| Message ID | 174481479787.2426861.10924329074660376176.stgit@mhiramat.tok.corp.google.com (mailing list archive) |
|---|---|
| State | Superseded |
| Headers | show |
| Series | tracing: Record trace_clock in last_boot_info | expand |
On Wed, 16 Apr 2025 23:46:38 +0900 "Masami Hiramatsu (Google)" <mhiramat@kernel.org> wrote: > Record trace_clock information in the trace_scratch area and show > it via `last_boot_info` so that reader can docode the timestamp > correctly. > With this change, the first line of the last_boot_info becomes > trace_clock which is used when the trace was recorded. E.g. > > /sys/kernel/tracing/instances/boot_mapped # cat last_boot_info > trace_clock: mono > ffffffff81000000 [kernel] > As this will not go in this cycle, which means tools may be built around this. Either we put it at the end, or we make it: # trace_clock: mono ffffffff81000000 [kernel] Where "info" like "trace_clock" will be in the comment section, and all non comments will be addresses. Tooling can then just ignore the comment section, or read it for more information. -- Steve
On Wed, 16 Apr 2025 11:12:27 -0400 Steven Rostedt <rostedt@goodmis.org> wrote: > On Wed, 16 Apr 2025 23:46:38 +0900 > "Masami Hiramatsu (Google)" <mhiramat@kernel.org> wrote: > > > Record trace_clock information in the trace_scratch area and show > > it via `last_boot_info` so that reader can docode the timestamp > > correctly. > > With this change, the first line of the last_boot_info becomes > > trace_clock which is used when the trace was recorded. E.g. > > > > /sys/kernel/tracing/instances/boot_mapped # cat last_boot_info > > trace_clock: mono > > ffffffff81000000 [kernel] > > > > As this will not go in this cycle, which means tools may be built > around this. > > Either we put it at the end, or we make it: > > # trace_clock: mono > ffffffff81000000 [kernel] > > Where "info" like "trace_clock" will be in the comment section, and all > non comments will be addresses. OK. Anyway, as we discussed offline, I'll update it to just set the value to the "trace_clock" file instead of using "last_boot_info". Thank you, > > Tooling can then just ignore the comment section, or read it for more > information. > > -- Steve
diff --git a/include/linux/trace_clock.h b/include/linux/trace_clock.h index 00e8f98c940f..2245848270d2 100644 --- a/include/linux/trace_clock.h +++ b/include/linux/trace_clock.h @@ -15,6 +15,8 @@ #include <asm/trace_clock.h> +#define TRACE_CLOCK_NAME_MAX 16 + extern u64 notrace trace_clock_local(void); extern u64 notrace trace_clock(void); extern u64 notrace trace_clock_jiffies(void); diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 8ddf6b17215c..e59826fa4cb3 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -6004,6 +6004,7 @@ struct trace_mod_entry { }; struct trace_scratch { + char clock[TRACE_CLOCK_NAME_MAX]; unsigned long text_addr; unsigned long nr_entries; struct trace_mod_entry entries[]; @@ -6114,6 +6115,8 @@ static void update_last_data(struct trace_array *tr) if (tr->scratch) { struct trace_scratch *tscratch = tr->scratch; + strscpy(tscratch->clock, trace_clocks[tr->clock_id].name, + TRACE_CLOCK_NAME_MAX); memset(tscratch->entries, 0, flex_array_size(tscratch, entries, tscratch->nr_entries)); tscratch->nr_entries = 0; @@ -7000,9 +7003,10 @@ static void show_last_boot_header(struct seq_file *m, struct trace_array *tr) * Otherwise it shows the KASLR address from the previous boot which * should not be the same as the current boot. */ - if (tscratch && (tr->flags & TRACE_ARRAY_FL_LAST_BOOT)) + if (tscratch && (tr->flags & TRACE_ARRAY_FL_LAST_BOOT)) { + seq_printf(m, "trace_clock: %s\n", tscratch->clock); seq_printf(m, "%lx\t[kernel]\n", tscratch->text_addr); - else + } else seq_puts(m, "# Current\n"); } @@ -7289,6 +7293,12 @@ int tracing_set_clock(struct trace_array *tr, const char *clockstr) tracing_reset_online_cpus(&tr->max_buffer); #endif + if (tr->scratch && !(tr->flags & TRACE_ARRAY_FL_LAST_BOOT)) { + struct trace_scratch *tscratch = tr->scratch; + + strscpy(tscratch->clock, trace_clocks[i].name, TRACE_CLOCK_NAME_MAX); + } + mutex_unlock(&trace_types_lock); return 0;