linux-kernel - Re: [PATCH v8 02/10] irqchip/riscv-intc: Allow large non-standard interrupt number

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]

Message-ID: <877cj8issa.ffs@tglx>
Date: Tue, 13 Feb 2024 11:04:53 +0100
From: Thomas Gleixner <tglx@...utronix.de>
To: Yu Chien Peter Lin <peterlin@...estech.com>, acme@...nel.org,
 adrian.hunter@...el.com, ajones@...tanamicro.com,
 alexander.shishkin@...ux.intel.com, andre.przywara@....com,
 anup@...infault.org, aou@...s.berkeley.edu, atishp@...shpatra.org,
 conor+dt@...nel.org, conor.dooley@...rochip.com, conor@...nel.org,
 devicetree@...r.kernel.org, evan@...osinc.com, geert+renesas@...der.be,
 guoren@...nel.org, heiko@...ech.de, irogers@...gle.com,
 jernej.skrabec@...il.com, jolsa@...nel.org, jszhang@...nel.org,
 krzysztof.kozlowski+dt@...aro.org, linux-arm-kernel@...ts.infradead.org,
 linux-kernel@...r.kernel.org, linux-perf-users@...r.kernel.org,
 linux-renesas-soc@...r.kernel.org, linux-riscv@...ts.infradead.org,
 linux-sunxi@...ts.linux.dev, locus84@...estech.com, magnus.damm@...il.com,
 mark.rutland@....com, mingo@...hat.com, n.shubin@...ro.com,
 namhyung@...nel.org, palmer@...belt.com, paul.walmsley@...ive.com,
 peterlin@...estech.com, peterz@...radead.org,
 prabhakar.mahadev-lad.rj@...renesas.com, rdunlap@...radead.org,
 robh+dt@...nel.org, samuel@...lland.org, sunilvl@...tanamicro.com,
 tim609@...estech.com, uwu@...nowy.me, wens@...e.org, will@...nel.org,
 inochiama@...look.com, unicorn_wang@...look.com, wefu@...hat.com
Cc: Randolph <randolph@...estech.com>, Atish Patra <atishp@...osinc.com>
Subject: Re: [PATCH v8 02/10] irqchip/riscv-intc: Allow large non-standard
 interrupt number

On Mon, Jan 29 2024 at 17:25, Yu Chien Peter Lin wrote:
>  static asmlinkage void riscv_intc_irq(struct pt_regs *regs)
>  {
>  	unsigned long cause = regs->cause & ~CAUSE_IRQ_FLAG;
>  
> -	if (unlikely(cause >= BITS_PER_LONG))
> -		panic("unexpected interrupt cause");
> -
> -	generic_handle_domain_irq(intc_domain, cause);
> +	if (generic_handle_domain_irq(intc_domain, cause))
> +		pr_warn_ratelimited("Failed to handle interrupt (cause: %ld)\n",
> +				    cause);

Either let the cause stick out or you need brackets. See:

  https://www.kernel.org/doc/html/latest/process/maintainer-tip.html#bracket-rules

>  }
>  
>  /*
> @@ -93,6 +95,14 @@ static int riscv_intc_domain_alloc(struct irq_domain *domain,
>  	if (ret)
>  		return ret;
>  
> +	/*
> +	 * Only allow hwirq for which we have corresponding standard or
> +	 * custom interrupt enable register.
> +	 */
> +	if ((riscv_intc_nr_irqs <= hwirq && hwirq < riscv_intc_custom_base) ||
> +	    (riscv_intc_custom_base + riscv_intc_custom_nr_irqs) <= hwirq)
> +		return -EINVAL;

Duh. This mix of ordering required to read this 3 times. What's wrong
with writing this consistently:

	if ((hwirq >= riscv_intc_nr_irqs && hwirq < riscv_intc_custom_base) ||
	    (hwirq >= iscv_intc_custom_base + riscv_intc_custom_nr_irqs)
		return -EINVAL;

Hmm?

> -	pr_info("%d local interrupts mapped\n", BITS_PER_LONG);
> +	pr_info("%d local interrupts mapped\n", riscv_intc_nr_irqs);
> +	if (riscv_intc_custom_nr_irqs)
> +		pr_info("%d custom local interrupts mapped\n",
> +			riscv_intc_custom_nr_irqs);

See bracket rules.
  
>  	return 0;
>  }
> @@ -166,6 +178,10 @@ static int __init riscv_intc_init(struct device_node *node,
>  		return 0;
>  	}
>  
> +	riscv_intc_nr_irqs = BITS_PER_LONG;
> +	riscv_intc_custom_base = riscv_intc_nr_irqs;

Why don't you initialize the static variables with constants right away?

> +	riscv_intc_custom_nr_irqs = 0;

It's already 0, no?

>  	return riscv_intc_init_common(of_node_to_fwnode(node));
>  }

Thanks,

        tglx