On Tue, 8 May 2018, Jan Hubicka wrote: > Hi, > this patch prevents lto-option to store some flags that does not make snese > to store, > in partiuclar dumpdir and -fresolution. These definitly should not be > preserved from > compile time to link time and in case of incremental linking they caused > trouble with > wrong resolution file being used in some cases. > > I guess this is just tip of iceberg - I think we should switch to > whitelisting options that needs saving rather than saving everything > with few exceptions. This is however a separate issue.
We probably should strip all CL_OPTIMIZATION and CL_TARGET options (if the target supports streaming target options and the option is saved, unfortunately we don't record a CL_SAVED flag). We should think about _not_ dropping diagnostic options and/or putting those into the optimization nodes. > Bootstrapped/regtested x86_64-linux, OK? > * lto-opts.c (lto_write_options): Skip OPT_dumpdir, OPT_fresolution_. > Index: lto-opts.c > =================================================================== > --- lto-opts.c (revision 260042) > +++ lto-opts.c (working copy) > @@ -109,6 +109,8 @@ > case OPT_SPECIAL_ignore: > case OPT_SPECIAL_program_name: > case OPT_SPECIAL_input_file: > + case OPT_dumpdir: > + case OPT_fresolution_: > continue; OK. Richard.