOW_OFFSET !=
> >> > >> > + KASAN_SHADOW_END - (1UL << (64 - KASAN_SHADOW_SCALE_SHIFT)));
> >> > >> > +
> >> > >> > for (i = 0; i < PTRS_PER_PTE; ++i)
> >> > >> > set_pte(kasan_early_shadow_pte + i,
> >> > >> > mk_pte(virt_to_page(kasan_early_shadow_page),
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH 1/2] riscv: Fix asan-stack clang build
2021-10-28 7:13 ` Alexandre Ghiti
@ 2021-10-28 15:11 ` Palmer Dabbelt
0 siblings, 0 replies; 10+ messages in thread
From: Palmer Dabbelt @ 2021-10-28 15:11 UTC (permalink / raw)
To: alexandre.ghiti
Cc: Paul Walmsley, aou, ryabinin.a.a, glider, andreyknvl, dvyukov,
linux-riscv, linux-kernel, kasan-dev, nathan
On Thu, 28 Oct 2021 00:13:06 PDT (-0700), alexandre.ghiti@canonical.com wrote:
> On Thu, Oct 28, 2021 at 8:45 AM Palmer Dabbelt <palmer@dabbelt.com> wrote:
>>
>> On Wed, 27 Oct 2021 22:34:32 PDT (-0700), alexandre.ghiti@canonical.com wrote:
>> > On Thu, Oct 28, 2021 at 7:30 AM Alexandre Ghiti
>> > <alexandre.ghiti@canonical.com> wrote:
>> >>
>> >> On Thu, Oct 28, 2021 at 7:02 AM Palmer Dabbelt <palmer@dabbelt.com> wrote:
>> >> >
>> >> > On Wed, 27 Oct 2021 21:15:28 PDT (-0700), alexandre.ghiti@canonical.com wrote:
>> >> > > On Thu, Oct 28, 2021 at 1:06 AM Palmer Dabbelt <palmer@dabbelt.com> wrote:
>> >> > >>
>> >> > >> On Tue, 26 Oct 2021 21:58:42 PDT (-0700), alexandre.ghiti@canonical.com wrote:
>> >> > >> > Nathan reported that because KASAN_SHADOW_OFFSET was not defined in
>> >> > >> > Kconfig, it prevents asan-stack from getting disabled with clang even
>> >> > >> > when CONFIG_KASAN_STACK is disabled: fix this by defining the
>> >> > >> > corresponding config.
>> >> > >> >
>> >> > >> > Reported-by: Nathan Chancellor <nathan@kernel.org>
>> >> > >> > Signed-off-by: Alexandre Ghiti <alexandre.ghiti@canonical.com>
>> >> > >> > ---
>> >> > >> > arch/riscv/Kconfig | 6 ++++++
>> >> > >> > arch/riscv/include/asm/kasan.h | 3 +--
>> >> > >> > arch/riscv/mm/kasan_init.c | 3 +++
>> >> > >> > 3 files changed, 10 insertions(+), 2 deletions(-)
>> >> > >> >
>> >> > >> > diff --git a/arch/riscv/Kconfig b/arch/riscv/Kconfig
>> >> > >> > index c1abbc876e5b..79250b1ed54e 100644
>> >> > >> > --- a/arch/riscv/Kconfig
>> >> > >> > +++ b/arch/riscv/Kconfig
>> >> > >> > @@ -162,6 +162,12 @@ config PAGE_OFFSET
>> >> > >> > default 0xffffffff80000000 if 64BIT && MAXPHYSMEM_2GB
>> >> > >> > default 0xffffffe000000000 if 64BIT && MAXPHYSMEM_128GB
>> >> > >> >
>> >> > >> > +config KASAN_SHADOW_OFFSET
>> >> > >> > + hex
>> >> > >> > + depends on KASAN_GENERIC
>> >> > >> > + default 0xdfffffc800000000 if 64BIT
>> >> > >> > + default 0xffffffff if 32BIT
>> >> > >>
>> >> > >> I thought I posted this somewhere, but this is exactly what my first
>> >> > >> guess was. The problem is that it's hanging on boot for me. I don't
>> >> > >> really have anything exotic going on, it's just a defconfig with
>> >> > >> CONFIG_KASAN=y running in QEMU.
>> >> > >>
>> >> > >> Does this boot for you?
>> >> > >
>> >> > > Yes with the 2nd patch of this series which fixes the issue
>> >> > > encountered here. And that's true I copied/pasted this part of your
>> >> > > patch which was better than what I had initially done, sorry I should
>> >> > > have mentioned you did that, please add a Codeveloped-by or something
>> >> > > like that.
>>
>> OK, those should probably be in the opposite order (though it looks like
>> they're inter-dependent, which makes things a bit trickier).
>>
>> >> >
>> >> > Not sure if I'm missing something, but it's still not booting for me.
>> >> > I've put what I'm testing on palmer/to-test, it's these two on top of
>> >> > fixes and merged into Linus' tree
>> >> >
>> >> > * 6d7d351902ff - (HEAD -> to-test, palmer/to-test) Merge remote-tracking branch 'palmer/fixes' into to-test (7 minutes ago) <Palmer Dabbelt>
>> >> > |\
>> >> > | * 782551edf8f8 - (palmer/fixes) riscv: Fix CONFIG_KASAN_STACK build (6 hours ago) <Alexandre Ghiti>
>> >> > | * 47383e5b3c4f - riscv: Fix asan-stack clang build (6 hours ago) <Alexandre Ghiti>
>> >> > | * 64a19591a293 - (riscv/fixes) riscv: fix misalgned trap vector base address (9 hours ago) <Chen Lu>
>> >> > * | 1fc596a56b33 - (palmer/master, linus/master, linus/HEAD, master) Merge tag 'trace-v5.15-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace (11 hours ago) <Linus Torvalds>
>> >> >
>> >> > Am I missing something else?
>> >>
>> >> Hmm, that's weird, I have just done the same: cherry-picked both my
>> >> commits on top of fixes (64a19591a293) and it boots fine with KASAN
>> >> enabled. Maybe a config thing? I pushed my branch here:
>> >> https://github.com/AlexGhiti/riscv-linux/tree/int/alex/kasan_stack_fixes_rebase
>> >
>> > I pushed the config I use and that boots in that branch, maybe there's
>> > another issue somewhere.
>>
>> CONFIG_KASAN_VMALLOC=n is what's causing the failure. I'm testing both
>> polarities of that, looks like your config has =y. I haven't looked any
>> further as I'm pretty much cooked for tonight, but if you don't have
>> time then I'll try to find some time tomorrow.
>>
>
> Arf, that was obvious and just under my nose: without KASAN_VMALLOC,
> kasan_populate_early_shadow is called and creates the same issue that
> the second patch fixes.
>
> I'll send a v2 today and try to swap both patches to avoid having a
> non-bootable kernel commit.
Thanks.
>
> Alex
>
>> >
>> >>
>> >> >
>> >> > >
>> >> > > Thanks,
>> >> > >
>> >> > > Alex
>> >> > >
>> >> > >>
>> >> > >> > +
>> >> > >> > config ARCH_FLATMEM_ENABLE
>> >> > >> > def_bool !NUMA
>> >> > >> >
>> >> > >> > diff --git a/arch/riscv/include/asm/kasan.h b/arch/riscv/include/asm/kasan.h
>> >> > >> > index a2b3d9cdbc86..b00f503ec124 100644
>> >> > >> > --- a/arch/riscv/include/asm/kasan.h
>> >> > >> > +++ b/arch/riscv/include/asm/kasan.h
>> >> > >> > @@ -30,8 +30,7 @@
>> >> > >> > #define KASAN_SHADOW_SIZE (UL(1) << ((CONFIG_VA_BITS - 1) - KASAN_SHADOW_SCALE_SHIFT))
>> >> > >> > #define KASAN_SHADOW_START KERN_VIRT_START
>> >> > >> > #define KASAN_SHADOW_END (KASAN_SHADOW_START + KASAN_SHADOW_SIZE)
>> >> > >> > -#define KASAN_SHADOW_OFFSET (KASAN_SHADOW_END - (1ULL << \
>> >> > >> > - (64 - KASAN_SHADOW_SCALE_SHIFT)))
>> >> > >> > +#define KASAN_SHADOW_OFFSET _AC(CONFIG_KASAN_SHADOW_OFFSET, UL)
>> >> > >> >
>> >> > >> > void kasan_init(void);
>> >> > >> > asmlinkage void kasan_early_init(void);
>> >> > >> > diff --git a/arch/riscv/mm/kasan_init.c b/arch/riscv/mm/kasan_init.c
>> >> > >> > index d7189c8714a9..8175e98b9073 100644
>> >> > >> > --- a/arch/riscv/mm/kasan_init.c
>> >> > >> > +++ b/arch/riscv/mm/kasan_init.c
>> >> > >> > @@ -17,6 +17,9 @@ asmlinkage void __init kasan_early_init(void)
>> >> > >> > uintptr_t i;
>> >> > >> > pgd_t *pgd = early_pg_dir + pgd_index(KASAN_SHADOW_START);
>> >> > >> >
>> >> > >> > + BUILD_BUG_ON(KASAN_SHADOW_OFFSET !=
>> >> > >> > + KASAN_SHADOW_END - (1UL << (64 - KASAN_SHADOW_SCALE_SHIFT)));
>> >> > >> > +
>> >> > >> > for (i = 0; i < PTRS_PER_PTE; ++i)
>> >> > >> > set_pte(kasan_early_shadow_pte + i,
>> >> > >> > mk_pte(virt_to_page(kasan_early_shadow_page),
^ permalink raw reply [flat|nested] 10+ messages in thread
end of thread, other threads:[~2021-10-28 15:11 UTC | newest]
Thread overview: 10+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-10-27 4:58 [PATCH 1/2] riscv: Fix asan-stack clang build Alexandre Ghiti
2021-10-27 4:58 ` [PATCH 2/2] riscv: Fix CONFIG_KASAN_STACK build Alexandre Ghiti
2021-10-27 23:06 ` [PATCH 1/2] riscv: Fix asan-stack clang build Palmer Dabbelt
2021-10-28 4:15 ` Alexandre Ghiti
2021-10-28 5:02 ` Palmer Dabbelt
2021-10-28 5:30 ` Alexandre Ghiti
2021-10-28 5:34 ` Alexandre Ghiti
2021-10-28 6:45 ` Palmer Dabbelt
2021-10-28 7:13 ` Alexandre Ghiti
2021-10-28 15:11 ` Palmer Dabbelt
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).