* new ubsan handler
@ 2022-12-17 19:35 Martin Uecker
2022-12-19 8:44 ` Martin Liška
0 siblings, 1 reply; 5+ messages in thread
From: Martin Uecker @ 2022-12-17 19:35 UTC (permalink / raw)
To: gcc
Hi all,
what is the process for adding a new UBsan handler?
We have the source in the GCC tree, but I guess this
should go via LLVM?
Martin
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: new ubsan handler
2022-12-17 19:35 new ubsan handler Martin Uecker
@ 2022-12-19 8:44 ` Martin Liška
2022-12-19 9:25 ` Martin Uecker
0 siblings, 1 reply; 5+ messages in thread
From: Martin Liška @ 2022-12-19 8:44 UTC (permalink / raw)
To: Martin Uecker, gcc
On 12/17/22 20:35, Martin Uecker wrote:
>
>
> Hi all,
>
> what is the process for adding a new UBsan handler?
Hello.
Yes, we sync the run-time library from LLVM project. So a new sanitizer
should go there first.
>
> We have the source in the GCC tree, but I guess this
> should go via LLVM?
And the compiler support (instrumentation) is independent.
Cheers,
Martin
>
>
> Martin
>
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: new ubsan handler
2022-12-19 8:44 ` Martin Liška
@ 2022-12-19 9:25 ` Martin Uecker
2022-12-19 14:22 ` Martin Liška
0 siblings, 1 reply; 5+ messages in thread
From: Martin Uecker @ 2022-12-19 9:25 UTC (permalink / raw)
To: Martin Liška, gcc
Am Montag, dem 19.12.2022 um 09:44 +0100 schrieb Martin Liška:
> On 12/17/22 20:35, Martin Uecker wrote:
> >
> >
> > Hi all,
> >
> > what is the process for adding a new UBsan handler?
>
> Hello.
>
> Yes, we sync the run-time library from LLVM project. So a new
> sanitizer
> should go there first.
>
> >
> > We have the source in the GCC tree, but I guess this
> > should go via LLVM?
>
> And the compiler support (instrumentation) is independent.
Thanks Martin!
Does it make sense to send the full patch (with changes
to the compiler and library) to gcc-patches first for
review?
I am working on a patch that calls a handler when bounds
for VM-types do not match on assignment.
Martin
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: new ubsan handler
2022-12-19 9:25 ` Martin Uecker
@ 2022-12-19 14:22 ` Martin Liška
2022-12-19 15:43 ` Martin Uecker
0 siblings, 1 reply; 5+ messages in thread
From: Martin Liška @ 2022-12-19 14:22 UTC (permalink / raw)
To: Martin Uecker, gcc
On 12/19/22 10:25, Martin Uecker wrote:
> Am Montag, dem 19.12.2022 um 09:44 +0100 schrieb Martin Liška:
>> On 12/17/22 20:35, Martin Uecker wrote:
>>>
>>>
>>> Hi all,
>>>
>>> what is the process for adding a new UBsan handler?
>>
>> Hello.
>>
>> Yes, we sync the run-time library from LLVM project. So a new
>> sanitizer
>> should go there first.
>>
>>>
>>> We have the source in the GCC tree, but I guess this
>>> should go via LLVM?
>>
>> And the compiler support (instrumentation) is independent.
>
> Thanks Martin!
>
> Does it make sense to send the full patch (with changes
> to the compiler and library) to gcc-patches first for
> review?
Yep, please CC me.
>
> I am working on a patch that calls a handler when bounds
> for VM-types do not match on assignment.
Nice. So basically something like -fsanitize=bounds/bounds-strict/object-size ?
Martin
>
>
> Martin
>
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: new ubsan handler
2022-12-19 14:22 ` Martin Liška
@ 2022-12-19 15:43 ` Martin Uecker
0 siblings, 0 replies; 5+ messages in thread
From: Martin Uecker @ 2022-12-19 15:43 UTC (permalink / raw)
To: Martin Liška, gcc
Am Montag, dem 19.12.2022 um 15:22 +0100 schrieb Martin Liška:
> On 12/19/22 10:25, Martin Uecker wrote:
> > Am Montag, dem 19.12.2022 um 09:44 +0100 schrieb Martin Liška:
> > > On 12/17/22 20:35, Martin Uecker wrote:
> > > >
> > > >
> > > > Hi all,
> > > >
> > > > what is the process for adding a new UBsan handler?
> > >
> > > Hello.
> > >
> > > Yes, we sync the run-time library from LLVM project. So a new
> > > sanitizer
> > > should go there first.
> > >
> > > >
> > > > We have the source in the GCC tree, but I guess this
> > > > should go via LLVM?
> > >
> > > And the compiler support (instrumentation) is independent.
> >
> > Thanks Martin!
> >
> > Does it make sense to send the full patch (with changes
> > to the compiler and library) to gcc-patches first for
> > review?
>
> Yep, please CC me.
>
> >
> > I am working on a patch that calls a handler when bounds
> > for VM-types do not match on assignment.
>
> Nice. So basically something like -fsanitize=bounds/bounds-
> strict/object-size ?
Yes, it would diagnose the following mismatch
and other similar cases:
int a[10];
int n = 5;
char (*p)[n] = &a;
Martin
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2022-12-19 15:43 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-12-17 19:35 new ubsan handler Martin Uecker
2022-12-19 8:44 ` Martin Liška
2022-12-19 9:25 ` Martin Uecker
2022-12-19 14:22 ` Martin Liška
2022-12-19 15:43 ` Martin Uecker
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 read-only IMAP folder(s) and NNTP newsgroup(s).