Michal Schmidt wrote:
You can take a look at the attached patch, which shrinks struct swsusp_info to 2028 bytes on i386. This is achieved by only saving the components (n, e, d). I am _not_ proposing to merge this one, because it requires a slightly modified libgcrypt to work. Original libgcrypt insists on knowing all the six components, even though it doesn't really need them. I consider that a bug in libgcrypt.

Here's the libgcrypt hack, in case someone's interested.
Michal
diff -Nurp libgcrypt11-1.2.3/cipher/pubkey.c 
libgcrypt11-1.2.3.hack/cipher/pubkey.c
--- libgcrypt11-1.2.3/cipher/pubkey.c   2005-07-29 15:45:48.000000000 +0200
+++ libgcrypt11-1.2.3.hack/cipher/pubkey.c      2007-01-20 02:54:44.000000000 
+0100
@@ -709,7 +709,7 @@ sexp_elements_extract (gcry_sexp_t key_s
     {
       list = gcry_sexp_find_token (key_sexp, name, 1);
       if (! list)
-       err = GPG_ERR_NO_OBJ;
+        elements[idx] = NULL;
       else
        {
          elements[idx] = gcry_sexp_nth_mpi (list, 1, GCRYMPI_FMT_USG);
diff -Nurp libgcrypt11-1.2.3/cipher/rsa.c libgcrypt11-1.2.3.hack/cipher/rsa.c
--- libgcrypt11-1.2.3/cipher/rsa.c      2005-07-29 15:45:48.000000000 +0200
+++ libgcrypt11-1.2.3.hack/cipher/rsa.c 2007-01-20 02:57:13.000000000 +0100
@@ -509,6 +509,9 @@ _gcry_rsa_decrypt (int algo, gcry_mpi_t 
   sk.q = skey[4];
   sk.u = skey[5];
 
+  if (! (flags & PUBKEY_FLAG_NO_BLINDING) && !sk.e)
+    return GPG_ERR_NO_OBJ;
+
   y = gcry_mpi_snew (gcry_mpi_get_nbits (sk.n));
 
   if (! (flags & PUBKEY_FLAG_NO_BLINDING))
-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
Suspend-devel mailing list
Suspend-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/suspend-devel

Reply via email to