Hi, On Wed, Nov 20, 2002 at 08:34:08AM +0100, Richard Levitte - VMS Whacker wrote: > In message <[EMAIL PROTECTED]> on Tue, 19 Nov 2002 18:23:37 +0100, >Peter Poeml <[EMAIL PROTECTED]> said: > poeml> .../openssl-0.9.7_beta4/test# LD_LIBRARY_PATH=.. sh ./tcrl > > Eh, tcrl isn't the same as trsa, so I'm not surprised you get > different results :-).
Oops :) > poeml> Anyway, with verbosity added these are the error messages: > poeml> > poeml> + echo 'd -> d' > poeml> d -> d > poeml> + ../apps/openssl rsa -in f.d -inform d -outform d > poeml> unable to load Private Key > poeml> 1397:error:0D07209B:asn1 encoding routines:ASN1_get_object:too >long:asn1_lib.c:130: > poeml> 1397:error:0D07209B:asn1 encoding routines:ASN1_get_object:too >long:asn1_lib.c:130: > poeml> 1397:error:0D068066:asn1 encoding routines:ASN1_CHECK_TLEN:bad object >header:tasn_dec.c:928: > poeml> 1397:error:0D07803A:asn1 encoding routines:ASN1_ITEM_EX_D2I:nested asn1 >error:tasn_dec.c:304:Type=RSA > poeml> 1397:error:0D09A00D:asn1 encoding routines:d2i_PrivateKey:ASN1 >lib:d2i_pr.c:96: > > Hmm, it's possible we've screwed up int vs. long (they differ in ia64, > if I understand correctly?)... Yes, like on alpha, a long is 8 bytes wide (64 bit) while int is 32 bit. Note also that pointers are also 64 bit and therefore larger than int, and all pointers will point above the 4G boundary. > poeml> I'll attach the suspicious file f.d to this mail. It seems that it is broken, >because my openssl-0.9.6 binary reports > poeml> > poeml> read RSA key > poeml> unable to load key > poeml> 3432:error:0906D06C:PEM routines:PEM_read_bio:no start >line:pem_lib.c:663:Expecting: ANY PRIVATE KEY I'm sorry again that I make it extra hard for you -- I had used "-inform p" while I meant "-inform d"... > Eh, f.d isn't in PEM format. Try the following command: > > openssl rsa -in f.d -inform d That yields the same result as I reported in the beginning: # LD_LIBRARY_PATH=`pwd`/.. ../apps/openssl rsa -in f.d -inform d unable to load Private Key 24687:error:0D07209B:asn1 encoding routines:ASN1_get_object:too long:asn1_lib.c:130: 24687:error:0D07209B:asn1 encoding routines:ASN1_get_object:too long:asn1_lib.c:130: 24687:error:0D068066:asn1 encoding routines:ASN1_CHECK_TLEN:bad object header:tasn_dec.c:928: 24687:error:0D07803A:asn1 encoding routines:ASN1_ITEM_EX_D2I:nested asn1 error:tasn_dec.c:304:Type=RSA 24687:error:0D09A00D:asn1 encoding routines:d2i_PrivateKey:ASN1 lib:d2i_pr.c:96: I checked the rest of the test suite, and found that there is one other test that fails -- test_sid: d -> d + ../apps/openssl sess_id -in f.d -inform d -outform d unable to load SSL_SESSION 8508:error:0D07209B:asn1 encoding routines:ASN1_get_object:too long:asn1_lib.c:130: 8508:error:0D06703C:asn1 encoding routines:a2i_ASN1_STRING:bad get asn1 object call:ssl_asn1.c:243:address=6917529027641429376 offset=0 + '[' 1 '!=' 0 ']' + exit 1 All other tests did successfully pass on ia64, and no test fails on other 64 bit archs: s390x and ppc64 (64bit, big endian), and x86_64 and alpha (64bit, little endian). Unfortunately I don't know what's so special with ia64, since word size, endianness and signedness of char are identical. Let me know if I can do anything else to help. Peter -- Thought is limitation. Free your mind.
msg13859/pgp00000.pgp
Description: PGP signature
