Skip to contentSkip to conversation

memaccess: reduce include dependencies

PATCH In review Jan Beulich 3 participants 7 messages Mar 9, 2020

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 memaccess: reduce include dependencies

The common header doesn't itself need to include public/vm_event.h nor public/memory.h. Drop their inclusion. This requires using the non- typedef names in two prototypes and an inline function; by not changing the callers and function definitions at the same time it'll remain certain that the build would fail if the typedef itself was changed.

xen/include/asm-arm/mem_access.h +3 -1
@@ -17,9 +17,11 @@
 #ifndef _ASM_ARM_MEM_ACCESS_H
 #define _ASM_ARM_MEM_ACCESS_H
 
+struct vm_event_st;
+
 static inline
 bool p2m_mem_access_emulate_check(struct vcpu *v,
-                                  const vm_event_response_t *rsp)
+                                  const struct vm_event_st *rsp)
 {
     /* Not supported on ARM. */
     return false;
xen/include/asm-x86/mem_access.h +4 -2
@@ -26,6 +26,8 @@
 #ifndef __ASM_X86_MEM_ACCESS_H__
 #define __ASM_X86_MEM_ACCESS_H__
 
+struct vm_event_st;
+
 /*
  * Setup vm_event request based on the access (gla is -1ull if not available).
  * Handles the rw2rx conversion. Boolean return value indicates if event type
@@ -36,12 +38,12 @@
  */
 bool p2m_mem_access_check(paddr_t gpa, unsigned long gla,
                           struct npfec npfec,
-                          vm_event_request_t **req_ptr);
+                          struct vm_event_st **req_ptr);
 
 /* Check for emulation and mark vcpu for skipping one instruction
  * upon rescheduling if required. */
 bool p2m_mem_access_emulate_check(struct vcpu *v,
-                                  const vm_event_response_t *rsp);
+                                  const struct vm_event_st *rsp);
 
 /* Sanity check for mem_access hardware support */
 bool p2m_mem_access_sanity_check(const struct domain *d);
xen/include/xen/mem_access.h +0 -2
@@ -24,8 +24,6 @@
 
 #include <xen/types.h>
 #include <xen/mm.h>
-#include <public/memory.h>
-#include <public/vm_event.h>
 #include <asm/mem_access.h>
 
 /*

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#1 lore raw
Quoted 6 lines · Jan Beulich

The common header doesn't itself need to include public/vm_event.h nor public/memory.h. Drop their inclusion. This requires using the non- typedef names in two prototypes and an inline function; by not changing the callers and function definitions at the same time it'll remain certain that the build would fail if the typedef itself was changed.

Just curious, what's the benefit of doing this?

Tamas

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#2 lore raw
Quoted 9 lines · Tamas K Lengyel
Quoted 6 lines · Jan Beulich

The common header doesn't itself need to include public/vm_event.h nor public/memory.h. Drop their inclusion. This requires using the non- typedef names in two prototypes and an inline function; by not changing the callers and function definitions at the same time it'll remain certain that the build would fail if the typedef itself was changed.

Just curious, what's the benefit of doing this?

Less dependencies that (almost) every CU gets, and hence statistically less rebuilds of (almost) everything when only a rather special purpose header file changes.

Jan

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#3 lore raw
Quoted 11 lines · Jan Beulich
Quoted 7 lines · Tamas K Lengyel
Quoted 5 lines · Jan Beulich

The common header doesn't itself need to include public/vm_event.h nor public/memory.h. Drop their inclusion. This requires using the non- typedef names in two prototypes and an inline function; by not changing the callers and function definitions at the same time it'll remain certain that the build would fail if the typedef itself was changed.

Just curious, what's the benefit of doing this?

Less dependencies that (almost) every CU gets, and hence statistically less rebuilds of (almost) everything when only a rather special purpose header file changes.

Also fractionally faster compile times generally (reduced parsing of unrelated header files), and overall, a reduction in the complexity of the tangled mess that is our header files.

~Andrew

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#4 lore raw
In reply to Jan Beulich
Quoted 27 lines · Jan Beulich

The common header doesn't itself need to include public/vm_event.h nor public/memory.h. Drop their inclusion. This requires using the non- typedef names in two prototypes and an inline function; by not changing the callers and function definitions at the same time it'll remain certain that the build would fail if the typedef itself was changed.

xen/include/asm-arm/mem_access.h +3 -1
@@ -17,9 +17,11 @@
 #ifndef _ASM_ARM_MEM_ACCESS_H
 #define _ASM_ARM_MEM_ACCESS_H
 
+struct vm_event_st;
+
 static inline
 bool p2m_mem_access_emulate_check(struct vcpu *v,
-                                  const vm_event_response_t *rsp)
+                                  const struct vm_event_st *rsp)
 {
     /* Not supported on ARM. */
     return false;

xen/include/asm-x86/mem_access.h line 29

Quoted 4 lines
xen/include/asm-x86/mem_access.h +1 -0
@@ -26,6 +26,8 @@
 #ifndef __ASM_X86_MEM_ACCESS_H__
 #define __ASM_X86_MEM_ACCESS_H__
 
+struct vm_event_st;

Wouldn't it make more sense to define this in xen/mem_access.h instead of having to do it in both asm versions? Nothing directly includes asm/mem_access.h, all users include xen/mem_access.h

Tamas

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#5 lore raw
Quoted 22 lines · Tamas K Lengyel
Quoted 21 lines · Jan Beulich
xen/include/asm-arm/mem_access.h +3 -1
@@ -17,9 +17,11 @@
 #ifndef _ASM_ARM_MEM_ACCESS_H
 #define _ASM_ARM_MEM_ACCESS_H
 
+struct vm_event_st;
+
 static inline
 bool p2m_mem_access_emulate_check(struct vcpu *v,
-                                  const vm_event_response_t *rsp)
+                                  const struct vm_event_st *rsp)
 {
     /* Not supported on ARM. */
     return false;
xen/include/asm-x86/mem_access.h +0 -0
@@ -26,6 +26,8 @@
 #ifndef __ASM_X86_MEM_ACCESS_H__
 #define __ASM_X86_MEM_ACCESS_H__
 
Quoted 5 lines

xen/include/asm-x86/mem_access.h line 29

Quoted 1 line
xen/include/asm-x86/mem_access.h +1 -0
+struct vm_event_st;

Wouldn't it make more sense to define this in xen/mem_access.h instead of having to do it in both asm versions? Nothing directly includes asm/mem_access.h, all users include xen/mem_access.h

If that's what you prefer - I can certainly do so. It'll look a little odd then, as the forward declaration has to come ahead of

#include <asm/mem_access.h>

Just let me know if you really prefer it that way.

Jan

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel

#6 lore raw
Quoted 30 lines · Jan Beulich
Quoted 22 lines · Tamas K Lengyel
Quoted 21 lines · Jan Beulich
xen/include/asm-arm/mem_access.h +3 -1
@@ -17,9 +17,11 @@
 #ifndef _ASM_ARM_MEM_ACCESS_H
 #define _ASM_ARM_MEM_ACCESS_H
 
+struct vm_event_st;
+
 static inline
 bool p2m_mem_access_emulate_check(struct vcpu *v,
-                                  const vm_event_response_t *rsp)
+                                  const struct vm_event_st *rsp)
 {
     /* Not supported on ARM. */
     return false;
xen/include/asm-x86/mem_access.h +0 -0
@@ -26,6 +26,8 @@
 #ifndef __ASM_X86_MEM_ACCESS_H__
 #define __ASM_X86_MEM_ACCESS_H__
 
Quoted 5 lines

xen/include/asm-x86/mem_access.h line 29

Quoted 1 line
xen/include/asm-x86/mem_access.h +1 -0
+struct vm_event_st;

Wouldn't it make more sense to define this in xen/mem_access.h instead of having to do it in both asm versions? Nothing directly includes asm/mem_access.h, all users include xen/mem_access.h

Quoted 6 lines

If that's what you prefer - I can certainly do so. It'll look a little odd then, as the forward declaration has to come ahead of

#include <asm/mem_access.h>

Just let me know if you really prefer it that way.

Well, I find it ugly either way. I would prefer if it's forward declared just at one spot, with a comment explaining why it's needed/done that way.

Thanks,
Tamas

_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel