On Fri, Dec 24, 2021 at 03:37:47AM +1100, Joel Sing wrote:
> On 21-12-23 13:14:26, Tobias Heider wrote:
> > On Mon, Dec 13, 2021 at 10:56:15AM -0700, Joel Sing wrote:
> > > CVSROOT: /cvs
> > > Module name: src
> > > Changes by: [email protected] 2021/12/13 10:56:15
> > >
> > > Modified files:
> > > lib/libcrypto : Makefile
> > > lib/libcrypto/asn1: tasn_typ.c
> > > Removed files:
> > > lib/libcrypto/asn1: a_bool.c
> > >
> > > Log message:
> > > Clean up d2i_ASN1_BOOLEAN() and i2d_ASN1_BOOLEAN().
> > >
> > > Convert these to templated ASN.1, given we already have ASN1_BOOLEAN_it.
> > >
> > > ok inoguchi@ tb@
> > >
> >
> > It looks like the new d2i_ASN1_BOOLEAN() is broken on several archs, see:
> >
> > http://bluhm.genua.de/regress/results/2021-12-22T10%3A17%3A02Z/logs/lib/libcrypto/asn1/make.log
> > http://bluhm.genua.de/regress/results/2021-12-21T02%3A17%3A18Z/logs/lib/libcrypto/asn1/make.log
>
> What architectures specifically?
The test fails on ppc64, macppc and i386.
sparc64 coredumps at:
#0 ASN1_item_d2i (pval=0xfffffffffffdced4, in=0xfffffffffffdcfa8, len=3,
it=0x28c07f1618) at /usr/src/lib/libcrypto/asn1/tasn_dec.c:119
119 return *pval;
If you look for the lib/libcrypto/asn1 row at
http://bluhm.genua.de/regress/results/regress.html you can view the
test output by clicking on the FAILs.
>
> > The diff below reverts the previous commit which fixes the test.
> >
> > ok to backout?
>
> Nothing really uses d2i_ASN1_BOOLEAN() directly - if there is a bug
> in the templated ASN1_BOOLEAN code (or the regress test) it needs
> to be found and fixed.
