linux-kernel - Re: [PATCH for-5.19 1/2] irqchip/loongson-liointc: Use architecture register to get coreid

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 for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Sat, 4 Jun 2022 21:18:56 +0800
From:   Huacai Chen <chenhuacai@...nel.org>
To:     Jiaxun Yang <jiaxun.yang@...goat.com>
Cc:     WANG Xuerui <kernel@...0n.name>, Marc Zyngier <maz@...nel.org>,
        LKML <linux-kernel@...r.kernel.org>,
        "open list:MIPS" <linux-mips@...r.kernel.org>
Subject: Re: [PATCH for-5.19 1/2] irqchip/loongson-liointc: Use architecture
 register to get coreid

Hi, Jiaxun,

On Sat, Jun 4, 2022 at 8:41 PM Jiaxun Yang <jiaxun.yang@...goat.com> wrote:
>
> fa84f89395e0 ("irqchip/loongson-liointc: Fix build error for
> LoongArch") replaced get_ebase_cpunum with physical processor
> id from SMP facilities. However that breaks MIPS non-SMP build
> and makes booting from other cores inpossible on non-SMP kernel.
>
> Thus we revert get_ebase_cpunum back and use get_csr_cpuid for
> LoongArch.
>
> Fixes: fa84f89395e0 ("irqchip/loongson-liointc: Fix build error for LoongArch")
> Signed-off-by: Jiaxun Yang <jiaxun.yang@...goat.com>
> ---
>  drivers/irqchip/irq-loongson-liointc.c | 10 +++++++++-
>  1 file changed, 9 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/irqchip/irq-loongson-liointc.c b/drivers/irqchip/irq-loongson-liointc.c
> index aed88857d90f..c11cf97bcd1a 100644
> --- a/drivers/irqchip/irq-loongson-liointc.c
> +++ b/drivers/irqchip/irq-loongson-liointc.c
> @@ -39,6 +39,14 @@
>
>  #define LIOINTC_ERRATA_IRQ     10
>
> +#if defined(CONFIG_MIPS)
> +#define liointc_core_id get_ebase_cpunum()
> +#elif defined(CONFIG_LOONGARCH)
> +#define liointc_core_id get_csr_cpuid()
> +#else
> +#define liointc_core_id 0
> +#endif
Thank you for your quick fix. But I think it is better to do like this:

#if defined(CONFIG_LOONGARCH)
#define liointc_core_id get_csr_cpuid()
#else
#define liointc_core_id get_ebase_cpunum()
#endif

Because this driver doesn't depend on COMPILE_TEST, it can only be
built under MIPS and LOONGARCH. Moreover, let the else branch be the
same as the old behavior looks more reasonable.

Huacai

> +
>  struct liointc_handler_data {
>         struct liointc_priv     *priv;
>         u32                     parent_int_map;
> @@ -57,7 +65,7 @@ static void liointc_chained_handle_irq(struct irq_desc *desc)
>         struct liointc_handler_data *handler = irq_desc_get_handler_data(desc);
>         struct irq_chip *chip = irq_desc_get_chip(desc);
>         struct irq_chip_generic *gc = handler->priv->gc;
> -       int core = cpu_logical_map(smp_processor_id()) % LIOINTC_NUM_CORES;
> +       int core = liointc_core_id % LIOINTC_NUM_CORES;
>         u32 pending;
>
>         chained_irq_enter(chip, desc);
> --
> 2.25.1
>