| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <4cbf5396-37e5-5e66-e26f-c43186e285da@linux.intel.com>
Date: Mon, 12 Aug 2024 15:19:23 +0300 (EEST)
From: Ilpo Järvinen <ilpo.jarvinen@...ux.intel.com>
To: Armin Wolf <W_Armin@....de>
cc: matan@...alib.org, ghostwind@...il.com,
Hans de Goede <hdegoede@...hat.com>, platform-driver-x86@...r.kernel.org,
LKML <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] platform/x86: lg-laptop: Add operation region support
On Fri, 9 Aug 2024, Armin Wolf wrote:
> The LEGX0820 ACPI device is expected to provide a custom operation
> region:
>
> OperationRegion (XIN1, 0x8F, Zero, 0x04B0)
> Field (XIN1, AnyAcc, Lock, Preserve)
> {
> DMSG, 8,
> HDAP, 8,
> Offset (0x03),
> AFNM, 8,
> Offset (0x10),
> P80B, 8,
> P81B, 8,
> P82B, 8,
> P83B, 8,
> P84B, 8,
> P85B, 8,
> P86B, 8,
> P87B, 8,
> Offset (0x20),
> DTTM, 8,
> TMP1, 8,
> LTP1, 8,
> HTP1, 8,
> TMP2, 8,
> LTP2, 8,
> HTP2, 8,
> Offset (0x3E8),
> PMSG, 1600
> }
>
> The PMSG field is used by AML code to log debug messages when DMSG is
> true. Since those debug messages are already logged using the standard
> ACPI Debug object, we set DMSG unconditionally to 0x00 and ignore any
> writes to PMSG.
>
> The TMPx, LTPx, HTPx and AFNM fields are used to inform the driver when
> the temperature/(presumably) trip points/fan mode changes. This only
> happens when the DTTM flag is set.
>
> Unfortunately we have to implement support for this operation region
> because the AML codes uses code constructs like this one:
>
> If (((\_SB.XINI.PLAV != Zero) && (\_SB.XINI.DTTM != Zero)))
>
> The PLAV field gets set to 1 when the driver registers its address space
> handler, so by default XIN1 should not be accessed.
>
> However ACPI does not use short-circuit evaluation when evaluating
> logical conditions. This causes the DTTM field to be accessed even
> when PLAV is 0, which results in an ACPI error.
> Since this check happens inside various thermal-related ACPI control
> methods, various thermal zone become unusable since any attempts to
attempt
> read their temperature results in an ACPI error.
>
> Fix this by providing support for this operation region. I suspect
> that the problem does not happen under Windows (which seemingly does
> not use short-circuit evaluation either) because the necessary driver
> comes preinstalled with the machine.
>
> Tested-by: Chris <ghostwind@...il.com>
> Signed-off-by: Armin Wolf <W_Armin@....de>
> ---
> drivers/platform/x86/lg-laptop.c | 128 +++++++++++++++++++++++++++++++
> 1 file changed, 128 insertions(+)
>
> diff --git a/drivers/platform/x86/lg-laptop.c b/drivers/platform/x86/lg-laptop.c
> index 9c7857842caf..6310a23f808c 100644
> --- a/drivers/platform/x86/lg-laptop.c
> +++ b/drivers/platform/x86/lg-laptop.c
> @@ -8,6 +8,8 @@
> #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
>
> #include <linux/acpi.h>
> +#include <linux/device.h>
> +#include <linux/dev_printk.h>
> #include <linux/dmi.h>
> #include <linux/input.h>
> #include <linux/input/sparse-keymap.h>
> @@ -31,6 +33,25 @@ MODULE_AUTHOR("Matan Ziv-Av");
> MODULE_DESCRIPTION("LG WMI Hotkey Driver");
> MODULE_LICENSE("GPL");
>
> +static bool fw_debug;
> +module_param(fw_debug, bool, 0);
> +MODULE_PARM_DESC(fw_debug, "Enable printing of firmware debug messages");
> +
> +#define LG_ADDRESS_SPACE_ID 0x8F
> +
> +#define LG_ADDRESS_SPACE_DEBUG_FLAG_ADR 0x00
> +#define LG_ADDRESS_SPACE_DEBUG_MSG_START_ADR 0x3E8
> +#define LG_ADDRESS_SPACE_DEBUG_MSG_END_ADR 0x5E8
> +
> +#define LG_ADDRESS_SPACE_DTTM_FLAG_ADR 0x20
> +#define LG_ADDRESS_SPACE_FAN_MODE_ADR 0x03
Any particular reason why there are not in numerical order?
> +#define LG_ADDRESS_SPACE_CPU_TEMP_ADR 0x21
> +#define LG_ADDRESS_SPACE_CPU_TRIP_LOW_ADR 0x22
> +#define LG_ADDRESS_SPACE_CPU_TRIP_HIGH_ADR 0x23
> +#define LG_ADDRESS_SPACE_MB_TEMP_ADR 0x24
> +#define LG_ADDRESS_SPACE_MB_TRIP_LOW_ADR 0x25
> +#define LG_ADDRESS_SPACE_MB_TRIP_HIGH_ADR 0x26
> +
> #define WMI_EVENT_GUID0 "E4FB94F9-7F2B-4173-AD1A-CD1D95086248"
> #define WMI_EVENT_GUID1 "023B133E-49D1-4E10-B313-698220140DC2"
> #define WMI_EVENT_GUID2 "37BE1AC0-C3F2-4B1F-BFBE-8FDEAF2814D6"
> @@ -646,6 +667,101 @@ static struct platform_driver pf_driver = {
> }
> };
>
> +static acpi_status lg_laptop_address_space_write(struct device *dev, acpi_physical_address address,
> + size_t size, u64 *value)
Why is this write function taking u64 _pointer_?
> +{
> + u8 byte;
> +
> + /* Ignore any debug messages */
> + if (address >= LG_ADDRESS_SPACE_DEBUG_MSG_START_ADR &&
> + address <= LG_ADDRESS_SPACE_DEBUG_MSG_END_ADR)
> + return AE_OK;
> +
> + if (size != sizeof(byte))
> + return AE_BAD_PARAMETER;
> +
> + byte = *value & 0xFF;
> +
> + switch (address) {
> + case LG_ADDRESS_SPACE_FAN_MODE_ADR:
> + dev_dbg(dev, "Fan mode set to mode %u\n", byte);
> + return AE_OK;
> + case LG_ADDRESS_SPACE_CPU_TEMP_ADR:
> + dev_dbg(dev, "CPU temperature is %u °C\n", byte);
> + return AE_OK;
> + case LG_ADDRESS_SPACE_CPU_TRIP_LOW_ADR:
> + dev_dbg(dev, "CPU lower trip point set to %u °C\n", byte);
> + return AE_OK;
> + case LG_ADDRESS_SPACE_CPU_TRIP_HIGH_ADR:
> + dev_dbg(dev, "CPU higher trip point set to %u °C\n", byte);
> + return AE_OK;
> + case LG_ADDRESS_SPACE_MB_TEMP_ADR:
> + dev_dbg(dev, "Motherboard temperature is %u °C\n", byte);
> + return AE_OK;
> + case LG_ADDRESS_SPACE_MB_TRIP_LOW_ADR:
> + dev_dbg(dev, "Motherboard lower trip point set to %u °C\n", byte);
> + return AE_OK;
> + case LG_ADDRESS_SPACE_MB_TRIP_HIGH_ADR:
> + dev_dbg(dev, "Motherboard higher trip point set to %u °C\n", byte);
> + return AE_OK;
> + default:
> + dev_notice_ratelimited(dev, "Ignoring write to unknown opregion address %llu\n",
> + address);
> + return AE_OK;
> + }
> +}
> +
> +static acpi_status lg_laptop_address_space_read(struct device *dev, acpi_physical_address address,
> + size_t size, u64 *value)
> +{
> + if (size != 1)
> + return AE_BAD_PARAMETER;
> +
> + switch (address) {
> + case LG_ADDRESS_SPACE_DEBUG_FLAG_ADR:
> + /* Debug messages are already printed using the standard ACPI Debug object */
> + *value = 0x00;
> + return AE_OK;
> + case LG_ADDRESS_SPACE_DTTM_FLAG_ADR:
> + *value = fw_debug;
> + return AE_OK;
> + default:
> + dev_notice_ratelimited(dev, "Attempt to read unknown opregion address %llu\n",
> + address);
> + return AE_BAD_PARAMETER;
> + }
> +}
> +
> +static acpi_status lg_laptop_address_space_handler(u32 function, acpi_physical_address address,
> + u32 bits, u64 *value, void *handler_context,
> + void *region_context)
> +{
> + struct device *dev = handler_context;
> + size_t size;
> +
> + if (bits % 8)
> + return AE_BAD_PARAMETER;
> +
> + size = bits / 8;
BITS_PER_BYTE x2
--
i.
> +
> + switch (function) {
> + case ACPI_READ:
> + return lg_laptop_address_space_read(dev, address, size, value);
> + case ACPI_WRITE:
> + return lg_laptop_address_space_write(dev, address, size, value);
> + default:
> + return AE_BAD_PARAMETER;
> + }
> +}
> +
> +static void lg_laptop_remove_address_space_handler(void *data)
> +{
> + struct acpi_device *device = data;
> +
> + acpi_remove_address_space_handler(device->handle, LG_ADDRESS_SPACE_ID,
> + &lg_laptop_address_space_handler);
> +}
> +
> static int acpi_add(struct acpi_device *device)
> {
> struct platform_device_info pdev_info = {
> @@ -653,6 +769,7 @@ static int acpi_add(struct acpi_device *device)
> .name = PLATFORM_NAME,
> .id = PLATFORM_DEVID_NONE,
> };
> + acpi_status status;
> int ret;
> const char *product;
> int year = 2017;
> @@ -660,6 +777,17 @@ static int acpi_add(struct acpi_device *device)
> if (pf_device)
> return 0;
>
> + status = acpi_install_address_space_handler(device->handle, LG_ADDRESS_SPACE_ID,
> + &lg_laptop_address_space_handler,
> + NULL, &device->dev);
> + if (ACPI_FAILURE(status))
> + return -ENODEV;
> +
> + ret = devm_add_action_or_reset(&device->dev, lg_laptop_remove_address_space_handler,
> + device);
> + if (ret < 0)
> + return ret;
> +
> ret = platform_driver_register(&pf_driver);
> if (ret)
> return ret;
> --
> 2.39.2
>
Powered by blists - more mailing lists