linux-kernel - Re: [PATCH] irqchip/loongson-liointc: Support to set IRQ_TYPE_EDGE

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] [day] [month] [year] [list]

Message-ID: <87cydmvdu3.ffs@tglx>
Date: Tue, 08 Apr 2025 18:05:08 +0200
From: Thomas Gleixner <tglx@...utronix.de>
To: Huacai Chen <chenhuacai@...il.com>
Cc: Huacai Chen <chenhuacai@...ngson.cn>, loongarch@...ts.linux.dev,
 linux-kernel@...r.kernel.org, Xuefeng Li <lixuefeng@...ngson.cn>, Jiaxun
 Yang <jiaxun.yang@...goat.com>, stable@...r.kernel.org, Yinbo Zhu
 <zhuyinbo@...ngson.cn>
Subject: Re: [PATCH] irqchip/loongson-liointc: Support to set
 IRQ_TYPE_EDGE_BOTH

On Tue, Apr 08 2025 at 20:37, Huacai Chen wrote:
> On Sun, Apr 6, 2025 at 10:20 PM Thomas Gleixner <tglx@...utronix.de> wrote:
>> This is the card change detection and it uses a GPIO. Insert raises one
>> edge and remove the opposite one.
>>
>> Which means whatever edge you chose randomly the detection will only
>> work in one direction. Please don't tell me that this is correct by any
>> meaning of correct. It's not.
>
> From experiments, either setting to EDGE_RISING or EDGE_FALLING, card
> detection (inserting and removing) works.

You might get lucky in that case because the switch bounces and there is
no debounce mechanism in place.

> Maybe the driver request "BOTH", but it really need "ANY"? I've
> searched git log, but I haven't get any useful information.

There is no maybe at all and speculation is not a technical argument.

It's well defined by the hardware:

Present           ________________
                 |               |
Not present _____|               |_______________

How can you detect that with a single edge?

If the switch bounces then the signal is:

Present           __  ___________  __
                 | | |          | | |
Not present _____| |_|          |_| |_______________

which makes it "work" by accident, but not by design.

There is ZERO guarantee that this will work with all other drivers which
request EDGE_BOTH.

Thanks,

        tglx