diff --git a/common/ksba-io-support.c b/common/ksba-io-support.c
index 5c7fd220c..c7dd81a61 100644
--- a/common/ksba-io-support.c
+++ b/common/ksba-io-support.c
@@ -1,753 +1,763 @@
 /* kska-io-support.c - Supporting functions for ksba reader and writer
  * Copyright (C) 2001-2005, 2007, 2010-2011, 2017  Werner Koch
  * Copyright (C) 2006  g10 Code GmbH
  *
  * This file is part of GnuPG.
  *
  * This file is free software; you can redistribute it and/or modify
  * it under the terms of either
  *
  *   - the GNU Lesser General Public License as published by the Free
  *     Software Foundation; either version 3 of the License, or (at
  *     your option) any later version.
  *
  * or
  *
  *   - the GNU General Public License as published by the Free
  *     Software Foundation; either version 2 of the License, or (at
  *     your option) any later version.
  *
  * or both in parallel, as here.
  *
  * This file is distributed in the hope that it will be useful,
  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  * GNU General Public License for more details.
  *
  * You should have received a copy of the GNU General Public License
  * along with this program; if not, see <https://www.gnu.org/licenses/>.
  */
 
 #include <config.h>
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
 #include <errno.h>
 #include <unistd.h>
 #include <time.h>
 #include <assert.h>
 #include <ksba.h>
 
 #include "util.h"
 #include "i18n.h"
 #include "ksba-io-support.h"
 
 
 #ifdef HAVE_DOSISH_SYSTEM
   #define LF "\r\n"
 #else
   #define LF "\n"
 #endif
 
 
 /* Data used by the reader callbacks.  */
 struct reader_cb_parm_s
 {
   estream_t fp;
 
   unsigned char line[1024];
   int linelen;
   int readpos;
   int have_lf;
   unsigned long line_counter;
 
   int allow_multi_pem;  /* Allow processing of multiple PEM objects. */
   int autodetect;       /* Try to detect the input encoding. */
   int assume_pem;       /* Assume input encoding is PEM. */
   int assume_base64;    /* Assume input is base64 encoded. */
 
   int identified;
   int is_pem;
   int is_base64;
   int stop_seen;
   int might_be_smime;
 
   int eof_seen;
 
   struct {
     int idx;
     unsigned char val;
     int stop_seen;
   } base64;
 };
 
 
 /* Data used by the writer callbacks.  */
 struct writer_cb_parm_s
 {
   estream_t stream;    /* Output stream.  */
 
   char *pem_name;      /* Malloced.  */
 
   int wrote_begin;
   int did_finish;
 
   struct {
     int idx;
     int quad_count;
     unsigned char radbuf[4];
   } base64;
 
 };
 
 
 /* Context for this module's functions.  */
 struct gnupg_ksba_io_s {
   union {
     struct reader_cb_parm_s rparm;
     struct writer_cb_parm_s wparm;
   } u;
 
   union {
     ksba_reader_t reader;
     ksba_writer_t writer;
   } u2;
 };
 
 
 /* The base-64 character list */
 static char bintoasc[64] =
        "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
        "abcdefghijklmnopqrstuvwxyz"
        "0123456789+/";
 /* The reverse base-64 list */
 static unsigned char asctobin[256] = {
   0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
   0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
   0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
   0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x3e, 0xff, 0xff, 0xff, 0x3f,
   0x34, 0x35, 0x36, 0x37, 0x38, 0x39, 0x3a, 0x3b, 0x3c, 0x3d, 0xff, 0xff,
   0xff, 0xff, 0xff, 0xff, 0xff, 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06,
   0x07, 0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f, 0x10, 0x11, 0x12,
   0x13, 0x14, 0x15, 0x16, 0x17, 0x18, 0x19, 0xff, 0xff, 0xff, 0xff, 0xff,
   0xff, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f, 0x20, 0x21, 0x22, 0x23, 0x24,
   0x25, 0x26, 0x27, 0x28, 0x29, 0x2a, 0x2b, 0x2c, 0x2d, 0x2e, 0x2f, 0x30,
   0x31, 0x32, 0x33, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
   0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
   0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
   0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
   0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
   0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
   0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
   0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
   0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
   0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
   0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
   0xff, 0xff, 0xff, 0xff
 };
 
 
 static int
 has_only_base64 (const unsigned char *line, int linelen)
 {
   if (linelen < 20)
     return 0;
   for (; linelen; line++, linelen--)
     {
       if (*line == '\n' || (linelen > 1 && *line == '\r' && line[1] == '\n'))
           break;
       if ( !strchr (bintoasc, *line) )
         return 0;
     }
   return 1;  /* yes */
 }
 
 static int
 is_empty_line (const unsigned char *line, int linelen)
 {
   if (linelen >= 2 && *line == '\r' && line[1] == '\n')
     return 1;
   if (linelen >= 1 && *line == '\n')
     return 1;
   return 0;
 }
 
 
 static int
 base64_reader_cb (void *cb_value, char *buffer, size_t count, size_t *nread)
 {
   struct reader_cb_parm_s *parm = cb_value;
   size_t n;
   int c, c2;
 
   *nread = 0;
   if (!buffer)
     return -1; /* not supported */
 
  next:
   if (!parm->linelen)
     {
       /* read an entire line or up to the size of the buffer */
       parm->line_counter++;
       parm->have_lf = 0;
       for (n=0; n < DIM(parm->line);)
         {
           c = es_getc (parm->fp);
           if (c == EOF)
             {
               parm->eof_seen = 1;
               if (es_ferror (parm->fp))
                 return -1;
               break;
             }
           parm->line[n++] = c;
           if (c == '\n')
             {
               parm->have_lf = 1;
               /* Fixme: we need to skip overlong lines while detecting
                  the dashed lines */
               break;
             }
         }
       parm->linelen = n;
       if (!n)
         return -1; /* eof */
       parm->readpos = 0;
     }
 
   if (!parm->identified)
     {
       if (!parm->autodetect)
         {
           if (parm->assume_pem)
             {
               /* wait for the header line */
               parm->linelen = parm->readpos = 0;
               if (!parm->have_lf
                   || strncmp ((char*)parm->line, "-----BEGIN ", 11)
                   || !strncmp ((char*)parm->line+11, "PGP ", 4))
                 goto next;
               parm->is_pem = 1;
             }
           else if (parm->assume_base64)
             parm->is_base64 = 1;
         }
       else if (parm->line_counter == 1 && !parm->have_lf)
         {
           /* first line too long - assume DER encoding */
           parm->is_pem = 0;
         }
       else if (parm->line_counter == 1 && parm->linelen && *parm->line == 0x30)
         {
           /* the very first byte does pretty much look like a SEQUENCE tag*/
           parm->is_pem = 0;
         }
       else if ( parm->have_lf
                 && !strncmp ((char*)parm->line, "-----BEGIN ", 11)
                 && strncmp ((char *)parm->line+11, "PGP ", 4) )
         {
           /* Fixme: we must only compare if the line really starts at
              the beginning */
           parm->is_pem = 1;
           parm->linelen = parm->readpos = 0;
         }
       else if ( parm->have_lf && parm->line_counter == 1
                 && parm->linelen >= 13
                 && !ascii_memcasecmp (parm->line, "Content-Type:", 13))
         { /* might be a S/MIME body */
           parm->might_be_smime = 1;
           parm->linelen = parm->readpos = 0;
           goto next;
         }
       else if (parm->might_be_smime == 1
                && is_empty_line (parm->line, parm->linelen))
         {
           parm->might_be_smime = 2;
           parm->linelen = parm->readpos = 0;
           goto next;
         }
       else if (parm->might_be_smime == 2)
         {
           parm->might_be_smime = 0;
           if ( !has_only_base64 (parm->line, parm->linelen))
             {
               parm->linelen = parm->readpos = 0;
               goto next;
             }
           parm->is_pem = 1;
         }
       else
         {
           parm->linelen = parm->readpos = 0;
           goto next;
         }
       parm->identified = 1;
       parm->base64.stop_seen = 0;
       parm->base64.idx = 0;
     }
 
 
   n = 0;
   if (parm->is_pem || parm->is_base64)
     {
       if (parm->is_pem && parm->have_lf
           && !strncmp ((char*)parm->line, "-----END ", 9))
         {
           parm->identified = 0;
           parm->linelen = parm->readpos = 0;
 
           /* If the caller want to read multiple PEM objects from one
              file, we have to reset our internal state and return a
              EOF immediately. The caller is the expected to use
              ksba_reader_clear to clear the EOF condition and continue
              to read.  If we don't want to do that we just return 0
              bytes which will force the ksba_reader to skip until
              EOF. */
           if (parm->allow_multi_pem)
             {
               parm->identified = 0;
               parm->autodetect = 0;
               parm->assume_pem = 1;
               parm->stop_seen = 0;
               return -1; /* Send EOF now. */
             }
         }
       else if (parm->stop_seen)
         { /* skip the rest of the line */
           parm->linelen = parm->readpos = 0;
         }
       else
         {
           int idx = parm->base64.idx;
           unsigned char val = parm->base64.val;
 
           while (n < count && parm->readpos < parm->linelen )
             {
               c = parm->line[parm->readpos++];
               if (c == '\n' || c == ' ' || c == '\r' || c == '\t')
                 continue;
-              if (c == '=')
-                { /* pad character: stop */
-                  if (idx == 1)
-                    buffer[n++] = val;
-                  parm->stop_seen = 1;
-                  break;
-                }
-              if( (c = asctobin[(c2=c)]) == 255 )
+              if ((c = asctobin[(c2=c)]) == 255)
                 {
+                  if (c2 == '=')
+                    { /* pad character: stop */
+                      if (idx == 1)
+                        buffer[n++] = val;
+                      parm->stop_seen = 1;
+                      break;
+                    }
+                  else if (c2 == '-'
+                           && parm->readpos == 1
+                           && parm->readpos-1+9 < parm->linelen
+                           && !strncmp ((char*)parm->line + parm->readpos-1,
+                                        "-----END ", 9))
+                    { /* END line seen (padding was not needed). */
+                      log_debug ("END seen\n");
+                      parm->stop_seen = 1;
+                      break;
+                    }
                   log_error (_("invalid radix64 character %02x skipped\n"),
                              c2);
                   continue;
                 }
               switch (idx)
                 {
                 case 0:
                   val = c << 2;
                   break;
                 case 1:
                   val |= (c>>4)&3;
                   buffer[n++] = val;
                   val = (c<<4)&0xf0;
                   break;
                 case 2:
                   val |= (c>>2)&15;
                   buffer[n++] = val;
                   val = (c<<6)&0xc0;
                   break;
                 case 3:
                   val |= c&0x3f;
                   buffer[n++] = val;
                   break;
                 }
               idx = (idx+1) % 4;
             }
           if (parm->readpos == parm->linelen)
             parm->linelen = parm->readpos = 0;
 
           parm->base64.idx = idx;
           parm->base64.val = val;
         }
     }
   else
     { /* DER encoded */
       while (n < count && parm->readpos < parm->linelen)
           buffer[n++] = parm->line[parm->readpos++];
       if (parm->readpos == parm->linelen)
         parm->linelen = parm->readpos = 0;
     }
 
   *nread = n;
   return 0;
 }
 
 
 
 static int
 simple_reader_cb (void *cb_value, char *buffer, size_t count, size_t *nread)
 {
   struct reader_cb_parm_s *parm = cb_value;
   size_t n;
   int c = 0;
 
   *nread = 0;
   if (!buffer)
     return -1; /* not supported */
 
   for (n=0; n < count; n++)
     {
       c = es_getc (parm->fp);
       if (c == EOF)
         {
           parm->eof_seen = 1;
           if (es_ferror (parm->fp))
             return -1;
           if (n)
             break; /* Return what we have before an EOF.  */
           return -1;
         }
       *(byte *)buffer++ = c;
     }
 
   *nread = n;
   return 0;
 }
 
 
 
 
 static int
 base64_writer_cb (void *cb_value, const void *buffer, size_t count)
 {
   struct writer_cb_parm_s *parm = cb_value;
   unsigned char radbuf[4];
   int i, c, idx, quad_count;
   const unsigned char *p;
   estream_t stream = parm->stream;
 
   if (!count)
     return 0;
 
   if (!parm->wrote_begin)
     {
       if (parm->pem_name)
         {
           es_fputs ("-----BEGIN ", stream);
           es_fputs (parm->pem_name, stream);
           es_fputs ("-----\n", stream);
         }
       parm->wrote_begin = 1;
       parm->base64.idx = 0;
       parm->base64.quad_count = 0;
     }
 
   idx = parm->base64.idx;
   quad_count = parm->base64.quad_count;
   for (i=0; i < idx; i++)
     radbuf[i] = parm->base64.radbuf[i];
 
   for (p=buffer; count; p++, count--)
     {
       radbuf[idx++] = *p;
       if (idx > 2)
         {
           idx = 0;
           c = bintoasc[(*radbuf >> 2) & 077];
           es_putc (c, stream);
           c = bintoasc[(((*radbuf<<4)&060)|((radbuf[1] >> 4)&017))&077];
           es_putc (c, stream);
           c = bintoasc[(((radbuf[1]<<2)&074)|((radbuf[2]>>6)&03))&077];
           es_putc (c, stream);
           c = bintoasc[radbuf[2]&077];
           es_putc (c, stream);
           if (++quad_count >= (64/4))
             {
               es_fputs (LF, stream);
               quad_count = 0;
             }
         }
     }
   for (i=0; i < idx; i++)
     parm->base64.radbuf[i] = radbuf[i];
   parm->base64.idx = idx;
   parm->base64.quad_count = quad_count;
 
   return es_ferror (stream)? gpg_error_from_syserror () : 0;
 }
 
 
 /* This callback is only used in stream mode.  However, we don't
    restrict it to this.  */
 static int
 plain_writer_cb (void *cb_value, const void *buffer, size_t count)
 {
   struct writer_cb_parm_s *parm = cb_value;
   estream_t stream = parm->stream;
 
   if (!count)
     return 0;
 
   es_write (stream, buffer, count, NULL);
 
   return es_ferror (stream)? gpg_error_from_syserror () : 0;
 }
 
 
 static int
 base64_finish_write (struct writer_cb_parm_s *parm)
 {
   unsigned char *radbuf;
   int c, idx, quad_count;
   estream_t stream = parm->stream;
 
   if (!parm->wrote_begin)
     return 0; /* Nothing written or we are not called in base-64 mode. */
 
   /* flush the base64 encoding */
   idx = parm->base64.idx;
   quad_count = parm->base64.quad_count;
   if (idx)
     {
       radbuf = parm->base64.radbuf;
 
       c = bintoasc[(*radbuf>>2)&077];
       es_putc (c, stream);
       if (idx == 1)
         {
           c = bintoasc[((*radbuf << 4) & 060) & 077];
           es_putc (c, stream);
           es_putc ('=', stream);
           es_putc ('=', stream);
         }
       else
         {
           c = bintoasc[(((*radbuf<<4)&060)|((radbuf[1]>>4)&017))&077];
           es_putc (c, stream);
           c = bintoasc[((radbuf[1] << 2) & 074) & 077];
           es_putc (c, stream);
           es_putc ('=', stream);
 
         }
       if (++quad_count >= (64/4))
         {
           es_fputs (LF, stream);
           quad_count = 0;
         }
     }
 
   if (quad_count)
     es_fputs (LF, stream);
 
   if (parm->pem_name)
     {
       es_fputs ("-----END ", stream);
       es_fputs (parm->pem_name, stream);
       es_fputs ("-----\n", stream);
     }
 
   return es_ferror (stream)? gpg_error_from_syserror () : 0;
 }
 
 
 
 
 /* Create a reader for the stream FP.  FLAGS can be used to specify
  * the expected input encoding.
  *
  * The function returns a gnupg_ksba_io_t object which must be passed to
  * the gpgme_destroy_reader function.  The created ksba_reader_t
  * object is stored at R_READER - the caller must not call the
  * ksba_reader_release function on.
  *
  * The supported flags are:
  *
  * GNUPG_KSBA_IO_PEM        - Assume the input is PEM encoded
  * GNUPG_KSBA_IO_BASE64     - Assume the input is Base64 encoded.
  * GNUPG_KSBA_IO_AUTODETECT - The reader tries to detect the encoding.
  * GNUPG_KSBA_IO_MULTIPEM   - The reader expects that the caller uses
  *                            ksba_reader_clear after EOF until no more
  *                            objects were found.
  *
  * Note that the PEM flag has a higher priority than the BASE64 flag
  * which in turn has a gight priority than the AUTODETECT flag.
  */
 gpg_error_t
 gnupg_ksba_create_reader (gnupg_ksba_io_t *ctx,
                           unsigned int flags, estream_t fp,
                           ksba_reader_t *r_reader)
 {
   int rc;
   ksba_reader_t r;
 
   *r_reader = NULL;
   *ctx = xtrycalloc (1, sizeof **ctx);
   if (!*ctx)
     return out_of_core ();
   (*ctx)->u.rparm.allow_multi_pem = !!(flags & GNUPG_KSBA_IO_MULTIPEM);
 
   rc = ksba_reader_new (&r);
   if (rc)
     {
       xfree (*ctx); *ctx = NULL;
       return rc;
     }
 
   (*ctx)->u.rparm.fp = fp;
   if ((flags & GNUPG_KSBA_IO_PEM))
     {
       (*ctx)->u.rparm.assume_pem = 1;
       (*ctx)->u.rparm.assume_base64 = 1;
       rc = ksba_reader_set_cb (r, base64_reader_cb, &(*ctx)->u.rparm);
     }
   else if ((flags & GNUPG_KSBA_IO_BASE64))
     {
       (*ctx)->u.rparm.assume_base64 = 1;
       rc = ksba_reader_set_cb (r, base64_reader_cb, &(*ctx)->u.rparm);
     }
   else if ((flags & GNUPG_KSBA_IO_AUTODETECT))
     {
       (*ctx)->u.rparm.autodetect = 1;
       rc = ksba_reader_set_cb (r, base64_reader_cb, &(*ctx)->u.rparm);
     }
   else
       rc = ksba_reader_set_cb (r, simple_reader_cb, &(*ctx)->u.rparm);
 
   if (rc)
     {
       ksba_reader_release (r);
       xfree (*ctx); *ctx = NULL;
       return rc;
     }
 
   (*ctx)->u2.reader = r;
   *r_reader = r;
   return 0;
 }
 
 
 /* Return True if an EOF as been seen.  */
 int
 gnupg_ksba_reader_eof_seen (gnupg_ksba_io_t ctx)
 {
   return ctx && ctx->u.rparm.eof_seen;
 }
 
 
 /* Destroy a reader object.  */
 void
 gnupg_ksba_destroy_reader (gnupg_ksba_io_t ctx)
 {
   if (!ctx)
     return;
 
   ksba_reader_release (ctx->u2.reader);
   xfree (ctx);
 }
 
 
 
 /* Create a writer for the given STREAM.  Depending on FLAGS an output
  * encoding is chosen.  In PEM mode PEM_NAME is used for the header
  * and footer lines; if PEM_NAME is NULL the string "CMS OBJECT" is
  * used.
  *
  * The function returns a gnupg_ksba_io_t object which must be passed to
  * the gpgme_destroy_writer function.  The created ksba_writer_t
  * object is stored at R_WRITER - the caller must not call the
  * ksba_reader_release function on it.
  *
  * The supported flags are:
  *
  * GNUPG_KSBA_IO_PEM    - Write output as PEM
  * GNUPG_KSBA_IO_BASE64 - Write output as plain Base64; note that the PEM
  *                        flag overrides this flag.
  *
  */
 gpg_error_t
 gnupg_ksba_create_writer (gnupg_ksba_io_t *ctx, unsigned int flags,
                           const char *pem_name, estream_t stream,
                           ksba_writer_t *r_writer)
 {
   int rc;
   ksba_writer_t w;
 
   *r_writer = NULL;
   *ctx = xtrycalloc (1, sizeof **ctx);
   if (!*ctx)
     return gpg_error_from_syserror ();
 
   rc = ksba_writer_new (&w);
   if (rc)
     {
       xfree (*ctx); *ctx = NULL;
       return rc;
     }
 
   if ((flags & GNUPG_KSBA_IO_PEM) || (flags & GNUPG_KSBA_IO_BASE64))
     {
       (*ctx)->u.wparm.stream = stream;
       if ((flags & GNUPG_KSBA_IO_PEM))
         {
           (*ctx)->u.wparm.pem_name = xtrystrdup (pem_name
                                                  ? pem_name
                                                  : "CMS OBJECT");
           if (!(*ctx)->u.wparm.pem_name)
             {
               rc = gpg_error_from_syserror ();
               ksba_writer_release (w);
               xfree (*ctx); *ctx = NULL;
               return rc;
             }
         }
       rc = ksba_writer_set_cb (w, base64_writer_cb, &(*ctx)->u.wparm);
     }
   else if (stream)
     {
       (*ctx)->u.wparm.stream = stream;
       rc = ksba_writer_set_cb (w, plain_writer_cb, &(*ctx)->u.wparm);
     }
   else
     rc = gpg_error (GPG_ERR_INV_ARG);
 
   if (rc)
     {
       ksba_writer_release (w);
       xfree (*ctx); *ctx = NULL;
       return rc;
     }
 
   (*ctx)->u2.writer = w;
   *r_writer = w;
   return 0;
 }
 
 
 /* Flush a writer.  This is for example required to write the padding
  * or the PEM footer.  */
 gpg_error_t
 gnupg_ksba_finish_writer (gnupg_ksba_io_t ctx)
 {
   struct writer_cb_parm_s *parm;
 
   if (!ctx)
     return gpg_error (GPG_ERR_INV_VALUE);
   parm = &ctx->u.wparm;
   if (parm->did_finish)
     return 0; /* Already done. */
   parm->did_finish = 1;
   if (!parm->stream)
     return 0; /* Callback was not used.  */
   return base64_finish_write (parm);
 }
 
 
 /* Destroy a writer object.  */
 void
 gnupg_ksba_destroy_writer (gnupg_ksba_io_t ctx)
 {
   if (!ctx)
     return;
 
   ksba_writer_release (ctx->u2.writer);
   xfree (ctx->u.wparm.pem_name);
   xfree (ctx);
 }