kernel.git/drivers/regulator/core.c, branch linux-5.5.y

regulator: core: Fix exported symbols to the exported GPL version

2020-02-24T07:38:38Z

[ Upstream commit 3d7610e8da993539346dce6f7c909fd3d56bf4d5 ] Change the exported symbols introduced by commit e9153311491da ("regulator: vctrl-regulator: Avoid deadlock getting and setting the voltage") from EXPORT_SYMBOL() to EXPORT_SYMBOL_GPL(), like is used for all the core parts. Fixes: e9153311491da ("regulator: vctrl-regulator: Avoid deadlock getting and setting the voltage") Reported-by: Dmitry Osipenko Signed-off-by: Enric Balletbo i Serra Link: https://lore.kernel.org/r/20200120123921.1204339-1-enric.balletbo@collabora.com Signed-off-by: Mark Brown Signed-off-by: Sasha Levin

regulator: vctrl-regulator: Avoid deadlock getting and setting the voltage

2020-02-24T07:38:35Z

[ Upstream commit e9153311491da9d9863ead9888a1613531cb4a1b ] `cat /sys/kernel/debug/regulator/regulator_summary` ends on a deadlock when you have a voltage controlled regulator (vctrl). The problem is that the vctrl_get_voltage() and vctrl_set_voltage() calls the regulator_get_voltage() and regulator_set_voltage() and that will try to lock again the dependent regulators (the regulator supplying the control voltage). Fix the issue by exporting the unlocked version of the regulator_get_voltage() and regulator_set_voltage() API so drivers that need it, like the voltage controlled regulator driver can use it. Fixes: f8702f9e4aa7 ("regulator: core: Use ww_mutex for regulators locking") Reported-by: Douglas Anderson Signed-off-by: Enric Balletbo i Serra Link: https://lore.kernel.org/r/20200116094543.2847321-1-enric.balletbo@collabora.com Signed-off-by: Mark Brown Signed-off-by: Sasha Levin

Merge tag 'regulator-fix-v5.5-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regulator

2019-12-17T21:08:41Z

Pull regulator fixes from Mark Brown: "A small set of fixes for mostly minor issues here, the only real code ones are Wen Yang's fixes for error handling in the core and Christian Marussi's list_voltage() change which is a fix for disruptively bad performance for regulators with continuous voltage control (which are rare)" * tag 'regulator-fix-v5.5-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regulator: regulator: rn5t618: fix module aliases regulator: max77650: add of_match table regulator: core: avoid unneeded .list_voltage calls regulator: s5m8767: Fix a warning message regulator: core: fix regulator_register() error paths to properly release rdev regulator: fix use after free issue

regulator: core: avoid unneeded .list_voltage calls

2019-12-09T18:41:40Z

Inside machine_constraints_voltage() a loop is in charge of verifying that each of the defined voltages are within the configured constraints and that those constraints are in fact compatible with the available voltages' list. When the registered regulator happens to be defined with a wide range of possible voltages the above O(n) loop can be costly. Moreover since this behaviour is triggered during the registration process, it means also that it can be easily triggered at probe time, slowing down considerably some module loading. On the other side if such wide range of voltage values happens to be also continuous and without discontinuity of any kind, the above potentially cumbersome operation is also useless. For these reasons, avoid such .list_voltage poll loop when regulator is described as 'continuous_voltage_range' as is, indeed, similarly already done inside regulator_is_supported_voltage(). Signed-off-by: Cristian Marussi Link: https://lore.kernel.org/r/20191209125239.46054-1-cristian.marussi@arm.com Signed-off-by: Mark Brown

regulator: core: fix regulator_register() error paths to properly release rdev

2019-12-03T12:18:33Z

There are several issues with the error handling code of the regulator_register() function: ret = device_register(&rdev->dev); if (ret != 0) { put_device(&rdev->dev); --> rdev released goto unset_supplies; } ... unset_supplies: ... unset_regulator_supplies(rdev); --> use-after-free ... clean: if (dangling_of_gpiod) gpiod_put(config->ena_gpiod); kfree(rdev); --> double free We add a variable to record the failure of device_register() and move put_device() down a bit to avoid the above issues. Fixes: c438b9d01736 ("regulator: core: Move registration of regulator device") Signed-off-by: Wen Yang Cc: Liam Girdwood Cc: Mark Brown Cc: linux-kernel@vger.kernel.org Link: https://lore.kernel.org/r/20191201030250.38074-1-wenyang@linux.alibaba.com Signed-off-by: Mark Brown

regulator: fix use after free issue

2019-11-27T12:53:48Z

This is caused by dereferencing 'rdev' after put_device() in the _regulator_get()/_regulator_put() functions. This patch just moves the put_device() down a bit to avoid the issue. Signed-off-by: Wen Yang Cc: Liam Girdwood Cc: Mark Brown Cc: linux-kernel@vger.kernel.org Link: https://lore.kernel.org/r/20191124145835.25999-1-wenyang@linux.alibaba.com Signed-off-by: Mark Brown

Merge branch 'regulator-5.5' into regulator-next

2019-11-22T19:56:20Z

regulator: core: Let boot-on regulators be powered off

2019-11-15T12:05:27Z

Boot-on regulators are always kept on because their use_count value is now incremented at boot time and never cleaned. Only increment count value for alway-on regulators. regulator_late_cleanup() is now able to power off boot-on regulators when unused. Fixes: 05f224ca6693 ("regulator: core: Clean enabling always-on regulators + their supplies") Signed-off-by: Pascal Paillet Link: https://lore.kernel.org/r/20191113102737.27831-1-p.paillet@st.com Signed-off-by: Mark Brown

regulator: core: Don't try to remove device links if add failed

2019-11-15T12:04:20Z

device_link_add() might not always succeed depending on the type of device link and the rest of the dependencies in the system. If device_link_add() didn't succeed, then we shouldn't try to remove the link later on as it might remove a link someone else created. Signed-off-by: Saravana Kannan Link: https://lore.kernel.org/r/20191115000438.45970-1-saravanak@google.com Signed-off-by: Mark Brown

regulator: core: Allow generic coupling only for always-on regulators

2019-10-28T13:15:52Z

The generic voltage balancer doesn't work correctly if one of regulator couples turns off. Currently there are no users in kernel for that case, although let's explicitly show that this case is unsupported for those who will try to use that feature. Link: https://lore.kernel.org/linux-samsung-soc/20191008170503.yd6GscYPLxjgrXqDuCO7AJc6i6egNZGJkVWHLlCxvA4@z/ Signed-off-by: Dmitry Osipenko Link: https://lore.kernel.org/r/20191025002240.25288-2-digetx@gmail.com Signed-off-by: Mark Brown