6.5-stable review patch. If anyone has any objections, please let me know.
------------------
From: Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx>
[ Upstream commit 046ece773cc77ef5d2a1431b188ac3d0840ed150 ]
In accordance with ACPI specificication and _DSD data buffer
representation the data there is an array of bytes. Hence,
accessing it with something longer will create a sparse data
which is against of how device property APIs work in general
and also not defined in the ACPI specification (see [1]).
Fix the code to emit an error if non-byte accessor is used to
retrieve _DSD buffer data.
Fixes: 369af6bf2c28 ("ACPI: property: Read buffer properties as integers")
Link: https://uefi.org/specs/ACPI/6.5/19_ASL_Reference.html#buffer-declare-buffer-object # [1]
Signed-off-by: Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx>
[ rjw: Add missing braces ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@xxxxxxxxx>
Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>
---
drivers/acpi/property.c | 19 ++++++++++---------
1 file changed, 10 insertions(+), 9 deletions(-)
diff --git a/drivers/acpi/property.c b/drivers/acpi/property.c
index 413e4fcadcaf7..99b4e33554355 100644
--- a/drivers/acpi/property.c
+++ b/drivers/acpi/property.c
@@ -1102,25 +1102,26 @@ static int acpi_data_prop_read(const struct acpi_device_data *data,
switch (proptype) {
case DEV_PROP_STRING:
break;
- case DEV_PROP_U8 ... DEV_PROP_U64:
+ default:
if (obj->type == ACPI_TYPE_BUFFER) {
if (nval > obj->buffer.length)
return -EOVERFLOW;
- break;
+ } else {
+ if (nval > obj->package.count)
+ return -EOVERFLOW;
}
- fallthrough;
- default:
- if (nval > obj->package.count)
- return -EOVERFLOW;
break;
}
if (nval == 0)
return -EINVAL;
- if (obj->type != ACPI_TYPE_BUFFER)
- items = obj->package.elements;
- else
+ if (obj->type == ACPI_TYPE_BUFFER) {
+ if (proptype != DEV_PROP_U8)
+ return -EPROTO;
items = obj;
+ } else {
+ items = obj->package.elements;
+ }
switch (proptype) {
case DEV_PROP_U8:
--
2.42.0
