linux-kernel - Re: [PATCH 02/13] leds: el15203000: Fix devm vs. non-devm ordering

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: Windows password security audit tool. GUI, reports in PDF.

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

Date:   Wed, 9 Nov 2022 18:25:12 +0800
From:   wangyufen <wangyufen@...wei.com>
To:     Oleh Kravchenko <oleg@....org.ua>
CC:     <linux-leds@...r.kernel.org>, <linux-kernel@...r.kernel.org>,
        <pavel@....cz>
Subject: Re: [PATCH 02/13] leds: el15203000: Fix devm vs. non-devm ordering


在 2022/11/9 17:39, Oleh Kravchenko 写道:
> Hello all!
>
>> 9 лист. 2022 р. о 10:48 Wang Yufen <wangyufen@...wei.com> написав(ла):
>>
>> When non-devm resources are allocated they mustn't be followed by devm
>> allocations, otherwise it will break the tear down ordering and might
>> lead to crashes or other bugs during ->remove() stage. Fix this by
>> wrapping mutex_destroy() call with devm_add_action_or_reset().
>>
>> Fixes: fc19967bcb8f ("leds: add LED driver for EL15203000 board")
>> Signed-off-by: Wang Yufen <wangyufen@...wei.com>
>> Cc: Oleh Kravchenko <oleg@....org.ua>
>> ---
>> drivers/leds/leds-el15203000.c | 18 +++++++++++-------
>> 1 file changed, 11 insertions(+), 7 deletions(-)
>>
>> diff --git a/drivers/leds/leds-el15203000.c b/drivers/leds/leds-el15203000.c
>> index 7e7b617..9be934e 100644
>> --- a/drivers/leds/leds-el15203000.c
>> +++ b/drivers/leds/leds-el15203000.c
>> @@ -287,10 +287,16 @@ static int el15203000_probe_dt(struct el15203000 *priv)
>> return ret;
>> }
>>
>> +static void el15203000_mutex_destroy(void *lock)
>> +{
>> + mutex_destroy(lock);
>> +}
>> +
>> static int el15203000_probe(struct spi_device *spi)
>> {
>> struct el15203000 *priv;
>> size_t count;
>> + int ret;
>>
>> count = device_get_child_node_count(&spi->dev);
>> if (!count) {
>> @@ -312,15 +318,14 @@ static int el15203000_probe(struct spi_device *spi)
>>
>> spi_set_drvdata(spi, priv);
>>
>> + ret = devm_add_action_or_reset(&spi->dev, el15203000_mutex_destroy,
>> +       &priv->lock);
>> + if (ret)
>> + return ret;
>> +
>> return el15203000_probe_dt(priv);
>> }
>>
>> -static void el15203000_remove(struct spi_device *spi)
> Is remove() callback from struct spi_driver deprecated?

It is not that remove() callback is deprecated,
it's that after wrapping mutex_destroy() call with devm_add_action_or_reset(),
remove() callback is unnecessary here.

>
>> -{
>> - struct el15203000 *priv = spi_get_drvdata(spi);
>> -
>> - mutex_destroy(&priv->lock);
>> -}
>>
>> static const struct of_device_id el15203000_dt_ids[] = {
>> { .compatible = "crane,el15203000", },
>> @@ -331,7 +336,6 @@ static void el15203000_remove(struct spi_device *spi)
>>
>> static struct spi_driver el15203000_driver = {
>> .probe = el15203000_probe,
>> - .remove = el15203000_remove,
>> .driver = {
>> .name = KBUILD_MODNAME,
>> .of_match_table = el15203000_dt_ids,
>> -- 
>> 1.8.3.1
>>
>