From: Tom Tromey <tromey@redhat.com>
To: Binutils Development <binutils@sourceware.org>
Subject: [PATCH 3/5] remove deleted BFDs from the archive cache
Date: Fri, 03 Aug 2012 14:55:00 -0000 [thread overview]
Message-ID: <87txwknhzj.fsf@fleche.redhat.com> (raw)
Right now if you bfd_close a member BFD, the pointer to the BFD
remains in the parent archive's cache. This means that if you try to
re-open that member, BFD will return a stale pointer.
This patch fixes the problem by deleting the entry from the hash.
I could not find a way to look up the member BFD in its parent cache
directly, hence the iteration.
* archive.c (delete_bfd_from_htab): New function.
(_bfd_delete_archive_data): Iterate over the parent archive's
hash table.
---
bfd/archive.c | 29 +++++++++++++++++++++++++++++
1 files changed, 29 insertions(+), 0 deletions(-)
diff --git a/bfd/archive.c b/bfd/archive.c
index 8d02257..f84a8fc 100644
--- a/bfd/archive.c
+++ b/bfd/archive.c
@@ -293,6 +293,24 @@ bfd_set_archive_head (bfd *output_archive, bfd *new_head)
return TRUE;
}
+/* A callback for htab_traverse that finds and clears an entry
+ corresponding to a given BFD. */
+
+static int
+delete_bfd_from_htab (void **slot, void *arg)
+{
+ struct ar_cache *cache = *slot;
+ bfd *abfd = arg;
+
+ if (cache->arbfd == abfd)
+ {
+ *slot = HTAB_DELETED_ENTRY;
+ return 0;
+ }
+
+ return 1;
+}
+
/* Free the archive hash table, if it exists. */
void
@@ -302,6 +320,17 @@ _bfd_delete_archive_data (bfd *abfd)
if (ardata && ardata->cache)
htab_delete (ardata->cache);
+
+ if (abfd->my_archive)
+ {
+ ardata = bfd_ardata (abfd->my_archive);
+ if (ardata && ardata->cache)
+ {
+ /* We have to traverse the hash table because there is no
+ way to find ABFD in it directly. */
+ htab_traverse_noresize (ardata->cache, delete_bfd_from_htab, abfd);
+ }
+ }
}
bfd *
--
1.7.7.6
next reply other threads:[~2012-08-03 14:55 UTC|newest]
Thread overview: 48+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-08-03 14:55 Tom Tromey [this message]
2012-08-03 16:12 ` Alan Modra
2012-08-03 20:56 ` Tom Tromey
2012-08-08 22:47 ` Tom Tromey
2012-08-09 10:35 ` Alan Modra
2012-08-15 20:14 ` H.J. Lu
2012-08-15 21:30 ` Tom Tromey
2012-08-16 14:47 ` Alan Modra
2012-08-16 16:31 ` H.J. Lu
2012-08-16 18:14 ` Tom Tromey
2012-08-16 18:33 ` H.J. Lu
2012-08-16 18:45 ` Tom Tromey
2012-08-16 18:49 ` H.J. Lu
2012-08-16 18:55 ` Tom Tromey
2012-08-16 20:33 ` Tom Tromey
2012-08-16 20:38 ` H.J. Lu
2012-08-16 21:11 ` H.J. Lu
2012-08-16 21:35 ` H.J. Lu
2012-08-17 1:01 ` Alan Modra
2012-08-17 1:06 ` H.J. Lu
2012-08-17 2:44 ` Alan Modra
2012-08-17 4:56 ` Hans-Peter Nilsson
2012-08-17 5:04 ` Alan Modra
2012-08-17 5:08 ` H.J. Lu
2012-08-17 5:30 ` H.J. Lu
2012-08-17 5:42 ` H.J. Lu
2012-08-17 5:45 ` H.J. Lu
2012-08-17 6:01 ` H.J. Lu
2012-08-17 13:10 ` H.J. Lu
2012-08-17 16:50 ` Hans-Peter Nilsson
2012-08-17 16:01 ` Hans-Peter Nilsson
2012-08-17 16:10 ` H.J. Lu
2012-08-17 16:14 ` Hans-Peter Nilsson
2012-08-17 16:13 ` Tom Tromey
2012-08-17 16:26 ` H.J. Lu
2012-08-17 16:51 ` Tom Tromey
2012-08-17 17:11 ` H.J. Lu
2012-08-17 17:22 ` Tom Tromey
2012-08-17 17:22 ` H.J. Lu
2012-08-17 19:03 ` Tom Tromey
2012-08-17 19:13 ` H.J. Lu
2012-08-18 4:37 ` Alan Modra
2012-08-21 9:55 ` Tom Tromey
2012-08-17 1:17 ` Alan Modra
2012-08-17 15:36 ` Tom Tromey
2012-08-17 15:38 ` H.J. Lu
2012-08-17 15:43 ` Tom Tromey
2012-08-17 0:48 ` H.J. Lu
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=87txwknhzj.fsf@fleche.redhat.com \
--to=tromey@redhat.com \
--cc=binutils@sourceware.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).