From f3612304ee04a1a36ded7604771ea56d818158cb Mon Sep 17 00:00:00 2001 From: Heiko Carstens Date: Fri, 17 Feb 2012 10:29:20 +0100 Subject: [PATCH 1/4] [S390] idle: avoid RCU usage in extended quiescent state Avoid calling wake_up() from our NMI "bottom halve" from RCU extended quiescent state in idle. wake_up() has RCU read-side critical sections but this will be completely ignored by RCU if the cpu is in extended quiescent state. Which means that whatever object is being accessed from within the read-side critical section can be freed concurrently from a different cpu. So make sure we leave extended quiescent state before calling wake_up(). Signed-off-by: Heiko Carstens Signed-off-by: Martin Schwidefsky --- arch/s390/kernel/process.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/arch/s390/kernel/process.c b/arch/s390/kernel/process.c index 3201ae447990c..4261aa799774f 100644 --- a/arch/s390/kernel/process.c +++ b/arch/s390/kernel/process.c @@ -76,7 +76,6 @@ static void default_idle(void) if (test_thread_flag(TIF_MCCK_PENDING)) { local_mcck_enable(); local_irq_enable(); - s390_handle_mcck(); return; } trace_hardirqs_on(); @@ -93,10 +92,12 @@ void cpu_idle(void) for (;;) { tick_nohz_idle_enter(); rcu_idle_enter(); - while (!need_resched()) + while (!need_resched() && !test_thread_flag(TIF_MCCK_PENDING)) default_idle(); rcu_idle_exit(); tick_nohz_idle_exit(); + if (test_thread_flag(TIF_MCCK_PENDING)) + s390_handle_mcck(); preempt_enable_no_resched(); schedule(); preempt_disable(); From 2320c5793790fcda80e6dcc088dbda86040235e5 Mon Sep 17 00:00:00 2001 From: Martin Schwidefsky Date: Fri, 17 Feb 2012 10:29:21 +0100 Subject: [PATCH 2/4] [S390] incorrect PageTables counter for kvm page tables The page_table_free_pgste function is used for kvm processes to free page tables that have the pgste extension. It calls pgtable_page_ctor instead of pgtable_page_dtor which increases NR_PAGETABLE instead of decreasing it. Signed-off-by: Martin Schwidefsky --- arch/s390/mm/pgtable.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/arch/s390/mm/pgtable.c b/arch/s390/mm/pgtable.c index 9a4d02f64f16e..51b0738e13d12 100644 --- a/arch/s390/mm/pgtable.c +++ b/arch/s390/mm/pgtable.c @@ -574,7 +574,7 @@ static inline void page_table_free_pgste(unsigned long *table) page = pfn_to_page(__pa(table) >> PAGE_SHIFT); mp = (struct gmap_pgtable *) page->index; BUG_ON(!list_empty(&mp->mapper)); - pgtable_page_ctor(page); + pgtable_page_dtor(page); atomic_set(&page->_mapcount, -1); kfree(mp); __free_page(page); From 656d9125376006cf696b0836f1c6723a892629ca Mon Sep 17 00:00:00 2001 From: Martin Schwidefsky Date: Fri, 17 Feb 2012 10:29:22 +0100 Subject: [PATCH 3/4] [S390] 3215 deadlock with tty_wakeup The 3215 driver calls tty_wakeup from irq context while holding the device spinlock. If printk is called by any function on the callchain starting from tty_wakeup the system deadlocks on the device spinlock. Using a tasklet to call tty_wakup solves the problem. Signed-off-by: Martin Schwidefsky --- drivers/s390/char/con3215.c | 22 +++++++++++++++------- 1 file changed, 15 insertions(+), 7 deletions(-) diff --git a/drivers/s390/char/con3215.c b/drivers/s390/char/con3215.c index 934458ad55e51..e71a50d4b2217 100644 --- a/drivers/s390/char/con3215.c +++ b/drivers/s390/char/con3215.c @@ -87,6 +87,7 @@ struct raw3215_info { struct tty_struct *tty; /* pointer to tty structure if present */ struct raw3215_req *queued_read; /* pointer to queued read requests */ struct raw3215_req *queued_write;/* pointer to queued write requests */ + struct tasklet_struct tlet; /* tasklet to invoke tty_wakeup */ wait_queue_head_t empty_wait; /* wait queue for flushing */ struct timer_list timer; /* timer for delayed output */ int line_pos; /* position on the line (for tabs) */ @@ -333,20 +334,24 @@ static inline void raw3215_try_io(struct raw3215_info *raw) } } +/* + * Call tty_wakeup from tasklet context + */ +static void raw3215_wakeup(unsigned long data) +{ + struct raw3215_info *raw = (struct raw3215_info *) data; + tty_wakeup(raw->tty); +} + /* * Try to start the next IO and wake up processes waiting on the tty. */ static void raw3215_next_io(struct raw3215_info *raw) { - struct tty_struct *tty; - raw3215_mk_write_req(raw); raw3215_try_io(raw); - tty = raw->tty; - if (tty != NULL && - RAW3215_BUFFER_SIZE - raw->count >= RAW3215_MIN_SPACE) { - tty_wakeup(tty); - } + if (raw->tty && RAW3215_BUFFER_SIZE - raw->count >= RAW3215_MIN_SPACE) + tasklet_schedule(&raw->tlet); } /* @@ -682,6 +687,7 @@ static int raw3215_probe (struct ccw_device *cdev) return -ENOMEM; } init_waitqueue_head(&raw->empty_wait); + tasklet_init(&raw->tlet, raw3215_wakeup, (unsigned long) raw); dev_set_drvdata(&cdev->dev, raw); cdev->handler = raw3215_irq; @@ -901,6 +907,7 @@ static int __init con3215_init(void) raw->flags |= RAW3215_FIXED; init_waitqueue_head(&raw->empty_wait); + tasklet_init(&raw->tlet, raw3215_wakeup, (unsigned long) raw); /* Request the console irq */ if (raw3215_startup(raw) != 0) { @@ -966,6 +973,7 @@ static void tty3215_close(struct tty_struct *tty, struct file * filp) tty->closing = 1; /* Shutdown the terminal */ raw3215_shutdown(raw); + tasklet_kill(&raw->tlet); tty->closing = 0; raw->tty = NULL; } From cf1eb40f8f5ea12c9e569e7282161fc7f194fd62 Mon Sep 17 00:00:00 2001 From: Martin Schwidefsky Date: Fri, 17 Feb 2012 10:29:23 +0100 Subject: [PATCH 4/4] [S390] correct ktime to tod clock comparator conversion The conversion of the ktime to a value suitable for the clock comparator does not take changes to wall_to_monotonic into account. In fact the conversion just needs the boot clock (sched_clock_base_cc) and the total_sleep_time. This is applicable to 3.2+ kernels. CC: stable@vger.kernel.org Signed-off-by: Martin Schwidefsky --- arch/s390/kernel/time.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/arch/s390/kernel/time.c b/arch/s390/kernel/time.c index fa02f443f5f65..14da278febbfa 100644 --- a/arch/s390/kernel/time.c +++ b/arch/s390/kernel/time.c @@ -113,11 +113,14 @@ static void fixup_clock_comparator(unsigned long long delta) static int s390_next_ktime(ktime_t expires, struct clock_event_device *evt) { + struct timespec ts; u64 nsecs; - nsecs = ktime_to_ns(ktime_sub(expires, ktime_get_monotonic_offset())); + ts.tv_sec = ts.tv_nsec = 0; + monotonic_to_bootbased(&ts); + nsecs = ktime_to_ns(ktime_add(timespec_to_ktime(ts), expires)); do_div(nsecs, 125); - S390_lowcore.clock_comparator = TOD_UNIX_EPOCH + (nsecs << 9); + S390_lowcore.clock_comparator = sched_clock_base_cc + (nsecs << 9); set_clock_comparator(S390_lowcore.clock_comparator); return 0; }