[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <7e80eaf4-4d78-420c-82c6-0604a5180f71@oss.nxp.com>
Date: Mon, 4 Aug 2025 13:02:38 +0300
From: Ghennadi Procopciuc <ghennadi.procopciuc@....nxp.com>
To: Daniel Lezcano <daniel.lezcano@...aro.org>, tglx@...utronix.de
Cc: S32@....com, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2 07/20] clocksource/drivers/vf-pit: Allocate the struct
timer at init time
On 8/4/2025 12:12 PM, Daniel Lezcano wrote:
> On 01/08/2025 09:33, Ghennadi Procopciuc wrote:
>> On 7/30/2025 11:27 AM, Daniel Lezcano wrote:
>>
>> [...]
>>
>
> [ ... ]
>
>>> + ret = -EINVAL;
>>> irq = irq_of_parse_and_map(np, 0);
>>> - if (irq <= 0)
>>> - return -EINVAL;
>>> + if (irq <= 0) {
>>> + pr_err("Failed to irq_of_parse_and_map\n");
>>> + goto out_iounmap;
>>> + }
>>> pit_clk = of_clk_get(np, 0);
>>> - if (IS_ERR(pit_clk))
>>> - return PTR_ERR(pit_clk);
>>> + if (IS_ERR(pit_clk)) {
>>> + ret = PTR_ERR(pit_clk);
>>> + goto out_iounmap;
>>
>> This does not revert the use of irq_of_parse_and_map. In my opinion, it
>> should be explicitly reverted as part of the cleanup phase by calling
>> irq_of_parse_and_map.
>
> AFAICT, it does not make sense to undo irq_of_parse_and_map() and there
> is no revert function for that.
>
> However, calling free_irq as commented in patch 20/20 makes sense.
>
>
I noticed that irq_of_parse_and_map internally calls both
of_irq_parse_one and irq_create_of_mapping. While the former does not
have a corresponding revert function, the latter can be reversed using
irq_dispose_mapping.
Examples:
drivers/char/ipmi/ipmi_powernv.c
drivers/clocksource/timer-microchip-pit64b.c
drivers/crypto/amcc/crypto4xx_core.c
drivers/memory/fsl_ifc.c
--
Regards,
Ghennadi
Powered by blists - more mailing lists