Skip to contentSkip to conversation

lib: s390x: add smp_cpu_setup_cur_psw_mask

PATCH In review Claudio Imbrenda 2 participants 3 messages Nov 28, 2022 KVM

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 lib: s390x: add smp_cpu_setup_cur_psw_mask

Since a lot of code starts new CPUs using the current PSW mask, add a wrapper to streamline the operation and hopefully make the code of the tests more readable.

---
 lib/s390x/smp.h | 9 +++++++++
 1 file changed, 9 insertions(+)
lib/s390x/smp.h +9 -0
index f4ae973d..0bcb1999 100644
@@ -47,4 +47,13 @@ void smp_setup(void);
 int smp_sigp(uint16_t idx, uint8_t order, unsigned long parm, uint32_t *status);
 struct lowcore *smp_get_lowcore(uint16_t idx);
 
+static inline void smp_cpu_setup_cur_psw_mask(uint16_t idx, void *addr)
+{
+	struct psw psw = {
+		.mask = extract_psw_mask(),
+		.addr = (unsigned long)addr,
+	};
+	smp_cpu_setup(idx, psw);
+}
+
 #endif
Signature · 1 line
2.38.1
Reviewed-by on v1 1/1
#1 lore raw
Quoted 17 lines · Claudio Imbrenda

Since a lot of code starts new CPUs using the current PSW mask, add a wrapper to streamline the operation and hopefully make the code of the tests more readable.

---
 lib/s390x/smp.h | 9 +++++++++
 1 file changed, 9 insertions(+)
lib/s390x/smp.h +0 -0
index f4ae973d..0bcb1999 100644
@@ -47,4 +47,13 @@ void smp_setup(void);
 int smp_sigp(uint16_t idx, uint8_t order, unsigned long parm, uint32_t *status);
 struct lowcore *smp_get_lowcore(uint16_t idx);
 

lib/s390x/smp.h void smp_setup(void); · line 59

Quoted 10 lines
lib/s390x/smp.h +9 -0
+static inline void smp_cpu_setup_cur_psw_mask(uint16_t idx, void *addr)
+{
+	struct psw psw = {
+		.mask = extract_psw_mask(),
+		.addr = (unsigned long)addr,
+	};
+	smp_cpu_setup(idx, psw);
+}
+
 #endif

Although I would have expected you to also use the function.

I'm wondering if just improving the ergonomics of creating a psw would suffice #define PSW(m, a) ((struct psw){ .mask = (uint64_t)m, .addr = (uint64_t)a })

Then it would look like

smp_cpu_setup(idx, PSW(extract_psw_mask(), addr))

and the macro might come in handy in other situations, too, but I haven't surveyed the code.

#2 lore raw
Quoted 33 lines · On Mon, 28 Nov 2022 16:10:32 +0100 Janis Schoetterl-Glausch…
Quoted 17 lines · Claudio Imbrenda

Since a lot of code starts new CPUs using the current PSW mask, add a wrapper to streamline the operation and hopefully make the code of the tests more readable.

---
 lib/s390x/smp.h | 9 +++++++++
 1 file changed, 9 insertions(+)
lib/s390x/smp.h +0 -0
index f4ae973d..0bcb1999 100644
@@ -47,4 +47,13 @@ void smp_setup(void);
 int smp_sigp(uint16_t idx, uint8_t order, unsigned long parm, uint32_t *status);
 struct lowcore *smp_get_lowcore(uint16_t idx);
 

lib/s390x/smp.h void smp_setup(void); · line 59

Quoted 10 lines
lib/s390x/smp.h +9 -0
+static inline void smp_cpu_setup_cur_psw_mask(uint16_t idx, void *addr)
+{
+	struct psw psw = {
+		.mask = extract_psw_mask(),
+		.addr = (unsigned long)addr,
+	};
+	smp_cpu_setup(idx, psw);
+}
+
 #endif  

Although I would have expected you to also use the function.

Quoted 8 lines

I'm wondering if just improving the ergonomics of creating a psw would suffice #define PSW(m, a) ((struct psw){ .mask = (uint64_t)m, .addr = (uint64_t)a })

Then it would look like

smp_cpu_setup(idx, PSW(extract_psw_mask(), addr))

and the macro might come in handy in other situations, too, but I haven't surveyed the code.

hmmm, this is actually not a bad idea at all

then it would be possible to also define

#define PSW_CUR_MASK(a) PSW(extract_psw_mask(), (a))

and have the code like this:

smp_cpu_setup(idx, PSW_CUR_MASK(addr));

I think I will do it like that (and also send a second patch where I actually put it to use)