Re: [PATCH i-g-t v3 05/31] lib/igt_tools_stub: add stubs needed by tools

[Sebastian Brzezinka <sebastian.brzezinka@intel.com>]

Hi Kamil,

On Thu Jul 16, 2026 at 12:13 PM CEST, Kamil Konieczny wrote:
> Hi all,
> On 2026-07-16 at 06:09:42 +0000, Krzysztof Karas wrote:
>> Hi Sebastian,
>> 
>> On 2026-07-15 at 14:15:15 +0200, Sebastian Brzezinka wrote:
>> > Add stubs for functions used by tools when linking without the
>> > full libigt.so: __igt_skip_check, igt_exit, igt_debugfs_open and
>> > igt_open_forcewake_handle_for_pcidev.
>> This patch does more than that: it unifies the comment style in
>> the file. This is a nit though, so consider it only if you would
>> prepare another version.
>> 
>> > 
>> > Signed-off-by: Sebastian Brzezinka <sebastian.brzezinka@intel.com>
>> > Acked-by: Ashutosh Dixit <ashutosh.dixit@intel.com>
>> > ---
>> >  lib/igt_tools_stub.c | 83 ++++++++++++++++++++++++++++++++++++++++----
>> >  1 file changed, 76 insertions(+), 7 deletions(-)
>> > 
>> > diff --git a/lib/igt_tools_stub.c b/lib/igt_tools_stub.c
>> > index 9a0ec6217..21dc53f3f 100644
>> > --- a/lib/igt_tools_stub.c
>> > +++ b/lib/igt_tools_stub.c
>> > @@ -21,12 +21,27 @@
>> >   * IN THE SOFTWARE.
>> >   *
>> >   */
>> > +#include <fcntl.h>
>> > +#include <limits.h>
>> > +#include <stdarg.h>
>> > +#include <stdio.h>
>> > +#include <stdlib.h>
>> > +#include <sys/stat.h>
>> > +#include <sys/sysmacros.h>
>> > +#include <unistd.h>
>> > +
>> >  #include "igt_core.h"
>> > +#include "igt_debugfs.h"
>> >  
>> > -/* Stub for igt_log, this stub will simply print the msg on stderr device.
>> > - * Domain and log level are ignored.
>> > +/*
>> > + * Forward declaration only - avoids pulling in the heavy igt_gt.h header.
>> > + * The type is used as an opaque pointer; this stub always returns -1.
>> >   */
>> This is a nit, so consider it only if you would prepare another
>> version of this series:
>> The comment should be moved right before the function declaration.
>> > +struct pci_device;
>> >  
>> > +int igt_open_forcewake_handle_for_pcidev(const struct pci_device *pci_dev);
>> > +
>> > +/* Stub for igt_log: print message to stderr; domain and log level are ignored. */
>> >  void igt_log(const char *domain, enum igt_log_level level, const char *format, ...)
>
> This will be a second documentation for a function.
> Question for Pawel: could it be a problem for our documentation generation?
> Do we need an exclusion for stub functions?
For what it's worth, this documentation style was already there. It's
not my invention, so it should be ok, as long as the previous comments
don't break the docs build.

-- 
Best regards,
Sebastian