[v2] efi/libstub/arm*: Set default address and size cells values for an empty dtb
diff mbox

Message ID 1486502360-18071-1-git-send-email-jhugo@codeaurora.org
State New, archived
Headers show

Commit Message

Jeffrey Hugo Feb. 7, 2017, 9:19 p.m. UTC
From: Sameer Goel <sgoel@codeaurora.org>

In cases where a device tree is not provided (ie ACPI based system), an
empty fdt is generated by efistub.  #address-cells and #size-cells are not
set in the empty fdt, so they default to 1 (4 byte wide).  This can be an
issue on 64-bit systems where values representing addresses, etc may be
8 bytes wide as the default value does not align with the general
requirements for an empty DTB, and is fragile when passed to other agents
as extra care is required to read the entire width of a value.

This issue is observed on Qualcomm Technologies QDF24XX platforms when
kexec-tools inserts 64-bit addresses into the "linux,elfcorehdr" and
"linux,usable-memory-range" properties of the fdt.  When the values are
later consumed, they are truncated to 32-bit.

Setting #address-cells and #size-cells to 2 at creation of the empty fdt
resolves the observed issue, and makes the fdt less fragile.

Signed-off-by: Sameer Goel <sgoel@codeaurora.org>
Signed-off-by: Jeffrey Hugo <jhugo@codeaurora.org>
---

[v2]
-Add braces to an if when the corresponding else has braces
-Remove print statements
-Reword commit text
-Removed gerrit tag

 drivers/firmware/efi/libstub/fdt.c | 28 ++++++++++++++++++++++++++--
 1 file changed, 26 insertions(+), 2 deletions(-)

Comments

Jeffrey Hugo Feb. 24, 2017, 2:33 p.m. UTC | #1
On 2/7/2017 2:19 PM, Jeffrey Hugo wrote:
> From: Sameer Goel <sgoel@codeaurora.org>
>
> In cases where a device tree is not provided (ie ACPI based system), an
> empty fdt is generated by efistub.  #address-cells and #size-cells are not
> set in the empty fdt, so they default to 1 (4 byte wide).  This can be an
> issue on 64-bit systems where values representing addresses, etc may be
> 8 bytes wide as the default value does not align with the general
> requirements for an empty DTB, and is fragile when passed to other agents
> as extra care is required to read the entire width of a value.
>
> This issue is observed on Qualcomm Technologies QDF24XX platforms when
> kexec-tools inserts 64-bit addresses into the "linux,elfcorehdr" and
> "linux,usable-memory-range" properties of the fdt.  When the values are
> later consumed, they are truncated to 32-bit.
>
> Setting #address-cells and #size-cells to 2 at creation of the empty fdt
> resolves the observed issue, and makes the fdt less fragile.
>
> Signed-off-by: Sameer Goel <sgoel@codeaurora.org>
> Signed-off-by: Jeffrey Hugo <jhugo@codeaurora.org>
> ---

Ping?

>
> [v2]
> -Add braces to an if when the corresponding else has braces
> -Remove print statements
> -Reword commit text
> -Removed gerrit tag
>
>  drivers/firmware/efi/libstub/fdt.c | 28 ++++++++++++++++++++++++++--
>  1 file changed, 26 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/firmware/efi/libstub/fdt.c b/drivers/firmware/efi/libstub/fdt.c
> index 921dfa0..22ea73b 100644
> --- a/drivers/firmware/efi/libstub/fdt.c
> +++ b/drivers/firmware/efi/libstub/fdt.c
> @@ -16,6 +16,22 @@
>
>  #include "efistub.h"
>
> +#define EFI_DT_ADDR_CELLS_DEFAULT 2
> +#define EFI_DT_SIZE_CELLS_DEFAULT 2
> +
> +static void fdt_update_cell_size(efi_system_table_t *sys_table, void *fdt)
> +{
> +	int offset;
> +
> +	offset = fdt_path_offset(fdt, "/");
> +	/* Set the #address-cells and #size-cells values for an empty tree */
> +
> +	fdt_setprop_u32(fdt, offset, "#address-cells",
> +			EFI_DT_ADDR_CELLS_DEFAULT);
> +
> +	fdt_setprop_u32(fdt, offset, "#size-cells", EFI_DT_SIZE_CELLS_DEFAULT);
> +}
> +
>  static efi_status_t update_fdt(efi_system_table_t *sys_table, void *orig_fdt,
>  			       unsigned long orig_fdt_size,
>  			       void *fdt, int new_fdt_size, char *cmdline_ptr,
> @@ -42,10 +58,18 @@ static efi_status_t update_fdt(efi_system_table_t *sys_table, void *orig_fdt,
>  		}
>  	}
>
> -	if (orig_fdt)
> +	if (orig_fdt) {
>  		status = fdt_open_into(orig_fdt, fdt, new_fdt_size);
> -	else
> +	} else {
>  		status = fdt_create_empty_tree(fdt, new_fdt_size);
> +		if (status == 0) {
> +			/*
> +			 * Any failure from the following function is non
> +			 * critical
> +			 */
> +			fdt_update_cell_size(sys_table, fdt);
> +		}
> +	}
>
>  	if (status != 0)
>  		goto fdt_set_fail;
>
Robin Murphy Feb. 24, 2017, 3:36 p.m. UTC | #2
On 07/02/17 21:19, Jeffrey Hugo wrote:
> From: Sameer Goel <sgoel@codeaurora.org>
> 
> In cases where a device tree is not provided (ie ACPI based system), an
> empty fdt is generated by efistub.  #address-cells and #size-cells are not
> set in the empty fdt, so they default to 1 (4 byte wide).  This can be an
> issue on 64-bit systems where values representing addresses, etc may be
> 8 bytes wide as the default value does not align with the general
> requirements for an empty DTB, and is fragile when passed to other agents
> as extra care is required to read the entire width of a value.
> 
> This issue is observed on Qualcomm Technologies QDF24XX platforms when
> kexec-tools inserts 64-bit addresses into the "linux,elfcorehdr" and
> "linux,usable-memory-range" properties of the fdt.  When the values are
> later consumed, they are truncated to 32-bit.
> 
> Setting #address-cells and #size-cells to 2 at creation of the empty fdt
> resolves the observed issue, and makes the fdt less fragile.

Hang on, if this code is shared with 32-bit ARM, doesn't this just move
the problem around? If arm64 kexec is blindly punting reg properties
into a DT assuming *-cells == 2, then wouldn't ARM kexec likely be doing
the same relying on *-cells == 1?

Robin.

> Signed-off-by: Sameer Goel <sgoel@codeaurora.org>
> Signed-off-by: Jeffrey Hugo <jhugo@codeaurora.org>
> ---
> 
> [v2]
> -Add braces to an if when the corresponding else has braces
> -Remove print statements
> -Reword commit text
> -Removed gerrit tag
> 
>  drivers/firmware/efi/libstub/fdt.c | 28 ++++++++++++++++++++++++++--
>  1 file changed, 26 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/firmware/efi/libstub/fdt.c b/drivers/firmware/efi/libstub/fdt.c
> index 921dfa0..22ea73b 100644
> --- a/drivers/firmware/efi/libstub/fdt.c
> +++ b/drivers/firmware/efi/libstub/fdt.c
> @@ -16,6 +16,22 @@
>  
>  #include "efistub.h"
>  
> +#define EFI_DT_ADDR_CELLS_DEFAULT 2
> +#define EFI_DT_SIZE_CELLS_DEFAULT 2
> +
> +static void fdt_update_cell_size(efi_system_table_t *sys_table, void *fdt)
> +{
> +	int offset;
> +
> +	offset = fdt_path_offset(fdt, "/");
> +	/* Set the #address-cells and #size-cells values for an empty tree */
> +
> +	fdt_setprop_u32(fdt, offset, "#address-cells",
> +			EFI_DT_ADDR_CELLS_DEFAULT);
> +
> +	fdt_setprop_u32(fdt, offset, "#size-cells", EFI_DT_SIZE_CELLS_DEFAULT);
> +}
> +
>  static efi_status_t update_fdt(efi_system_table_t *sys_table, void *orig_fdt,
>  			       unsigned long orig_fdt_size,
>  			       void *fdt, int new_fdt_size, char *cmdline_ptr,
> @@ -42,10 +58,18 @@ static efi_status_t update_fdt(efi_system_table_t *sys_table, void *orig_fdt,
>  		}
>  	}
>  
> -	if (orig_fdt)
> +	if (orig_fdt) {
>  		status = fdt_open_into(orig_fdt, fdt, new_fdt_size);
> -	else
> +	} else {
>  		status = fdt_create_empty_tree(fdt, new_fdt_size);
> +		if (status == 0) {
> +			/*
> +			 * Any failure from the following function is non
> +			 * critical
> +			 */
> +			fdt_update_cell_size(sys_table, fdt);
> +		}
> +	}
>  
>  	if (status != 0)
>  		goto fdt_set_fail;
>
Mark Rutland Feb. 24, 2017, 3:36 p.m. UTC | #3
Hi,

On Tue, Feb 07, 2017 at 02:19:20PM -0700, Jeffrey Hugo wrote:
> From: Sameer Goel <sgoel@codeaurora.org>
> 
> In cases where a device tree is not provided (ie ACPI based system), an
> empty fdt is generated by efistub.  #address-cells and #size-cells are not
> set in the empty fdt, so they default to 1 (4 byte wide).  This can be an
> issue on 64-bit systems where values representing addresses, etc may be
> 8 bytes wide as the default value does not align with the general
> requirements for an empty DTB, and is fragile when passed to other agents
> as extra care is required to read the entire width of a value.
> 
> This issue is observed on Qualcomm Technologies QDF24XX platforms when
> kexec-tools inserts 64-bit addresses into the "linux,elfcorehdr" and
> "linux,usable-memory-range" properties of the fdt.  When the values are
> later consumed, they are truncated to 32-bit.
> 
> Setting #address-cells and #size-cells to 2 at creation of the empty fdt
> resolves the observed issue, and makes the fdt less fragile.
> 
> Signed-off-by: Sameer Goel <sgoel@codeaurora.org>
> Signed-off-by: Jeffrey Hugo <jhugo@codeaurora.org>

Technically the kdump ABI isn't set in stone yet, and this isn't a
problem until that goes in.

So while this generally looks ok, it's possible that this may be
unnecessary, and on reflection I do symptahise with Ard's earlier
comment that this shouldn't otherwise be necessary for the empty dt.

So I'd prefer if this were folded into the kdump series if it's
necessary. That way this goes in if it's necessary, and we can drop it
otherwise.

Thanks,
Mark.

> ---
> 
> [v2]
> -Add braces to an if when the corresponding else has braces
> -Remove print statements
> -Reword commit text
> -Removed gerrit tag
> 
>  drivers/firmware/efi/libstub/fdt.c | 28 ++++++++++++++++++++++++++--
>  1 file changed, 26 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/firmware/efi/libstub/fdt.c b/drivers/firmware/efi/libstub/fdt.c
> index 921dfa0..22ea73b 100644
> --- a/drivers/firmware/efi/libstub/fdt.c
> +++ b/drivers/firmware/efi/libstub/fdt.c
> @@ -16,6 +16,22 @@
>  
>  #include "efistub.h"
>  
> +#define EFI_DT_ADDR_CELLS_DEFAULT 2
> +#define EFI_DT_SIZE_CELLS_DEFAULT 2
> +
> +static void fdt_update_cell_size(efi_system_table_t *sys_table, void *fdt)
> +{
> +	int offset;
> +
> +	offset = fdt_path_offset(fdt, "/");
> +	/* Set the #address-cells and #size-cells values for an empty tree */
> +
> +	fdt_setprop_u32(fdt, offset, "#address-cells",
> +			EFI_DT_ADDR_CELLS_DEFAULT);
> +
> +	fdt_setprop_u32(fdt, offset, "#size-cells", EFI_DT_SIZE_CELLS_DEFAULT);
> +}
> +
>  static efi_status_t update_fdt(efi_system_table_t *sys_table, void *orig_fdt,
>  			       unsigned long orig_fdt_size,
>  			       void *fdt, int new_fdt_size, char *cmdline_ptr,
> @@ -42,10 +58,18 @@ static efi_status_t update_fdt(efi_system_table_t *sys_table, void *orig_fdt,
>  		}
>  	}
>  
> -	if (orig_fdt)
> +	if (orig_fdt) {
>  		status = fdt_open_into(orig_fdt, fdt, new_fdt_size);
> -	else
> +	} else {
>  		status = fdt_create_empty_tree(fdt, new_fdt_size);
> +		if (status == 0) {
> +			/*
> +			 * Any failure from the following function is non
> +			 * critical
> +			 */
> +			fdt_update_cell_size(sys_table, fdt);
> +		}
> +	}
>  
>  	if (status != 0)
>  		goto fdt_set_fail;
> -- 
> Qualcomm Datacenter Technologies as an affiliate of Qualcomm Technologies, Inc.
> Qualcomm Technologies, Inc. is a member of the
> Code Aurora Forum, a Linux Foundation Collaborative Project.
>
Jeffrey Hugo Feb. 24, 2017, 3:54 p.m. UTC | #4
On 2/24/2017 8:36 AM, Mark Rutland wrote:
> Hi,
>
> On Tue, Feb 07, 2017 at 02:19:20PM -0700, Jeffrey Hugo wrote:
>> From: Sameer Goel <sgoel@codeaurora.org>
>>
>> In cases where a device tree is not provided (ie ACPI based system), an
>> empty fdt is generated by efistub.  #address-cells and #size-cells are not
>> set in the empty fdt, so they default to 1 (4 byte wide).  This can be an
>> issue on 64-bit systems where values representing addresses, etc may be
>> 8 bytes wide as the default value does not align with the general
>> requirements for an empty DTB, and is fragile when passed to other agents
>> as extra care is required to read the entire width of a value.
>>
>> This issue is observed on Qualcomm Technologies QDF24XX platforms when
>> kexec-tools inserts 64-bit addresses into the "linux,elfcorehdr" and
>> "linux,usable-memory-range" properties of the fdt.  When the values are
>> later consumed, they are truncated to 32-bit.
>>
>> Setting #address-cells and #size-cells to 2 at creation of the empty fdt
>> resolves the observed issue, and makes the fdt less fragile.
>>
>> Signed-off-by: Sameer Goel <sgoel@codeaurora.org>
>> Signed-off-by: Jeffrey Hugo <jhugo@codeaurora.org>
>
> Technically the kdump ABI isn't set in stone yet, and this isn't a
> problem until that goes in.
>
> So while this generally looks ok, it's possible that this may be
> unnecessary, and on reflection I do symptahise with Ard's earlier
> comment that this shouldn't otherwise be necessary for the empty dt.
>
> So I'd prefer if this were folded into the kdump series if it's
> necessary. That way this goes in if it's necessary, and we can drop it
> otherwise.

Ok, I'll go coordinate with our folks looking at kdump.

>
> Thanks,
> Mark.
>
>> ---
>>
>> [v2]
>> -Add braces to an if when the corresponding else has braces
>> -Remove print statements
>> -Reword commit text
>> -Removed gerrit tag
>>
>>  drivers/firmware/efi/libstub/fdt.c | 28 ++++++++++++++++++++++++++--
>>  1 file changed, 26 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/firmware/efi/libstub/fdt.c b/drivers/firmware/efi/libstub/fdt.c
>> index 921dfa0..22ea73b 100644
>> --- a/drivers/firmware/efi/libstub/fdt.c
>> +++ b/drivers/firmware/efi/libstub/fdt.c
>> @@ -16,6 +16,22 @@
>>
>>  #include "efistub.h"
>>
>> +#define EFI_DT_ADDR_CELLS_DEFAULT 2
>> +#define EFI_DT_SIZE_CELLS_DEFAULT 2
>> +
>> +static void fdt_update_cell_size(efi_system_table_t *sys_table, void *fdt)
>> +{
>> +	int offset;
>> +
>> +	offset = fdt_path_offset(fdt, "/");
>> +	/* Set the #address-cells and #size-cells values for an empty tree */
>> +
>> +	fdt_setprop_u32(fdt, offset, "#address-cells",
>> +			EFI_DT_ADDR_CELLS_DEFAULT);
>> +
>> +	fdt_setprop_u32(fdt, offset, "#size-cells", EFI_DT_SIZE_CELLS_DEFAULT);
>> +}
>> +
>>  static efi_status_t update_fdt(efi_system_table_t *sys_table, void *orig_fdt,
>>  			       unsigned long orig_fdt_size,
>>  			       void *fdt, int new_fdt_size, char *cmdline_ptr,
>> @@ -42,10 +58,18 @@ static efi_status_t update_fdt(efi_system_table_t *sys_table, void *orig_fdt,
>>  		}
>>  	}
>>
>> -	if (orig_fdt)
>> +	if (orig_fdt) {
>>  		status = fdt_open_into(orig_fdt, fdt, new_fdt_size);
>> -	else
>> +	} else {
>>  		status = fdt_create_empty_tree(fdt, new_fdt_size);
>> +		if (status == 0) {
>> +			/*
>> +			 * Any failure from the following function is non
>> +			 * critical
>> +			 */
>> +			fdt_update_cell_size(sys_table, fdt);
>> +		}
>> +	}
>>
>>  	if (status != 0)
>>  		goto fdt_set_fail;
>> --
>> Qualcomm Datacenter Technologies as an affiliate of Qualcomm Technologies, Inc.
>> Qualcomm Technologies, Inc. is a member of the
>> Code Aurora Forum, a Linux Foundation Collaborative Project.
>>
AKASHI Takahiro March 2, 2017, 10:23 a.m. UTC | #5
Mark,

On Fri, Feb 24, 2017 at 03:36:44PM +0000, Mark Rutland wrote:
> Hi,
> 
> On Tue, Feb 07, 2017 at 02:19:20PM -0700, Jeffrey Hugo wrote:
> > From: Sameer Goel <sgoel@codeaurora.org>
> > 
> > In cases where a device tree is not provided (ie ACPI based system), an
> > empty fdt is generated by efistub.  #address-cells and #size-cells are not
> > set in the empty fdt, so they default to 1 (4 byte wide).  This can be an
> > issue on 64-bit systems where values representing addresses, etc may be
> > 8 bytes wide as the default value does not align with the general
> > requirements for an empty DTB, and is fragile when passed to other agents
> > as extra care is required to read the entire width of a value.
> > 
> > This issue is observed on Qualcomm Technologies QDF24XX platforms when
> > kexec-tools inserts 64-bit addresses into the "linux,elfcorehdr" and
> > "linux,usable-memory-range" properties of the fdt.  When the values are
> > later consumed, they are truncated to 32-bit.
> > 
> > Setting #address-cells and #size-cells to 2 at creation of the empty fdt
> > resolves the observed issue, and makes the fdt less fragile.
> > 
> > Signed-off-by: Sameer Goel <sgoel@codeaurora.org>
> > Signed-off-by: Jeffrey Hugo <jhugo@codeaurora.org>
> 
> Technically the kdump ABI isn't set in stone yet, and this isn't a
> problem until that goes in.
> 
> So while this generally looks ok, it's possible that this may be
> unnecessary, and on reflection I do symptahise with Ard's earlier
> comment that this shouldn't otherwise be necessary for the empty dt.
> 
> So I'd prefer if this were folded into the kdump series if it's
> necessary. That way this goes in if it's necessary, and we can drop it
> otherwise.

I don't have any problem in folding this patch into my kdump series,
but I don't know how we should address Robin's comment:

> Hang on, if this code is shared with 32-bit ARM, doesn't this just move
> the problem around?

Since the values of *-cells under root node should reflect the hardware,
the kernel has no way to determine whether they be 1 or 2 here.

Thinking of this circumstance, I'd prefer to always use 64-bit values
for the address and the size. This aligns with other properties under /chosen
node, like initrd-* and uefi-*, whose values are also 64-bit wide.

(I know that the current kernel cannot boot if the entire memory is
located at >4GB, but it's a different issue.)

Thanks,
-Takahiro AKASHI

> Thanks,
> Mark.
> 
> > ---
> > 
> > [v2]
> > -Add braces to an if when the corresponding else has braces
> > -Remove print statements
> > -Reword commit text
> > -Removed gerrit tag
> > 
> >  drivers/firmware/efi/libstub/fdt.c | 28 ++++++++++++++++++++++++++--
> >  1 file changed, 26 insertions(+), 2 deletions(-)
> > 
> > diff --git a/drivers/firmware/efi/libstub/fdt.c b/drivers/firmware/efi/libstub/fdt.c
> > index 921dfa0..22ea73b 100644
> > --- a/drivers/firmware/efi/libstub/fdt.c
> > +++ b/drivers/firmware/efi/libstub/fdt.c
> > @@ -16,6 +16,22 @@
> >  
> >  #include "efistub.h"
> >  
> > +#define EFI_DT_ADDR_CELLS_DEFAULT 2
> > +#define EFI_DT_SIZE_CELLS_DEFAULT 2
> > +
> > +static void fdt_update_cell_size(efi_system_table_t *sys_table, void *fdt)
> > +{
> > +	int offset;
> > +
> > +	offset = fdt_path_offset(fdt, "/");
> > +	/* Set the #address-cells and #size-cells values for an empty tree */
> > +
> > +	fdt_setprop_u32(fdt, offset, "#address-cells",
> > +			EFI_DT_ADDR_CELLS_DEFAULT);
> > +
> > +	fdt_setprop_u32(fdt, offset, "#size-cells", EFI_DT_SIZE_CELLS_DEFAULT);
> > +}
> > +
> >  static efi_status_t update_fdt(efi_system_table_t *sys_table, void *orig_fdt,
> >  			       unsigned long orig_fdt_size,
> >  			       void *fdt, int new_fdt_size, char *cmdline_ptr,
> > @@ -42,10 +58,18 @@ static efi_status_t update_fdt(efi_system_table_t *sys_table, void *orig_fdt,
> >  		}
> >  	}
> >  
> > -	if (orig_fdt)
> > +	if (orig_fdt) {
> >  		status = fdt_open_into(orig_fdt, fdt, new_fdt_size);
> > -	else
> > +	} else {
> >  		status = fdt_create_empty_tree(fdt, new_fdt_size);
> > +		if (status == 0) {
> > +			/*
> > +			 * Any failure from the following function is non
> > +			 * critical
> > +			 */
> > +			fdt_update_cell_size(sys_table, fdt);
> > +		}
> > +	}
> >  
> >  	if (status != 0)
> >  		goto fdt_set_fail;
> > -- 
> > Qualcomm Datacenter Technologies as an affiliate of Qualcomm Technologies, Inc.
> > Qualcomm Technologies, Inc. is a member of the
> > Code Aurora Forum, a Linux Foundation Collaborative Project.
> >
Timur Tabi March 2, 2017, 2:38 p.m. UTC | #6
AKASHI Takahiro wrote:
> Since the values of *-cells under root node should reflect the hardware,
> the kernel has no way to determine whether they be 1 or 2 here.

It's safe to set the *-cells to 2 on all platforms, as long as everyone who 
reads/writes from/to the device tree respects those values (which is what 
they're supposed to do anyway).
Goel, Sameer March 2, 2017, 3:23 p.m. UTC | #7
On 2/24/2017 8:36 AM, Robin Murphy wrote:
> On 07/02/17 21:19, Jeffrey Hugo wrote:
>> From: Sameer Goel <sgoel@codeaurora.org>
>>
>> In cases where a device tree is not provided (ie ACPI based system), an
>> empty fdt is generated by efistub.  #address-cells and #size-cells are not
>> set in the empty fdt, so they default to 1 (4 byte wide).  This can be an
>> issue on 64-bit systems where values representing addresses, etc may be
>> 8 bytes wide as the default value does not align with the general
>> requirements for an empty DTB, and is fragile when passed to other agents
>> as extra care is required to read the entire width of a value.
>>
>> This issue is observed on Qualcomm Technologies QDF24XX platforms when
>> kexec-tools inserts 64-bit addresses into the "linux,elfcorehdr" and
>> "linux,usable-memory-range" properties of the fdt.  When the values are
>> later consumed, they are truncated to 32-bit.
>>
>> Setting #address-cells and #size-cells to 2 at creation of the empty fdt
>> resolves the observed issue, and makes the fdt less fragile.
> 
> Hang on, if this code is shared with 32-bit ARM, doesn't this just move
> the problem around? If arm64 kexec is blindly punting reg properties
> into a DT assuming *-cells == 2, then wouldn't ARM kexec likely be doing
> the same relying on *-cells == 1?
> 
> Robin.
Kdump additions to the kexec code for arm64 is assuming 8 byte resource values 
for updating the chosen node with the address and size values. 
The 32 bit kexec is not updating the chosen node to set up the crashkernel start 
property and is passing the required parameters as command line args.

Any tool making the device tree update should be checking for the cell sizes before 
assuming a 32 bit or a 64 bit value.

Sameer   

> 
>> Signed-off-by: Sameer Goel <sgoel@codeaurora.org>
>> Signed-off-by: Jeffrey Hugo <jhugo@codeaurora.org>
>> ---
>>
>> [v2]
>> -Add braces to an if when the corresponding else has braces
>> -Remove print statements
>> -Reword commit text
>> -Removed gerrit tag
>>
>>  drivers/firmware/efi/libstub/fdt.c | 28 ++++++++++++++++++++++++++--
>>  1 file changed, 26 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/firmware/efi/libstub/fdt.c b/drivers/firmware/efi/libstub/fdt.c
>> index 921dfa0..22ea73b 100644
>> --- a/drivers/firmware/efi/libstub/fdt.c
>> +++ b/drivers/firmware/efi/libstub/fdt.c
>> @@ -16,6 +16,22 @@
>>  
>>  #include "efistub.h"
>>  
>> +#define EFI_DT_ADDR_CELLS_DEFAULT 2
>> +#define EFI_DT_SIZE_CELLS_DEFAULT 2
>> +
>> +static void fdt_update_cell_size(efi_system_table_t *sys_table, void *fdt)
>> +{
>> +	int offset;
>> +
>> +	offset = fdt_path_offset(fdt, "/");
>> +	/* Set the #address-cells and #size-cells values for an empty tree */
>> +
>> +	fdt_setprop_u32(fdt, offset, "#address-cells",
>> +			EFI_DT_ADDR_CELLS_DEFAULT);
>> +
>> +	fdt_setprop_u32(fdt, offset, "#size-cells", EFI_DT_SIZE_CELLS_DEFAULT);
>> +}
>> +
>>  static efi_status_t update_fdt(efi_system_table_t *sys_table, void *orig_fdt,
>>  			       unsigned long orig_fdt_size,
>>  			       void *fdt, int new_fdt_size, char *cmdline_ptr,
>> @@ -42,10 +58,18 @@ static efi_status_t update_fdt(efi_system_table_t *sys_table, void *orig_fdt,
>>  		}
>>  	}
>>  
>> -	if (orig_fdt)
>> +	if (orig_fdt) {
>>  		status = fdt_open_into(orig_fdt, fdt, new_fdt_size);
>> -	else
>> +	} else {
>>  		status = fdt_create_empty_tree(fdt, new_fdt_size);
>> +		if (status == 0) {
>> +			/*
>> +			 * Any failure from the following function is non
>> +			 * critical
>> +			 */
>> +			fdt_update_cell_size(sys_table, fdt);
>> +		}
>> +	}
>>  
>>  	if (status != 0)
>>  		goto fdt_set_fail;
>>
> 
> 
> _______________________________________________
> linux-arm-kernel mailing list
> linux-arm-kernel@lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
>
Timur Tabi March 8, 2017, 11:32 p.m. UTC | #8
On Thu, Mar 2, 2017 at 9:23 AM, Goel, Sameer <sgoel@codeaurora.org> wrote:

> Any tool making the device tree update should be checking for the cell sizes before
> assuming a 32 bit or a 64 bit value.

To be clear, Akashi's patches already look at the cell sizes:

+static int __init early_init_dt_scan_usablemem(unsigned long node,
+               const char *uname, int depth, void *data)
+{
+       struct memblock_region *usablemem = data;
+       const __be32 *reg;
+       int len;
+
+       if (depth != 1 || strcmp(uname, "chosen") != 0)
+               return 0;
+
+       reg = of_get_flat_dt_prop(node, "linux,usable-memory-range", &len);
+       if (!reg || (len < (dt_root_addr_cells + dt_root_size_cells)))
+               return 1;
+
+       usablemem->base = dt_mem_next_cell(dt_root_addr_cells, &reg);
+       usablemem->size = dt_mem_next_cell(dt_root_size_cells, &reg);
+
+       return 1;
+}

This function uses dt_root_addr_cells and dt_root_size_cells.  With
Sameer's patch, those values should be 2 and 2, respectively, on ARM64
systems.

Patch
diff mbox

diff --git a/drivers/firmware/efi/libstub/fdt.c b/drivers/firmware/efi/libstub/fdt.c
index 921dfa0..22ea73b 100644
--- a/drivers/firmware/efi/libstub/fdt.c
+++ b/drivers/firmware/efi/libstub/fdt.c
@@ -16,6 +16,22 @@ 
 
 #include "efistub.h"
 
+#define EFI_DT_ADDR_CELLS_DEFAULT 2
+#define EFI_DT_SIZE_CELLS_DEFAULT 2
+
+static void fdt_update_cell_size(efi_system_table_t *sys_table, void *fdt)
+{
+	int offset;
+
+	offset = fdt_path_offset(fdt, "/");
+	/* Set the #address-cells and #size-cells values for an empty tree */
+
+	fdt_setprop_u32(fdt, offset, "#address-cells",
+			EFI_DT_ADDR_CELLS_DEFAULT);
+
+	fdt_setprop_u32(fdt, offset, "#size-cells", EFI_DT_SIZE_CELLS_DEFAULT);
+}
+
 static efi_status_t update_fdt(efi_system_table_t *sys_table, void *orig_fdt,
 			       unsigned long orig_fdt_size,
 			       void *fdt, int new_fdt_size, char *cmdline_ptr,
@@ -42,10 +58,18 @@  static efi_status_t update_fdt(efi_system_table_t *sys_table, void *orig_fdt,
 		}
 	}
 
-	if (orig_fdt)
+	if (orig_fdt) {
 		status = fdt_open_into(orig_fdt, fdt, new_fdt_size);
-	else
+	} else {
 		status = fdt_create_empty_tree(fdt, new_fdt_size);
+		if (status == 0) {
+			/*
+			 * Any failure from the following function is non
+			 * critical
+			 */
+			fdt_update_cell_size(sys_table, fdt);
+		}
+	}
 
 	if (status != 0)
 		goto fdt_set_fail;