From: Namhyung Kim <namhyung@kernel.org>
To: Arnaldo Carvalho de Melo <acme@kernel.org>,
Ian Rogers <irogers@google.com>
Cc: Jiri Olsa <jolsa@kernel.org>,
Adrian Hunter <adrian.hunter@intel.com>,
Peter Zijlstra <peterz@infradead.org>,
Ingo Molnar <mingo@kernel.org>,
LKML <linux-kernel@vger.kernel.org>,
linux-perf-users@vger.kernel.org,
Linus Torvalds <torvalds@linux-foundation.org>,
Stephane Eranian <eranian@google.com>,
Masami Hiramatsu <mhiramat@kernel.org>,
linux-toolchains@vger.kernel.org,
linux-trace-devel@vger.kernel.org
Subject: [PATCH 17/23] perf annotate-data: Handle this-cpu variables in kernel
Date: Mon, 18 Mar 2024 22:51:09 -0700 [thread overview]
Message-ID: <20240319055115.4063940-18-namhyung@kernel.org> (raw)
In-Reply-To: <20240319055115.4063940-1-namhyung@kernel.org>
On x86, the kernel gets the current task using the current macro like
below:
#define current get_current()
static __always_inline struct task_struct *get_current(void)
{
return this_cpu_read_stable(pcpu_hot.current_task);
}
So it returns the current_task field of struct pcpu_hot which is the
first member. On my build, it's located at 0x32940.
$ nm vmlinux | grep pcpu_hot
0000000000032940 D pcpu_hot
And the current macro generates the instructions like below:
mov %gs:0x32940, %rcx
So the %gs segment register points to the beginning of the per-cpu
region of this cpu and it points the variable with a constant.
Let's update the instruction location info to have a segment register
and handle %gs in kernel to look up a global variable. Pretend it as
a global variable by changing the register number to DWARF_REG_PC.
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
---
tools/perf/util/annotate-data.c | 21 +++++++++++++++++++--
tools/perf/util/annotate.c | 7 +++++++
2 files changed, 26 insertions(+), 2 deletions(-)
diff --git a/tools/perf/util/annotate-data.c b/tools/perf/util/annotate-data.c
index f5329a78a97d..d57622ddd5d3 100644
--- a/tools/perf/util/annotate-data.c
+++ b/tools/perf/util/annotate-data.c
@@ -790,7 +790,7 @@ static void delete_var_types(struct die_var_type *var_types)
/* It's at the target address, check if it has a matching type */
static bool check_matching_type(struct type_state *state,
struct data_loc_info *dloc, int reg,
- Dwarf_Die *type_die)
+ Dwarf_Die *cu_die, Dwarf_Die *type_die)
{
Dwarf_Word size;
u32 insn_offset = dloc->ip - dloc->ms->sym->start;
@@ -864,6 +864,23 @@ static bool check_matching_type(struct type_state *state,
return true;
}
+ if (map__dso(dloc->ms->map)->kernel && arch__is(dloc->arch, "x86")) {
+ u64 addr;
+ int offset;
+
+ if (dloc->op->segment == INSN_SEG_X86_GS && dloc->op->imm) {
+ pr_debug_dtp(" this-cpu var\n");
+
+ addr = dloc->op->offset;
+
+ if (get_global_var_type(cu_die, dloc, dloc->ip, addr,
+ &offset, type_die)) {
+ dloc->type_offset = offset;
+ return true;
+ }
+ }
+ }
+
pr_debug_dtp("\n");
return false;
}
@@ -897,7 +914,7 @@ static bool find_data_type_insn(struct data_loc_info *dloc, int reg,
if (this_ip == dloc->ip) {
found = check_matching_type(&state, dloc, reg,
- type_die);
+ cu_die, type_die);
goto out;
}
diff --git a/tools/perf/util/annotate.c b/tools/perf/util/annotate.c
index 3aa3a3b987ad..e4121acb4f88 100644
--- a/tools/perf/util/annotate.c
+++ b/tools/perf/util/annotate.c
@@ -3921,6 +3921,13 @@ struct annotated_data_type *hist_entry__get_data_type(struct hist_entry *he)
op_loc->offset, dl);
}
+ /* This CPU access in kernel - pretend PC-relative addressing */
+ if (map__dso(ms->map)->kernel && arch__is(arch, "x86") &&
+ op_loc->segment == INSN_SEG_X86_GS && op_loc->imm) {
+ dloc.var_addr = op_loc->offset;
+ op_loc->reg1 = DWARF_REG_PC;
+ }
+
mem_type = find_data_type(&dloc);
if (mem_type)
istat->good++;
--
2.44.0.291.gc1ea87d7ee-goog
next prev parent reply other threads:[~2024-03-19 5:51 UTC|newest]
Thread overview: 43+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-03-19 5:50 [PATCHSET 00/23] Remaining bits of data type profiling (v7) Namhyung Kim
2024-03-19 5:50 ` [PATCH 01/23] perf dwarf-aux: Remove unused pc argument Namhyung Kim
2024-03-19 13:43 ` Arnaldo Carvalho de Melo
2024-03-19 17:39 ` Namhyung Kim
2024-03-19 5:50 ` [PATCH 02/23] perf dwarf-aux: Add die_collect_vars() Namhyung Kim
2024-03-19 13:45 ` Arnaldo Carvalho de Melo
2024-03-19 5:50 ` [PATCH 03/23] perf dwarf-aux: Handle type transfer for memory access Namhyung Kim
2024-03-19 13:55 ` Arnaldo Carvalho de Melo
2024-03-19 17:41 ` Namhyung Kim
2024-03-19 5:50 ` [PATCH 04/23] perf dwarf-aux: Add die_find_func_rettype() Namhyung Kim
2024-03-19 13:56 ` Arnaldo Carvalho de Melo
2024-03-19 17:42 ` Namhyung Kim
2024-03-19 18:19 ` Arnaldo Carvalho de Melo
2024-03-19 20:33 ` Namhyung Kim
2024-03-19 5:50 ` [PATCH 05/23] perf map: Add map__objdump_2rip() Namhyung Kim
2024-03-19 5:50 ` [PATCH 06/23] perf annotate-data: Introduce struct data_loc_info Namhyung Kim
2024-03-19 5:50 ` [PATCH 07/23] perf annotate: Add annotate_get_basic_blocks() Namhyung Kim
2024-03-19 5:51 ` [PATCH 08/23] perf annotate-data: Add debug messages Namhyung Kim
2024-03-19 14:05 ` Arnaldo Carvalho de Melo
2024-03-19 5:51 ` [PATCH 09/23] perf annotate-data: Maintain variable type info Namhyung Kim
2024-03-19 14:07 ` Arnaldo Carvalho de Melo
2024-03-19 17:44 ` Namhyung Kim
2024-03-19 18:12 ` Arnaldo Carvalho de Melo
2024-03-19 20:34 ` Namhyung Kim
2024-03-19 5:51 ` [PATCH 10/23] perf annotate-data: Add update_insn_state() Namhyung Kim
2024-03-19 5:51 ` [PATCH 11/23] perf annotate-data: Add get_global_var_type() Namhyung Kim
2024-03-19 5:51 ` [PATCH 12/23] perf annotate-data: Handle global variable access Namhyung Kim
2024-03-19 5:51 ` [PATCH 13/23] perf annotate-data: Handle call instructions Namhyung Kim
2024-03-19 5:51 ` [PATCH 14/23] perf annotate-data: Implement instruction tracking Namhyung Kim
2024-03-19 5:51 ` [PATCH 15/23] perf annotate-data: Check register state for type Namhyung Kim
2024-03-19 5:51 ` [PATCH 16/23] perf annotate: Parse x86 segment register location Namhyung Kim
2024-03-19 5:51 ` Namhyung Kim [this message]
2024-03-19 5:51 ` [PATCH 18/23] perf annotate-data: Track instructions with a this-cpu variable Namhyung Kim
2024-03-19 5:51 ` [PATCH 19/23] perf annotate-data: Support general per-cpu access Namhyung Kim
2024-03-19 5:51 ` [PATCH 20/23] perf annotate-data: Handle ADD instructions Namhyung Kim
2024-03-19 5:51 ` [PATCH 21/23] perf annotate-data: Add stack canary type Namhyung Kim
2024-03-19 5:51 ` [PATCH 22/23] perf annotate-data: Add a cache for global variable types Namhyung Kim
2024-03-19 18:05 ` Namhyung Kim
2024-03-19 18:07 ` Arnaldo Carvalho de Melo
2024-03-19 18:09 ` Arnaldo Carvalho de Melo
2024-03-19 18:09 ` Namhyung Kim
2024-03-19 5:51 ` [PATCH 23/23] perf annotate-data: Do not retry for invalid types Namhyung Kim
2024-03-19 14:17 ` [PATCHSET 00/23] Remaining bits of data type profiling (v7) Arnaldo Carvalho de Melo
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20240319055115.4063940-18-namhyung@kernel.org \
--to=namhyung@kernel.org \
--cc=acme@kernel.org \
--cc=adrian.hunter@intel.com \
--cc=eranian@google.com \
--cc=irogers@google.com \
--cc=jolsa@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-perf-users@vger.kernel.org \
--cc=linux-toolchains@vger.kernel.org \
--cc=linux-trace-devel@vger.kernel.org \
--cc=mhiramat@kernel.org \
--cc=mingo@kernel.org \
--cc=peterz@infradead.org \
--cc=torvalds@linux-foundation.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).