| From | Greg Kroah-Hartman <> | Subject | [ 202/221] libceph: avoid using freed osd in __kick_osd_requests() | Date | Tue, 15 Jan 2013 10:52:09 -0800 |
| |
3.7-stable review patch. If anyone has any objections, please let me know.
------------------
From: Alex Elder <elder@inktank.com>
(cherry picked from commit 685a7555ca69030739ddb57a47f0ea8ea80196a4)
If an osd has no requests and no linger requests, __reset_osd() will just remove it with a call to __remove_osd(). That drops a reference to the osd, and therefore the osd may have been free by the time __reset_osd() returns. That function offers no indication this may have occurred, and as a result the osd will continue to be used even when it's no longer valid.
Change__reset_osd() so it returns an error (ENODEV) when it deletes the osd being reset. And change __kick_osd_requests() so it returns immediately (before referencing osd again) if __reset_osd() returns *any* error.
Signed-off-by: Alex Elder <elder@inktank.com> Reviewed-by: Sage Weil <sage@inktank.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> --- net/ceph/osd_client.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
--- a/net/ceph/osd_client.c +++ b/net/ceph/osd_client.c @@ -581,7 +581,7 @@ static void __kick_osd_requests(struct c dout("__kick_osd_requests osd%d\n", osd->o_osd); err = __reset_osd(osdc, osd); - if (err == -EAGAIN) + if (err) return; list_for_each_entry(req, &osd->o_requests, r_osd_item) { @@ -752,6 +752,7 @@ static int __reset_osd(struct ceph_osd_c if (list_empty(&osd->o_requests) && list_empty(&osd->o_linger_requests)) { __remove_osd(osdc, osd); + ret = -ENODEV; } else if (memcmp(&osdc->osdmap->osd_addr[osd->o_osd], &osd->o_con.peer_addr, sizeof(osd->o_con.peer_addr)) == 0 &&
|