[4.17] EFI: don't convert memory marked for runtime use to ordinary RAM
PATCH v2 1/1 [4.17] EFI: don't convert memory marked for runtime use to ordinary RAM
efi_init_memory() in both relevant places is treating EFI_MEMORY_RUNTIME higher priority than the type of the range. To avoid accessing memory at runtime which was re-used for other purposes, make efi_arch_process_memory_map() follow suit. While on x86 in theory the same would apply to EfiACPIReclaimMemory, we don't actually "reclaim" E820_ACPI memory there (and it would be a bug if the Dom0 kernel tried to do so, bypassing Xen's memory management), hence that type's handling can be left alone.
---
v2: Amend description.
xen/arch/arm/efi/efi-boot.h +9 -7
@@ -183,13 +183,15 @@ static EFI_STATUS __init efi_process_memfor ( Index =0 ; Index < (mmap_size / desc_size); Index++ ) { -if ( desc_ptr->Attribute &EFI_MEMORY_WB && - (desc_ptr->Type == EfiConventionalMemory || - desc_ptr->Type == EfiLoaderCode || - desc_ptr->Type == EfiLoaderData || - (!map_bs && - (desc_ptr->Type == EfiBootServicesCode || - desc_ptr->Type == EfiBootServicesData))) ) +if ( desc_ptr->Attribute &EFI_MEMORY_RUNTIME ) +/* nothing */ ; +else if ( (desc_ptr->Attribute &EFI_MEMORY_WB ) && + (desc_ptr->Type == EfiConventionalMemory || + desc_ptr->Type == EfiLoaderCode || + desc_ptr->Type == EfiLoaderData || + (!map_bs && + (desc_ptr->Type == EfiBootServicesCode || + desc_ptr->Type == EfiBootServicesData))) ) {if ( !meminfo_add_bank (&bootinfo.mem, desc_ptr) ) {
xen/arch/x86/efi/efi-boot.h +3 -1
@@ -185,7 +185,9 @@ static void __init efi_arch_process_memo/* fall through */ case EfiLoaderCode:case EfiLoaderData: -if ( desc->Attribute &EFI_MEMORY_WB ) +if ( desc->Attribute &EFI_MEMORY_RUNTIME ) + type =E820_RESERVED ; +else if ( desc->Attribute &EFI_MEMORY_WB ) type =E820_RAM ;else case EfiUnusableMemory:
Hi Jan,
Quoted 14 lines
-----Original Message-----
From: Jan Beulich <[email protected]>
Subject: [PATCH v2][4.17] EFI: don't convert memory marked for runtime use to ordinary RAM
efi_init_memory()in both relevant places is treatingEFI_MEMORY_RUNTIMEhigher priority than the type of the range. To avoid accessing memory at runtime which was re-used for other purposes, makeefi_arch_process_memory_map()follow suit. While on x86 in theory the same would apply to EfiACPIReclaimMemory, we don't actually "reclaim"E820_ACPImemory there (and it would be a bug if the Dom0 kernel tried to do so, bypassing Xen's memory management), hence that type's handling can be left alone.
Quoted 7 lines
Fixes bf6501a62e80 ("x86-64: EFI boot code")Fixes facac0af87ef ("x86-64: EFI runtime code")Fixes 6d70ea10d49f ("Add ARM EFI boot support")Signed-off-by Jan Beulich <[email protected]>Reviewed-by Luca Fancellu <[email protected]> # ArmTested-by Luca Fancellu <[email protected]> # ArmReviewed-by Bertrand Marquis <[email protected]> #arm
Kind regards,
Henry
Quoted 2 lines
---
v2: Amend description.
Quoted 7 lines · Jan Beulich
efi_init_memory()in both relevant places is treatingEFI_MEMORY_RUNTIMEhigher priority than the type of the range. To avoid accessing memory at runtime which was re-used for other purposes, makeefi_arch_process_memory_map()follow suit. While on x86 in theory the same would apply to EfiACPIReclaimMemory, we don't actually "reclaim"E820_ACPImemory there (and it would be a bug if the Dom0 kernel tried to do so, bypassing Xen's memory management), hence that type's handling
Strictly speaking I don't think dom0 needs to bypass Xen's memory management, just overwriting the page would be bad enough for runtime services to not work correctly I would think.
Quoted 9 lines
can be left alone.
Fixes bf6501a62e80 ("x86-64: EFI boot code")Fixes facac0af87ef ("x86-64: EFI runtime code")Fixes 6d70ea10d49f ("Add ARM EFI boot support")Signed-off-by Jan Beulich <[email protected]>Reviewed-by Luca Fancellu <[email protected]> # ArmTested-by Luca Fancellu <[email protected]> # ArmReviewed-by Bertrand Marquis <[email protected]> #arm
Thanks, Roger.
Quoted 9 lines · Roger Pau Monné
Quoted 7 lines · Jan Beulich
efi_init_memory()in both relevant places is treatingEFI_MEMORY_RUNTIMEhigher priority than the type of the range. To avoid accessing memory at runtime which was re-used for other purposes, makeefi_arch_process_memory_map()follow suit. While on x86 in theory the same would apply to EfiACPIReclaimMemory, we don't actually "reclaim"E820_ACPImemory there (and it would be a bug if the Dom0 kernel tried to do so, bypassing Xen's memory management), hence that type's handling
Quoted 3 lines
Strictly speaking I don't think dom0 needs to bypass Xen's memory management, just overwriting the page would be bad enough for runtime services to not work correctly I would think.
Then how about:
"While on x86 in theory the same would apply to EfiACPIReclaimMemory, we don't
actually "reclaim" or clobber E820_ACPI memory there (and it would be a bug if the Dom0 kernel tried to reclaim the range, bypassing Xen's memory management, plus it would be at least bogus if it clobbered that space), hence that type's handling can be left alone."
I didn't think the clobbering aspect needed pointing out, as the same applies to all other memory which Dom0 is able to access beyond its actual allocation.
Quoted 11 lines
Quoted 9 lines
can be left alone.
Fixes bf6501a62e80 ("x86-64: EFI boot code")Fixes facac0af87ef ("x86-64: EFI runtime code")Fixes 6d70ea10d49f ("Add ARM EFI boot support")Signed-off-by Jan Beulich <[email protected]>Reviewed-by Luca Fancellu <[email protected]> # ArmTested-by Luca Fancellu <[email protected]> # ArmReviewed-by Bertrand Marquis <[email protected]> #armAcked-by Roger Pau Monné <[email protected]>
Thanks.
Jan
Quoted 16 lines · Jan Beulich
Quoted 9 lines · Roger Pau Monné
Quoted 7 lines · Jan Beulich
efi_init_memory()in both relevant places is treatingEFI_MEMORY_RUNTIMEhigher priority than the type of the range. To avoid accessing memory at runtime which was re-used for other purposes, makeefi_arch_process_memory_map()follow suit. While on x86 in theory the same would apply to EfiACPIReclaimMemory, we don't actually "reclaim"E820_ACPImemory there (and it would be a bug if the Dom0 kernel tried to do so, bypassing Xen's memory management), hence that type's handlingQuoted 3 lines
Strictly speaking I don't think dom0 needs to bypass Xen's memory management, just overwriting the page would be bad enough for runtime services to not work correctly I would think.
Then how about:
Quoted 8 lines
"While on x86 in theory the same would apply to EfiACPIReclaimMemory, we don't
actually "reclaim" or clobberE820_ACPImemory there (and it would be a bug if the Dom0 kernel tried to reclaim the range, bypassing Xen's memory management, plus it would be at least bogus if it clobbered that space), hence that type's handling can be left alone."I didn't think the clobbering aspect needed pointing out, as the same applies to all other memory which Dom0 is able to access beyond its actual allocation.
I think it makes it clear that just clobbering it from dom0 could cause issues to runtime services. I guess it can be extrapolated that clobbering is also bad if reclaiming is.
Thanks, Roger.