View previous topic :: View next topic |
Author |
Message |
rmh3093 Advocate
Joined: 06 Aug 2003 Posts: 2138 Location: Albany, NY
|
Posted: Sun Mar 04, 2007 5:23 pm Post subject: 2.6.21-rc2-skunk2.5 "Hindu Kush" :: Genetic RSDL |
|
|
Project Home: http://sourceforge.net/projects/skunk-sources/
layman -a custom-kernels
emerge skunk-sources
Kernel Base: 2.6.21-rc2
Series: http://skunk-sources.svn.sourceforge.net/viewvc/*checkout*/skunk-sources/2.6.21-rc2-skunk2.x/series
Broken-out Patches: http://skunk-sources.svn.sourceforge.net/viewvc/skunk-sources/2.6.21-rc2-skunk2.x/patches/
Release Notes:: | 1] Genetic Ingosched and Genetic Anticipatory I/O work
2] NEW!!! Genetic RSDL
3] RSDL v0.30
4] It is highly recommend that you enable the following kernel config options: HIGH_RES_TIMERS, PREEMPT_RCU, HPET_TIMER, RTC_CLASS, PREEMPT, RTC, HPET
5] Suspend2 is back... please test it out
6] tp_smapi is now inlucded
7] implemented genetic_ns_award_top_parents() in lib/fingerprint.c
8] fixed genetic-rsdl's genes min values |
2.6.21-rc2-skunk2.5: | skunk-sources.patch
#
gentoo_patches.patch
gentoo_patches/4105_dm-bbr.patch
gentoo_patches/4200_fbsplash-0.9.2-r5.patch
gentoo_patches/4205_vesafb-tng-1.0-rc2.patch
gentoo_patches/4300_squashfs-3.2.patch
gentoo_patches/4400_speakup-20070204.patch
gentoo_patches/4405_alpha-sysctl-uac.patch
#
rsdl/create-rsdl.c.patch
rsdl/lists-add_list_splice_tail-1.patch
rsdl/sched-remove_sleepavg_from_proc.patch
rsdl/sched-remove_noninteractive_flag.patch
rsdl/sched-implement_180_bit_sched_bitmap.patch
rsdl/sched-implement_rsdl_cpu_scheduler.patch
rsdl/sched-document_rsdl_cpu_scheduler.patch
rsdl/sched-rsdl-0.26-0.28.patch
rsdl/sched-rsdl-0.28-0.29.patch
rsdl/2.6.21-rc3-mm2-rsdl-0.29-0.30.patch
#
ingosched/create-ingosched.c.patch
ingosched/create-genetic-ingosched.c.patch
#
genetics.patch
genetics/genetic-lib/genetic-lib.patch
genetics/genetic-lib/genetic-debug.patch
genetics/genetic-lib/fingerprinting.patch
genetics/genetic-io/io-fingerprinting.patch
genetics/genetic-io/genetic-io-sched.patch
genetics/genetic-io/create-genetic-as-iosched.c.patch
genetics/genetic-io/genetic-as-sched.patch
genetics/genetic-io/genetic-aio.patch
genetics/genetic-cpu/genetic-cpu-constants.patch
genetics/genetic-cpu/genetic-cpu-sched.patch
genetics/natural-selection-award-top-parents.patch
#
splitsched.patch
splitsched/splitsched-ingosched.patch
splitsched/splitsched-genetic-ingosched.patch
splitsched/splitsched-rsdl.patch
splitsched/splitsched-localversion.patch
#
drivers.patch
drivers/ipw3945-1.2.0_for_2.6.19.patch
drivers/iwlwifi-0.0.11.patch
drivers/linux-phc-0.3.0-pre1-kernel-vanilla-2.6.19-rc1.patch
drivers/tp_smapi-0.31.patch
#
security.patch
security/rt-lsm-0.8.7-kernel.patch
#
misc.patch
misc/kconfig-expose_vmsplit_option.patch
misc/chmp5.patch
misc/hz-raise_max-2.patch
misc/hz-default_864.patch
misc/config-nr-tty-devices-r5.patch
misc/paravirt_ops.patch
misc/badram.patch
#
skunksources.patch
skunksources/daconfig.patch
skunksources/skunk-logo.patch
skunksources/extraversion.patch
#
lkml/cpu_idle-fix-build-break.patch
lkml/log2_up.patch
lkml/i386-alternative_io-implementation.patch
lkml/revert-optimize-and-simplify-get_cycles_sync.patch
lkml/pi-state-locking-fix.patch
lkml/futex-restartable-futex_wait.patch
lkml/kref-refcounting-breakage-in-mainline.patch
lkml/hrtimer-prevent-overrun-DoS-in-hrtimer_forward.patch
lkml/fix-BUG_ON-should-check-end_page-minus-one.patch
lkml/dio-invalidate-clean-pages-before-dio-write.patch
lkml/splice-partial-write-handling-fix.patch
lkml/unconditionally-call-mark_page_accessed.patch
lkml/hrtimer-fixup-unlocked-access-to-wall_to_monotonic.patch
#
genetics/add-genetic-rsdl-localversion.patch
genetics/create-genetic-rsdl.c.patch
genetics/genetic-rsdl.patch
genetics/make-lib-genetic-use-ilog2_up.patch
#
suspend2.patch
suspend2/suspend2-2.2.9.9-for-2.6.21-rc3.patch
suspend2/suspend2-compile-fix.patch
#
skunk-defaults.patch |
_________________ Do not meddle in the affairs of wizards, for they are subtle and quick to anger.
Last edited by rmh3093 on Wed Mar 14, 2007 10:20 pm; edited 21 times in total |
|
Back to top |
|
|
martin.k Guru
Joined: 28 Nov 2004 Posts: 493 Location: Wylatowo, Polska
|
Posted: Sun Mar 04, 2007 7:46 pm Post subject: |
|
|
That's what I like
Is it a rollup from Jake's 2.6.17 original stuff or a brand new stuff from him ???
Gonna try soon _________________ linux-2.6.17 +ck +R4 +lockless +genetic-as +... więcej nie pamiętam
LRU #299256 |
|
Back to top |
|
|
rmh3093 Advocate
Joined: 06 Aug 2003 Posts: 2138 Location: Albany, NY
|
Posted: Sun Mar 04, 2007 8:36 pm Post subject: |
|
|
martin.k wrote: |
That's what I like
Is it a rollup from Jake's 2.6.17 original stuff or a brand new stuff from him ???
Gonna try soon |
no genetic.patch is not a roll up its empty, like origin.patch is in mm-sources, i use it with quilt when i manage my patches, its like a marker i can jump to easily, the important genetic patches are the ones that follow _________________ Do not meddle in the affairs of wizards, for they are subtle and quick to anger. |
|
Back to top |
|
|
rmh3093 Advocate
Joined: 06 Aug 2003 Posts: 2138 Location: Albany, NY
|
Posted: Mon Mar 05, 2007 4:33 pm Post subject: |
|
|
bump.... 2.6.21-rc2-skunk1.1 is out....
and like always, if there is a driver/patch you want included just ask _________________ Do not meddle in the affairs of wizards, for they are subtle and quick to anger. |
|
Back to top |
|
|
tranquilcool Veteran
Joined: 25 Mar 2005 Posts: 1246
|
Posted: Mon Mar 05, 2007 7:35 pm Post subject: |
|
|
Doesn't compile.
this is the error (genetic schedular);
CC kernel/ingosched.o
kernel/ingosched.c: In function âtrigger_load_balanceâ:
kernel/ingosched.c:3541: error: âstruct rqâ has no member named âin_nohz_recentlyâ
kernel/ingosched.c:3541: error: âstruct rqâ has no member named âidle_at_tickâ
kernel/ingosched.c:3542: error: âstruct rqâ has no member named âin_nohz_recentlyâ
kernel/ingosched.c:3544: error: ânohzâ undeclared (first use in this function)
kernel/ingosched.c:3544: error: (Each undeclared identifier is reported only once
kernel/ingosched.c:3544: error: for each function it appears in.)
kernel/ingosched.c:3561: warning: implicit declaration of function âresched_cpuâ
kernel/ingosched.c:3569: error: âstruct rqâ has no member named âidle_at_tickâ
kernel/ingosched.c:3579: error: âstruct rqâ has no member named âidle_at_tickâ
make[1]: *** [kernel/ingosched.o] Error 1
make: *** [kernel] Error 2
OK compiling right now with the tickless timer unset. _________________ this is a strange strange world. |
|
Back to top |
|
|
rmh3093 Advocate
Joined: 06 Aug 2003 Posts: 2138 Location: Albany, NY
|
Posted: Mon Mar 05, 2007 7:47 pm Post subject: |
|
|
tranquilcool wrote: | Doesn't compile.
this is the error (genetic schedular);
CC kernel/ingosched.o
kernel/ingosched.c: In function âtrigger_load_balanceâ:
kernel/ingosched.c:3541: error: âstruct rqâ has no member named âin_nohz_recentlyâ
kernel/ingosched.c:3541: error: âstruct rqâ has no member named âidle_at_tickâ
kernel/ingosched.c:3542: error: âstruct rqâ has no member named âin_nohz_recentlyâ
kernel/ingosched.c:3544: error: ânohzâ undeclared (first use in this function)
kernel/ingosched.c:3544: error: (Each undeclared identifier is reported only once
kernel/ingosched.c:3544: error: for each function it appears in.)
kernel/ingosched.c:3561: warning: implicit declaration of function âresched_cpuâ
kernel/ingosched.c:3569: error: âstruct rqâ has no member named âidle_at_tickâ
kernel/ingosched.c:3579: error: âstruct rqâ has no member named âidle_at_tickâ
make[1]: *** [kernel/ingosched.o] Error 1
make: *** [kernel] Error 2
OK compiling right now with the tickless timer unset. |
are you sure you have 2.6.21-rc2-skunk1.1 because I am 99% sure that compile error should not happen, I think you have skunk1.0 or something, delete anything skunks-sources-2.6.21 related from /usr/portage/distfiles and re-emerge
EDIT: this is straight out of my ingosched.c:
ingosched.c line:402: | #ifdef CONFIG_SMP
unsigned long cpu_load[3];
unsigned char idle_at_tick;
#ifdef CONFIG_NO_HZ
unsigned char in_nohz_recently;
#endif
#endif
|
_________________ Do not meddle in the affairs of wizards, for they are subtle and quick to anger. |
|
Back to top |
|
|
tranquilcool Veteran
Joined: 25 Mar 2005 Posts: 1246
|
Posted: Mon Mar 05, 2007 8:44 pm Post subject: |
|
|
EFLAGS: 00010286 (2.6.21-rc2-skunk1.1 #2)
EIP is at sysfs_readdir+0x137/0x1e4
eax: 00000000 ebx: e59c1bc0 ecx: ffffffff edx: 00000004
esi: f38dbae4 edi: 00000000 ebp: 00000000 esp: df2dff38
ds: 007b es: 007b fs: 00d8 gs: 0033 ss: 0068
Process ksensors (pid: 9771, ti=df2de000 task=e92ce550 task.ti=df2de000)
Stack: 00000006 00000000 000024ac 00000000 0000000a 78164705 df2dff94 dda19c00
f38dbb0c f38dbae0 00000009 7836f7a0 dda19c00 f38dcd34 f38dcd9c 78164821
df2dff94 78164705 fffffff7 0818d564 dda19c00 00001000 78164a69 0818d5f0
Call Trace:
[<78164705>] filldir+0x0/0xb9
[<78164821>] vfs_readdir+0x63/0x8d
[<78164705>] filldir+0x0/0xb9
[<78164a69>] sys_getdents+0x5f/0x9e
[<78103ada>] sysenter_past_esp+0x5f/0x85
=======================
Code: 04 89 59 0c 8b 76 04 e9 b3 00 00 00 8d 46 fc 89 44 24 24 83 78 14 00 0f 84 a0 00 00 00 e8 79 e8 ff ff 89 c5 31 c0 83 c9 ff 89 ef <f2> ae f7 d1 49 89 4c 24 28 8b 54 24 24 8b 42 20 85 c0 74 08 8b
EIP: [<7818d510>] sysfs_readdir+0x137/0x1e4 SS:ESP 0068:df2dff38
BUG: scheduling while atomic: swapper/0x10000100/0
[<783641a4>] __sched_text_start+0x54/0x5b8
[<78226234>] extract_entropy+0x43/0x89
[<78118c2c>] __cond_resched+0x13/0x30
[<783647d4>] cond_resched+0x23/0x2c
[<781588c2>] __kmalloc+0x38/0x67
[<782160b9>] update_phenotype_top_performer+0x185/0x37c
[<78216f5f>] genetic_switch_child+0x975/0xaa4
[<7812ac7f>] hrtimer_wakeup+0x15/0x18
[<782165ea>] genetic_switch_child+0x0/0xaa4
[<781211cd>] run_timer_softirq+0xe0/0x12f
[<781195be>] scheduler_tick+0x9/0x57
[<7811eb40>] __do_softirq+0x35/0x6b
[<7811eb98>] do_softirq+0x22/0x26
[<7811072e>] smp_apic_timer_interrupt+0x67/0x72
[<7810450c>] apic_timer_interrupt+0x28/0x30
[<78102289>] mwait_idle_with_hints+0x36/0x3a
[<78102374>] cpu_idle+0x39/0x4e
[<78456af1>] start_kernel+0x35c/0x364
[<7845642b>] unknown_bootoption+0x0/0x202
=======================
am pretty sure i have the latest skunk-sources. this are some of the dmesg messages.
any clue? _________________ this is a strange strange world. |
|
Back to top |
|
|
rmh3093 Advocate
Joined: 06 Aug 2003 Posts: 2138 Location: Albany, NY
|
Posted: Mon Mar 05, 2007 10:16 pm Post subject: |
|
|
tranquilcool wrote: | EFLAGS: 00010286 (2.6.21-rc2-skunk1.1 #2)
EIP is at sysfs_readdir+0x137/0x1e4
eax: 00000000 ebx: e59c1bc0 ecx: ffffffff edx: 00000004
esi: f38dbae4 edi: 00000000 ebp: 00000000 esp: df2dff38
ds: 007b es: 007b fs: 00d8 gs: 0033 ss: 0068
Process ksensors (pid: 9771, ti=df2de000 task=e92ce550 task.ti=df2de000)
Stack: 00000006 00000000 000024ac 00000000 0000000a 78164705 df2dff94 dda19c00
f38dbb0c f38dbae0 00000009 7836f7a0 dda19c00 f38dcd34 f38dcd9c 78164821
df2dff94 78164705 fffffff7 0818d564 dda19c00 00001000 78164a69 0818d5f0
Call Trace:
[<78164705>] filldir+0x0/0xb9
[<78164821>] vfs_readdir+0x63/0x8d
[<78164705>] filldir+0x0/0xb9
[<78164a69>] sys_getdents+0x5f/0x9e
[<78103ada>] sysenter_past_esp+0x5f/0x85
=======================
Code: 04 89 59 0c 8b 76 04 e9 b3 00 00 00 8d 46 fc 89 44 24 24 83 78 14 00 0f 84 a0 00 00 00 e8 79 e8 ff ff 89 c5 31 c0 83 c9 ff 89 ef <f2> ae f7 d1 49 89 4c 24 28 8b 54 24 24 8b 42 20 85 c0 74 08 8b
EIP: [<7818d510>] sysfs_readdir+0x137/0x1e4 SS:ESP 0068:df2dff38
BUG: scheduling while atomic: swapper/0x10000100/0
[<783641a4>] __sched_text_start+0x54/0x5b8
[<78226234>] extract_entropy+0x43/0x89
[<78118c2c>] __cond_resched+0x13/0x30
[<783647d4>] cond_resched+0x23/0x2c
[<781588c2>] __kmalloc+0x38/0x67
[<782160b9>] update_phenotype_top_performer+0x185/0x37c
[<78216f5f>] genetic_switch_child+0x975/0xaa4
[<7812ac7f>] hrtimer_wakeup+0x15/0x18
[<782165ea>] genetic_switch_child+0x0/0xaa4
[<781211cd>] run_timer_softirq+0xe0/0x12f
[<781195be>] scheduler_tick+0x9/0x57
[<7811eb40>] __do_softirq+0x35/0x6b
[<7811eb98>] do_softirq+0x22/0x26
[<7811072e>] smp_apic_timer_interrupt+0x67/0x72
[<7810450c>] apic_timer_interrupt+0x28/0x30
[<78102289>] mwait_idle_with_hints+0x36/0x3a
[<78102374>] cpu_idle+0x39/0x4e
[<78456af1>] start_kernel+0x35c/0x364
[<7845642b>] unknown_bootoption+0x0/0x202
=======================
am pretty sure i have the latest skunk-sources. this are some of the dmesg messages.
any clue? |
yeah i get that BUG too with genetic aio, idk if it has to do with the -rc prepatch or an -mm patch or if it has to do with the kobj implementation in genetic-aio or if it has to do with my port of genetic aio to >=2.6.20 (my porting might not be 100% proper); however it dosent crash on my computer so maybe its OK for now, does your system seem stable?, you should be able to use dynticks(no_hz) , i have it enabled on my system...... is your computer a uni-processor machine? _________________ Do not meddle in the affairs of wizards, for they are subtle and quick to anger. |
|
Back to top |
|
|
rmh3093 Advocate
Joined: 06 Aug 2003 Posts: 2138 Location: Albany, NY
|
Posted: Mon Mar 05, 2007 10:26 pm Post subject: |
|
|
tranquilcool wrote: | Doesn't compile.
this is the error (genetic schedular);
CC kernel/ingosched.o
kernel/ingosched.c: In function âtrigger_load_balanceâ:
kernel/ingosched.c:3541: error: âstruct rqâ has no member named âin_nohz_recentlyâ
kernel/ingosched.c:3541: error: âstruct rqâ has no member named âidle_at_tickâ
kernel/ingosched.c:3542: error: âstruct rqâ has no member named âin_nohz_recentlyâ
kernel/ingosched.c:3544: error: ânohzâ undeclared (first use in this function)
kernel/ingosched.c:3544: error: (Each undeclared identifier is reported only once
kernel/ingosched.c:3544: error: for each function it appears in.)
kernel/ingosched.c:3561: warning: implicit declaration of function âresched_cpuâ
kernel/ingosched.c:3569: error: âstruct rqâ has no member named âidle_at_tickâ
kernel/ingosched.c:3579: error: âstruct rqâ has no member named âidle_at_tickâ
make[1]: *** [kernel/ingosched.o] Error 1
make: *** [kernel] Error 2
OK compiling right now with the tickless timer unset. |
try and apply this patch, it should fix your issues with nohz + !smp
EDIT: I revided the patch below so it actually applies
Code: | Index: linux/kernel/ingosched.c
===================================================================
--- linux.orig/kernel/ingosched.c
+++ linux/kernel/ingosched.c
@@ -3313,6 +3313,67 @@ static void run_rebalance_domains(struct
}
#endif
}
+/*
+ * Trigger the SCHED_SOFTIRQ if it is time to do periodic load balancing.
+ *
+ * In case of CONFIG_NO_HZ, this is the place where we nominate a new
+ * idle load balancing owner or decide to stop the periodic load balancing,
+ * if the whole system is idle.
+ */
+static inline void trigger_load_balance(int cpu)
+{
+ struct rq *rq = cpu_rq(cpu);
+#ifdef CONFIG_NO_HZ
+ /*
+ * If we were in the nohz mode recently and busy at the current
+ * scheduler tick, then check if we need to nominate new idle
+ * load balancer.
+ */
+ if (rq->in_nohz_recently && !rq->idle_at_tick) {
+ rq->in_nohz_recently = 0;
+
+ if (atomic_read(&nohz.load_balancer) == cpu) {
+ cpu_clear(cpu, nohz.cpu_mask);
+ atomic_set(&nohz.load_balancer, -1);
+ }
+
+ if (atomic_read(&nohz.load_balancer) == -1) {
+ /*
+ * simple selection for now: Nominate the
+ * first cpu in the nohz list to be the next
+ * ilb owner.
+ *
+ * TBD: Traverse the sched domains and nominate
+ * the nearest cpu in the nohz.cpu_mask.
+ */
+ int ilb = first_cpu(nohz.cpu_mask);
+
+ if (ilb != NR_CPUS)
+ resched_cpu(ilb);
+ }
+ }
+
+ /*
+ * If this cpu is idle and doing idle load balancing for all the
+ * cpus with ticks stopped, is it time for that to stop?
+ */
+ if (rq->idle_at_tick && atomic_read(&nohz.load_balancer) == cpu &&
+ cpus_weight(nohz.cpu_mask) == num_online_cpus()) {
+ resched_cpu(cpu);
+ return;
+ }
+
+ /*
+ * If this cpu is idle and the idle load balancing is done by
+ * someone else, then no need raise the SCHED_SOFTIRQ
+ */
+ if (rq->idle_at_tick && atomic_read(&nohz.load_balancer) != cpu &&
+ cpu_isset(cpu, nohz.cpu_mask))
+ return;
+#endif
+ if (time_after_eq(jiffies, rq->next_balance))
+ raise_softirq(SCHED_SOFTIRQ);
+}
#else
/*
* on UP we do not need to balance between CPUs:
@@ -3523,68 +3584,6 @@ out_unlock:
}
/*
- * Trigger the SCHED_SOFTIRQ if it is time to do periodic load balancing.
- *
- * In case of CONFIG_NO_HZ, this is the place where we nominate a new
- * idle load balancing owner or decide to stop the periodic load balancing,
- * if the whole system is idle.
- */
-static inline void trigger_load_balance(int cpu)
-{
- struct rq *rq = cpu_rq(cpu);
-#ifdef CONFIG_NO_HZ
- /*
- * If we were in the nohz mode recently and busy at the current
- * scheduler tick, then check if we need to nominate new idle
- * load balancer.
- */
- if (rq->in_nohz_recently && !rq->idle_at_tick) {
- rq->in_nohz_recently = 0;
-
- if (atomic_read(&nohz.load_balancer) == cpu) {
- cpu_clear(cpu, nohz.cpu_mask);
- atomic_set(&nohz.load_balancer, -1);
- }
-
- if (atomic_read(&nohz.load_balancer) == -1) {
- /*
- * simple selection for now: Nominate the
- * first cpu in the nohz list to be the next
- * ilb owner.
- *
- * TBD: Traverse the sched domains and nominate
- * the nearest cpu in the nohz.cpu_mask.
- */
- int ilb = first_cpu(nohz.cpu_mask);
-
- if (ilb != NR_CPUS)
- resched_cpu(ilb);
- }
- }
-
- /*
- * If this cpu is idle and doing idle load balancing for all the
- * cpus with ticks stopped, is it time for that to stop?
- */
- if (rq->idle_at_tick && atomic_read(&nohz.load_balancer) == cpu &&
- cpus_weight(nohz.cpu_mask) == num_online_cpus()) {
- resched_cpu(cpu);
- return;
- }
-
- /*
- * If this cpu is idle and the idle load balancing is done by
- * someone else, then no need raise the SCHED_SOFTIRQ
- */
- if (rq->idle_at_tick && atomic_read(&nohz.load_balancer) != cpu &&
- cpu_isset(cpu, nohz.cpu_mask))
- return;
-#endif
- if (time_after_eq(jiffies, rq->next_balance))
- raise_softirq(SCHED_SOFTIRQ);
-}
-
-/*
* This function gets called by the timer code, with HZ frequency.
* We call it with interrupts disabled.
*
|
_________________ Do not meddle in the affairs of wizards, for they are subtle and quick to anger. |
|
Back to top |
|
|
tranquilcool Veteran
Joined: 25 Mar 2005 Posts: 1246
|
Posted: Mon Mar 05, 2007 10:52 pm Post subject: |
|
|
doesn't patch clean
patching file kernel/ingosched.c
Hunk #1 succeeded at 3313 with fuzz 1.
patch unexpectedly ends in middle of line
Hunk #2 FAILED at 3584.
1 out of 2 hunks FAILED -- saving rejects to file kernel/ingosched.c.rej _________________ this is a strange strange world. |
|
Back to top |
|
|
buddabrod Apprentice
Joined: 15 Oct 2006 Posts: 241 Location: Germany
|
Posted: Mon Mar 05, 2007 11:02 pm Post subject: |
|
|
Would be nice to add to the first post, that this applies to -mm kernel, not the mainline kernel |
|
Back to top |
|
|
rmh3093 Advocate
Joined: 06 Aug 2003 Posts: 2138 Location: Albany, NY
|
Posted: Mon Mar 05, 2007 11:05 pm Post subject: |
|
|
tranquilcool wrote: | doesn't patch clean
patching file kernel/ingosched.c
Hunk #1 succeeded at 3313 with fuzz 1.
patch unexpectedly ends in middle of line
Hunk #2 FAILED at 3584.
1 out of 2 hunks FAILED -- saving rejects to file kernel/ingosched.c.rej |
did you get my revised copy of the patch... i will upload the working patch to svn as soon as its back up _________________ Do not meddle in the affairs of wizards, for they are subtle and quick to anger. |
|
Back to top |
|
|
buddabrod Apprentice
Joined: 15 Oct 2006 Posts: 241 Location: Germany
|
Posted: Mon Mar 05, 2007 11:20 pm Post subject: |
|
|
Doesn't compile here at all with rsdl
Code: | CHK include/linux/version.h
CHK include/linux/utsrelease.h
CHK include/linux/compile.h
dnsdomainname: No address associated with name
CC kernel/rsdl.o
In file included from include/linux/mm.h:4,
from kernel/rsdl.c:23:
include/linux/sched.h:207:1: warning: "TASK_DEAD" redefined
include/linux/sched.h:154:1: warning: this is the location of the previous definition
include/linux/sched.h:569:1: warning: "MAX_PRIO" redefined
include/linux/sched.h:539:1: warning: this is the location of the previous definition
kernel/rsdl.c: In function 'trigger_load_balance':
kernel/rsdl.c:3328: error: 'struct rq' has no member named 'in_nohz_recently'
kernel/rsdl.c:3328: error: 'struct rq' has no member named 'idle_at_tick'
kernel/rsdl.c:3329: error: 'struct rq' has no member named 'in_nohz_recently'
kernel/rsdl.c:3331: error: 'nohz' undeclared (first use in this function)
kernel/rsdl.c:3331: error: (Each undeclared identifier is reported only once
kernel/rsdl.c:3331: error: for each function it appears in.)
kernel/rsdl.c:3348: warning: implicit declaration of function 'resched_cpu'
kernel/rsdl.c:3356: error: 'struct rq' has no member named 'idle_at_tick'
kernel/rsdl.c:3366: error: 'struct rq' has no member named 'idle_at_tick'
make[1]: *** [kernel/rsdl.o] Error 1
make: *** [kernel] Error 2
|
Without tickless system it seems to compile atm, but there are lots of warnings: Code: | include/linux/sched.h:207:1: warning: "TASK_DEAD" redefined
include/linux/sched.h:154:1: warning: this is the location of the previous definition
include/linux/sched.h:569:1: warning: "MAX_PRIO" redefined
include/linux/sched.h:539:1: warning: this is the location of the previous definition
|
Edit: Finally it didn't even compile without that stuff, there is the end: Code: | from drivers/media/video/saa7134/saa7134-cards.c:24:
include/linux/sched.h:207:1: warning: "TASK_DEAD" redefined
include/linux/sched.h:154:1: warning: this is the location of the previous definition
include/linux/sched.h:569:1: warning: "MAX_PRIO" redefined
include/linux/sched.h:539:1: warning: this is the location of the previous definition
CC drivers/media/video/saa7134/saa7134-core.o
In file included from include/linux/utsname.h:35,
from include/asm/elf.h:12,
from include/linux/elf.h:7,
from include/linux/module.h:15,
from drivers/media/video/saa7134/saa7134-core.c:25:
include/linux/sched.h:207:1: warning: "TASK_DEAD" redefined
include/linux/sched.h:154:1: warning: this is the location of the previous definition
include/linux/sched.h:569:1: warning: "MAX_PRIO" redefined
include/linux/sched.h:539:1: warning: this is the location of the previous definition
drivers/media/video/saa7134/saa7134-core.c:979:24: error: macro "request_submodules" passed 1 arguments, but takes just 0
drivers/media/video/saa7134/saa7134-core.c: In function 'saa7134_initdev':
drivers/media/video/saa7134/saa7134-core.c:979: error: 'request_submodules' undeclared (first use in this function)
drivers/media/video/saa7134/saa7134-core.c:979: error: (Each undeclared identifier is reported only once
drivers/media/video/saa7134/saa7134-core.c:979: error: for each function it appears in.)
make[4]: *** [drivers/media/video/saa7134/saa7134-core.o] Error 1
make[3]: *** [drivers/media/video/saa7134] Error 2
make[2]: *** [drivers/media/video] Error 2
make[1]: *** [drivers/media] Error 2
make: *** [drivers] Error 2
|
As i said, only with rsdl
Last edited by buddabrod on Mon Mar 05, 2007 11:25 pm; edited 1 time in total |
|
Back to top |
|
|
rmh3093 Advocate
Joined: 06 Aug 2003 Posts: 2138 Location: Albany, NY
|
Posted: Mon Mar 05, 2007 11:25 pm Post subject: |
|
|
buddabrod wrote: | Doesn't compile here at all with rsdl
Code: | CHK include/linux/version.h
CHK include/linux/utsrelease.h
CHK include/linux/compile.h
dnsdomainname: No address associated with name
CC kernel/rsdl.o
In file included from include/linux/mm.h:4,
from kernel/rsdl.c:23:
include/linux/sched.h:207:1: warning: "TASK_DEAD" redefined
include/linux/sched.h:154:1: warning: this is the location of the previous definition
include/linux/sched.h:569:1: warning: "MAX_PRIO" redefined
include/linux/sched.h:539:1: warning: this is the location of the previous definition
kernel/rsdl.c: In function 'trigger_load_balance':
kernel/rsdl.c:3328: error: 'struct rq' has no member named 'in_nohz_recently'
kernel/rsdl.c:3328: error: 'struct rq' has no member named 'idle_at_tick'
kernel/rsdl.c:3329: error: 'struct rq' has no member named 'in_nohz_recently'
kernel/rsdl.c:3331: error: 'nohz' undeclared (first use in this function)
kernel/rsdl.c:3331: error: (Each undeclared identifier is reported only once
kernel/rsdl.c:3331: error: for each function it appears in.)
kernel/rsdl.c:3348: warning: implicit declaration of function 'resched_cpu'
kernel/rsdl.c:3356: error: 'struct rq' has no member named 'idle_at_tick'
kernel/rsdl.c:3366: error: 'struct rq' has no member named 'idle_at_tick'
make[1]: *** [kernel/rsdl.o] Error 1
make: *** [kernel] Error 2
|
Without tickless system it seems to compile atm, but there are lots of warnings: Code: | include/linux/sched.h:207:1: warning: "TASK_DEAD" redefined
include/linux/sched.h:154:1: warning: this is the location of the previous definition
include/linux/sched.h:569:1: warning: "MAX_PRIO" redefined
include/linux/sched.h:539:1: warning: this is the location of the previous definition
|
|
i need to make that patch apply to rsdl.c not just ingosched.c..... give me a few min.... _________________ Do not meddle in the affairs of wizards, for they are subtle and quick to anger. |
|
Back to top |
|
|
buddabrod Apprentice
Joined: 15 Oct 2006 Posts: 241 Location: Germany
|
Posted: Mon Mar 05, 2007 11:26 pm Post subject: |
|
|
Yuppie thank you |
|
Back to top |
|
|
rmh3093 Advocate
Joined: 06 Aug 2003 Posts: 2138 Location: Albany, NY
|
Posted: Tue Mar 06, 2007 12:49 am Post subject: |
|
|
buddabrod wrote: | Yuppie thank you |
bump... 2.6.21-rc2-skunk1.2 should take care off all NO_HZ issues and a few others _________________ Do not meddle in the affairs of wizards, for they are subtle and quick to anger. |
|
Back to top |
|
|
tranquilcool Veteran
Joined: 25 Mar 2005 Posts: 1246
|
Posted: Tue Mar 06, 2007 10:33 am Post subject: |
|
|
rmh3093 wrote: | tranquilcool wrote: | EFLAGS: 00010286 (2.6.21-rc2-skunk1.1 #2)
EIP is at sysfs_readdir+0x137/0x1e4
eax: 00000000 ebx: e59c1bc0 ecx: ffffffff edx: 00000004
esi: f38dbae4 edi: 00000000 ebp: 00000000 esp: df2dff38
ds: 007b es: 007b fs: 00d8 gs: 0033 ss: 0068
Process ksensors (pid: 9771, ti=df2de000 task=e92ce550 task.ti=df2de000)
Stack: 00000006 00000000 000024ac 00000000 0000000a 78164705 df2dff94 dda19c00
f38dbb0c f38dbae0 00000009 7836f7a0 dda19c00 f38dcd34 f38dcd9c 78164821
df2dff94 78164705 fffffff7 0818d564 dda19c00 00001000 78164a69 0818d5f0
Call Trace:
[<78164705>] filldir+0x0/0xb9
[<78164821>] vfs_readdir+0x63/0x8d
[<78164705>] filldir+0x0/0xb9
[<78164a69>] sys_getdents+0x5f/0x9e
[<78103ada>] sysenter_past_esp+0x5f/0x85
=======================
Code: 04 89 59 0c 8b 76 04 e9 b3 00 00 00 8d 46 fc 89 44 24 24 83 78 14 00 0f 84 a0 00 00 00 e8 79 e8 ff ff 89 c5 31 c0 83 c9 ff 89 ef <f2> ae f7 d1 49 89 4c 24 28 8b 54 24 24 8b 42 20 85 c0 74 08 8b
EIP: [<7818d510>] sysfs_readdir+0x137/0x1e4 SS:ESP 0068:df2dff38
BUG: scheduling while atomic: swapper/0x10000100/0
[<783641a4>] __sched_text_start+0x54/0x5b8
[<78226234>] extract_entropy+0x43/0x89
[<78118c2c>] __cond_resched+0x13/0x30
[<783647d4>] cond_resched+0x23/0x2c
[<781588c2>] __kmalloc+0x38/0x67
[<782160b9>] update_phenotype_top_performer+0x185/0x37c
[<78216f5f>] genetic_switch_child+0x975/0xaa4
[<7812ac7f>] hrtimer_wakeup+0x15/0x18
[<782165ea>] genetic_switch_child+0x0/0xaa4
[<781211cd>] run_timer_softirq+0xe0/0x12f
[<781195be>] scheduler_tick+0x9/0x57
[<7811eb40>] __do_softirq+0x35/0x6b
[<7811eb98>] do_softirq+0x22/0x26
[<7811072e>] smp_apic_timer_interrupt+0x67/0x72
[<7810450c>] apic_timer_interrupt+0x28/0x30
[<78102289>] mwait_idle_with_hints+0x36/0x3a
[<78102374>] cpu_idle+0x39/0x4e
[<78456af1>] start_kernel+0x35c/0x364
[<7845642b>] unknown_bootoption+0x0/0x202
=======================
am pretty sure i have the latest skunk-sources. this are some of the dmesg messages.
any clue? |
yeah i get that BUG too with genetic aio, idk if it has to do with the -rc prepatch or an -mm patch or if it has to do with the kobj implementation in genetic-aio or if it has to do with my port of genetic aio to >=2.6.20 (my porting might not be 100% proper); however it dosent crash on my computer so maybe its OK for now, does your system seem stable?, you should be able to use dynticks(no_hz) , i have it enabled on my system...... is your computer a uni-processor machine? |
skunk1.2 compiles fine with the tickless timer and boots.
the errors are still there.
my computer is a uni-processor.
skunk1.1 used to hardlock after a few minutes of use. now let's see if skunk1.2 is going to be more stable.
kernel is really fast. very very fast. am just hoping the errors are ironed out.
thanks for always playing forward with the mm series. good job. _________________ this is a strange strange world. |
|
Back to top |
|
|
tranquilcool Veteran
Joined: 25 Mar 2005 Posts: 1246
|
Posted: Tue Mar 06, 2007 10:56 am Post subject: |
|
|
well skunk1.2(genetic) still hardlocks after say 5 minutes of use.
am still convinced once the problems are ironed out skunk1.2 will be one of the fastest kernel here. _________________ this is a strange strange world. |
|
Back to top |
|
|
buddabrod Apprentice
Joined: 15 Oct 2006 Posts: 241 Location: Germany
|
Posted: Tue Mar 06, 2007 11:38 am Post subject: |
|
|
Here it works without locking or stuff, but the cursor or mouse hangs under heavy i/o load, no matter if i use cfq or deadline or tickless timer +genetic or not. |
|
Back to top |
|
|
rmh3093 Advocate
Joined: 06 Aug 2003 Posts: 2138 Location: Albany, NY
|
Posted: Tue Mar 06, 2007 12:58 pm Post subject: |
|
|
tranquilcool wrote: | well skunk1.2(genetic) still hardlocks after say 5 minutes of use.
am still convinced once the problems are ironed out skunk1.2 will be one of the fastest kernel here. |
the hardlocks I think might be caused by the experimental genetic patches I am using... ( the ones with kobj integration), the genetic patches I am using are the handy work of someone other than Jake.... if I cant get the bug worked out of it I will go back to the last patches Jake released for 2.6.17... do you get hardlocks with with out genetic ingo cpu or just with it _________________ Do not meddle in the affairs of wizards, for they are subtle and quick to anger. |
|
Back to top |
|
|
rmh3093 Advocate
Joined: 06 Aug 2003 Posts: 2138 Location: Albany, NY
|
Posted: Tue Mar 06, 2007 1:31 pm Post subject: |
|
|
buddabrod wrote: | Here it works without locking or stuff, but the cursor or mouse hangs under heavy i/o load, no matter if i use cfq or deadline or tickless timer +genetic or not. |
see if this patch helps fix the heavy i/o load hangs:
Code: | Index: linux/kernel/ingosched.c
===================================================================
--- linux.orig/kernel/ingosched.c
+++ linux/kernel/ingosched.c
@@ -3600,7 +3600,7 @@ void scheduler_tick(void)
update_cpu_clock(p, rq, now);
- if (idle_at_tick)
+ if (!idle_at_tick)
task_running_tick(rq, p);
#ifdef CONFIG_SMP
update_load(rq);
Index: linux/kernel/rsdl.c
===================================================================
--- linux.orig/kernel/rsdl.c
+++ linux/kernel/rsdl.c
@@ -3390,7 +3390,7 @@ void scheduler_tick(void)
update_cpu_clock(p, rq, now);
- if (idle_at_tick)
+ if (!idle_at_tick)
task_running_tick(rq, p);
#ifdef CONFIG_SMP
update_load(rq);
|
_________________ Do not meddle in the affairs of wizards, for they are subtle and quick to anger. |
|
Back to top |
|
|
vipernicus Veteran
Joined: 17 Jan 2005 Posts: 1462 Location: Your College IT Dept.
|
Posted: Tue Mar 06, 2007 2:11 pm Post subject: |
|
|
rmh3093 wrote: | buddabrod wrote: | Here it works without locking or stuff, but the cursor or mouse hangs under heavy i/o load, no matter if i use cfq or deadline or tickless timer +genetic or not. |
see if this patch helps fix the heavy i/o load hangs:
Code: | Index: linux/kernel/ingosched.c
===================================================================
--- linux.orig/kernel/ingosched.c
+++ linux/kernel/ingosched.c
@@ -3600,7 +3600,7 @@ void scheduler_tick(void)
update_cpu_clock(p, rq, now);
- if (idle_at_tick)
+ if (!idle_at_tick)
task_running_tick(rq, p);
#ifdef CONFIG_SMP
update_load(rq);
Index: linux/kernel/rsdl.c
===================================================================
--- linux.orig/kernel/rsdl.c
+++ linux/kernel/rsdl.c
@@ -3390,7 +3390,7 @@ void scheduler_tick(void)
update_cpu_clock(p, rq, now);
- if (idle_at_tick)
+ if (!idle_at_tick)
task_running_tick(rq, p);
#ifdef CONFIG_SMP
update_load(rq);
|
|
What version of rsdl.c are you running, that code isn't correct at all.
Should be:
Code: | update_cpu_clock(p, rq, now);
if (p == rq->idle)
/* Task on the idle queue */
wake_priority_sleeper(rq);
else
task_running_tick(rq, p);
#ifdef CONFIG_SMP
update_load(rq); |
_________________ Viper-Sources Maintainer || nesl247 Projects || vipernicus.org blog |
|
Back to top |
|
|
rmh3093 Advocate
Joined: 06 Aug 2003 Posts: 2138 Location: Albany, NY
|
Posted: Tue Mar 06, 2007 2:50 pm Post subject: |
|
|
vipernicus wrote: | rmh3093 wrote: | buddabrod wrote: | Here it works without locking or stuff, but the cursor or mouse hangs under heavy i/o load, no matter if i use cfq or deadline or tickless timer +genetic or not. |
see if this patch helps fix the heavy i/o load hangs:
Code: | Index: linux/kernel/ingosched.c
===================================================================
--- linux.orig/kernel/ingosched.c
+++ linux/kernel/ingosched.c
@@ -3600,7 +3600,7 @@ void scheduler_tick(void)
update_cpu_clock(p, rq, now);
- if (idle_at_tick)
+ if (!idle_at_tick)
task_running_tick(rq, p);
#ifdef CONFIG_SMP
update_load(rq);
Index: linux/kernel/rsdl.c
===================================================================
--- linux.orig/kernel/rsdl.c
+++ linux/kernel/rsdl.c
@@ -3390,7 +3390,7 @@ void scheduler_tick(void)
update_cpu_clock(p, rq, now);
- if (idle_at_tick)
+ if (!idle_at_tick)
task_running_tick(rq, p);
#ifdef CONFIG_SMP
update_load(rq);
|
|
What version of rsdl.c are you running, that code isn't correct at all.
Should be:
Code: | update_cpu_clock(p, rq, now);
if (p == rq->idle)
/* Task on the idle queue */
wake_priority_sleeper(rq);
else
task_running_tick(rq, p);
#ifdef CONFIG_SMP
update_load(rq); |
|
I think its -mm code...
Quote: | On Tuesday 06 March 2007 18:02, Andrew Morton wrote:
> On Tue, 6 Mar 2007 17:25:36 +1100 Con Kolivas <kernel@kolivas.org> wrote:
> > Signed-off-by: Con Kolivas <kernel@kolivas.org>
> > ---
> > kernel/sched.c | 2 +-
> > 1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > Index: linux-2.6.21-rc2-mm1-base/kernel/sched.c
> > ===================================================================
> > --- linux-2.6.21-rc2-mm1-base.orig/kernel/sched.c 2007-03-06
> > 17:19:17.000000000 +1100 +++
> > linux-2.6.21-rc2-mm1-base/kernel/sched.c 2007-03-06 17:20:40.000000000
> > +1100 @@ -3444,7 +3444,7 @@ void scheduler_tick(void)
> >
> > update_cpu_clock(p, rq, now);
> >
> > - if (idle_at_tick)
> > + if (!idle_at_tick)
> > task_running_tick(rq, p);
> > #ifdef CONFIG_SMP
> > update_load(rq);
>
> Looks right, thanks. The original patch had
>
> - if (p == rq->idle)
> + if (idle_at_tick)
> /* Task on the idle queue */
> wake_priority_sleeper(rq);
> else
> task_running_tick(rq, p);
>
> but it got damaged by smt-nice removal.
I gathered something like that happened. If it wasn't clear this change caused
massive scheduler damage with no cpu accounting whatsoever occurring. I
recommend putting it in your hotfixes/ dir if you're not planning an -mm2
soon. |
EDIT: -mm2 is out so hopefully this is all straightened out _________________ Do not meddle in the affairs of wizards, for they are subtle and quick to anger. |
|
Back to top |
|
|
buddabrod Apprentice
Joined: 15 Oct 2006 Posts: 241 Location: Germany
|
Posted: Tue Mar 06, 2007 3:56 pm Post subject: |
|
|
Thank you, that did the trick and it's working fine here^^ |
|
Back to top |
|
|
rmh3093 Advocate
Joined: 06 Aug 2003 Posts: 2138 Location: Albany, NY
|
Posted: Tue Mar 06, 2007 4:07 pm Post subject: |
|
|
buddabrod wrote: | Thank you, that did the trick and it's working fine here^^ |
haha... my computer didnt hardlock until I applied that patch... weird im testing 2.6.21-rc2-skunk2.0 now.... see what that brings up _________________ Do not meddle in the affairs of wizards, for they are subtle and quick to anger. |
|
Back to top |
|
|
|
|
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum
|
|