[<prev] [next>] [day] [month] [year] [list]
Message-Id: <20210429143811.2030717-1-mkorpershoek@baylibre.com>
Date: Thu, 29 Apr 2021 16:38:11 +0200
From: Mattijs Korpershoek <mkorpershoek@...libre.com>
To: Lee Jones <lee.jones@...aro.org>,
Matthias Brugger <matthias.bgg@...il.com>,
Chen Zhong <chen.zhong@...iatek.com>
Cc: fparent@...libre.com, dmitry.torokhov@...il.com,
Mattijs Korpershoek <mkorpershoek@...libre.com>,
linux-arm-kernel@...ts.infradead.org,
linux-mediatek@...ts.infradead.org, linux-kernel@...r.kernel.org
Subject: [PATCH] mfd: mt6397: rename mtk-pmic-keys cells to avoid OOPS
When booting with CONFIG_KEYBOARD_MTK_PMIC=y, we have the following oops:
[ 0.384359] Unable to handle kernel NULL pointer dereference at virtual address 00000000000000c0
[ 0.385471] Mem abort info:
[ 0.385826] ESR = 0x96000004
[ 0.386214] EC = 0x25: DABT (current EL), IL = 32 bits
[ 0.386886] SET = 0, FnV = 0
[ 0.387282] EA = 0, S1PTW = 0
[ 0.387681] Data abort info:
[ 0.388047] ISV = 0, ISS = 0x00000004
[ 0.388533] CM = 0, WnR = 0
[ 0.388910] [00000000000000c0] user address but active_mm is swapper
[ 0.389712] Internal error: Oops: 96000004 [#1] PREEMPT SMP
[ 0.390419] Modules linked in:
[ 0.390812] CPU: 4 PID: 1 Comm: swapper/0 Not tainted 5.12.0-mtk-04316-g7f74036c9b47 #58
[ 0.391835] Hardware name: Pumpkin MT8183 (DT)
[ 0.392399] pstate: 60000005 (nZCv daif -PAN -UAO -TCO BTYPE=--)
[ 0.393160] pc : mtk_pmic_keys_probe+0x8c/0x57c
[ 0.393744] lr : mtk_pmic_keys_probe+0x70/0x57c
[ 0.394318] sp : ffff80001008bb10
[ 0.394738] x29: ffff80001008bb10 x28: 0000000000000000
[ 0.395414] x27: ffff8000117f04b0 x26: ffff80001121aa88
[ 0.396089] x25: ffff0000031ea400 x24: ffff0000031ea410
[ 0.396763] x23: 0000000000000000 x22: ffff00000342ad80
[ 0.397438] x21: ffff000003589780 x20: 0000000000000000
[ 0.398112] x19: ffff80001129d0d0 x18: 0000000000003ca8
[ 0.398786] x17: ffff800010bad620 x16: ffff800010bacc44
[ 0.399460] x15: 0000000000000040 x14: ffffffffffffffff
[ 0.400134] x13: ffff800011df2728 x12: 0000000000000030
[ 0.400808] x11: 0101010101010101 x10: 7f7f7f7f7f7f7f7f
[ 0.401482] x9 : 0000000000000000 x8 : ffff000003589800
[ 0.402155] x7 : 0000000000000000 x6 : 000000000000003f
[ 0.402829] x5 : 0000000000000040 x4 : ffff0000031ea610
[ 0.403503] x3 : ffff0000031ea6b8 x2 : 0000000000000000
[ 0.404176] x1 : ffff000002d8d800 x0 : ffff0000031ea410
[ 0.404850] Call trace:
[ 0.405162] mtk_pmic_keys_probe+0x8c/0x57c
[ 0.405694] platform_probe+0x68/0xe0
[ 0.406162] really_probe+0x1c0/0x480
[ 0.406631] driver_probe_device.part.0+0xd8/0x118
[ 0.407239] device_driver_attach+0xcc/0xd4
[ 0.407773] __driver_attach+0x60/0x12c
[ 0.408261] bus_for_each_dev+0x70/0xc4
[ 0.408749] driver_attach+0x24/0x30
[ 0.409204] bus_add_driver+0x118/0x200
[ 0.409692] driver_register+0x78/0x130
[ 0.410180] __platform_driver_register+0x28/0x34
[ 0.410779] pmic_keys_pdrv_init+0x1c/0x28
[ 0.411301] do_one_initcall+0x54/0x1c0
[ 0.411791] kernel_init_freeable+0x1d4/0x23c
[ 0.412347] kernel_init+0x14/0x118
[ 0.412792] ret_from_fork+0x10/0x34
[ 0.413251] Code: aa1803e0 f90006b8 f94006c1 f9000aa1 (f9406281)
[ 0.414032] ---[ end trace ad605d8d1790625c ]---
[ 0.414630] Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b
[ 0.415600] SMP: stopping secondary CPUs
[ 0.416103] Kernel Offset: 0x80000 from 0xffff800010000000
[ 0.416796] PHYS_OFFSET: 0x40000000
[ 0.417238] CPU features: 0x00240002,2188200c
[ 0.417791] Memory Limit: none
[ 0.418181] ---[ end Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b ]---
This happens when we don't have a matching device-tree entry
(such as mediatek,mt6323-keys) for the driver.
In that case, we dereference of_id->data with of_id == NULL
The mfd core warns us about this:
[ 0.352175] mtk-pmic-keys: Failed to locate of_node [id: -1]
mtk_pmic_keys gets probed() anyways because its mfd_cell->name:
}, {
.name = "mtk-pmic-keys",
.num_resources = ARRAY_SIZE(mt6323_keys_resources),
.resources = mt6323_keys_resources,
.of_compatible = "mediatek,mt6323-keys"
}, {
.name is used by mfd_add_device() and becomes the name of the platform
device:
pdev = platform_device_alloc(cell->name, platform_id);
Later, in initcall, we do:
[ 0.296114] calling pmic_keys_pdrv_init+0x0/0x28 @ 1
[ 0.296142] bus: 'platform': add driver mtk-pmic-keys
Then, to check if we should call mtk_pmic_keys_probe(), we call
platform_match(drv, dev)
which will return true since drv->name and pdev->name match.
Since both the platform device is named mtk-pmic-keys and the driver is
called mtk-pmic-keys, we call probe(), which triggers the oops.
Fix it by renaming the mtk-pmic-key subdevices to right side of
their compatible.
This way, probe() will only be called when we have a valid dts entry for
mtk-pmic-keys.
Note: this is also more consistent naming with the other subdevices
such as mt6397-rtc and mt6397-codec.
Fixes: 55d1d1547ab5 ("mfd: mt6397: Add PMIC keys support to MT6397 driver")
Signed-off-by: Mattijs Korpershoek <mkorpershoek@...libre.com>
---
This is a replacement for:
https://patchwork.kernel.org/project/linux-mediatek/patch/20210428164219.1115537-2-mkorpershoek@baylibre.com/
drivers/mfd/mt6397-core.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/mfd/mt6397-core.c b/drivers/mfd/mt6397-core.c
index 7518d74c3b4c..a83fbc486d26 100644
--- a/drivers/mfd/mt6397-core.c
+++ b/drivers/mfd/mt6397-core.c
@@ -72,7 +72,7 @@ static const struct mfd_cell mt6323_devs[] = {
.name = "mt6323-led",
.of_compatible = "mediatek,mt6323-led"
}, {
- .name = "mtk-pmic-keys",
+ .name = "mt6323-keys",
.num_resources = ARRAY_SIZE(mt6323_keys_resources),
.resources = mt6323_keys_resources,
.of_compatible = "mediatek,mt6323-keys"
@@ -118,7 +118,7 @@ static const struct mfd_cell mt6397_devs[] = {
.name = "mt6397-pinctrl",
.of_compatible = "mediatek,mt6397-pinctrl",
}, {
- .name = "mtk-pmic-keys",
+ .name = "mt6397-keys",
.num_resources = ARRAY_SIZE(mt6397_keys_resources),
.resources = mt6397_keys_resources,
.of_compatible = "mediatek,mt6397-keys"
base-commit: 4a0225c3d208cfa6e4550f2210ffd9114a952a81
--
2.27.0
Powered by blists - more mailing lists