linux-kernel - Re: [PATCH v6 7/7] arm64: kgdb: Roundup cpus using IPI as NMI

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:   Thu, 29 Oct 2020 16:22:34 +0000
From:   Daniel Thompson <daniel.thompson@...aro.org>
To:     Sumit Garg <sumit.garg@...aro.org>
Cc:     maz@...nel.org, catalin.marinas@....com, will@...nel.org,
        linux-arm-kernel@...ts.infradead.org, tglx@...utronix.de,
        jason@...edaemon.net, linux@...linux.org.uk,
        tsbogend@...ha.franken.de, mpe@...erman.id.au, davem@...emloft.net,
        mingo@...hat.com, bp@...en8.de, x86@...nel.org,
        mark.rutland@....com, julien.thierry.kdev@...il.com,
        dianders@...omium.org, jason.wessel@...driver.com,
        msys.mizuma@...il.com, ito-yuichi@...itsu.com,
        kgdb-bugreport@...ts.sourceforge.net, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v6 7/7] arm64: kgdb: Roundup cpus using IPI as NMI

On Thu, Oct 29, 2020 at 08:26:27PM +0530, Sumit Garg wrote:
> arm64 platforms with GICv3 or later supports pseudo NMIs which can be
> leveraged to roundup CPUs which are stuck in hard lockup state with
> interrupts disabled that wouldn't be possible with a normal IPI.
> 
> So instead switch to roundup CPUs using IPI turned as NMI. And in
> case a particular arm64 platform doesn't supports pseudo NMIs,
> it will switch back to default kgdb CPUs roundup mechanism.
> 
> Signed-off-by: Sumit Garg <sumit.garg@...aro.org>
> ---
>  arch/arm64/include/asm/kgdb.h |  9 +++++++++
>  arch/arm64/kernel/ipi_nmi.c   |  5 +++++
>  arch/arm64/kernel/kgdb.c      | 35 +++++++++++++++++++++++++++++++++++
>  3 files changed, 49 insertions(+)
> 
> diff --git a/arch/arm64/include/asm/kgdb.h b/arch/arm64/include/asm/kgdb.h
> index 21fc85e..c3d2425 100644
> --- a/arch/arm64/include/asm/kgdb.h
> +++ b/arch/arm64/include/asm/kgdb.h
> @@ -24,6 +24,15 @@ static inline void arch_kgdb_breakpoint(void)
>  extern void kgdb_handle_bus_error(void);
>  extern int kgdb_fault_expected;
>  
> +#ifdef CONFIG_KGDB
> +extern bool kgdb_ipi_nmicallback(int cpu, void *regs);
> +#else
> +static inline bool kgdb_ipi_nmicallback(int cpu, void *regs)
> +{
> +	return false;
> +}
> +#endif
> +
>  #endif /* !__ASSEMBLY__ */
>  
>  /*
> diff --git a/arch/arm64/kernel/ipi_nmi.c b/arch/arm64/kernel/ipi_nmi.c
> index 597dcf7..6ace182 100644
> --- a/arch/arm64/kernel/ipi_nmi.c
> +++ b/arch/arm64/kernel/ipi_nmi.c
> @@ -8,6 +8,7 @@
>  
>  #include <linux/interrupt.h>
>  #include <linux/irq.h>
> +#include <linux/kgdb.h>
>  #include <linux/nmi.h>
>  #include <linux/smp.h>
>  
> @@ -45,10 +46,14 @@ bool arch_trigger_cpumask_backtrace(const cpumask_t *mask, bool exclude_self)
>  static irqreturn_t ipi_nmi_handler(int irq, void *data)
>  {
>  	irqreturn_t ret = IRQ_NONE;
> +	unsigned int cpu = smp_processor_id();
>  
>  	if (nmi_cpu_backtrace(get_irq_regs()))
>  		ret = IRQ_HANDLED;
>  
> +	if (kgdb_ipi_nmicallback(cpu, get_irq_regs()))
> +		ret = IRQ_HANDLED;
> +
>  	return ret;

It would be better to declare existing return value for
kgdb_nmicallback() to be dangerously stupid and fix it so it returns an
irqreturn_t (that's easy since most callers do not need to check the
return value).

Then this code simply becomes:

	return kgdb_nmicallback(cpu, get_irq_regs());


>  }
>  
> diff --git a/arch/arm64/kernel/kgdb.c b/arch/arm64/kernel/kgdb.c
> index 1a157ca3..c26e710 100644
> --- a/arch/arm64/kernel/kgdb.c
> +++ b/arch/arm64/kernel/kgdb.c
> @@ -17,6 +17,7 @@
>  
>  #include <asm/debug-monitors.h>
>  #include <asm/insn.h>
> +#include <asm/nmi.h>
>  #include <asm/traps.h>
>  
>  struct dbg_reg_def_t dbg_reg_def[DBG_MAX_REG_NUM] = {
> @@ -353,3 +354,37 @@ int kgdb_arch_remove_breakpoint(struct kgdb_bkpt *bpt)
>  	return aarch64_insn_write((void *)bpt->bpt_addr,
>  			*(u32 *)bpt->saved_instr);
>  }
> +
> +bool kgdb_ipi_nmicallback(int cpu, void *regs)
> +{
> +	if (atomic_read(&kgdb_active) != -1) {
> +		kgdb_nmicallback(cpu, regs);
> +		return true;
> +	}
> +
> +	return false;
> +}

I *really* don't like this function.

If the return code of kgdb_nmicallback() is broken then fix it, don't
just wrap it and invent a new criteria for the return code.

To be honest I don't actually think the logic in kgdb_nmicallback() is
broken. As mentioned above the return value has a weird definition (0
for "handled it OK" and 1 for "nothing for me to do") but the logic to
calculate the return code looks OK.


> +
> +static void kgdb_smp_callback(void *data)
> +{
> +	unsigned int cpu = smp_processor_id();
> +
> +	if (atomic_read(&kgdb_active) != -1)
> +		kgdb_nmicallback(cpu, get_irq_regs());
> +}

This is Unused. I presume it is litter from a previous revision of the
code and can be deleted?


> +
> +bool kgdb_arch_roundup_cpus(void)
> +{
> +	struct cpumask mask;
> +
> +	if (!arm64_supports_nmi())
> +		return false;
> +
> +	cpumask_copy(&mask, cpu_online_mask);
> +	cpumask_clear_cpu(raw_smp_processor_id(), &mask);
> +	if (cpumask_empty(&mask))
> +		return false;

Why do we need to fallback if there is no work to do? There will still
be no work to do when we call the fallback.


Daniel.