Re: [PATCH] mm, arch: Complete pagefault_disable abstraction

From: Peter Zijlstra
Date: Wed Oct 05 2011 - 09:12:21 EST


For those with a strong stomach, here's what we do with kmap_atomic:

---
Subject: mm, rt: kmap_atomic scheduling
From: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
Date: Thu, 28 Jul 2011 10:43:51 +0200

In fact, with migrate_disable() existing one could play games with
kmap_atomic. You could save/restore the kmap_atomic slots on context
switch (if there are any in use of course), this should be esp easy now
that we have a kmap_atomic stack.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@xxxxxxxxx>
[dvhart@xxxxxxxxxxxxxxx: build fix]
Link: http://lkml.kernel.org/r/1311842631.5890.208.camel@twins
---
arch/x86/kernel/process_32.c | 36 ++++++++++++++++++++++++++++++++++++
include/linux/sched.h | 5 +++++
mm/memory.c | 2 ++
3 files changed, 43 insertions(+)

Index: linux-2.6/arch/x86/kernel/process_32.c
===================================================================
--- linux-2.6.orig/arch/x86/kernel/process_32.c
+++ linux-2.6/arch/x86/kernel/process_32.c
@@ -38,6 +38,7 @@
#include <linux/uaccess.h>
#include <linux/io.h>
#include <linux/kdebug.h>
+#include <linux/highmem.h>

#include <asm/pgtable.h>
#include <asm/system.h>
@@ -346,6 +347,41 @@ __switch_to(struct task_struct *prev_p,
task_thread_info(next_p)->flags & _TIF_WORK_CTXSW_NEXT))
__switch_to_xtra(prev_p, next_p, tss);

+#if defined CONFIG_PREEMPT_RT_FULL && defined CONFIG_HIGHMEM
+ /*
+ * Save @prev's kmap_atomic stack
+ */
+ prev_p->kmap_idx = __this_cpu_read(__kmap_atomic_idx);
+ if (unlikely(prev_p->kmap_idx)) {
+ int i;
+
+ for (i = 0; i < prev_p->kmap_idx; i++) {
+ int idx = i + KM_TYPE_NR * smp_processor_id();
+
+ pte_t *ptep = kmap_pte - idx;
+ prev_p->kmap_pte[i] = *ptep;
+ kpte_clear_flush(ptep, __fix_to_virt(FIX_KMAP_BEGIN + idx));
+ }
+
+ __this_cpu_write(__kmap_atomic_idx, 0);
+ }
+
+ /*
+ * Restore @next_p's kmap_atomic stack
+ */
+ if (unlikely(next_p->kmap_idx)) {
+ int i;
+
+ __this_cpu_write(__kmap_atomic_idx, next_p->kmap_idx);
+
+ for (i = 0; i < next_p->kmap_idx; i++) {
+ int idx = i + KM_TYPE_NR * smp_processor_id();
+
+ set_pte(kmap_pte - idx, next_p->kmap_pte[i]);
+ }
+ }
+#endif
+
/* If we're going to preload the fpu context, make sure clts
is run while we're batching the cpu state updates. */
if (preload_fpu)
Index: linux-2.6/include/linux/sched.h
===================================================================
--- linux-2.6.orig/include/linux/sched.h
+++ linux-2.6/include/linux/sched.h
@@ -63,6 +63,7 @@ struct sched_param {
#include <linux/nodemask.h>
#include <linux/mm_types.h>

+#include <asm/kmap_types.h>
#include <asm/system.h>
#include <asm/page.h>
#include <asm/ptrace.h>
@@ -1594,6 +1595,10 @@ struct task_struct {
struct rcu_head put_rcu;
int softirq_nestcnt;
#endif
+#if defined CONFIG_PREEMPT_RT_FULL && defined CONFIG_HIGHMEM
+ int kmap_idx;
+ pte_t kmap_pte[KM_TYPE_NR];
+#endif
};

#ifdef CONFIG_PREEMPT_RT_FULL
Index: linux-2.6/mm/memory.c
===================================================================
--- linux-2.6.orig/mm/memory.c
+++ linux-2.6/mm/memory.c
@@ -3431,6 +3431,7 @@ unlock:
#ifdef CONFIG_PREEMPT_RT_FULL
void pagefault_disable(void)
{
+ migrate_disable();
current->pagefault_disabled++;
/*
* make sure to have issued the store before a pagefault
@@ -3448,6 +3449,7 @@ void pagefault_enable(void)
*/
barrier();
current->pagefault_disabled--;
+ migrate_enable();
}
EXPORT_SYMBOL_GPL(pagefault_enable);
#endif

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/