Hotplug CPU prep III: daemonize idle tasks
PATCH 1/1 Hotplug CPU prep III: daemonize idle tasks
This patch introduces __daemonize(task), so idle tasks can be detached after cloning (required for late creation of idle tasks).
This is independent of the last two patches.
Name: Daemonize idle task
Author: Rusty Russell
D: This patch allows daemonize() to be called on another process (if
D: not started yet), and calls it on the idle task.
include/linux/sched.h +6 -1
@@ -654,7 +654,12 @@extern void exit_sighand (struct task_struct *);extern void reparent_to_init (void ); -extern void daemonize (void ); +extern void __daemonize (struct task_struct *); +static inline void daemonize (void ) +{ +__daemonize (current); +} +extern task_t *child_reaper;extern int do_execve (char *,char **,char **,struct pt_regs *);
kernel/exit.c +10 -12
@@ -201,32 +201,30 @@* Put all the gunge required to become a kernel thread without * attached user resources in one place where it belongs. */ - -void daemonize (void ) +void __daemonize (struct task_struct *tsk) {struct fs_struct *fs; -/* * If we were started as result of loading a module, close all of the * user space pages. We don't need them, and if we didn't close them * they would be locked into memory. */ -exit_mm (current); +exit_mm (tsk); - current->session =1 ; - current->pgrp =1 ; - current->tty =NULL ; + tsk->session =1 ; + tsk->pgrp =1 ; + tsk->tty =NULL ;/* Become as one with the init task */ -exit_fs (current);/* current->fs->count--; */ +exit_fs (tsk);/* current->fs->count--; */ fs = init_task.fs; - current->fs = fs; + tsk->fs = fs;atomic_inc (&fs->count); -exit_files (current); - current->files = init_task.files; -atomic_inc (¤t->files->count); +exit_files (tsk); + tsk->files = init_task.files; +atomic_inc (&tsk->files->count); }/*
kernel/sched.c +2 -0
@@ -1555,6 +1555,8 @@runqueue_t *idle_rq =cpu_rq (cpu), *rq =cpu_rq (idle->thread_info->cpu);unsigned long flags; +if (idle != &init_task) +__daemonize (idle);__save_flags (flags);__cli ();double_rq_lock (idle_rq, rq);
--
Anyone who quotes me in their sig is an idiot. -- Rusty Russell.
I tried a version of __daemonize in 2.4. It panics in schedule()
prepare_to_switch();
{
struct mm_struct *mm = next->mm;
struct mm_struct *oldmm = prev->active_mm;
if (!mm) {
if (next->active_mm) BUG();I got hit by the BUG() in 2.4, I think prepare_to_switch has moved to specific archs in 2.5. A quick look showed that this issue no longer exists in 2.5.
Comments,
Balbir Singh.
|-----Original Message----- |From: [email protected] |[mailto:[email protected]]On Behalf Of Rusty Russell |Sent: Thursday, May 09, 2002 2:20 PM |To: [email protected] |Cc: [email protected] |Subject: [PATCH] Hotplug CPU prep III: daemonize idle tasks | | |This patch introduces __daemonize(task), so idle tasks can be detached |after cloning (required for late creation of idle tasks). | |This is independent of the last two patches. | |Name: Daemonize idle task |Author: Rusty Russell | |D: This patch allows daemonize() to be called on another process (if |D: not started yet), and calls it on the idle task. | |diff -urN -I \$.*\$ --exclude TAGS -X |/home/rusty/devel/kernel/kernel-patches/current-dontdiff --minimal |working-2.5.14-dofork+clonepid/include/linux/sched.h |tmp/include/linux/sched.h |--- working-2.5.14-dofork+clonepid/include/linux/sched.h Thu |May 9 18:25:21 2002 |+++ tmp/include/linux/sched.h Thu May 9 18:36:17 2002 |@@ -654,7 +654,12 @@ | extern void exit_sighand(struct task_struct *); | | extern void reparent_to_init(void); |-extern void daemonize(void); |+extern void __daemonize(struct task_struct *); |+static inline void daemonize(void) |+{ |+ __daemonize(current); |+} |+ | extern task_t *child_reaper; | | extern int do_execve(char *, char **, char **, struct pt_regs *); |diff -urN -I \$.*\$ --exclude TAGS -X |/home/rusty/devel/kernel/kernel-patches/current-dontdiff --minimal |working-2.5.14-dofork+clonepid/kernel/exit.c tmp/kernel/exit.c |--- working-2.5.14-dofork+clonepid/kernel/exit.c Mon Apr 29 |16:00:29 2002 |+++ tmp/kernel/exit.c Thu May 9 18:36:17 2002 |@@ -201,32 +201,30 @@ | * Put all the gunge required to become a kernel thread without | * attached user resources in one place where it belongs. | */ |- |-void daemonize(void) |+void __daemonize(struct task_struct *tsk) | { | struct fs_struct *fs; | |- | /* | * If we were started as result of loading a module, close |all of the | * user space pages. We don't need them, and if we didn't |close them | * they would be locked into memory. | */ |- exit_mm(current); |+ exit_mm(tsk); | |- current->session = 1; |- current->pgrp = 1; |- current->tty = NULL; |+ tsk->session = 1; |+ tsk->pgrp = 1; |+ tsk->tty = NULL; | | /* Become as one with the init task */ | |- exit_fs(current); /* current->fs->count--; */ |+ exit_fs(tsk); /* current->fs->count--; */ | fs = init_task.fs; |- current->fs = fs; |+ tsk->fs = fs; | atomic_inc(&fs->count); |- exit_files(current); |- current->files = init_task.files; |- atomic_inc(¤t->files->count); |+ exit_files(tsk); |+ tsk->files = init_task.files; |+ atomic_inc(&tsk->files->count); | } | | /* |diff -urN -I \$.*\$ --exclude TAGS -X |/home/rusty/devel/kernel/kernel-patches/current-dontdiff --minimal |working-2.5.14-dofork+clonepid/kernel/sched.c tmp/kernel/sched.c |--- working-2.5.14-dofork+clonepid/kernel/sched.c Wed May 1 |15:09:29 2002 |+++ tmp/kernel/sched.c Thu May 9 18:45:17 2002 |@@ -1555,6 +1555,8 @@ | runqueue_t *idle_rq = cpu_rq(cpu), *rq = |cpu_rq(idle->thread_info->cpu); | unsigned long flags; | |+ if (idle != &init_task) |+ __daemonize(idle); | __save_flags(flags); | __cli(); | double_rq_lock(idle_rq, rq); | |-- | Anyone who quotes me in their sig is an idiot. -- Rusty Russell. |- |To unsubscribe from this list: send the line "unsubscribe linux-kernel" in |the body of a message to [email protected] |More majordomo info at http://vger.kernel.org/majordomo-info.html |Please read the FAQ at http://www.tux.org/lkml/
**************************Disclaimer************************************
Information contained in this E-MAIL being proprietary to Wipro Limited is 'privileged' and 'confidential' and intended for use only by the individual or entity to which it is addressed. You are notified that any use, copying or dissemination of the information contained in the E-MAIL in any manner whatsoever is strictly prohibited.
********************************************************************
In message <[email protected]> you write:
Quoted 10 lines
I tried a version of
__daemonizein 2.4. It panics inschedule()prepare_to_switch(); { struct mm_struct *mm = next->mm; struct mm_struct *oldmm = prev->active_mm; if (!mm) { if (next->active_mm) BUG();
Quoted 3 lines
I got hit by the
BUG()in 2.4, I thinkprepare_to_switchhas moved to specific archs in 2.5. A quick look showed that this issue no longer exists in 2.5.
Uhhh... there are many other problems if you are trying to do this on a "live" process. idle process creation is a special case because it has never been run yet.
This is not a generic "daemonize this task" mechanism! Rusty.
--
Anyone who quotes me in their sig is an idiot. -- Rusty Russell.
Should __daemonize() check this? check if in the arguments tsk == current or tsk == idle_task. This would prevent people from using __daemonize() for the wrong things.
Balbir
|-----Original Message----- |From: [email protected] |[mailto:[email protected]]On Behalf Of Rusty Russell |Sent: Monday, May 13, 2002 7:31 AM |To: BALBIR SINGH |Cc: [email protected]; [email protected] |Subject: Re: [PATCH] Hotplug CPU prep III: daemonize idle tasks | | |In message <[email protected]> you write: |> I tried a version of __daemonize in 2.4. It panics in |> schedule() |> |> prepare_to_switch(); |> { |> struct mm_struct *mm = next->mm; |> struct mm_struct *oldmm = prev->active_mm; |> if (!mm) { |> if (next->active_mm) BUG(); |> |> I got hit by the BUG() in 2.4, I think prepare_to_switch has |> moved to specific archs in 2.5. A quick look showed that |> this issue no longer exists in 2.5. | |Uhhh... there are many other problems if you are trying to do this on |a "live" process. idle process creation is a special case because it |has never been run yet. | |This is not a generic "daemonize this task" mechanism! |Rusty. |-- | Anyone who quotes me in their sig is an idiot. -- Rusty Russell. |- |To unsubscribe from this list: send the line "unsubscribe linux-kernel" in |the body of a message to [email protected] |More majordomo info at http://vger.kernel.org/majordomo-info.html |Please read the FAQ at http://www.tux.org/lkml/
**************************Disclaimer************************************
Information contained in this E-MAIL being proprietary to Wipro Limited is 'privileged' and 'confidential' and intended for use only by the individual or entity to which it is addressed. You are notified that any use, copying or dissemination of the information contained in the E-MAIL in any manner whatsoever is strictly prohibited.
********************************************************************