On Wed, May 5, 2021 at 3:19 PM Alex Bennée <alex.ben...@linaro.org> wrote:
>
> Mahmoud Mandour <ma.mando...@gmail.com> writes:
>
> > Added a table-like output which contains the total number of calls
> > for each used syscall along with the number of errors that occurred.
> >
> > Per-call tracing is still available through supplying the argument
> > ``print`` to the plugin.
> >
> > Signed-off-by: Mahmoud Mandour <ma.mando...@gmail.com>
> > ---
> <snip>
> > +
> > +void print_entry(gpointer val, gpointer user_data)
> > {
> > g_autofree gchar *out;
> > - out = g_strdup_printf("syscall #%" PRIi64 " returned -> %" PRIi64
> "\n",
> > - num, ret);
> > + SyscallStats *entry = (SyscallStats *) val;
> > + int64_t syscall_num = entry->num;
> > + out = g_strdup_printf(
> > + "%-13" PRIi64 "%-6" PRIi64 " %" PRIi64 "\n",
> > + syscall_num, entry->calls, entry->errors);
> > qemu_plugin_outs(out);
> > }
>
> This still fails to compile due to a missing static:
>
> [2/10] Compiling C object tests/plugin/libsyscall.so.p/syscall.c.o
> FAILED: tests/plugin/libsyscall.so.p/syscall.c.o
> cc -Itests/plugin/libsyscall.so.p -Itests/plugin -I../../tests/plugin
> -I../../include/qemu -I/usr/include/glib-2.0
> -I/usr/lib/x86_64-linux-gnu/glib-2.0/include -fdiagnostics-color=auto -pipe
> -Wall -Winvalid-pch -Werror -std=gnu99 -O2 -g -isystem
> /home/alex/lsrc/qemu.git/linux-headers -isystem linux-headers -iquote .
> -iquote /home/alex/lsrc/qemu.git -iquote /home/alex/lsrc/qemu.git/include
> -iquote /home/alex/lsrc/qemu.git/disas/libvixl -iquote
> /home/alex/lsrc/qemu.git/tcg/i386 -pthread -U_FORTIFY_SOURCE
> -D_FORTIFY_SOURCE=2 -m64 -mcx16 -D_GNU_SOURCE -D_FILE_OFFSET_BITS=64
> -D_LARGEFILE_SOURCE -Wstrict-prototypes -Wredundant-decls -Wundef
> -Wwrite-strings -Wmissing-prototypes -fno-strict-aliasing -fno-common
> -fwrapv -Wold-style-declaration -Wold-style-definition -Wtype-limits
> -Wformat-security -Wformat-y2k -Winit-self -Wignored-qualifiers
> -Wempty-body -Wnested-externs -Wendif-labels -Wexpansion-to-defined
> -Wimplicit-fallthrough=2 -Wno-missing-include-dirs
> -Wno-shift-negative-value -Wno-psabi -fstack-protector-strong
> -DLEGACY_RDMA_REG_MR -fPIC -MD -MQ tests/plugin/libsyscall.so.p/syscall.c.o
> -MF tests/plugin/libsyscall.so.p/syscall.c.o.d -o
> tests/plugin/libsyscall.so.p/syscall.c.o -c ../../tests/plugin/syscall.c
> ../../tests/plugin/syscall.c:80:6: error: no previous prototype for
> ‘print_entry’ [-Werror=missing-prototypes]
> void print_entry(gpointer val, gpointer user_data)
> ^~~~~~~~~~~
> cc1: all warnings being treated as errors
> ninja: build stopped: subcommand failed.
> make: *** [Makefile:152: run-ninja] Error 1
>
> Thanks for the notice. I got it working because I was building it using
the makefile in the
contrib/plugins makefile and it does not report warnings. I fixed it.
> >
> > +static gint comp_func(gconstpointer ea, gconstpointer eb)
> > +{
> > + SyscallStats *ent_a = (SyscallStats *) ea;
> > + SyscallStats *ent_b = (SyscallStats *) eb;
> > +
> > + return ent_a->calls > ent_b->calls ? -1 : 1;
> > +}
> > +
> > /*
> ************************************************************************* */
> > +static void plugin_exit(qemu_plugin_id_t id, void *p)
> > +{
> > + if (!statistics) {
> > + return;
> > + }
> > +
> > + g_mutex_lock(&lock);
> > + GList *entries = g_hash_table_get_values(statistics);
> > + entries = g_list_sort(entries, comp_func);
> > + qemu_plugin_outs("syscall no. calls errors\n");
> >
> > -static void plugin_exit(qemu_plugin_id_t id, void *p) {}
> > + g_list_foreach(entries, print_entry, NULL);
> > +
> > + g_list_free(entries);
> > + g_hash_table_destroy(statistics);
> > + g_mutex_unlock(&lock);
> > +}
>
> Hmm it looks like we see multiple plugin_exit's when running multiple
> threads:
>
> ./qemu-aarch64 -d plugin -D output -plugin ./tests/plugin/libsyscall.so
> ./tests/tcg/aarch64-linux-user/linux-test
>
> I don't see it with testthread or the new signals test though which is
> confusing. Something linux-test is doing must be different.
>
The output of other plugins is similar so apparently there's something
different with linux-test, yes.
Since this is not a bug related to this particular patch, should I fix the
compilation error and resend the patch?
>
> >
> > QEMU_PLUGIN_EXPORT int qemu_plugin_install(qemu_plugin_id_t id,
> > const qemu_info_t *info,
> > int argc, char **argv)
> > {
> > + if (argc == 0) {
> > + statistics = g_hash_table_new_full(NULL, g_direct_equal, NULL,
> g_free);
> > + } else {
> > + for (int i = 0; i < argc; i++) {
> > + if (g_strcmp0(argv[i], "print") != 0) {
> > + fprintf(stderr, "unsupported argument: %s\n", argv[i]);
> > + return -1;
> > + }
> > + }
> > + }
> > +
> > qemu_plugin_register_vcpu_syscall_cb(id, vcpu_syscall);
> > qemu_plugin_register_vcpu_syscall_ret_cb(id, vcpu_syscall_ret);
> > qemu_plugin_register_atexit_cb(id, plugin_exit, NULL);
>
> Otherwise looking pretty good - certainly a more useful default ;-)
>
> Reviewed-by: Alex Bennée <alex.ben...@linaro.org>
>
>
> --
> Alex Bennée
>
Thanks,
Mahmoud
