| 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
| ||
|
Date: Fri, 14 Feb 2020 21:02:03 +0000
From: Mubin Usman Sayyed <MUBINUSM@...inx.com>
To: Thomas Gleixner <tglx@...utronix.de>,
"jason@...edaemon.net" <jason@...edaemon.net>,
"maz@...nel.org" <maz@...nel.org>,
Michal Simek <michals@...inx.com>,
"linux-arm-kernel@...ts.infradead.org"
<linux-arm-kernel@...ts.infradead.org>
CC: "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
Siva Durga Prasad Paladugu <sivadur@...inx.com>,
Anirudha Sarangi <anirudh@...inx.com>
Subject: RE: [PATCH v3] irqchip: xilinx: Add support for multiple instances
Hi Thomas,
Thanks for the review. Please see my inline replies.
> -----Original Message-----
> From: Thomas Gleixner <tglx@...utronix.de>
> Sent: Thursday, February 13, 2020 5:31 PM
> To: Mubin Usman Sayyed <MUBINUSM@...inx.com>;
> jason@...edaemon.net; maz@...nel.org; Michal Simek
> <michals@...inx.com>; linux-arm-kernel@...ts.infradead.org
> Cc: linux-kernel@...r.kernel.org; Siva Durga Prasad Paladugu
> <sivadur@...inx.com>; Anirudha Sarangi <anirudh@...inx.com>; Mubin
> Usman Sayyed <MUBINUSM@...inx.com>
> Subject: Re: [PATCH v3] irqchip: xilinx: Add support for multiple instances
>
> Mubin,
>
> Mubin Usman Sayyed <mubin.usman.sayyed@...inx.com> writes:
>
> > From: Mubin Sayyed <mubin.usman.sayyed@...inx.com>
> >
> > This patch adds support for multiple instances of
>
> git grep 'This patch' Documentation/process/submitting-patches.rst
[Mubin]: I will re-phrase it in next version.
>
> > xilinx interrupt controller. Below configurations are supported by
> > driver,
> >
> > - peripheral->xilinx-intc->xilinx-intc->gic
> > - peripheral->xilinx-intc->xilinx-intc
>
> This is really not much of an explanation.
[Mubin]: Will elaborate in next version
>
> > Signed-off-by: Anirudha Sarangi <anirudha.sarangi@...inx.com>
> > Signed-off-by: Mubin Sayyed <mubin.usman.sayyed@...inx.com>
>
> This Signed-off-by chain is incorrect. See chapter 11 and 12 in the same
> document.
[Mubin]: Sure, I will check and fix it. I ran checkpatch, it didn't reported errors/warning related to that.
>
> > @@ -38,29 +38,32 @@ struct xintc_irq_chip {
> > void __iomem *base;
> > struct irq_domain *root_domain;
> > u32 intr_mask;
> > + struct irq_chip *intc_dev;
> > + u32 nr_irq;
> > };
> >
> > -static struct xintc_irq_chip *xintc_irqc;
> > +static struct xintc_irq_chip *primary_intc;
> >
> > -static void xintc_write(int reg, u32 data)
> > +static void xintc_write(struct xintc_irq_chip *irqc, int reg, u32
> > +data)
> > {
> > if (static_branch_unlikely(&xintc_is_be))
> > - iowrite32be(data, xintc_irqc->base + reg);
> > + iowrite32be(data, irqc->base + reg);
> > else
> > - iowrite32(data, xintc_irqc->base + reg);
> > + iowrite32(data, irqc->base + reg);
> > }
> >
> > -static unsigned int xintc_read(int reg)
> > +static u32 xintc_read(struct xintc_irq_chip *irqc, int reg)
> > {
> > if (static_branch_unlikely(&xintc_is_be))
> > - return ioread32be(xintc_irqc->base + reg);
> > + return ioread32be(irqc->base + reg);
> > else
> > - return ioread32(xintc_irqc->base + reg);
> > + return ioread32(irqc->base + reg);
> > }
> >
> > static void intc_enable_or_unmask(struct irq_data *d) {
> > - unsigned long mask = 1 << d->hwirq;
> > + unsigned long mask = BIT(d->hwirq);
> > + struct xintc_irq_chip *irqc = irq_data_get_irq_chip_data(d);
>
> Please order your local variables in reverse fir tree order:
>
> struct xintc_irq_chip *irqc = irq_data_get_irq_chip_data(d);
> unsigned long mask = BIT(d->hwirq);
>
> which is the preferred coding style in this subsystem and way simpler to
> read.
[Mubin]: I will fix all such instances in next version
>
> > static void intc_mask_ack(struct irq_data *d) {
> > - unsigned long mask = 1 << d->hwirq;
> > + unsigned long mask = BIT(d->hwirq);
> > + struct xintc_irq_chip *irqc = irq_data_get_irq_chip_data(d);
>
> Ditto.
>
> > pr_debug("irq-xilinx: disable_and_ack: %ld\n", d->hwirq);
> > - xintc_write(CIE, mask);
> > - xintc_write(IAR, mask);
> > + xintc_write(irqc, CIE, mask);
> > + xintc_write(irqc, IAR, mask);
> > }
> > +static unsigned int xintc_get_irq_local(struct xintc_irq_chip *irqc)
> > +{
> > + u32 hwirq;
> > + unsigned int irq = 0;
>
> Same.
>
> > + hwirq = xintc_read(irqc, IVR);
> > + if (hwirq != -1U)
> > + irq = irq_find_mapping(irqc->root_domain, hwirq);
> > +
> > + pr_debug("irq-xilinx: hwirq=%d, irq=%d\n", hwirq, irq);
>
> Are these pr_debugs all over the please really required? I can understand
> that you use them for development, but are they useful once the stuff
> works?
[Mubin] They might be useful to debug interrupt issues. Do you want me to remove them?
>
> > + return irq;
> > +}
> > +
> > unsigned int xintc_get_irq(void)
> > {
> > - unsigned int hwirq, irq = -1;
> > + u32 hwirq;
> > + unsigned int irq = -1;
>
> See above.
>
> > - hwirq = xintc_read(IVR);
> > + hwirq = xintc_read(primary_intc, IVR);
> > if (hwirq != -1U)
> > - irq = irq_find_mapping(xintc_irqc->root_domain, hwirq);
> > + irq = irq_find_mapping(primary_intc->root_domain, hwirq);
> >
> > pr_debug("irq-xilinx: hwirq=%d, irq=%d\n", hwirq, irq);
> >
> > @@ -138,12 +164,14 @@ static const struct irq_domain_ops
> > xintc_irq_domain_ops = { static void xil_intc_irq_handler(struct
> > irq_desc *desc) {
> > struct irq_chip *chip = irq_desc_get_chip(desc);
> > + struct xintc_irq_chip *irqc =
> > + irq_data_get_irq_handler_data(&desc->irq_data);
>
> Please avoid these ugly line breaks and put the initialization of the variable in
> to the code below the declaration.
[Mubin]: Will do in next version
>
> > /* Turn on the Master Enable. */
> > - xintc_write(MER, MER_HIE | MER_ME);
> > - if (!(xintc_read(MER) & (MER_HIE | MER_ME))) {
> > + xintc_write(irqc, MER, MER_HIE | MER_ME);
> > + if (!(xintc_read(irqc, MER) & (MER_HIE | MER_ME))) {
> > static_branch_enable(&xintc_is_be);
>
> I see it's existing logic, but this lacks a comment how it's determined that
> xintc is big endian. Looks like some weird "write works?"
> probing. Why?
>
> > + xintc_write(irqc, MER, MER_HIE | MER_ME);
>
> So this writes MER_HIE | MER_ME into MER
>
> > + if (!(xintc_read(irqc, MER) & (MER_HIE | MER_ME))) {
>
> but this checks just whether ONE of the bits is set. Shouldn't it check for MER
> == (MER_HIE | MER_ME), i.e. read back what was written?
[Mubin]: Agreed, will fix it in v4.
Thanks,
Mubin
>
> Thanks,
>
> tglx
Powered by blists - more mailing lists