kill: add a feature decoding signal masks #3105

masatake · 2024-07-01T11:12:55Z

-d/--decode-mask option works well with the output of ps command:

   $ ps s $$
     UID     PID          PENDING          BLOCKED          IGNORED           CAUGHT STAT TTY        TIME COMMAND
    1000 1588189 0000000000000000 0000000000010000 0000000000384004 000000004b813efb S    pts/56     0:00 bash

   $ ./kill --decode-mask 000000004b813efb
   HUP
   INT
   ILL
   TRAP
   ABRT
   ...

If you know the pid you are interested in, you can skip running ps comman with -D/--decode-process-state option:

   $ ./kill --decode-process-state $$
   SigPnd:
   ShdPnd:
   SigBlk:
	   CHLD
   SigIgn:
	   QUIT
	   TERM
	   TSTP
	   TTIN
	   TTOU
   SigCgt:
	   HUP
	   INT
	   ILL
	   TRAP
	   ABRT
	   BUS
	   FPE
	   USR1
           ...

add test cases

karelzak · 2024-07-01T11:50:24Z

Ah, I like this idea :-)

What do you think about adding support for decoding the masks to the --list? It already supports conversion from signal number to signal name (--list 9) and also supports masks prefixed by 0x seems like a native thing; kill --list 0x000000004b813efb.

And --show-process-state rather than --decode-process-state, and report it in more human-readable way

non-empty items only
use human-readable names for the masks (SigBlk --> Blocking)
not separate signals by \n

$ kill --show-process-state $$`
Blocking: CHLD
Ignoring: QUIT TERM TSTP TTIN TTOU
Catching: HUP INT ILL TRAP ABRT BUS FPE USR1

I can also imagine lssig listing all about signals from /proc :-)))

masatake · 2024-07-01T16:37:43Z

@karelzak Thank you for the comment. I updated the change:

       $ ps s $$
         UID     PID          PENDING          BLOCKED          IGNORED           CAUGHT STAT TTY        TIME COMMAND
        1000 1588189 0000000000000000 0000000000010000 0000000000384004 000000004b813efb S    pts/56     0:00 bash
    
       $ ./kill --list 0x000000004b813efb
       HUP
       INT
       ILL
       TRAP
       ABRT
       ...
       $ ./kill --show-process-state 1
       Blocked: HUP INT USR1 USR2 TERM CHLD WINCH PWR RT0 RT1 RT2 RT3 RT4 RT5 RT6 RT7 RT13 RT14 RT15 RT16 RT17 RT18 RT20 RT21 RT22 RT23 RT24 RT25 RT26 RT27 RT28 RT29 
       Ignored: PIPE 
       Caught: QUIT ILL ABRT BUS FPE SEGV

I can also imagine lssig listing all about signals from /proc :-)))

Eventually, we will want to implement lsps with -o and -Q options.

t-8ch · 2024-07-01T20:00:54Z

misc-utils/kill.c

 			exit(EXIT_SUCCESS);
 		}
 		if (!strcmp(arg, "-L") || !strcmp(arg, "--table")) {
 			print_all_signals(stdout, 1);
 			exit(EXIT_SUCCESS);
 		}
+		if (!strcmp(arg, "-d") || !strcmp(arg, "--show-process-state")) {
+			uint64_t pid;


Why not pid_t?

Thank you. Fixed in locally.

t-8ch · 2024-07-01T20:02:03Z

misc-utils/kill.c

+	FILE *fp;
+	char buf[BUFSIZ];
+
+	const struct sigfield {


Thank you. Fixed in locally.

t-8ch · 2024-07-01T20:02:56Z

misc-utils/kill.c

+
+			if (strncmp(buf, key, keylen) == 0) {
+				char *val = buf + keylen;
+				uint64_t sigmask;


Aren't sigmasks int?

At least on x86_64, int looks too small.

[yamato@dev64]~/var/util-linux/util-linux% cat /tmp/foo.c int main(void) {return sizeof(int);} [yamato@dev64]~/var/util-linux/util-linux% gcc /tmp/foo.c [yamato@dev64]~/var/util-linux/util-linux% ./a.out; echo $? 4 [yamato@dev64]~/var/util-linux/util-linux% ./kill -l 0xffffffffffffffff | cat -n | tail 55 RT21 56 RT22 57 RT23 58 RT24 59 RT25 60 RT26 61 RT27 62 RT28 63 RT29 64 RT30

Am I missing something?

Ah, I looked at sigmask(3) which is deprecated.

pthread_sigmask(3) proposes sigset_t. But those are 128 bytes big on amd64 glibc and a structure to boot.
So a uint64_t is probably the best to work with.

karelzak · 2024-07-02T09:50:51Z

misc-utils/kill.c

+		{ "ShdPnd:\t", "Pending (process)" },
+		{ "SigBlk:\t", "Blocked"           },
+		{ "SigIgn:\t", "Ignored"           },
+		{ "SigCgt:\t", "Caught"            },


Let's translate the strings:

N_("Caught")

Fixed in locally.

karelzak · 2024-07-02T09:51:29Z

misc-utils/kill.c

+					continue;
+				}
+				if (sigmask != 0) {
+					printf("%s: ", sigfields[i].label);


printf("%s: ", _(sigfields[i].label));

Fixed in locally.

karelzak · 2024-07-02T09:56:48Z

misc-utils/kill.c

 	fputs(_(" -L, --table            list signal names and numbers\n"), out);
 	fputs(_(" -r, --require-handler  do not send signal if signal handler is not present\n"), out);
+	fputs(_(" -d, --decode-process-state <pid>\n"


--show-process-state

Fixed in locally.

karelzak · 2024-07-02T10:01:03Z

misc-utils/kill.c

+				errx(EXIT_FAILURE, _("too many arguments"));
+			arg = argv[1];
+			if (ul_strtou64(arg, &pid, 10) < 0)
+				errx(EXIT_FAILURE, _("wrong pid specification: %s"), arg);


pid = strtou64_or_err(argv[1], _("invalid pid argument"));

With the change to pid_t: strtopid_or_err()

I didn't know strtopid_or_err. Thank you.

Fixed in locally.

-l/--list option is extended to work well with the output of ps command: $ ps s $$ UID PID PENDING BLOCKED IGNORED CAUGHT STAT TTY TIME COMMAND 1000 1588189 0000000000000000 0000000000010000 0000000000384004 000000004b813efb S pts/56 0:00 bash $ ./kill --list 0x000000004b813efb HUP INT ILL TRAP ABRT ... If you know the pid you are interested in, you can skip running ps comman with -d/--show-process-state option: $ ./kill --show-process-state $$ Blocked: INT Ignored: TERM TSTP TTIN TTOU Caught: HUP INT PIPE ALRM CHLD WINCH Signed-off-by: Masatake YAMATO <yamato@redhat.com>

masatake · 2024-07-03T21:15:47Z

lsns/filter test case failed on rpm-build:fedora-rawhide-ppc64le. Mysterious. I will record the failure in #2967.

masatake · 2024-07-07T18:02:20Z

Added a test case.

Signed-off-by: Masatake YAMATO <yamato@redhat.com>

masatake force-pushed the kill--decode branch 2 times, most recently from 83d3ae7 to a67a7d6 Compare July 1, 2024 16:27

t-8ch reviewed Jul 1, 2024

View reviewed changes

karelzak reviewed Jul 2, 2024

View reviewed changes

masatake force-pushed the kill--decode branch from a67a7d6 to 42da770 Compare July 2, 2024 14:05

masatake force-pushed the kill--decode branch from 42da770 to 873370b Compare July 3, 2024 03:00

masatake mentioned this pull request Jul 3, 2024

lsns: failure of lsns related test cases #2967

Open

masatake force-pushed the kill--decode branch from 66f39be to fa0efd8 Compare July 7, 2024 17:59

masatake changed the title ~~[RFC] kill: add a feature decoding signal masks~~ kill: add a feature decoding signal masks Jul 7, 2024

masatake marked this pull request as ready for review July 7, 2024 18:01

masatake force-pushed the kill--decode branch 4 times, most recently from 23ba951 to ac801df Compare July 7, 2024 20:48

masatake marked this pull request as draft July 8, 2024 05:44

kill: (test) add a case for testing -l 0xSIGMASK and -d $PID options

e58bd8e

Signed-off-by: Masatake YAMATO <yamato@redhat.com>

masatake force-pushed the kill--decode branch from ac801df to e58bd8e Compare July 8, 2024 08:55

masatake marked this pull request as ready for review July 8, 2024 11:48

karelzak merged commit 30603d6 into util-linux:master Jul 15, 2024
33 checks passed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

kill: add a feature decoding signal masks #3105

kill: add a feature decoding signal masks #3105

masatake commented Jul 1, 2024 •

edited

Loading

karelzak commented Jul 1, 2024

masatake commented Jul 1, 2024

t-8ch Jul 1, 2024

masatake Jul 2, 2024

t-8ch Jul 1, 2024

masatake Jul 2, 2024

t-8ch Jul 1, 2024

masatake Jul 2, 2024

t-8ch Jul 2, 2024

karelzak Jul 2, 2024

masatake Jul 2, 2024

karelzak Jul 2, 2024

masatake Jul 2, 2024

karelzak Jul 2, 2024

masatake Jul 2, 2024

karelzak Jul 2, 2024

t-8ch Jul 2, 2024

masatake Jul 2, 2024

masatake Jul 2, 2024

masatake commented Jul 3, 2024

masatake commented Jul 7, 2024

kill: add a feature decoding signal masks #3105

kill: add a feature decoding signal masks #3105

Conversation

masatake commented Jul 1, 2024 • edited Loading

karelzak commented Jul 1, 2024

masatake commented Jul 1, 2024

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

masatake commented Jul 3, 2024

masatake commented Jul 7, 2024

masatake commented Jul 1, 2024 •

edited

Loading