[cocci] Spatch 1.1.1 backward incompatibility

[cocci] Spatch 1.1.1 backward incompatibility

[Christoph Böhmwalder]

Hi,

I recently(?) got a new version of coccinelle from Fedora:


$ spatch --version
spatch version 1.1.1 compiled with OCaml version 5.0.1+dev0-2022-12-15
Flags passed to the configure script: --build=x86_64-redhat-linux-gnu
--host=x86_64-redhat-linux-gnu --program-prefix=
--disable-dependency-tracking --prefix=/usr --exec-prefix=/usr
--bindir=/usr/bin --sbindir=/usr/sbin --sysconfdir=/etc
--datadir=/usr/share --includedir=/usr/include --libdir=/usr/lib64
--libexecdir=/usr/libexec --localstatedir=/var --runstatedir=/run
--sharedstatedir=/var/lib --mandir=/usr/share/man
--infodir=/usr/share/info --with-python=/usr/bin/python3
--with-menhir=/usr/bin/menhir
OCaml scripting support: yes
Python scripting support: yes
Syntax of regular expressions: PCRE


And it broke a lot of our semantic patches. I now get lots of these
warnings:

warning: rule starting on line 248, node 45:
kunmap_local(...)[1,2,20,21,25] in recv_dless_read may be inconsistently
modified
warning: rule starting on line 248, node 42: decl:mapped[1,2,20,21,22]
in recv_dless_read may be inconsistently modified
warning: rule starting on line 273, node 18: src = ... ;[1,2,8,9,11] in
drbd_csum_bio may be inconsistently modified
warning: rule starting on line 273, node 20:
kunmap_local(...)[1,2,8,9,13] in drbd_csum_bio may be inconsistently
modified

(By the way: these are reported on stdout, I'm not sure that's intended
either)

And the patch output is noticeably different.

This all works fine on 1.1.0, and curiously it also works on the 1.1.1
version in Ubuntu:

$ spatch --version
spatch version 1.1.1 compiled with OCaml version 4.13.1
Flags passed to the configure script: --prefix=/usr --sysconfdir=/etc
--libdir=/usr/lib --enable-ocaml --enable-python --enable-opt
OCaml scripting support: yes
Python scripting support: yes
Syntax of regular expressions: PCRE

I'm trying to come up with a minimal reproducer right now, but just
throwing this out there in case anyone has an idea.

Thanks!

-- 
Christoph Böhmwalder
LINBIT | Keeping the Digital World Running
DRBD HA —  Disaster Recovery — Software defined Storage

Re: [cocci] Spatch 1.1.1 backward incompatibility

[Julia Lawall]

[-- Attachment #1: Type: text/plain, Size: 2238 bytes --]



On Fri, 16 Feb 2024, Christoph Böhmwalder wrote:

> Hi,
>
> I recently(?) got a new version of coccinelle from Fedora:
>
>
> $ spatch --version
> spatch version 1.1.1 compiled with OCaml version 5.0.1+dev0-2022-12-15
> Flags passed to the configure script: --build=x86_64-redhat-linux-gnu
> --host=x86_64-redhat-linux-gnu --program-prefix=
> --disable-dependency-tracking --prefix=/usr --exec-prefix=/usr
> --bindir=/usr/bin --sbindir=/usr/sbin --sysconfdir=/etc
> --datadir=/usr/share --includedir=/usr/include --libdir=/usr/lib64
> --libexecdir=/usr/libexec --localstatedir=/var --runstatedir=/run
> --sharedstatedir=/var/lib --mandir=/usr/share/man
> --infodir=/usr/share/info --with-python=/usr/bin/python3
> --with-menhir=/usr/bin/menhir
> OCaml scripting support: yes
> Python scripting support: yes
> Syntax of regular expressions: PCRE
>
>
> And it broke a lot of our semantic patches. I now get lots of these
> warnings:
>
> warning: rule starting on line 248, node 45:
> kunmap_local(...)[1,2,20,21,25] in recv_dless_read may be inconsistently
> modified
> warning: rule starting on line 248, node 42: decl:mapped[1,2,20,21,22]
> in recv_dless_read may be inconsistently modified
> warning: rule starting on line 273, node 18: src = ... ;[1,2,8,9,11] in
> drbd_csum_bio may be inconsistently modified
> warning: rule starting on line 273, node 20:
> kunmap_local(...)[1,2,8,9,13] in drbd_csum_bio may be inconsistently
> modified
>
> (By the way: these are reported on stdout, I'm not sure that's intended
> either)
>
> And the patch output is noticeably different.
>
> This all works fine on 1.1.0, and curiously it also works on the 1.1.1
> version in Ubuntu:
>
> $ spatch --version
> spatch version 1.1.1 compiled with OCaml version 4.13.1
> Flags passed to the configure script: --prefix=/usr --sysconfdir=/etc
> --libdir=/usr/lib --enable-ocaml --enable-python --enable-opt
> OCaml scripting support: yes
> Python scripting support: yes
> Syntax of regular expressions: PCRE
>
> I'm trying to come up with a minimal reproducer right now, but just
> throwing this out there in case anyone has an idea.

Thanks for the report.  If convenient you could try the github version.

I really need to make a new release...

julia

Re: [cocci] Spatch 1.1.1 backward incompatibility

[Christoph Böhmwalder]

Am 16.02.24 um 10:51 schrieb Julia Lawall:
> 
> Thanks for the report.  If convenient you could try the github version.

The latest GitHub version (0df3efff) has the same problem. It also seems
to have some other issue, because it took 1385.5s (instead of the usual
~25) to process the patch.

I also noticed that Fedora used OCaml 5.0, so I tried with the same
commit the Fedora version is based on (0afff7f), but compiled with OCaml
4.14. It didn't make a difference, so it's not an issue with OCaml 5.

I'm trying to bisect this now, will report if I find anything.

> 
> I really need to make a new release...

I was going to ask about that... :)

> 
> julia
-- 
Christoph Böhmwalder
LINBIT | Keeping the Digital World Running
DRBD HA —  Disaster Recovery — Software defined Storage

Re: [cocci] Spatch 1.1.1 backward incompatibility

[Markus Elfring]

> The latest GitHub version (0df3efff) has the same problem. It also seems
> to have some other issue, because it took 1385.5s (instead of the usual
> ~25) to process the patch.

How do you think about to share any more details about the SmPL scripts
which became affected in your test environments?

Regards,
Markus

Re: [cocci] Spatch 1.1.1 backward incompatibility

[Julia Lawall]

[-- Attachment #1: Type: text/plain, Size: 1047 bytes --]



On Fri, 16 Feb 2024, Christoph Böhmwalder wrote:

> Am 16.02.24 um 10:51 schrieb Julia Lawall:
> >
> > Thanks for the report.  If convenient you could try the github version.
>
> The latest GitHub version (0df3efff) has the same problem. It also seems
> to have some other issue, because it took 1385.5s (instead of the usual
> ~25) to process the patch.
>
> I also noticed that Fedora used OCaml 5.0, so I tried with the same
> commit the Fedora version is based on (0afff7f), but compiled with OCaml
> 4.14. It didn't make a difference, so it's not an issue with OCaml 5.
>
> I'm trying to bisect this now, will report if I find anything.

OK, thanks.  It may be sufficient to bisect your example.  You can try for
example

--debug --show-trying

to see at what function it is getting stuck.

julia

>
> >
> > I really need to make a new release...
>
> I was going to ask about that... :)
>
> >
> > julia
> --
> Christoph Böhmwalder
> LINBIT | Keeping the Digital World Running
> DRBD HA —  Disaster Recovery — Software defined Storage
>