memory: mtk-smi: Fix a null dereference for the ostd
authorYong Wu <yong.wu@mediatek.com>
Mon, 8 Nov 2021 08:24:29 +0000 (16:24 +0800)
committerKrzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Mon, 15 Nov 2021 14:07:24 +0000 (15:07 +0100)
commit8c5ba21c16bd7f8e23b8740dead6eaf164b8caa0
tree379010d3013db347a5d38d08b1fbe4a0dd1267b0
parentfa55b7dcdc43c1aa1ba12bca9d2dd4318c2a0dbf
memory: mtk-smi: Fix a null dereference for the ostd

We add the ostd setting for mt8195. It introduces a KE for the
previous SoC which doesn't have ostd setting. This is the log:

Unable to handle kernel NULL pointer dereference at virtual address
0000000000000080
...
pc : mtk_smi_larb_config_port_gen2_general+0x64/0x130
lr : mtk_smi_larb_resume+0x54/0x98
...
Call trace:
 mtk_smi_larb_config_port_gen2_general+0x64/0x130
 pm_generic_runtime_resume+0x2c/0x48
 __genpd_runtime_resume+0x30/0xa8
 genpd_runtime_resume+0x94/0x2c8
 __rpm_callback+0x44/0x150
 rpm_callback+0x6c/0x78
 rpm_resume+0x310/0x558
 __pm_runtime_resume+0x3c/0x88

In the code: larbostd = larb->larb_gen->ostd[larb->larbid],
if "larb->larb_gen->ostd" is null, the "larbostd" is the offset(e.g.
0x80 above), it's also a valid value, then accessing "larbostd[i]" in the
"for" loop will cause the KE above. To avoid this issue, initialize
"larbostd" to NULL when the SoC doesn't have ostd setting.

Fixes: fe6dd2a4017d ("memory: mtk-smi: mt8195: Add initial setting for smi-larb")
Signed-off-by: Yong Wu <yong.wu@mediatek.com>
Link: https://lore.kernel.org/r/20211108082429.15080-1-yong.wu@mediatek.com
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
drivers/memory/mtk-smi.c