Message-ID: <20260102115250.0000045c@huawei.com>
Date: Fri, 2 Jan 2026 11:52:50 +0000
From: Jonathan Cameron <jonathan.cameron@...wei.com>
To: "Rafael J. Wysocki" <rafael@...nel.org>
CC: Linux ACPI <linux-acpi@...r.kernel.org>, Dan Carpenter
	<dan.carpenter@...aro.org>, LKML <linux-kernel@...r.kernel.org>, Linux PCI
	<linux-pci@...r.kernel.org>, Bjorn Helgaas <helgaas@...nel.org>, "Srinivas
 Pandruvada" <srinivas.pandruvada@...ux.intel.com>, Hans de Goede
	<hansg@...nel.org>, Mario Limonciello <mario.limonciello@....com>
Subject: Re: [PATCH v1] ACPI: bus: Adjust acpi_osc_handshake() parameter
 list

On Fri, 26 Dec 2025 14:48:45 +0100
"Rafael J. Wysocki" <rafael@...nel.org> wrote:

> From: Rafael J. Wysocki <rafael.j.wysocki@...el.com>
> 
> For the sake of interface cleanliness, it is better to avoid using
> ACPICA data types in the parameter lists of helper functions that
> don't belong to ACPICA, so adjust the parameter list of recently
> introduced acpi_osc_handshake() to take a capabilities buffer pointer
> and the size of the buffer (in u32 size units) as parameters directly
> instead of a struct acpi_buffer pointer.
> 
> This is also somewhat more straightforward on the caller side because
> they won't need to create struct acpi_buffer objects themselves to pass
> them to the helper function and it guarantees that the size of the
> buffer in bytes will always be a multiple of 4 (the size of u32).
> 
> Moreover, it addresses a premature cap pointer dereference and
> eliminates a sizeof(32) that should have been sizeof(u32) [1].
> 
> Fixes: e5322888e6bf ("ACPI: bus: Rework the handling of \_SB._OSC platform features")
> Reported-by: kernel test robot <lkp@...el.com>
> Reported-by: Dan Carpenter <dan.carpenter@...aro.org>
> Closes: https://lore.kernel.org/linux-acpi/202512242052.W4GhDauV-lkp@intel.com/
> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@...el.com>
A couple of minor comments inline.  I see you have it queued up already, but
FWIW nothing here major enough to warrant reverting that.
Reviewed-by: Jonathan Cameron <jonathan.cameron@...wei.com>

> ---
>  drivers/acpi/bus.c |   30 ++++++++++++------------------
>  1 file changed, 12 insertions(+), 18 deletions(-)
> 
> --- a/drivers/acpi/bus.c
> +++ b/drivers/acpi/bus.c
> @@ -326,31 +326,33 @@ out:
>  EXPORT_SYMBOL(acpi_run_osc);
>  
>  static int acpi_osc_handshake(acpi_handle handle, const char *uuid_str,
> -			      int rev, struct acpi_buffer *cap)
> +			      int rev, u32 *capbuf, size_t bufsize)

Size parameters in number of u32s is to me a little confusing but
I guess this is only used locally so that's probably fine.
I'd have been tempted to call it dwords or something like that.


>  {
>  	union acpi_object in_params[4], *out_obj;
> -	size_t bufsize = cap->length / sizeof(u32);
>  	struct acpi_object_list input;
> +	struct acpi_buffer cap = {
> +		.pointer = capbuf,
> +		.length = bufsize * sizeof(32),

You fixed this up already but just for completeness.
sizeof(u32)

> +	};
>  	struct acpi_buffer output;
> -	u32 *capbuf, *retbuf, test;
> +	u32 *retbuf, test;
>  	guid_t guid;
>  	int ret, i;
>  
> -	if (!cap || cap->length < 2 * sizeof(32) || guid_parse(uuid_str, &guid))
> +	if (!capbuf || bufsize < 2 || guid_parse(uuid_str, &guid))
>  		return -EINVAL;
>  
>  	/* First evaluate _OSC with OSC_QUERY_ENABLE set. */
> -	capbuf = cap->pointer;
>  	capbuf[OSC_QUERY_DWORD] = OSC_QUERY_ENABLE;
>  
> -	ret = acpi_eval_osc(handle, &guid, rev, cap, in_params, &output);
> +	ret = acpi_eval_osc(handle, &guid, rev, &cap, in_params, &output);
>  	if (ret)
>  		return ret;
>  
>  	out_obj = output.pointer;
>  	retbuf = (u32 *)out_obj->buffer.pointer;
>  
> -	if (acpi_osc_error_check(handle, &guid, rev, cap, retbuf)) {
> +	if (acpi_osc_error_check(handle, &guid, rev, &cap, retbuf)) {
>  		ret = -ENODATA;
>  		goto out;
>  	}
> @@ -403,7 +405,7 @@ static int acpi_osc_handshake(acpi_handl
>  		 */
>  		acpi_handle_err(handle, "_OSC: errors while processing control request\n");
>  		acpi_handle_err(handle, "_OSC: some features may be missing\n");
> -		acpi_osc_error_check(handle, &guid, rev, cap, retbuf);
> +		acpi_osc_error_check(handle, &guid, rev, &cap, retbuf);
>  	}
>  
>  out:
> @@ -446,10 +448,6 @@ static void acpi_bus_osc_negotiate_platf
>  {
>  	static const u8 sb_uuid_str[] = "0811B06E-4A27-44F9-8D60-3CBBC22E7B48";
>  	u32 capbuf[2], feature_mask;
> -	struct acpi_buffer cap = {
> -		.pointer = capbuf,
> -		.length = sizeof(capbuf),
> -	};
>  	acpi_handle handle;
>  
>  	feature_mask = OSC_SB_PR3_SUPPORT | OSC_SB_HOTPLUG_OST_SUPPORT |
> @@ -497,7 +495,7 @@ static void acpi_bus_osc_negotiate_platf
>  
>  	acpi_handle_info(handle, "platform _OSC: OS support mask [%08x]\n", feature_mask);
>  
> -	if (acpi_osc_handshake(handle, sb_uuid_str, 1, &cap))
> +	if (acpi_osc_handshake(handle, sb_uuid_str, 1, capbuf, 2))

As below. Maybe ARRAY_SIZE(capbuf) instead of that 2.

>  		return;
>  
>  	feature_mask = capbuf[OSC_SUPPORT_DWORD];
> @@ -532,10 +530,6 @@ static void acpi_bus_osc_negotiate_usb_c
>  {
>  	static const u8 sb_usb_uuid_str[] = "23A0D13A-26AB-486C-9C5F-0FFA525A575A";
>  	u32 capbuf[3], control;
> -	struct acpi_buffer cap = {
> -		.pointer = capbuf,
> -		.length = sizeof(capbuf),
> -	};
>  	acpi_handle handle;
>  
>  	if (!osc_sb_native_usb4_support_confirmed)
> @@ -550,7 +544,7 @@ static void acpi_bus_osc_negotiate_usb_c
>  	capbuf[OSC_SUPPORT_DWORD] = 0;
>  	capbuf[OSC_CONTROL_DWORD] = control;
>  
> -	if (acpi_osc_handshake(handle, sb_usb_uuid_str, 1, &cap))
> +	if (acpi_osc_handshake(handle, sb_usb_uuid_str, 1, capbuf, 3))

Maybe ARRAY_SIZE(capbuf) just to avoid any chance they get out of sync?

>  		return;
>  
>  	osc_sb_native_usb4_control = capbuf[OSC_CONTROL_DWORD];
> 
> 
> 
> 

Powered by blists - more mailing lists