s390: reintroduce diag 44 calls for cpu_relax()

Christian Borntraeger reported that the now missing diag 44 calls (voluntary
time slice end) does cause a performance regression for stop_machine() calls
if a machine has more virtual cpus than the host has physical cpus.

This patch mainly reverts 57f2ffe14fd125c2 ("s390: remove diag 44 calls from
cpu_relax()") with the exception that we still do not issue diag 44 calls if
running with smt enabled. Due to group scheduling algorithms when running in
LPAR this would lead to significant latencies.
However, when running in LPAR we do not have more virtual than physical cpus.

Reported-and-tested-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>

diff --git a/arch/s390/include/asm/processor.h b/arch/s390/include/asm/processor.h
index bed05ea7ec27eb236cc65b2007379fa6f61fb865..e7cbbdcdee1337cf8cefa5fbe74dab2e174eadf2 100644 (file)
--- a/arch/s390/include/asm/processor.h
+++ b/arch/s390/include/asm/processor.h
@@ -215,10 +215,7 @@ static inline unsigned short stap(void)
 /*
  * Give up the time slice of the virtual PU.
  */
-static inline void cpu_relax(void)
-{
-       barrier();
-}
+void cpu_relax(void);
 
 #define cpu_relax_lowlatency()  barrier()
 

diff --git a/arch/s390/kernel/processor.c b/arch/s390/kernel/processor.c
index dbdd33ee010204f0e03c9c1fd48ad30acad59b4c..26108232fcaaf049f4e4caa3938e537af5066ac6 100644 (file)
--- a/arch/s390/kernel/processor.c
+++ b/arch/s390/kernel/processor.c
@@ -8,16 +8,24 @@
 
 #include <linux/kernel.h>
 #include <linux/init.h>
-#include <linux/smp.h>
 #include <linux/seq_file.h>
 #include <linux/delay.h>
 #include <linux/cpu.h>
 #include <asm/elf.h>
 #include <asm/lowcore.h>
 #include <asm/param.h>
+#include <asm/smp.h>
 
 static DEFINE_PER_CPU(struct cpuid, cpu_id);
 
+void cpu_relax(void)
+{
+       if (!smp_cpu_mtid && MACHINE_HAS_DIAG44)
+               asm volatile("diag 0,0,0x44");
+       barrier();
+}
+EXPORT_SYMBOL(cpu_relax);
+
 /*
  * cpu_init - initializes state that is per-CPU.
  */