Skip to contentSkip to conversation

x86/efi: Free EFI memory map only when installing a new one.

PATCH Superseded Ard Biesheuvel 4 participants 4 messages Jun 10, 2024 – Jun 12, 2024

Subsystems (2)

v1 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 x86/efi: Free EFI memory map only when installing a new one.

From: Ard Biesheuvel <[email protected]>

The logic in __efi_memmap_init() is shared between two different execution flows:
- mapping the EFI memory map early or late into the kernel VA space, so that its entries can be accessed;
- cloning the EFI memory map in order to insert new entries that are created as a result of creating a memory reservation (efi_arch_mem_reserve())

In the former case, the underlying memory containing the kernel's view of the EFI memory map (which may be heavily modified by the kernel itself on x86) is not modified at all, and the only thing that changes is the virtual mapping of this memory, which is different between early and late boot.

In the latter case, an entirely new allocation is created that carries a new, updated version of the kernel's view of the EFI memory map. When installing this new version, the old version will no longer be referenced, and if the memory was allocated by the kernel, it will leak unless it gets freed.

The logic that implements this freeing currently lives on the code path that is shared between these two use cases, but it should only apply to the latter. So move it to the correct spot.

---
 arch/x86/platform/efi/memmap.c | 5 +++++
 drivers/firmware/efi/memmap.c  | 5 -----
 2 files changed, 5 insertions(+), 5 deletions(-)
arch/x86/platform/efi/memmap.c +5 -0
index 4ef20b49eb5e..4990244e5168 100644
@@ -97,6 +97,11 @@ int __init efi_memmap_install(struct efi_memory_map_data *data)
 	if (efi_enabled(EFI_PARAVIRT))
 		return 0;
 
+	if (efi.memmap.flags & (EFI_MEMMAP_MEMBLOCK | EFI_MEMMAP_SLAB))
+		__efi_memmap_free(efi.memmap.phys_map,
+				  efi.memmap.desc_size * efi.memmap.nr_map,
+				  efi.memmap.flags);
+
 	return __efi_memmap_init(data);
 }
 
drivers/firmware/efi/memmap.c +0 -5
index 3365944f7965..3759e95a7407 100644
@@ -51,11 +51,6 @@ int __init __efi_memmap_init(struct efi_memory_map_data *data)
 		return -ENOMEM;
 	}
 
-	if (efi.memmap.flags & (EFI_MEMMAP_MEMBLOCK | EFI_MEMMAP_SLAB))
-		__efi_memmap_free(efi.memmap.phys_map,
-				  efi.memmap.desc_size * efi.memmap.nr_map,
-				  efi.memmap.flags);
-
 	map.phys_map = data->phys_map;
 	map.nr_map = data->size / data->desc_size;
 	map.map_end = map.map + data->size;
Signature · 1 line
2.45.2.505.gda0bf45e8d-goog
maintainer
#1 lore raw

+ Dan because of

f0ef6523475f ("efi: Fix efi_memmap_alloc() leaks")

Thanks Ard.

Quoted 72 lines · Ard Biesheuvel

From: Ard Biesheuvel <[email protected]>

The logic in __efi_memmap_init() is shared between two different execution flows:
- mapping the EFI memory map early or late into the kernel VA space, so that its entries can be accessed;
- cloning the EFI memory map in order to insert new entries that are created as a result of creating a memory reservation (efi_arch_mem_reserve())

In the former case, the underlying memory containing the kernel's view of the EFI memory map (which may be heavily modified by the kernel itself on x86) is not modified at all, and the only thing that changes is the virtual mapping of this memory, which is different between early and late boot.

In the latter case, an entirely new allocation is created that carries a new, updated version of the kernel's view of the EFI memory map. When installing this new version, the old version will no longer be referenced, and if the memory was allocated by the kernel, it will leak unless it gets freed.

The logic that implements this freeing currently lives on the code path that is shared between these two use cases, but it should only apply to the latter. So move it to the correct spot.

---
 arch/x86/platform/efi/memmap.c | 5 +++++
 drivers/firmware/efi/memmap.c  | 5 -----
 2 files changed, 5 insertions(+), 5 deletions(-)
arch/x86/platform/efi/memmap.c +5 -0
index 4ef20b49eb5e..4990244e5168 100644
@@ -97,6 +97,11 @@ int __init efi_memmap_install(struct efi_memory_map_data *data)
 	if (efi_enabled(EFI_PARAVIRT))
 		return 0;
 
+	if (efi.memmap.flags & (EFI_MEMMAP_MEMBLOCK | EFI_MEMMAP_SLAB))
+		__efi_memmap_free(efi.memmap.phys_map,
+				  efi.memmap.desc_size * efi.memmap.nr_map,
+				  efi.memmap.flags);
+
 	return __efi_memmap_init(data);
 }
 
drivers/firmware/efi/memmap.c +0 -5
index 3365944f7965..3759e95a7407 100644
@@ -51,11 +51,6 @@ int __init __efi_memmap_init(struct efi_memory_map_data *data)
 		return -ENOMEM;
 	}
 
-	if (efi.memmap.flags & (EFI_MEMMAP_MEMBLOCK | EFI_MEMMAP_SLAB))
-		__efi_memmap_free(efi.memmap.phys_map,
-				  efi.memmap.desc_size * efi.memmap.nr_map,
-				  efi.memmap.flags);
-
 	map.phys_map = data->phys_map;
 	map.nr_map = data->size / data->desc_size;
 	map.map_end = map.map + data->size;
Signature · 1 line
2.45.2.505.gda0bf45e8d-goog
Signature · 4 lines
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette
#2 lore raw
In reply to Ard Biesheuvel
Quoted 43 lines · Ard Biesheuvel

From: Ard Biesheuvel <[email protected]>

The logic in __efi_memmap_init() is shared between two different execution flows:
- mapping the EFI memory map early or late into the kernel VA space, so that its entries can be accessed;
- cloning the EFI memory map in order to insert new entries that are created as a result of creating a memory reservation (efi_arch_mem_reserve())

In the former case, the underlying memory containing the kernel's view of the EFI memory map (which may be heavily modified by the kernel itself on x86) is not modified at all, and the only thing that changes is the virtual mapping of this memory, which is different between early and late boot.

In the latter case, an entirely new allocation is created that carries a new, updated version of the kernel's view of the EFI memory map. When installing this new version, the old version will no longer be referenced, and if the memory was allocated by the kernel, it will leak unless it gets freed.

The logic that implements this freeing currently lives on the code path that is shared between these two use cases, but it should only apply to the latter. So move it to the correct spot.

---
 arch/x86/platform/efi/memmap.c | 5 +++++
 drivers/firmware/efi/memmap.c  | 5 -----
 2 files changed, 5 insertions(+), 5 deletions(-)
arch/x86/platform/efi/memmap.c +0 -0
index 4ef20b49eb5e..4990244e5168 100644

arch/x86/platform/efi/memmap.c int __init efi_memmap_install(struct efi_memory_map_data *data) · line 105

Quoted 9 lines
arch/x86/platform/efi/memmap.c +5 -0
@@ -97,6 +97,11 @@ int __init efi_memmap_install(struct efi_memory_map_data *data)
        if (efi_enabled(EFI_PARAVIRT))
                return 0;
 
+       if (efi.memmap.flags & (EFI_MEMMAP_MEMBLOCK | EFI_MEMMAP_SLAB))
+               __efi_memmap_free(efi.memmap.phys_map,
+                                 efi.memmap.desc_size * efi.memmap.nr_map,
+                                 efi.memmap.flags);
+
        return __efi_memmap_init(data);

Nice fix! but would it be better to save the old addr/size/flag and free them only when __efi_memmap_init succeed?

Quoted 21 lines
arch/x86/platform/efi/memmap.c +0 -0
 }
 
drivers/firmware/efi/memmap.c +0 -6
index 3365944f7965..3759e95a7407 100644
@@ -51,11 +51,6 @@ int __init __efi_memmap_init(struct efi_memory_map_data *data)
                return -ENOMEM;
        }
 
-       if (efi.memmap.flags & (EFI_MEMMAP_MEMBLOCK | EFI_MEMMAP_SLAB))
-               __efi_memmap_free(efi.memmap.phys_map,
-                                 efi.memmap.desc_size * efi.memmap.nr_map,
-                                 efi.memmap.flags);
-
        map.phys_map = data->phys_map;
        map.nr_map = data->size / data->desc_size;
        map.map_end = map.map + data->size;
--

2.45.2.505.gda0bf45e8d-goog

maintainer
#3 lore raw
Quoted 48 lines · Dave Young
Quoted 46 lines · Ard Biesheuvel

From: Ard Biesheuvel <[email protected]>

The logic in __efi_memmap_init() is shared between two different execution flows:
- mapping the EFI memory map early or late into the kernel VA space, so that its entries can be accessed;
- cloning the EFI memory map in order to insert new entries that are created as a result of creating a memory reservation (efi_arch_mem_reserve())

In the former case, the underlying memory containing the kernel's view of the EFI memory map (which may be heavily modified by the kernel itself on x86) is not modified at all, and the only thing that changes is the virtual mapping of this memory, which is different between early and late boot.

In the latter case, an entirely new allocation is created that carries a new, updated version of the kernel's view of the EFI memory map. When installing this new version, the old version will no longer be referenced, and if the memory was allocated by the kernel, it will leak unless it gets freed.

The logic that implements this freeing currently lives on the code path that is shared between these two use cases, but it should only apply to the latter. So move it to the correct spot.

---
 arch/x86/platform/efi/memmap.c | 5 +++++
 drivers/firmware/efi/memmap.c  | 5 -----
 2 files changed, 5 insertions(+), 5 deletions(-)
arch/x86/platform/efi/memmap.c +0 -0
index 4ef20b49eb5e..4990244e5168 100644
@@ -97,6 +97,11 @@ int __init efi_memmap_install(struct efi_memory_map_data *data)
        if (efi_enabled(EFI_PARAVIRT))
                return 0;
 
Quoted 10 lines

arch/x86/platform/efi/memmap.c int __init efi_memmap_install(struct efi_memory_map_data *data) · line 105

Quoted 6 lines
arch/x86/platform/efi/memmap.c +5 -0
+       if (efi.memmap.flags & (EFI_MEMMAP_MEMBLOCK | EFI_MEMMAP_SLAB))
+               __efi_memmap_free(efi.memmap.phys_map,
+                                 efi.memmap.desc_size * efi.memmap.nr_map,
+                                 efi.memmap.flags);
+
        return __efi_memmap_init(data);

Nice fix! but would it be better to save the old addr/size/flag and free them only when __efi_memmap_init succeed?

Yeah. I guess the same applies to the original code, but now is as good a time as ever to fix that as well.