Attached Hurd patch
El dom., 27 oct. 2019 a las 20:13, Almudena Garcia (<
liberamenso10...@gmail.com>) escribió:
> > Please really add it to thread_sched_info_t.
>
> Fixed. Now working in Hurd patch
>
> El dom., 27 oct. 2019 a las 19:56, Samuel Thibault (<
> samuel.thiba...@gnu.org>) escribió:
>
>> Almudena Garcia, le dim. 27 oct. 2019 19:44:23 +0100, a ecrit:
>> > > + #if THREAD_BASIC_INFO_COUNT > 10
>> >
>> > That will be always true. What you want to check is
>> *thread_info_count.
>> >
>> > --- gnumach/kern/thread.c 2019-09-03 01:22:10.932747830 +0200
>> > +++ GNUMach_SMP/kern/thread.c 2019-10-27 19:42:15.360761959 +0100
>> > @@ -1500,11 +1500,11 @@
>> > if (flavor == THREAD_BASIC_INFO) {
>> > thread_basic_info_t basic_info;
>> >
>> > - /* Allow *thread_info_count to be one smaller than the
>> > - usual amount, because creation_time is a new member
>> > - that some callers might not know about. */
>> > + /* Allow *thread_info_count to be two smaller than the
>> > + usual amount, because creation_time and last_processor
>> > + are new members that some callers might not know
>> about. */
>> >
>> > - if (*thread_info_count < THREAD_BASIC_INFO_COUNT - 1) {
>> > + if (*thread_info_count < THREAD_BASIC_INFO_COUNT - 2) {
>> [...]
>> > + /* Check if the basic_info includes all new members
>> (including last_processor) */
>> > + if (*thread_info_count == THREAD_BASIC_INFO_COUNT) {
>>
>> Yes, that's the idea (though you want a >= instead of ==, so that it
>> will work after adding another field).
>>
>> That being said I now realize you added the field to the
>> thread_basic_info_t, while the information really belongs to the
>> thread_sched_info_t. Please really add it to thread_sched_info_t. You
>> have access to it from procfs's process_file_gc_stat by adding
>>
>> thread_basic_info_t thsi = proc_stat_thread_sched_info (ps);
>>
>> Samuel
>>
>>
--- hurd/procfs/process.c 2019-10-26 23:12:40.495359917 +0200
+++ hurd~/procfs/process.c 2019-10-27 20:18:26.114993614 +0100
@@ -221,14 +221,26 @@
struct procinfo *pi = proc_stat_proc_info (ps);
task_basic_info_t tbi = proc_stat_task_basic_info (ps);
thread_basic_info_t thbi = proc_stat_thread_basic_info (ps);
+ thread_sched_info_t thsi = proc_stat_thread_sched_info (ps);
const char *fn = args_filename (proc_stat_args (ps));
+ integer_t last_processor = 0;
+
vm_address_t start_code = 1; /* 0 would make killall5.c consider it
a kernel process, thus use 1 as
default. */
vm_address_t end_code = 1;
process_t p;
error_t err = proc_pid2proc (ps->context->server, ps->pid, &p);
+
+
+ if(sizeof(thsi) >= THREAD_SCHED_INFO_COUNT){
+ last_processor = thsi->last_processor;
+ }
+ else{
+ last_processor = 0;
+ }
+
if (! err)
{
boolean_t essential = 0;
@@ -286,7 +298,7 @@
(long unsigned) proc_stat_thread_rpc (ps), /* close enough */
0L, 0L,
0,
- 0,
+ (long unsigned) last_processor,
0, 0,
0LL);
}
