| 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
| ||
|
Message-ID: <CAAofZF4w2KBJHTA+ufAA-4L5OXGyNxnG+d3YOoQG68gZ86Af6w@mail.gmail.com>
Date: Fri, 7 Feb 2025 11:17:53 +0100
From: Marco Crivellari <marco.crivellari@...e.com>
To: Huacai Chen <chenhuacai@...nel.org>
Cc: loongarch@...ts.linux.dev, linux-kernel@...r.kernel.org,
bibo mao <maobibo@...ngson.cn>, WANG Xuerui <kernel@...0n.name>,
Tiezhu Yang <yangtiezhu@...ngson.cn>, Frederic Weisbecker <frederic@...nel.org>,
Youling Tang <tangyouling@...ngson.cn>, Peter Zijlstra <peterz@...radead.org>,
Jiaxun Yang <jiaxun.yang@...goat.com>
Subject: Re: [PATCH] loongson: Fix idle VS timer enqueue
Hi,
Thanks for your replies!
I will make a new version with the aforementioned changes.
Marco
On Fri, Feb 7, 2025 at 9:18 AM Huacai Chen <chenhuacai@...nel.org> wrote:
>
> Hi, Marco,
>
> Thank you for your patch, the logic is correct but the code can be improved.
>
> On Fri, Feb 7, 2025 at 12:30 AM Marco Crivellari
> <marco.crivellari@...e.com> wrote:
> >
> > From: Peter Zijlstra <peterz@...radead.org>
> >
> > Loongson re-enables interrupts on its idle routine and performs a
> > TIF_NEED_RESCHED check afterwards before putting the CPU to sleep.
> >
> > The IRQs firing between the check and the idling instruction may set the
> > TIF_NEED_RESCHED flag. In order to deal with the such a race, IRQs
> > interrupting __arch_cpu_idle() rollback their return address to the
> > beginning of __arch_cpu_idle() so that TIF_NEED_RESCHED is checked
> > again before going back to sleep.
> >
> > However idle IRQs can also queue timers that may require a tick
> > reprogramming through a new generic idle loop iteration but those timers
> > would go unnoticed here because __arch_cpu_idle() only checks
> > TIF_NEED_RESCHED. It doesn't check for pending timers.
> >
> > Fix this with fast-forwarding idle IRQs return value to the end of the
> > idle routine instead of the beginning, so that the generic idle loop
> > handles both TIF_NEED_RESCHED and pending timers.
> >
> > Fixes: 0603839b18f4 ("LoongArch: Add exception/interrupt handling")
> > Not-yet-signed-off-by: Peter Zijlstra <peterz@...radead.org>
> > Cc: WANG Xuerui <kernel@...0n.name>
> > Signed-off-by: Frederic Weisbecker <frederic@...nel.org>
> > Signed-off-by: Marco Crivellari <marco.crivellari@...e.com>
> > ---
> > arch/loongarch/kernel/genex.S | 33 +++++++++++++++++++--------------
> > arch/loongarch/kernel/idle.c | 3 +--
> > arch/loongarch/kernel/reset.c | 23 ++++++++++++++---------
> > 3 files changed, 34 insertions(+), 25 deletions(-)
> >
> > diff --git a/arch/loongarch/kernel/genex.S b/arch/loongarch/kernel/genex.S
> > index 86d5d90ebefe..ab2e9a262ee2 100644
> > --- a/arch/loongarch/kernel/genex.S
> > +++ b/arch/loongarch/kernel/genex.S
> > @@ -18,28 +18,33 @@
> >
> > .align 5
> > SYM_FUNC_START(__arch_cpu_idle)
> > - /* start of rollback region */
> > - LONG_L t0, tp, TI_FLAGS
> > - nop
> > - andi t0, t0, _TIF_NEED_RESCHED
> > - bnez t0, 1f
> > - nop
> > - nop
> > - nop
> > + /* start of idle interrupt region */
> > + li.w t0, CSR_CRMD_IE
> It is better to replace it with "ori t0, zero, CSR_CRMD_IE", ensuring
> single instruction.
>
> > + /* idle instruction needs irq enabled */
> > + csrxchg t0, t0, LOONGARCH_CSR_CRMD
> > + /*
> > + * If an interrupt lands here; between enabling interrupts above and
> > + * going idle on the next instruction, we must *NOT* go idle since the
> > + * interrupt could have set TIF_NEED_RESCHED or caused an timer to need
> > + * reprogramming. Fall through -- see handle_vint() below -- and have
> > + * the idle loop take care of things.
> > + */
> > idle 0
> > - /* end of rollback region */
> > -1: jr ra
> > + nop
> Remove the "nop", of course it should be together with the later modification.
>
> > + /* end of idle interrupt region */
> > +SYM_INNER_LABEL(__arch_cpu_idle_exit, SYM_L_LOCAL)
> > + jr ra
> > SYM_FUNC_END(__arch_cpu_idle)
> >
> > SYM_CODE_START(handle_vint)
> > UNWIND_HINT_UNDEFINED
> > BACKUP_T0T1
> > SAVE_ALL
> > - la_abs t1, __arch_cpu_idle
> > + la_abs t1, __arch_cpu_idle_exit
> > LONG_L t0, sp, PT_ERA
> > - /* 32 byte rollback region */
> > - ori t0, t0, 0x1f
> > - xori t0, t0, 0x1f
> > + /* 16 byte idle interrupt region */
> > + ori t0, t0, 0x0f
> > + addi.d t0, t0, 1
> Replace these two instructions with "ori t0, t0, 0xc", and "16 bytes"
> can be replaced with "3 instructions".
>
> > bne t0, t1, 1f
> > LONG_S t0, sp, PT_ERA
> > 1: move a0, sp
> > diff --git a/arch/loongarch/kernel/idle.c b/arch/loongarch/kernel/idle.c
> > index 0b5dd2faeb90..54b247d8cdb6 100644
> > --- a/arch/loongarch/kernel/idle.c
> > +++ b/arch/loongarch/kernel/idle.c
> > @@ -11,7 +11,6 @@
> >
> > void __cpuidle arch_cpu_idle(void)
> > {
> > - raw_local_irq_enable();
> > - __arch_cpu_idle(); /* idle instruction needs irq enabled */
> > + __arch_cpu_idle();
> > raw_local_irq_disable();
> > }
> > diff --git a/arch/loongarch/kernel/reset.c b/arch/loongarch/kernel/reset.c
> > index 1ef8c6383535..d78e8a1064a1 100644
> > --- a/arch/loongarch/kernel/reset.c
> > +++ b/arch/loongarch/kernel/reset.c
> > @@ -20,6 +20,11 @@
> > void (*pm_power_off)(void);
> > EXPORT_SYMBOL(pm_power_off);
> >
> > +static __always_inline void native_halt(void)
> > +{
> > + asm volatile("idle 0" : : : "memory");
> > +}
> Just use the inline assembly at the call sites, please.
>
> Huacai
>
> > +
> > void machine_halt(void)
> > {
> > #ifdef CONFIG_SMP
> > @@ -32,9 +37,9 @@ void machine_halt(void)
> > pr_notice("\n\n** You can safely turn off the power now **\n\n");
> > console_flush_on_panic(CONSOLE_FLUSH_PENDING);
> >
> > - while (true) {
> > - __arch_cpu_idle();
> > - }
> > + while (1) {
> > + native_halt();
> > + };
> > }
> >
> > void machine_power_off(void)
> > @@ -52,9 +57,9 @@ void machine_power_off(void)
> > efi.reset_system(EFI_RESET_SHUTDOWN, EFI_SUCCESS, 0, NULL);
> > #endif
> >
> > - while (true) {
> > - __arch_cpu_idle();
> > - }
> > + while (1) {
> > + native_halt();
> > + };
> > }
> >
> > void machine_restart(char *command)
> > @@ -73,7 +78,7 @@ void machine_restart(char *command)
> > if (!acpi_disabled)
> > acpi_reboot();
> >
> > - while (true) {
> > - __arch_cpu_idle();
> > - }
> > + while (1) {
> > + native_halt();
> > + };
> > }
> > --
> > 2.48.1
> >
--
Powered by blists - more mailing lists