Bug 2015623
| Summary: | mptcpize fails to force mptcp usage on some programs | ||
|---|---|---|---|
| Product: | Red Hat Enterprise Linux 9 | Reporter: | Paolo Abeni <pabeni> |
| Component: | mptcpd | Assignee: | Davide Caratti <dcaratti> |
| Status: | CLOSED ERRATA | QA Contact: | xmu |
| Severity: | unspecified | Docs Contact: | |
| Priority: | unspecified | ||
| Version: | 9.0 | CC: | chorn, dcaratti, xmu |
| Target Milestone: | rc | Keywords: | Triaged |
| Target Release: | --- | Flags: | pm-rhel:
mirror+
|
| Hardware: | Unspecified | ||
| OS: | Unspecified | ||
| Whiteboard: | |||
| Fixed In Version: | mptcpd-0.8-2.el9 | Doc Type: | If docs needed, set a value |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2022-05-17 12:54:56 UTC | Type: | Bug |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Embargoed: | |||
*** Bug 2027232 has been marked as a duplicate of this bug. *** @xmu, this can be verified by looking whether the listening socket got the TCP_ULP attached. See below: # mptcpize run ncat -l 0.0.0.0 1234 & # ss -santi | grep -A1 1234 LISTEN 0 10 0.0.0.0:1234 0.0.0.0:* cubic cwnd:10 tcp-ulp-mptcp flags:m token:0000(id:0)/0000(id:0) seq:0 sfseq:0 ssnoff:0 maplen:0 <ctrl+c> # mptcpize run iperf3 -s -p 1234 & # ss -santi | grep -A1 1234 LISTEN 0 4096 *:1234 *:* cubic cwnd:10 I'm in the progress of doing a build that backports a fix for this problem, so MPTCP testing with native applications will be much easier. I see you provided 'qa_ack+', can you please set a value >= 16 for the ITM? thanks! set ITM18 > Fixed In Version: mptcpd-0.8-2.el9
Fixes the issue works for me, tnx
case /kernel/networking/mptcp/mptcpd/mptcpd_commands pass with mptcpd-0.8-2.el9 and kernel-5.14.0-21.el9.x86_64 Hi Davide, I tested "mptcpize run" fololowed comment #2, how to test "mptcp enable" for patch "mptcpize: use explicit file copy instead of rename()"? Thanks hello Xiumei, thanks for confirming! (In reply to xmu from comment #9) > Hi Davide, > I tested "mptcpize run" fololowed comment #2, > how to test "mptcp enable" for patch "mptcpize: use explicit file copy > instead of rename()"? > Thanks I *think* it should be sufficient to mptcpize a command that is in '/tmp' rather than '/usr/bin'. On mptcpd-0.8-1.el9 that should fail to mptcpize. 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 (new packages: mptcpd), 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-2022:2417 |
mptcpize run <command> fails if <command> uses 'socket(AF_INET, SOCK_STREAM, IPPROTO_IP)' instead of 'socket(AF_INET, SOCK_STREAM, IPPROTO_TCP)' mptcpize enable <unit> fails if the /tmp directory is located on a different fs WRT the unit file itself. Solved by the following upstream commits: commit 591b3b168d949c45d5b5994332a3007767845434 Author: Paolo Abeni <paolo.abeni> Date: Thu Oct 14 05:05:54 2021 +0200 mptcpize: use explicit file copy instead of rename() (#161) commit 5633f08e35552295b2f9414ff32ca4e8e081b2f2 Author: Paolo Abeni <paolo.abeni> Date: Tue Oct 12 19:24:33 2021 +0200 mptcpize: force MPTCP usage for IPPROTO_IP, too (#159)