Re: [PATCH] eal: remove unnecessary argv[0] handling

[Thomas Monjalon <thomas@monjalon.net>]

09/02/2022 20:29, Stephen Hemminger:
> On Wed, 09 Feb 2022 19:51:15 +0100
> Thomas Monjalon <thomas@monjalon.net> wrote:
> 
> > 09/02/2022 16:45, Stephen Hemminger:
> > > On Wed, 09 Feb 2022 15:58:26 +0100
> > > Thomas Monjalon <thomas@monjalon.net> wrote:
> > >   
> > > > 02/02/2022 20:47, Stephen Hemminger:  
> > > > > The rte_eal_init function looks at argv[0] to determine
> > > > > the program name to pass to the log init function.
> > > > > This is both unnecessary and in a corner case a problem.
> > > > > 
> > > > > Parsing argv[0] is unnecessary because the function openlog()
> > > > > already determines the log identifier from program name if
> > > > > NULL is passed, see openlog man page:
> > > > >     The string pointed to by ident is prepended to every message, and is
> > > > >     typically set to the program name. If ident is NULL, the program name
> > > > >     is used.  (POSIX.1-2008 does not specify the behavior when ident is
> > > > >     NULL.)    
> > > > 
> > > > What about POSIX warning?
> > > > Did you test it? Which libc? What about musl?  
> > > 
> > > Source for musl shows that openlog accepts NULL as ident.
> > > It then generates log messages with ident in the message.  
> > 
> > In this case, it will be empty string.
> > It seems this solution does not work with musl.
> 
> Good news: musl is available as debian package
> Bad news: DPDK build is broken with musl
> 
> 
> Library IPSec_MB found: YES
> 
> drivers/crypto/ipsec_mb/meson.build:19:4: ERROR: Could not get define 'IMB_VERSION_STR'

This error is not related to musl.

Compilation on Alpine (with musl) is tested regularly,
it works for me.