Re: [PATCH] eal/windows: define standard file numbers

[Dmitry Kozlyuk <dmitry.kozliuk@gmail.com>]

2024-10-10 12:39 (UTC+0100), Bruce Richardson:
> On Thu, Oct 10, 2024 at 01:43:41PM +0300, Dmitry Kozlyuk wrote:
> > 2024-10-10 10:54 (UTC+0100), Bruce Richardson:  
> > > The macros for STD*_FILENO are missing on windows. Add defines for them
> > > to the DPDK-local unistd.h file.
> > > 
> > > Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
> > > ---
> > >  lib/eal/windows/include/unistd.h | 11 +++++++++++
> > >  1 file changed, 11 insertions(+)
> > > 
> > > diff --git a/lib/eal/windows/include/unistd.h b/lib/eal/windows/include/unistd.h
> > > index 6b33005b24..78150c6480 100644
> > > --- a/lib/eal/windows/include/unistd.h
> > > +++ b/lib/eal/windows/include/unistd.h
> > > @@ -12,4 +12,15 @@
> > >  
> > >  #include <io.h>
> > >  
> > > +/*
> > > + * Windows appears to be missing STD*_FILENO macros, so define here.
> > > + * For simplicity, assume that if STDIN_FILENO is missing, all are,
> > > + * rather than checking each individually.
> > > + */
> > > +#ifndef STDIN_FILENO
> > > +#define STDIN_FILENO _fileno(stdin)
> > > +#define STDOUT_FILENO _fileno(stdout)
> > > +#define STDERR_FILENO _fileno(stderr)
> > > +#endif
> > > +
> > >  #endif /* _UNISTD_H_ */  
> > 
> > Why is this needed?
> > There are 3 users of STD*_FILENO within DPDK:
> > * lib/eal/unix/eal_debug.c - not for Windows, obviously
> > * app/dumpcap - not for Windows, requires multi-process
> > * app/proc-info - not for Windows, requires multi-process
> > * examples/l2fwd-cat - not for Windows, requires <pqos.h>  
> 
> https://patches.dpdk.org/project/dpdk/patch/20240822104109.116208-2-bruce.richardson@intel.com/
> 
> This merged patch is now throwing errors about the missing defines on
> Windows.

Thanks, I was searching some outdated source.

Currently, <rte_os_shim.h> contains at least `read()`, `write()`,
and `unlink()` which POSIX places in <unistd.h>.
I wonder whether all that stuff should be moved into EAL's <unistd.h>
or, on the opposite, <rte_os_shim.h> should consume all other shims.
But that would be a broader refactoring, so for the fix:

Acked-by: Dmitry Kozlyuk <dmitry.kozliuk@gmail.com>