[PATCH] arch/x86/um: Disable UBSAN sanitization

[PATCH] arch/x86/um: Disable UBSAN sanitization

[Roberto Sassu]

From: Roberto Sassu <roberto.sassu@huawei.com>

Disable UBSAN sanitization on UML, since UML does not support it.

This fixes the error message when building the kernel:

  CALL    scripts/checksyscalls.sh
  VDSO    arch/x86/um/vdso/vdso.so.dbg
arch/x86/um/vdso/vdso.so.dbg: undefined symbols found

Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
---
 arch/x86/um/vdso/Makefile | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/arch/x86/um/vdso/Makefile b/arch/x86/um/vdso/Makefile
index b86d634730b2..ca79c0de582e 100644
--- a/arch/x86/um/vdso/Makefile
+++ b/arch/x86/um/vdso/Makefile
@@ -3,8 +3,10 @@
 # Building vDSO images for x86.
 #
 
-# do not instrument on vdso because KASAN is not compatible with user mode
+# do not instrument on vdso because KASAN/UBSAN are not compatible with user
+# mode
 KASAN_SANITIZE			:= n
+UBSAN_SANITIZE			:= n
 
 # Prevents link failures: __sanitizer_cov_trace_pc() is not linked in.
 KCOV_INSTRUMENT                := n
-- 
2.34.1

Re: [PATCH] arch/x86/um: Disable UBSAN sanitization

[Johannes Berg]

On Mon, 2024-05-13 at 14:27 +0200, Roberto Sassu wrote:
> From: Roberto Sassu <roberto.sassu@huawei.com>
> 
> Disable UBSAN sanitization on UML, since UML does not support it.
> 

Luckily, that isn't actually true, nor does it actually do this at all.
Please fix the commit message.

johannes

Re: [PATCH] arch/x86/um: Disable UBSAN sanitization

[Roberto Sassu]

On Mon, 2024-05-13 at 14:29 +0200, Johannes Berg wrote:
> On Mon, 2024-05-13 at 14:27 +0200, Roberto Sassu wrote:
> > From: Roberto Sassu <roberto.sassu@huawei.com>
> > 
> > Disable UBSAN sanitization on UML, since UML does not support it.
> > 
> 
> Luckily, that isn't actually true, nor does it actually do this at all.
> Please fix the commit message.

Thanks, I was actually wondering. I based that statement based on
ARCH_HAS_UBSAN=n.

Any other solution would be ok.

Thanks

Roberto

Re: [PATCH] arch/x86/um: Disable UBSAN sanitization

[Johannes Berg]

On Mon, 2024-05-13 at 14:42 +0200, Roberto Sassu wrote:
> On Mon, 2024-05-13 at 14:29 +0200, Johannes Berg wrote:
> > On Mon, 2024-05-13 at 14:27 +0200, Roberto Sassu wrote:
> > > From: Roberto Sassu <roberto.sassu@huawei.com>
> > > 
> > > Disable UBSAN sanitization on UML, since UML does not support it.
> > > 
> > 
> > Luckily, that isn't actually true, nor does it actually do this at all.
> > Please fix the commit message.
> 
> Thanks, I was actually wondering. I based that statement based on
> ARCH_HAS_UBSAN=n.
> 
> Any other solution would be ok.

Not sure I get it. What you're doing in the patch is perfectly fine and
almost certainly required, but you're definitely not disabling UBSAN on
ARCH=um as you described in the commit message?

johannes

Re: [PATCH] arch/x86/um: Disable UBSAN sanitization

[Roberto Sassu]

On Mon, 2024-05-13 at 14:52 +0200, Johannes Berg wrote:
> On Mon, 2024-05-13 at 14:42 +0200, Roberto Sassu wrote:
> > On Mon, 2024-05-13 at 14:29 +0200, Johannes Berg wrote:
> > > On Mon, 2024-05-13 at 14:27 +0200, Roberto Sassu wrote:
> > > > From: Roberto Sassu <roberto.sassu@huawei.com>
> > > > 
> > > > Disable UBSAN sanitization on UML, since UML does not support it.
> > > > 
> > > 
> > > Luckily, that isn't actually true, nor does it actually do this at all.
> > > Please fix the commit message.
> > 
> > Thanks, I was actually wondering. I based that statement based on
> > ARCH_HAS_UBSAN=n.
> > 
> > Any other solution would be ok.
> 
> Not sure I get it. What you're doing in the patch is perfectly fine and
> almost certainly required, but you're definitely not disabling UBSAN on
> ARCH=um as you described in the commit message?

Ok, I guess the right word is instrumentation (got it from commit
d4be85d068b44). And the reason is that the vDSO is executing in user
space. Will fix it.

Thanks

Roberto

Re: [PATCH] arch/x86/um: Disable UBSAN sanitization

[Johannes Berg]

On Mon, 2024-05-13 at 14:58 +0200, Roberto Sassu wrote:
> On Mon, 2024-05-13 at 14:52 +0200, Johannes Berg wrote:
> > On Mon, 2024-05-13 at 14:42 +0200, Roberto Sassu wrote:
> > > On Mon, 2024-05-13 at 14:29 +0200, Johannes Berg wrote:
> > > > On Mon, 2024-05-13 at 14:27 +0200, Roberto Sassu wrote:
> > > > > From: Roberto Sassu <roberto.sassu@huawei.com>
> > > > > 
> > > > > Disable UBSAN sanitization on UML, since UML does not support it.
> > > > > 
> > > > 
> > > > Luckily, that isn't actually true, nor does it actually do this at all.
> > > > Please fix the commit message.
> > > 
> > > Thanks, I was actually wondering. I based that statement based on
> > > ARCH_HAS_UBSAN=n.
> > > 
> > > Any other solution would be ok.
> > 
> > Not sure I get it. What you're doing in the patch is perfectly fine and
> > almost certainly required, but you're definitely not disabling UBSAN on
> > ARCH=um as you described in the commit message?
> 
> Ok, I guess the right word is instrumentation (got it from commit
> d4be85d068b44). And the reason is that the vDSO is executing in user
> space. Will fix it.

No, UBSAN is fine, but you're only disabling it for the vDSO :) The
commit message doesn't even mention the vDSO though.

johannes

Re: [PATCH] arch/x86/um: Disable UBSAN sanitization

[Roberto Sassu]

On Mon, 2024-05-13 at 15:08 +0200, Johannes Berg wrote:
> On Mon, 2024-05-13 at 14:58 +0200, Roberto Sassu wrote:
> > On Mon, 2024-05-13 at 14:52 +0200, Johannes Berg wrote:
> > > On Mon, 2024-05-13 at 14:42 +0200, Roberto Sassu wrote:
> > > > On Mon, 2024-05-13 at 14:29 +0200, Johannes Berg wrote:
> > > > > On Mon, 2024-05-13 at 14:27 +0200, Roberto Sassu wrote:
> > > > > > From: Roberto Sassu <roberto.sassu@huawei.com>
> > > > > > 
> > > > > > Disable UBSAN sanitization on UML, since UML does not support it.
> > > > > > 
> > > > > 
> > > > > Luckily, that isn't actually true, nor does it actually do this at all.
> > > > > Please fix the commit message.
> > > > 
> > > > Thanks, I was actually wondering. I based that statement based on
> > > > ARCH_HAS_UBSAN=n.
> > > > 
> > > > Any other solution would be ok.
> > > 
> > > Not sure I get it. What you're doing in the patch is perfectly fine and
> > > almost certainly required, but you're definitely not disabling UBSAN on
> > > ARCH=um as you described in the commit message?
> > 
> > Ok, I guess the right word is instrumentation (got it from commit
> > d4be85d068b44). And the reason is that the vDSO is executing in user
> > space. Will fix it.
> 
> No, UBSAN is fine, but you're only disabling it for the vDSO :) The
> commit message doesn't even mention the vDSO though.

You are right, the commit message was misleading without vDSO.

Thanks

Roberto