linux-kernel - Re: [PATCH v24 09/11] acpi/arm64: Add memory-mapped timer support in GTDT driver

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, 19 Apr 2017 16:45:54 +0100
From:   Mark Rutland <mark.rutland@....com>
To:     Lorenzo Pieralisi <lorenzo.pieralisi@....com>
Cc:     linaro-acpi@...ts.linaro.org, catalin.marinas@....com,
        will.deacon@....com, rruigrok@...eaurora.org, wim@...ana.be,
        fu.wei@...aro.org, wei@...hat.com, al.stone@...aro.org,
        tn@...ihalf.com, timur@...eaurora.org, daniel.lezcano@...aro.org,
        linux-acpi@...r.kernel.org, linux@...ck-us.net, lenb@...nel.org,
        harba@...eaurora.org, julien.grall@....com,
        linux-watchdog@...r.kernel.org, arnd@...db.de,
        marc.zyngier@....com, jcm@...hat.com, cov@...eaurora.org,
        tglx@...utronix.de, linux-arm-kernel@...ts.infradead.org,
        graeme.gregory@...aro.org, rjw@...ysocki.net,
        linux-kernel@...r.kernel.org, leo.duran@....com,
        hanjun.guo@...aro.org, Suravee.Suthikulpanit@....com,
        sudeep.holla@....com, christoffer.dall@...aro.org
Subject: Re: [PATCH v24 09/11] acpi/arm64: Add memory-mapped timer support in
 GTDT driver

On Wed, Apr 19, 2017 at 04:07:25PM +0100, Mark Rutland wrote:
> On Tue, Apr 18, 2017 at 06:21:07PM +0100, Lorenzo Pieralisi wrote:
> > On Sat, Apr 15, 2017 at 02:40:12AM +0800, fu.wei@...aro.org wrote:
 
> > If yes, why can't it simply be written like this ?
> > 
> > for (; i >= 0; i--, gtdt_frame--) {
> > 	frame = &timer_mem->frame[gtdt_frame->frame_number];
> > 
> > 	/* not sure this check is actually needed */
> > 	if (gtdt_frame->common_flags & ACPI_GTDT_GT_IS_SECURE_TIMER)
> > 		continue;
> > 
> > 	if (frame->phys_irq > 0)
> > 		acpi_unregister_gsi(gtdt_frame->timer_interrupt);
> > 	if (frame->virt_irq > 0)
> > 		acpi_unregister_gsi(gtdt_frame->virtual_timer_interrupt);
> > }
> 
> A reverse loop of this form will work.
> 
> That requires some restructuring, and care to avoid going out of bounds
> instantaneously with the gtdt_frame--, so as to not invoke nasal demons.
> 
> I've attacked this locally, and will send this out after testing. I'll
> drop the new ACPI API patch.

FWIW, I've set this up so the cleanup path is:

do {
        if (gtdt_frame->common_flags & ACPI_GTDT_GT_IS_SECURE_TIMER ||
            gtdt_frame->frame_number >= ARCH_TIMER_MEM_MAX_FRAMES)
                continue;

        frame = &timer_mem->frame[gtdt_frame->frame_number];

        if (frame->phys_irq > 0)
                acpi_unregister_gsi(gtdt_frame->timer_interrupt);
        frame->phys_irq = 0;

        if (frame->virt_irq > 0)
                acpi_unregister_gsi(gtdt_frame->virtual_timer_interrupt);
        frame->virt_irq = 0;
} while (i-- >= 0 && gtdt_frame--);

... the zeroing is to account for duplicate frames, which I now check for in
the probe path (as we do for DT).

Can I take it per your comment on the prior version that with this change I can
take your ack?

I also assume that you're happy for all of the drivers/acpi/arm64/ patches in the
series to go via the clocksource tree?

Thanks,
Mark.