include/linux/platform_data/x86/intel_pmc_ipc.h | 1 + 1 file changed, 1 insertion(+)
From: Yongxin Liu <yongxin.liu@windriver.com>
The intel_pmc_ipc() function uses ACPI_ALLOCATE_BUFFER to allocate memory
for the ACPI evaluation result but never frees it, causing a 192-byte
memory leak on each call.
This leak is triggered during network interface initialization when the
stmmac driver calls intel_mac_finish() -> intel_pmc_ipc().
unreferenced object 0xffff96a848d6ea80 (size 192):
comm "dhcpcd", pid 541, jiffies 4294684345
hex dump (first 32 bytes):
04 00 00 00 05 00 00 00 98 ea d6 48 a8 96 ff ff ...........H....
00 00 00 00 00 00 00 00 01 00 00 00 00 00 00 00 ................
backtrace (crc b1564374):
kmemleak_alloc+0x2d/0x40
__kmalloc_noprof+0x2fa/0x730
acpi_ut_initialize_buffer+0x83/0xc0
acpi_evaluate_object+0x29a/0x2f0
intel_pmc_ipc+0xfd/0x170
intel_mac_finish+0x168/0x230
stmmac_mac_finish+0x3d/0x50
phylink_major_config+0x22b/0x5b0
phylink_mac_initial_config.constprop.0+0xf1/0x1b0
phylink_start+0x8e/0x210
__stmmac_open+0x12c/0x2b0
stmmac_open+0x23c/0x380
__dev_open+0x11d/0x2c0
__dev_change_flags+0x1d2/0x250
netif_change_flags+0x2b/0x70
dev_change_flags+0x40/0xb0
Add kfree() to properly release the allocated buffer.
Cc: stable@vger.kernel.org
Fixes: 7e2f7e25f6ff ("arch: x86: add IPC mailbox accessor function and add SoC register access")
Signed-off-by: Yongxin Liu <yongxin.liu@windriver.com>
---
include/linux/platform_data/x86/intel_pmc_ipc.h | 1 +
1 file changed, 1 insertion(+)
diff --git a/include/linux/platform_data/x86/intel_pmc_ipc.h b/include/linux/platform_data/x86/intel_pmc_ipc.h
index 1d34435b7001..2fd5e684ce26 100644
--- a/include/linux/platform_data/x86/intel_pmc_ipc.h
+++ b/include/linux/platform_data/x86/intel_pmc_ipc.h
@@ -89,6 +89,7 @@ static inline int intel_pmc_ipc(struct pmc_ipc_cmd *ipc_cmd, struct pmc_ipc_rbuf
return -EINVAL;
}
+ kfree (obj);
return 0;
#else
return -ENODEV;
--
2.46.2
On Mon, 24 Nov 2025, yongxin.liu@windriver.com wrote:
> From: Yongxin Liu <yongxin.liu@windriver.com>
>
> The intel_pmc_ipc() function uses ACPI_ALLOCATE_BUFFER to allocate memory
> for the ACPI evaluation result but never frees it, causing a 192-byte
> memory leak on each call.
>
> This leak is triggered during network interface initialization when the
> stmmac driver calls intel_mac_finish() -> intel_pmc_ipc().
>
> unreferenced object 0xffff96a848d6ea80 (size 192):
> comm "dhcpcd", pid 541, jiffies 4294684345
> hex dump (first 32 bytes):
> 04 00 00 00 05 00 00 00 98 ea d6 48 a8 96 ff ff ...........H....
> 00 00 00 00 00 00 00 00 01 00 00 00 00 00 00 00 ................
> backtrace (crc b1564374):
> kmemleak_alloc+0x2d/0x40
> __kmalloc_noprof+0x2fa/0x730
> acpi_ut_initialize_buffer+0x83/0xc0
> acpi_evaluate_object+0x29a/0x2f0
> intel_pmc_ipc+0xfd/0x170
> intel_mac_finish+0x168/0x230
> stmmac_mac_finish+0x3d/0x50
> phylink_major_config+0x22b/0x5b0
> phylink_mac_initial_config.constprop.0+0xf1/0x1b0
> phylink_start+0x8e/0x210
> __stmmac_open+0x12c/0x2b0
> stmmac_open+0x23c/0x380
> __dev_open+0x11d/0x2c0
> __dev_change_flags+0x1d2/0x250
> netif_change_flags+0x2b/0x70
> dev_change_flags+0x40/0xb0
>
> Add kfree() to properly release the allocated buffer.
>
> Cc: stable@vger.kernel.org
> Fixes: 7e2f7e25f6ff ("arch: x86: add IPC mailbox accessor function and add SoC register access")
> Signed-off-by: Yongxin Liu <yongxin.liu@windriver.com>
> ---
> include/linux/platform_data/x86/intel_pmc_ipc.h | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/include/linux/platform_data/x86/intel_pmc_ipc.h b/include/linux/platform_data/x86/intel_pmc_ipc.h
> index 1d34435b7001..2fd5e684ce26 100644
> --- a/include/linux/platform_data/x86/intel_pmc_ipc.h
> +++ b/include/linux/platform_data/x86/intel_pmc_ipc.h
> @@ -89,6 +89,7 @@ static inline int intel_pmc_ipc(struct pmc_ipc_cmd *ipc_cmd, struct pmc_ipc_rbuf
> return -EINVAL;
> }
>
> + kfree (obj);
Good catch but this fix doesn't address all possible paths. So please use
cleanup.h instead:
union acpi_object *obj __free(kfree) = buffer.pointer;
And don't forget to add the #include.
> return 0;
> #else
> return -ENODEV;
>
--
i.
> Good catch but this fix doesn't address all possible paths. So please use
> cleanup.h instead:
>
> union acpi_object *obj __free(kfree) = buffer.pointer;
>
> And don't forget to add the #include.
https://www.kernel.org/doc/html/latest/process/maintainer-netdev.html
1.6.5. Using device-managed and cleanup.h constructs¶
Low level cleanup constructs (such as __free()) can be used when
building APIs and helpers, especially scoped iterators. However,
direct use of __free() within networking core and drivers is
discouraged. Similar guidance applies to declaring variables
mid-function.
Andrew
---
pw-bot: cr
© 2016 - 2025 Red Hat, Inc.