RHEL Engineering is moving the tracking of its product development work on RHEL 6 through RHEL 9 to Red Hat Jira (issues.redhat.com). If you're a Red Hat customer, please continue to file support cases via the Red Hat customer portal. If you're not, please head to the "RHEL project" in Red Hat Jira and file new tickets here. Individual Bugzilla bugs in the statuses "NEW", "ASSIGNED", and "POST" are being migrated throughout September 2023. Bugs of Red Hat partners with an assigned Engineering Partner Manager (EPM) are migrated in late September as per pre-agreed dates. Bugs against components "kernel", "kernel-rt", and "kpatch" are only migrated if still in "NEW" or "ASSIGNED". If you cannot log in to RH Jira, please consult article #7032570. That failing, please send an e-mail to the RH Jira admins at rh-issues@redhat.com to troubleshoot your issue as a user management inquiry. The email creates a ServiceNow ticket with Red Hat. Individual Bugzilla bugs that are migrated will be moved to status "CLOSED", resolution "MIGRATED", and set with "MigratedToJIRA" in "Keywords". The link to the successor Jira issue will be found under "Links", have a little "two-footprint" icon next to it, and direct you to the "RHEL project" in Red Hat Jira (issue links are of type "https://issues.redhat.com/browse/RHEL-XXXX", where "X" is a digit). This same link will be available in a blue banner at the top of the page informing you that that bug has been migrated.
Bug 1870279 - Transfer via socat fails with openssl enabled
Summary: Transfer via socat fails with openssl enabled
Keywords:
Status: CLOSED ERRATA
Alias: None
Product: Red Hat Enterprise Linux 8
Classification: Red Hat
Component: socat
Version: CentOS Stream
Hardware: x86_64
OS: Linux
low
low
Target Milestone: rc
: 8.0
Assignee: Ondrej Moriš
QA Contact: BaseOS QE Security Team
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2020-08-19 16:08 UTC by Pavel Safronov
Modified: 2021-11-09 20:50 UTC (History)
5 users (show)

Fixed In Version: socat-1.7.4.1-1.el8
Doc Type: No Doc Update
Doc Text:
Clone Of:
Environment:
Last Closed: 2021-11-09 17:48:19 UTC
Type: Bug
Target Upstream Version:
Embargoed:

Attachments (Terms of Use)
don't disable SSL_MODE_AUTO_RETRY (1.40 KB, patch)
2020-08-24 16:36 UTC, Carl George 🤠 		no flags Details | Diff
View All


Links
System ID Private Priority Status Summary Last Updated
Red Hat Issue Tracker CRYPTO-5247 0 None None None 2021-11-09 18:08:00 UTC
Red Hat Product Errata RHBA-2021:4167 0 None None None 2021-11-09 17:48:26 UTC

Description Pavel Safronov 2020-08-19 16:08:07 UTC 
Description of problem:

The receiver side of the transfer with ssl enabled always hits "SSL_read(): Connection reset by peer" when it gets close to EOF and leaves the transfer partially finished.

Version-Release number of selected component (if applicable):
1.7.3.3-2.el8


How reproducible:
Almost always. 90% of my attempts failed


Steps to Reproduce:
1. Run the server `socat openssl-listen:12345,reuseaddr,pf=ip6,cert=server.pem - > /dev/null`
2. Run client `yes | head -n 1000000 | socat  -u - openssl-connect:[::1]:12345,cn=host:/example.com,cert=server.pem,cafile=ca.pem` (on the same machine or somewhere else, doesn't matter)
3. You should see an error like this on the receiving side as a result
2020/08/19 08:52:06 socat[2164968] E SSL_read(): Connection reset by peer


Actual results:
Getting an error "E SSL_read(): Connection reset by peer" and transfer fails


Expected results:
No error, transfer succeeds

Additional info:
Allegedly this is the consequence of the modification of file xio-openssl.c introduced in version 1.7.3.3 which forcefully disables SSL_MODE_AUTO_RETRY. After commenting out line "SSL_CTX_clear_mode(*ctx, SSL_MODE_AUTO_RETRY);" and rebuilding the client socat binary I don't have an error anymore.
Also it works well with version 1.7.3.2 of socat.
Comment 1 Pavel Safronov 2020-08-21 17:58:36 UTC 
This patch below is fixing the problem by rolling back the change. So basically we don't disable SSL_MODE_AUTO_RETRY anymore.

Tests run results:

For both tests I run the following command multiple times and check the server output
yes | head -n 1000000 | socat  -u - openssl-connect:[::1]:12345,cn=host:/example.com,cert=server.pem,cafile=ca.pem


With patch:
# while true; do socat openssl-listen:12345,reuseaddr,pf=ip6,cert=server.pem - > /dev/null; date; done
Fri Aug 21 10:53:11 PDT 2020
Fri Aug 21 10:53:13 PDT 2020
Fri Aug 21 10:53:15 PDT 2020
Fri Aug 21 10:53:17 PDT 2020
Fri Aug 21 10:53:19 PDT 2020
Fri Aug 21 10:53:22 PDT 2020
Fri Aug 21 10:53:24 PDT 2020

Without patch:
# while true; do ./socat openssl-listen:12345,reuseaddr,pf=ip6,cert=server.pem - > /dev/null; date; done
Fri Aug 21 10:50:01 PDT 2020
2020/08/21 10:50:02 socat[976825] E SSL_read(): Connection reset by peer
Fri Aug 21 10:50:02 PDT 2020
2020/08/21 10:50:05 socat[977067] E SSL_read(): Connection reset by peer
Fri Aug 21 10:50:05 PDT 2020
2020/08/21 10:50:07 socat[977300] E SSL_read(): Connection reset by peer
Fri Aug 21 10:50:07 PDT 2020
2020/08/21 10:50:09 socat[977442] E SSL_read(): Connection reset by peer
Fri Aug 21 10:50:09 PDT 2020
2020/08/21 10:50:10 socat[977660] E SSL_read(): Connection reset by peer
Fri Aug 21 10:50:10 PDT 2020
2020/08/21 10:50:12 socat[977718] E SSL_read(): Connection reset by peer
Fri Aug 21 10:50:12 PDT 2020
Fri Aug 21 10:50:14 PDT 2020



Here is the patch:

diff -Naur socat-1.7.3.3-orig/CHANGES socat-1.7.3.3/CHANGES
--- socat-1.7.3.3-orig/CHANGES	2019-04-05 13:10:24.000000000 -0700
+++ socat-1.7.3.3/CHANGES	2020-08-21 09:59:35.233714747 -0700
@@ -79,9 +79,6 @@
 	RES_AAONLY, RES_PRIMARY are deprecated. You can still enable them with
 	configure option --enable-res-deprecated.
 
-	New versions of OpenSSL preset SSL_MODE_AUTO_RETRY which may hang socat.
-	Solution: clear SSL_MODE_AUTO_RETRY when it is set.
-
 	Renamed configure.in to configure.ac and set an appropriate symlink for
 	older environments.
 	Related Gentoo bug 426262: Warning on configure.in
diff -Naur socat-1.7.3.3-orig/xio-openssl.c socat-1.7.3.3/xio-openssl.c
--- socat-1.7.3.3-orig/xio-openssl.c	2019-04-04 01:59:55.000000000 -0700
+++ socat-1.7.3.3/xio-openssl.c	2020-08-21 09:58:27.445138134 -0700
@@ -1023,18 +1023,6 @@
    }
 #endif
 
-   /* It seems that OpenSSL-1.1.1 presets the mode differently.
-      Without correction socat might hang in SSL_read() */
-   {
-      long mode = 0;
-      mode = SSL_CTX_get_mode(*ctx);
-      if (mode & SSL_MODE_AUTO_RETRY) {
-	 Info("SSL_CTX mode has SSL_MODE_AUTO_RETRY set. Correcting..");
-	 Debug1("SSL_CTX_clean_mode(%p, SSL_MODE_AUTO_RETRY)", *ctx);
-	 SSL_CTX_clear_mode(*ctx, SSL_MODE_AUTO_RETRY);
-      }
-   }
-
    if (opt_cafile != NULL || opt_capath != NULL) {
       if (sycSSL_CTX_load_verify_locations(*ctx, opt_cafile, opt_capath) != 1) {
 	 int result;
Comment 2 Pavel Safronov 2020-08-21 18:07:20 UTC 
Created a pull request to centos8 socat repo with the patch https://git.centos.org/rpms/socat/pull-request/1
Comment 3 Carl George 🤠 2020-08-24 16:36:47 UTC 
Created attachment 1712396 [details]
don't disable SSL_MODE_AUTO_RETRY

Here is the patch file from Pavel's pull request.  It is a revert of this upstream commit.

https://repo.or.cz/socat.git/commitdiff/bc3723e970bf5e47d681f719f67899046554459c?hp=7a621dd3a245a46c65b726b86a72564f84dced74
Comment 17 errata-xmlrpc 2021-11-09 17:48:19 UTC 
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory (socat bug fix and enhancement update), and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://access.redhat.com/errata/RHBA-2021:4167
Note You need to log in before you can comment on or make changes to this bug.