linux-kernel - Re: [PATCH 4/8] clk: davinci: add a reset lookup table for psc0

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date:   Fri, 23 Mar 2018 10:07:30 +0100
From:   Bartosz Golaszewski <brgl@...ev.pl>
To:     David Lechner <david@...hnology.com>
Cc:     Bartosz Golaszewski <bgolaszewski@...libre.com>,
        Sekhar Nori <nsekhar@...com>,
        Kevin Hilman <khilman@...nel.org>,
        Rob Herring <robh+dt@...nel.org>,
        Mark Rutland <mark.rutland@....com>,
        Russell King <linux@...linux.org.uk>,
        Michael Turquette <mturquette@...libre.com>,
        Stephen Boyd <sboyd@...nel.org>,
        Ohad Ben-Cohen <ohad@...ery.com>,
        Bjorn Andersson <bjorn.andersson@...aro.org>,
        Philipp Zabel <p.zabel@...gutronix.de>,
        arm-soc <linux-arm-kernel@...ts.infradead.org>,
        linux-devicetree <devicetree@...r.kernel.org>,
        LKML <linux-kernel@...r.kernel.org>, linux-clk@...r.kernel.org,
        linux-remoteproc@...r.kernel.org
Subject: Re: [PATCH 4/8] clk: davinci: add a reset lookup table for psc0

2018-03-21 17:17 GMT+01:00 David Lechner <david@...hnology.com>:
> On 03/21/2018 11:08 AM, Bartosz Golaszewski wrote:
>>
>> 2018-03-21 17:01 GMT+01:00 David Lechner <david@...hnology.com>:
>>>
>>> On 03/21/2018 07:08 AM, Bartosz Golaszewski wrote:
>>>>
>>>>
>>>> From: Bartosz Golaszewski <bgolaszewski@...libre.com>
>>>>
>>>> In order to be able to use the reset framework in legacy boot mode as
>>>> well, add the reset lookup table to the psc driver for da850 variant.
>>>>
>>>> Signed-off-by: Bartosz Golaszewski <bgolaszewski@...libre.com>
>>>> ---
>>>>    drivers/clk/davinci/psc-da850.c | 8 ++++++++
>>>>    drivers/clk/davinci/psc.c       | 1 +
>>>>    2 files changed, 9 insertions(+)
>>>>
>>>> diff --git a/drivers/clk/davinci/psc-da850.c
>>>> b/drivers/clk/davinci/psc-da850.c
>>>> index ccc7eb17bf3a..395db4b2c0ee 100644
>>>> --- a/drivers/clk/davinci/psc-da850.c
>>>> +++ b/drivers/clk/davinci/psc-da850.c
>>>> @@ -6,6 +6,7 @@
>>>>     */
>>>>      #include <linux/clk-provider.h>
>>>> +#include <linux/reset-controller.h>
>>>>    #include <linux/clk.h>
>>>>    #include <linux/clkdev.h>
>>>>    #include <linux/init.h>
>>>> @@ -66,8 +67,15 @@ LPSC_CLKDEV3(ecap_clkdev,    "fck",  "ecap.0",
>>>>                                  "fck",  "ecap.1",
>>>>                                  "fck",  "ecap.2");
>>>>    +static struct reset_control_lookup da850_psc0_reset_lookup_table[] =
>>>> {
>>>> +       RESET_LOOKUP("davinci-rproc.0", NULL, 15),
>>>> +};
>>>> +
>>>>    static int da850_psc0_init(struct device *dev, void __iomem *base)
>>>>    {
>>>> +       reset_controller_add_lookup("da850-psc0",
>>>> +                                   da850_psc0_reset_lookup_table,
>>>> +
>>>> ARRAY_SIZE(da850_psc0_reset_lookup_table));
>>>
>>>
>>>
>>> Could there be a race condition here since you are adding the lookup
>>> *before*
>>> you are adding the actual provider? It seems like
>>> reset_controller_add_lookup()
>>> should be after davinci_psc_register_clocks().
>>>
>>
>> I don't think so, because reset_controller_add_lookup() only adds the
>> lookup structure to the list in reset/core.c. The actual reset
>> controller struct is only located and used when reset_control_get_*()
>> is called, so after probing the user. And it's all protected with
>> mutexes.
>>
>> This made me think though - maybe if we can't locate the controller,
>> we should return -EPROBE_DEFER from probe in davinci-rproc?
>>
>> Bart
>>
>
> Yes, especially since we know that the PSC driver itself does get
> deferred already.

On the other hand if clk_get() succeeded, than the psc driver is
already initialized and the subsequent call to reset_control_get()
must succeed. But this is probably too machine-specific for a driver.

Bart