linux-kernel - Regression in v5.0-rc with regmap-irq level-irq changes

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-next>] [day] [month] [year] [list]

Message-ID: <20190115192840.GK5544@atomide.com>
Date:   Tue, 15 Jan 2019 11:28:40 -0800
From:   Tony Lindgren <tony@...mide.com>
To:     Matti Vaittinen <matti.vaittinen@...rohmeurope.com>,
        Mark Brown <broonie@...nel.org>
Cc:     Belisko Marek <marek.belisko@...il.com>,
        "Dr. H. Nikolaus Schaller" <hns@...delico.com>,
        Jon Hunter <jonathanh@...dia.com>,
        Laxman Dewangan <ldewangan@...dia.com>,
        Keerthy <j-keerthy@...com>,
        Peter Ujfalusi <peter.ujfalusi@...com>,
        Tero Kristo <t-kristo@...com>,
        Thierry Reding <treding@...dia.com>,
        linux-kernel@...r.kernel.org, linux-arm-kernel@...ts.infradead.org,
        linux-omap@...r.kernel.org
Subject: Regression in v5.0-rc with regmap-irq level-irq changes

Hi,

Looks like commit 1c2928e3e321 ("regmap: regmap-irq/gpio-max77620: add
level-irq support") caused a regression at least with palmas PMIC where
child interrupt no longer work and we get something like this on
omap5-uevm for example:

genirq: Setting trigger mode 3 for irq 157 failed
(regmap_irq_set_type+0x0/0x15c)
genirq: Setting trigger mode 12 for irq 154 failed
(regmap_irq_set_type+0x0/0x15c)
genirq: Setting trigger mode 8 for irq 149 failed
(regmap_irq_set_type+0x0/0x15c)

These are for child devices like power button, USB and RTC interrupts.
I'd also assume GPIOs might be broken now.

Any ideas what's needed? Some triggering mapping data to palmas.c
driver?

To me it seems the child device interrupt triggering can be whatever
the hardware supports including IRQ_TYPE_NONE for non-configurable
interrupts. Additionally palmas.c supports inverting the SoC
interrupt also.

Not sure if other things are broken now too, it might be worth
grepping the logs at kernelci.org for similar genirq errors above?

Regards,

Tony