| Message ID | 20240113-ss-fix-ext-col-disabled-v1-1-cf99a7381dec@kernel.org (mailing list archive) |
|---|---|
| State | Accepted |
| Commit | 87d804ca0854b09c07872e9fd6036bf5d3e0cc85 |
| Delegated to: | Stephen Hemminger |
| Headers | show |
| Series | [iproute2] ss: show extra info when '--processes' is not used | expand |
| Context | Check | Description |
|---|---|---|
| netdev/tree_selection | success | Not a local patch |
On Sat, 13 Jan 2024 18:10:21 +0100 "Matthieu Baerts (NGI0)" <matttbe@kernel.org> wrote: > A recent modification broke "extra" options for all protocols showing > info about the processes when '-p' / '--processes' option was not used > as well. In other words, all the additional bits displayed at the end or > at the next line were no longer printed if the user didn't ask to show > info about processes as well. > > The reason is that, the "current_field" pointer never switched to the > "Ext" column. If the user didn't ask to display the processes, nothing > happened when trying to print extra bits using the "out()" function, > because the current field was still pointing to the "Process" one, now > marked as disabled. > > Before the commit mentioned below, it was not an issue not to switch to > the "Ext" or "Process" columns because they were never marked as > "disabled". > > Here is a quick list of options that were no longer displayed if '-p' / > '--processes' was not set: > > - AF_INET(6): > -o, --options > -e, --extended > --tos > --cgroup > --inet-sockopt > -m, --memory > -i, --info > > - AF_PACKET: > -e, --extended > > - AF_XDP: > -e, --extended > > - AF_UNIX: > -m, --memory > -e, --extended > > - TIPC: > --tipcinfo > > That was just by quickly reading the code, I probably missed some. But > this shows that the impact can be quite important for all scripts using > 'ss' to monitor connections or to report info. > > Fixes: 1607bf53 ("ss: prevent "Process" column from being printed unless requested") > Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org> > --- This needs more review and testing before being merged. "once burned, twice shy"
Hello: This patch was applied to iproute2/iproute2.git (main) by Stephen Hemminger <stephen@networkplumber.org>: On Sat, 13 Jan 2024 18:10:21 +0100 you wrote: > A recent modification broke "extra" options for all protocols showing > info about the processes when '-p' / '--processes' option was not used > as well. In other words, all the additional bits displayed at the end or > at the next line were no longer printed if the user didn't ask to show > info about processes as well. > > The reason is that, the "current_field" pointer never switched to the > "Ext" column. If the user didn't ask to display the processes, nothing > happened when trying to print extra bits using the "out()" function, > because the current field was still pointing to the "Process" one, now > marked as disabled. > > [...] Here is the summary with links: - [iproute2] ss: show extra info when '--processes' is not used https://git.kernel.org/pub/scm/network/iproute2/iproute2.git/commit/?id=87d804ca0854 You are awesome, thank you!
diff --git a/misc/ss.c b/misc/ss.c index 900fefa4..5296cabe 100644 --- a/misc/ss.c +++ b/misc/ss.c @@ -2427,6 +2427,8 @@ static void proc_ctx_print(struct sockstat *s) free(buf); } } + + field_next(); } static void inet_stats_print(struct sockstat *s, bool v6only)
A recent modification broke "extra" options for all protocols showing info about the processes when '-p' / '--processes' option was not used as well. In other words, all the additional bits displayed at the end or at the next line were no longer printed if the user didn't ask to show info about processes as well. The reason is that, the "current_field" pointer never switched to the "Ext" column. If the user didn't ask to display the processes, nothing happened when trying to print extra bits using the "out()" function, because the current field was still pointing to the "Process" one, now marked as disabled. Before the commit mentioned below, it was not an issue not to switch to the "Ext" or "Process" columns because they were never marked as "disabled". Here is a quick list of options that were no longer displayed if '-p' / '--processes' was not set: - AF_INET(6): -o, --options -e, --extended --tos --cgroup --inet-sockopt -m, --memory -i, --info - AF_PACKET: -e, --extended - AF_XDP: -e, --extended - AF_UNIX: -m, --memory -e, --extended - TIPC: --tipcinfo That was just by quickly reading the code, I probably missed some. But this shows that the impact can be quite important for all scripts using 'ss' to monitor connections or to report info. Fixes: 1607bf53 ("ss: prevent "Process" column from being printed unless requested") Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org> --- Notes: Note that this issue has quite an annoying impact on our side with the MPTCP subsystem: because '-p' is not used with 'ss', this commit broke 2 selftests (13 subtests). Also, 'ss' is used in case of errors to help better understanding issues, and it is not so useful if it is missing the most important bits: MPTCP info. I know that typically there is no bug-fix version with IPRoute2, but could you please consider one in this case? That would avoid troubles for those relying on 'ss' for the monitoring or the reporting when this specific version of IPRoute2 is used. In our case, it means we have to patch our selftests in 20+ places to support this "broken" version. Plus making sure this is backported correctly, resolving conflicts if needed, etc. It would be really nice if we could avoid that by making a v6.7.1 version including this fix :) --- misc/ss.c | 2 ++ 1 file changed, 2 insertions(+) --- base-commit: 05a4fc72587fed4ad5a0a93c59394b3e39f30381 change-id: 20240113-ss-fix-ext-col-disabled-3f489367a5e7 Best regards,