[DTrace-devel] [PATCH 5/7] Add support for built-in variable args

Kris Van Hees kris.van.hees at oracle.com
Fri May 27 18:59:33 UTC 2022


On Fri, May 27, 2022 at 10:22:37AM -0700, Eugene Loh via DTrace-devel wrote:
> 
> On 5/27/22 11:32, Kris Van Hees wrote:
> > On Fri, May 27, 2022 at 09:29:41AM -0700, Eugene Loh via DTrace-devel wrote:
> > 
> > > *)  test/unittest/disasm/tst.ann-bvar.sh is missing uregs.
> > That is because uregs isn't supported yet.  But since this test is never
> > actually executing the generated BPF program, I guess I could include it
> > anyhow.
> Terrific.  That also deals with the case of uregs support being added but
> updating this test getting overlooked.

No, no!  I am *not* implementing uregs yet.  I merely stated that I can include
uregs in this test *because* it is not executing the BPF code so the lack of
support won't cause the test to fail.

> > > *)  Why are there no tests (other than disasm and err)?  One argument could
> > > be
> > >      because SDT args are not yet populated... patch 6.  But syscall
> > > apparently
> > >      has typed args.  After all,
> > >      test/unittest/variables/bvar/err.D_ARGS_IDX.args-too-many.d requires it.
> > Well, as you point out, there are tests.  And there are no tests to actually
> > test the proper functioning in this patch because this patch is not actually
> > providing any implementation of a provider that makes use of args[].
> > 
> > Arguably, the variables/bvar/tst.args-match-argN.d test could be moved into
> > this patch, if you prefer that.
> Sounds great.
> 
> _______________________________________________
> DTrace-devel mailing list
> DTrace-devel at oss.oracle.com
> https://oss.oracle.com/mailman/listinfo/dtrace-devel



More information about the DTrace-devel mailing list