Skip to contentSkip to conversation

x86/HVM: use single (atomic) MOV for aligned emulated writes

PATCH In review Jan Beulich 3 participants 6 messages Sep 16, 2019 – Jan 14, 2020

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 x86/HVM: use single (atomic) MOV for aligned emulated writes

Using memcpy() may result in multiple individual byte accesses (dependening how memcpy() is implemented and how the resulting insns, e.g. REP MOVSB, get carried out in hardware), which isn't what we want/need for carrying out guest insns as correctly as possible. Fall back to memcpy() only for accesses not 2, 4, or 8 bytes in size.

---
TBD: Besides it still being open whether the linear_write() path also
needs playing with), a question also continues to be whether we'd want to extend this to reads as well. linear_{read,write}() currently don't use hvmemul_map_linear_addr(), i.e. in both cases I'd need to also fiddle with __hvm_copy() (perhaps by making the construct below a helper function).

xen/arch/x86/hvm/emulate.c +8 -1
@@ -1324,7 +1324,14 @@ static int hvmemul_write(
     if ( !mapping )
         return linear_write(addr, bytes, p_data, pfec, hvmemul_ctxt);
 
-    memcpy(mapping, p_data, bytes);
+    /* Where possible use single (and hence generally atomic) MOV insns. */
+    switch ( bytes )
+    {
+    case 2: write_u16_atomic(mapping, *(uint16_t *)p_data); break;
+    case 4: write_u32_atomic(mapping, *(uint32_t *)p_data); break;
+    case 8: write_u64_atomic(mapping, *(uint64_t *)p_data); break;
+    default: memcpy(mapping, p_data, bytes);                break;
+    }
 
     hvmemul_unmap_linear_addr(mapping, addr, bytes, hvmemul_ctxt);
 

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

#1 lore raw

[Xen-devel] Ping: [PATCH] x86/HVM: use single (atomic) MOV for aligned emulated writes

Quoted 25 lines · Jan Beulich

Using memcpy() may result in multiple individual byte accesses (dependening how memcpy() is implemented and how the resulting insns, e.g. REP MOVSB, get carried out in hardware), which isn't what we want/need for carrying out guest insns as correctly as possible. Fall back to memcpy() only for accesses not 2, 4, or 8 bytes in size.

---
TBD: Besides it still being open whether the linear_write() path also
needs playing with, a question also continues to be whether we'd want to extend this to reads as well. linear_{read,write}() currently don't use hvmemul_map_linear_addr(), i.e. in both cases I'd need to also fiddle with __hvm_copy() (perhaps by making the construct below a helper function).

xen/arch/x86/hvm/emulate.c +1 -1
@@ -1324,7 +1324,14 @@ static int hvmemul_write(
     if ( !mapping )
         return linear_write(addr, bytes, p_data, pfec, hvmemul_ctxt);
 
-    memcpy(mapping, p_data, bytes);
+    /* Where possible use single (and hence generally atomic) MOV insns. */

xen/arch/x86/hvm/emulate.c static int hvmemul_write( · line 1337

Quoted 10 lines
xen/arch/x86/hvm/emulate.c +7 -0
+    switch ( bytes )
+    {
+    case 2: write_u16_atomic(mapping, *(uint16_t *)p_data); break;
+    case 4: write_u32_atomic(mapping, *(uint32_t *)p_data); break;
+    case 8: write_u64_atomic(mapping, *(uint64_t *)p_data); break;
+    default: memcpy(mapping, p_data, bytes);                break;
+    }
 
     hvmemul_unmap_linear_addr(mapping, addr, bytes, hvmemul_ctxt);
 

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

Acked-by
#2 lore raw

Re: [Xen-devel] Ping: [PATCH] x86/HVM: use single (atomic) MOV for aligned emulated writes

Quoted 9 lines · Jan Beulich
Quoted 8 lines · Jan Beulich

Using memcpy() may result in multiple individual byte accesses (dependening how memcpy() is implemented and how the resulting insns, e.g. REP MOVSB, get carried out in hardware), which isn't what we want/need for carrying out guest insns as correctly as possible. Fall back to memcpy() only for accesses not 2, 4, or 8 bytes in size.

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

#3 lore raw

Re: [Xen-devel] Ping: [PATCH] x86/HVM: use single (atomic) MOV for aligned emulated writes

Quoted 9 lines · Andrew Cooper
Quoted 7 lines · Jan Beulich
Quoted 6 lines · Jan Beulich

Using memcpy() may result in multiple individual byte accesses (dependening how memcpy() is implemented and how the resulting insns, e.g. REP MOVSB, get carried out in hardware), which isn't what we want/need for carrying out guest insns as correctly as possible. Fall back to memcpy() only for accesses not 2, 4, or 8 bytes in size.

Quoted 4 lines
Quoted 2 lines
Quoted 2 lines

Should xen/arch/x86/mm/shadow/hvm.c:hvm_emulate_write() be similarly changed?

Thanks,
Jason

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

#4 lore raw
Quoted 11 lines · Jason Andryuk
Quoted 9 lines · Andrew Cooper
Quoted 7 lines · Jan Beulich
Quoted 6 lines · Jan Beulich

Using memcpy() may result in multiple individual byte accesses (dependening how memcpy() is implemented and how the resulting insns, e.g. REP MOVSB, get carried out in hardware), which isn't what we want/need for carrying out guest insns as correctly as possible. Fall back to memcpy() only for accesses not 2, 4, or 8 bytes in size.

Quoted 6 lines
Quoted 4 lines
Quoted 2 lines
Quoted 2 lines

Should xen/arch/x86/mm/shadow/hvm.c:hvm_emulate_write() be similarly changed?

Probably. Care to make a patch?

Jan

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

#5 lore raw
Quoted 13 lines · Jan Beulich
Quoted 11 lines · Jason Andryuk
Quoted 9 lines · Andrew Cooper
Quoted 7 lines · Jan Beulich
Quoted 6 lines · Jan Beulich

Using memcpy() may result in multiple individual byte accesses (dependening how memcpy() is implemented and how the resulting insns, e.g. REP MOVSB, get carried out in hardware), which isn't what we want/need for carrying out guest insns as correctly as possible. Fall back to memcpy() only for accesses not 2, 4, or 8 bytes in size.

Quoted 8 lines
Quoted 6 lines
Quoted 4 lines
Quoted 2 lines
Quoted 2 lines

Should xen/arch/x86/mm/shadow/hvm.c:hvm_emulate_write() be similarly changed?

Probably. Care to make a patch?

Sure :)

-Jason

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