Skip to contentSkip to conversation

xen/dt-overlay: support phandle-based targeting in overlay_get_nodes_info

PATCH In review Michal Orzel 3 participants 6 messages Apr 28

lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 xen/dt-overlay: support phandle-based targeting in overlay_get_nodes_info

overlay_get_nodes_info() is called before fdt_overlay_apply() to extract target paths from the overlay. This fails for overlays using phandle-based targeting (target = <&label>) because DTC compiles these as unresolved fixups (target = <0xffffffff>), causing fdt_overlay_target_offset() to return -FDT_ERR_BADPHANDLE. Prior to this change users were forced to manually modify the dtbo (even for hwdom) to switch from target to target-phandle by manually inspecting also the host DTB.

Introduce overlay_get_target_path() which directly handles the two targeting cases that occur before fixup resolution:
- target-path: the string property is returned directly.
- target = <&label>: the label is found in the overlay's __fixups__ node, then resolved to a path via the base DTB's __symbols__ node.

Libfdt fdt_for_each_property_offset() violates MISRA R20.7. Despite libfdt being excluded from the analysis, this causes ECLAIR scan to report a regression because the violation introduced by a macro is reported at the call site. Deviate R20.7 for libfdt.h.

---
In my pre-push testing I realized ECLAIR allcode reported a violation related to the use of fdt_for_each_property_offset. Libfdt is excluded from analysis but this does not apply to e.g. macros used in other places.

Changes in v2:
- strip out from series that is ready to be committed
- deviate 20.7 for libfdt.h. Discussed with Nicole.

https://gitlab.com/xen-project/people/morzel/xen/-/pipelines/2484326830

---
 .../eclair_analysis/ECLAIR/deviations.ecl     |  1 +
 xen/common/device-tree/dt-overlay.c           | 65 +++++++++++++++++--
 2 files changed, 60 insertions(+), 6 deletions(-)
automation/eclair_analysis/ECLAIR/deviations.ecl +1 -0
index 30c323906924..5f0b73062474 100644
@@ -631,6 +631,7 @@ not in scope for compliance are allowed, as that is imported code."
 -file_tag+={acpi_cpu_idle, "^xen/arch/x86/acpi/cpu_idle\\.c$"}
 -config=MC3A2.R20.7,reports+={safe, "any_area(any_loc(file(gnu_efi_include)||any_exp(macro(^NextMemoryDescriptor$))))"}
 -config=MC3A2.R20.7,reports+={safe, "any_area(any_loc(file(acpi_cpu_idle)))"}
+-config=MC3A2.R20.7,reports+={safe, "any_area(any_loc(any_exp(file(^xen/include/xen/libfdt/libfdt.h$))))"}
 -doc_end
 
 -doc_begin="To avoid compromising readability, the macros alternative_(v)?call[0-9] are allowed
xen/common/device-tree/dt-overlay.c +59 -6

83 lines not on this page · raw message

Signature · 1 line
2.43.0
#1 lore raw
Quoted 18 lines · Michal Orzel

overlay_get_nodes_info() is called before fdt_overlay_apply() to extract target paths from the overlay. This fails for overlays using phandle-based targeting (target = <&label>) because DTC compiles these as unresolved fixups (target = <0xffffffff>), causing fdt_overlay_target_offset() to return -FDT_ERR_BADPHANDLE. Prior to this change users were forced to manually modify the dtbo (even for hwdom) to switch from target to target-phandle by manually inspecting also the host DTB.

Introduce overlay_get_target_path() which directly handles the two targeting cases that occur before fixup resolution:
- target-path: the string property is returned directly.
- target = <&label>: the label is found in the overlay's __fixups__ node, then resolved to a path via the base DTB's __symbols__ node.

Quoted 8 lines

Libfdt fdt_for_each_property_offset() violates MISRA R20.7. Despite libfdt being excluded from the analysis, this causes ECLAIR scan to report a regression because the violation introduced by a macro is reported at the call site. Deviate R20.7 for libfdt.h.

NIT: shouldn’t these be in chronological order?

Cheers,
Luca

#2 lore raw
Quoted 23 lines · Luca Fancellu
Quoted 23 lines · Michal Orzel

overlay_get_nodes_info() is called before fdt_overlay_apply() to extract target paths from the overlay. This fails for overlays using phandle-based targeting (target = <&label>) because DTC compiles these as unresolved fixups (target = <0xffffffff>), causing fdt_overlay_target_offset() to return -FDT_ERR_BADPHANDLE. Prior to this change users were forced to manually modify the dtbo (even for hwdom) to switch from target to target-phandle by manually inspecting also the host DTB.

Introduce overlay_get_target_path() which directly handles the two targeting cases that occur before fixup resolution:
- target-path: the string property is returned directly.
- target = <&label>: the label is found in the overlay's __fixups__ node, then resolved to a path via the base DTB's __symbols__ node.

Libfdt fdt_for_each_property_offset() violates MISRA R20.7. Despite libfdt being excluded from the analysis, this causes ECLAIR scan to report a regression because the violation introduced by a macro is reported at the call site. Deviate R20.7 for libfdt.h.

Quoted 5 lines
Quoted 3 lines

NIT: shouldn’t these be in chronological order?

Yes, they should. It's b4 that messes things up. Anyway, to be done on commit.

~Michal

#3 lore raw
In reply to Michal Orzel
Quoted 43 lines · Michal Orzel

overlay_get_nodes_info() is called before fdt_overlay_apply() to extract target paths from the overlay. This fails for overlays using phandle-based targeting (target = <&label>) because DTC compiles these as unresolved fixups (target = <0xffffffff>), causing fdt_overlay_target_offset() to return -FDT_ERR_BADPHANDLE. Prior to this change users were forced to manually modify the dtbo (even for hwdom) to switch from target to target-phandle by manually inspecting also the host DTB.

Introduce overlay_get_target_path() which directly handles the two targeting cases that occur before fixup resolution:
- target-path: the string property is returned directly.
- target = <&label>: the label is found in the overlay's __fixups__ node, then resolved to a path via the base DTB's __symbols__ node.

Libfdt fdt_for_each_property_offset() violates MISRA R20.7. Despite libfdt being excluded from the analysis, this causes ECLAIR scan to report a regression because the violation introduced by a macro is reported at the call site. Deviate R20.7 for libfdt.h.

---
In my pre-push testing I realized ECLAIR allcode reported a violation related to the use of fdt_for_each_property_offset. Libfdt is excluded from analysis but this does not apply to e.g. macros used in other places.

Changes in v2:
- strip out from series that is ready to be committed
- deviate 20.7 for libfdt.h. Discussed with Nicole.

https://gitlab.com/xen-project/people/morzel/xen/-/pipelines/2484326830

---
 .../eclair_analysis/ECLAIR/deviations.ecl     |  1 +
 xen/common/device-tree/dt-overlay.c           | 65 +++++++++++++++++--
 2 files changed, 60 insertions(+), 6 deletions(-)
automation/eclair_analysis/ECLAIR/deviations.ecl +0 -0
index 30c323906924..5f0b73062474 100644
@@ -631,6 +631,7 @@ not in scope for compliance are allowed, as that is 

imported code."

Quoted 9 lines

-file_tag+={acpi_cpu_idle, "^xen/arch/x86/acpi/cpu_idle\\.c$"} -config=MC3A2.R20.7,reports+={safe,
"any_area(any_loc(file(gnu_efi_include)||any_exp(macro(^NextMemoryDescriptor$))))"}
-config=MC3A2.R20.7,reports+={safe,
"any_area(any_loc(file(acpi_cpu_idle)))"} +-config=MC3A2.R20.7,reports+={safe, "any_area(any_loc(any_exp(file(^xen/include/xen/libfdt/libfdt.h$))))"}
-doc_end

Nit: I think using a file_tag as above is better, but in any case use libfdt\\.h in the regex

Quoted 2 lines

-doc_begin="To avoid compromising readability, the macros
alternative_(v)?call[0-9] are allowed

Signature · 4 lines
Nicola Vetrini, B.Sc.
Software Engineer
BUGSENG (https://bugseng.com)
LinkedIn: https://www.linkedin.com/in/nicola-vetrini-a42471253
#4 lore raw
Quoted 53 lines · Nicola Vetrini
Quoted 52 lines · Michal Orzel

overlay_get_nodes_info() is called before fdt_overlay_apply() to extract
target paths from the overlay. This fails for overlays using phandle-based
targeting (target = <&label>) because DTC compiles these as unresolved fixups (target = <0xffffffff>), causing fdt_overlay_target_offset() to return -FDT_ERR_BADPHANDLE. Prior to this change users were forced to manually modify the dtbo (even for hwdom) to switch from target to target-phandle by manually inspecting also the host DTB.

Introduce overlay_get_target_path() which directly handles the two targeting cases that occur before fixup resolution:
- target-path: the string property is returned directly.
- target = <&label>: the label is found in the overlay's __fixups__ node, then resolved to a path via the base DTB's __symbols__ node.

Libfdt fdt_for_each_property_offset() violates MISRA R20.7. Despite libfdt being excluded from the analysis, this causes ECLAIR scan to report a regression because the violation introduced by a macro is reported at the call site. Deviate R20.7 for libfdt.h.

---
In my pre-push testing I realized ECLAIR allcode reported a violation related to
the use of fdt_for_each_property_offset. Libfdt is excluded from analysis but
this does not apply to e.g. macros used in other places.

Changes in v2:
- strip out from series that is ready to be committed
- deviate 20.7 for libfdt.h. Discussed with Nicole.

https://gitlab.com/xen-project/people/morzel/xen/-/pipelines/2484326830

---
 .../eclair_analysis/ECLAIR/deviations.ecl     |  1 +
 xen/common/device-tree/dt-overlay.c           | 65 +++++++++++++++++--
 2 files changed, 60 insertions(+), 6 deletions(-)
automation/eclair_analysis/ECLAIR/deviations.ecl +0 -0
index 30c323906924..5f0b73062474 100644
@@ -631,6 +631,7 @@ not in scope for compliance are allowed, as that is 

imported code."
-file_tag+={acpi_cpu_idle, "^xen/arch/x86/acpi/cpu_idle\\.c$"} -config=MC3A2.R20.7,reports+={safe,
"any_area(any_loc(file(gnu_efi_include)||any_exp(macro(^NextMemoryDescriptor$))))"}
-config=MC3A2.R20.7,reports+={safe,

Quoted 8 lines
Quoted 5 lines

"any_area(any_loc(file(acpi_cpu_idle)))"} +-config=MC3A2.R20.7,reports+={safe, "any_area(any_loc(any_exp(file(^xen/include/xen/libfdt/libfdt.h$))))"}
-doc_end

Nit: I think using a file_tag as above is better, but in any case use libfdt\\.h in the regex

Ok. Provided I will do this on commit, may I convert this as your Rb/Ab?

~Michal

Reviewed-by
#5 lore raw
Quoted 64 lines · Michal
Quoted 55 lines · Nicola Vetrini
Quoted 54 lines · Michal Orzel

overlay_get_nodes_info() is called before fdt_overlay_apply() to extract
target paths from the overlay. This fails for overlays using phandle-based
targeting (target = <&label>) because DTC compiles these as unresolved fixups (target = <0xffffffff>), causing fdt_overlay_target_offset() to return -FDT_ERR_BADPHANDLE. Prior to this change users were forced to manually modify the dtbo (even for hwdom) to switch from target to target-phandle by manually inspecting also the host DTB.

Introduce overlay_get_target_path() which directly handles the two targeting cases that occur before fixup resolution:
- target-path: the string property is returned directly.
- target = <&label>: the label is found in the overlay's __fixups__ node, then resolved to a path via the base DTB's __symbols__ node.

Libfdt fdt_for_each_property_offset() violates MISRA R20.7. Despite libfdt being excluded from the analysis, this causes ECLAIR scan to report a regression because the violation introduced by a macro is reported at the call site. Deviate R20.7 for libfdt.h.

---
In my pre-push testing I realized ECLAIR allcode reported a violation related to
the use of fdt_for_each_property_offset. Libfdt is excluded from analysis but
this does not apply to e.g. macros used in other places.

Changes in v2:
- strip out from series that is ready to be committed
- deviate 20.7 for libfdt.h. Discussed with Nicole.

https://gitlab.com/xen-project/people/morzel/xen/-/pipelines/2484326830

---
 .../eclair_analysis/ECLAIR/deviations.ecl     |  1 +
 xen/common/device-tree/dt-overlay.c           | 65 
+++++++++++++++++--
 2 files changed, 60 insertions(+), 6 deletions(-)
automation/eclair_analysis/ECLAIR/deviations.ecl +0 -0
index 30c323906924..5f0b73062474 100644
@@ -631,6 +631,7 @@ not in scope for compliance are allowed, as that 

is
imported code."
-file_tag+={acpi_cpu_idle, "^xen/arch/x86/acpi/cpu_idle\\.c$"} -config=MC3A2.R20.7,reports+={safe,
"any_area(any_loc(file(gnu_efi_include)||any_exp(macro(^NextMemoryDescriptor$))))"}
-config=MC3A2.R20.7,reports+={safe,

Quoted 8 lines
Quoted 5 lines

"any_area(any_loc(file(acpi_cpu_idle)))"} +-config=MC3A2.R20.7,reports+={safe, "any_area(any_loc(any_exp(file(^xen/include/xen/libfdt/libfdt.h$))))"}
-doc_end

Nit: I think using a file_tag as above is better, but in any case use libfdt\\.h in the regex

Quoted 3 lines

Ok. Provided I will do this on commit, may I convert this as your Rb/Ab?

~Michal

Yes, sure

Signature · 4 lines
Nicola Vetrini, B.Sc.
Software Engineer
BUGSENG (https://bugseng.com)
LinkedIn: https://www.linkedin.com/in/nicola-vetrini-a42471253