regulator: core: Use different devices for resource allocation and DT lookup
authorChiYuan Huang <cy_huang@richtek.com>
Tue, 6 Dec 2022 07:22:21 +0000 (15:22 +0800)
committerMark Brown <broonie@kernel.org>
Thu, 8 Dec 2022 13:02:37 +0000 (13:02 +0000)
commit8f3cbcd6b440032ebc7f7d48a1689dcc70a4eb98
treef132fd4e1de3e7dc4990b6b5609ed701a91ce979
parent2a17ddfdca25ac4cbb0d317362608ec32f98fd82
regulator: core: Use different devices for resource allocation and DT lookup

Following by the below discussion, there's the potential UAF issue
between regulator and mfd.
https://lore.kernel.org/all/20221128143601.1698148-1-yangyingliang@huawei.com/

From the analysis of Yingliang

CPU A |CPU B
mt6370_probe() |
  devm_mfd_add_devices() |
|mt6370_regulator_probe()
|  regulator_register()
|    //allocate init_data and add it to devres
|    regulator_of_get_init_data()
i2c_unregister_device() |
  device_del() |
    devres_release_all() |
      // init_data is freed |
      release_nodes() |
|  // using init_data causes UAF
|  regulator_register()

It's common to use mfd core to create child device for the regulator.
In order to do the DT lookup for init data, the child that registered
the regulator would pass its parent as the parameter. And this causes
init data resource allocated to its parent, not itself. The issue happen
when parent device is going to release and regulator core is still doing
some operation of init data constraint for the regulator of child device.

To fix it, this patch expand 'regulator_register' API to use the
different devices for init data allocation and DT lookup.

Reported-by: Yang Yingliang <yangyingliang@huawei.com>
Signed-off-by: ChiYuan Huang <cy_huang@richtek.com>
Link: https://lore.kernel.org/r/1670311341-32664-1-git-send-email-u0084500@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>
drivers/platform/x86/intel/int3472/clk_and_regulator.c
drivers/regulator/core.c
drivers/regulator/devres.c
drivers/regulator/of_regulator.c
drivers/regulator/stm32-vrefbuf.c
include/linux/regulator/driver.h