[PATCH 1/3] irqchip/irq-sifive-plic: Fixup wrong size of xxx_PER_HART and reg base
Anup Patel
anup at brainfault.org
Sun Oct 25 05:17:47 EDT 2020
On Sat, Oct 24, 2020 at 8:40 AM Guo Ren <guoren at kernel.org> wrote:
>
> On Fri, Oct 23, 2020 at 8:31 PM Anup Patel <anup at brainfault.org> wrote:
> >
> > On Fri, Oct 23, 2020 at 3:48 PM <guoren at kernel.org> wrote:
> > >
> > > From: Guo Ren <guoren at linux.alibaba.com>
> > >
> > > ENABLE and CONTEXT registers contain M & S status for per-hart, so
> > > ref to the specification the correct definition is double to the
> > > current value.
> > >
> > > The value of hart_base and enable_base should be calculated by real
> > > physical hartid not software id. Sometimes the CPU node's <reg>
> > > from dts is not equal to the sequence index.
> > >
> > > Signed-off-by: Guo Ren <guoren at linux.alibaba.com>
> > > ---
> > > drivers/irqchip/irq-sifive-plic.c | 12 ++++++------
> > > 1 file changed, 6 insertions(+), 6 deletions(-)
> > >
> > > diff --git a/drivers/irqchip/irq-sifive-plic.c b/drivers/irqchip/irq-sifive-plic.c
> > > index eaa3e9f..2e56576 100644
> > > --- a/drivers/irqchip/irq-sifive-plic.c
> > > +++ b/drivers/irqchip/irq-sifive-plic.c
> > > @@ -44,16 +44,16 @@
> > > * Each hart context has a vector of interrupt enable bits associated with it.
> > > * There's one bit for each interrupt source.
> > > */
> > > -#define ENABLE_BASE 0x2000
> > > -#define ENABLE_PER_HART 0x80
> > > +#define ENABLE_BASE 0x2080
> > > +#define ENABLE_PER_HART 0x100
> > >
> > > /*
> > > * Each hart context has a set of control registers associated with it. Right
> > > * now there's only two: a source priority threshold over which the hart will
> > > * take an interrupt, and a register to claim interrupts.
> > > */
> > > -#define CONTEXT_BASE 0x200000
> > > -#define CONTEXT_PER_HART 0x1000
> > > +#define CONTEXT_BASE 0x201000
> > > +#define CONTEXT_PER_HART 0x2000
> > > #define CONTEXT_THRESHOLD 0x00
> > > #define CONTEXT_CLAIM 0x04
> > >
> > > @@ -358,10 +358,10 @@ static int __init plic_init(struct device_node *node,
> > > cpumask_set_cpu(cpu, &priv->lmask);
> > > handler->present = true;
> > > handler->hart_base =
> > > - priv->regs + CONTEXT_BASE + i * CONTEXT_PER_HART;
> > > + priv->regs + CONTEXT_BASE + hartid * CONTEXT_PER_HART;
> > > raw_spin_lock_init(&handler->enable_lock);
> > > handler->enable_base =
> > > - priv->regs + ENABLE_BASE + i * ENABLE_PER_HART;
> > > + priv->regs + ENABLE_BASE + hartid * ENABLE_PER_HART;
> > > handler->priv = priv;
> > > done:
> > > for (hwirq = 1; hwirq <= nr_irqs; hwirq++)
> > > --
> > > 2.7.4
> > >
> >
> > There is no one-to-one mapping between PLIC context and HARTID. Instead,
> > we have many-to-one mapping between PLIC contexts and HARTID. In other
> > words, we have one PLIC context for each interrupt capable mode (i.e.
> > M/S-mode) of each HART.
> >
> > For example, on SiFive Unleashed we have 5 HARTs but HARTID=0 has
> > only M-mode capable of taking interrupts so we have total (1 + 2x4) = 9
> > PLIC contexts.
> That's OK, but what the bug I want to point out is enable_base &
> context_base should be calculated by 'hartid' not 'i'.
There is no relation between PLIC context number and HART IDs. The
PLIC context to HART mapping is discovered from the "interrupts-extended"
DT property of PLIC DT node. The "i" in the loop is PLIC context number.
The PLIC spec does not mandate any ordering/pattern of PLIC context to
HART mappings. Also, the interrupts-extended DT property is generic
enough to represent any kind of PLIC context to HART mappings.
Your patch breaks SiFive Unleashed board and NoMMU kernel because
of incorrect assumptions about PLIC contexts to HART mappings.
>
> For example, how we deal with below dts configuration:
> cpus {
> #address-cells = <1>;
> #size-cells = <0>;
> timebase-frequency = <3000000>;
> cpu at 0 {
> device_type = "cpu";
> reg = <2>; //********* different from index
> status = "okay";
> compatible = "riscv";
> riscv,isa = "rv64imafdcsu";
> mmu-type = "riscv,sv39";
> cpu0_intc: interrupt-controller {
> #interrupt-cells = <1>;
> compatible = "riscv,cpu-intc";
> interrupt-controller;
> };
> };
> cpu at 1 {
> device_type = "cpu";
> reg = <3>; //********* different from index
> status = "fail";
> compatible = "riscv";
> riscv,isa = "rv64imafdcsu";
> mmu-type = "riscv,sv39";
> cpu1_intc: interrupt-controller {
> #interrupt-cells = <1>;
> compatible = "riscv,cpu-intc";
> interrupt-controller;
> };
> };
> }
For above DTS configuration, we need a PLIC with just 4 contexts
(irrespective to the HART IDs).
The PLIC context to HART mapping can be described using the
interrupts-extended DT property of PLIC DT node as follows:
interrupts-extended = <&cpu0_intc 11 &cpu0_intc 9
&cpu1_intc 11 &cpu1_intc 9>;
>
> >
> > I would also like to highlight that this patch is forcing PLIC driver to always
> > use PLIC S-mode context for each HART which breaks the Linux RISC-V
> > NoMMU kernel.
> Yes, I forgot M-mode and I will correct it.
>
> >
> > There is no issue with the existing defines because these are aligned with
> > above and latest PLIC spec.
> > (Refer, https://github.com/riscv/riscv-plic-spec/blob/master/riscv-plic.adoc)
> >
> > NACK to this patch from my side.
>
> Here is my new patch which fixup m-mode linux:
>
> diff --git a/drivers/irqchip/irq-sifive-plic.c
> b/drivers/irqchip/irq-sifive-plic.c
> index 4048657..e34e1d9 100644
> --- a/drivers/irqchip/irq-sifive-plic.c
> +++ b/drivers/irqchip/irq-sifive-plic.c
> @@ -45,7 +45,13 @@
> * There's one bit for each interrupt source.
> */
> #define ENABLE_BASE 0x2000
> -#define ENABLE_PER_HART 0x80
> +#define ENABLE_PER_HART 0x100
> +#ifdef CONFIG_RISCV_M_MODE
> +#define ENABLE_OFFSET 0
> +#else
> +#define ENABLE_OFFSET 0x80
> +#endif
> +
>
> /*
> * Each hart context has a set of control registers associated with it. Right
> @@ -53,9 +59,14 @@
> * take an interrupt, and a register to claim interrupts.
> */
> #define CONTEXT_BASE 0x200000
> -#define CONTEXT_PER_HART 0x1000
> +#define CONTEXT_PER_HART 0x2000
> #define CONTEXT_THRESHOLD 0x00
> #define CONTEXT_CLAIM 0x04
> +#ifdef CONFIG_RISCV_M_MODE
> +#define CONTEXT_OFFSET 0
> +#else
> +#define CONTEXT_OFFSET 0x1000
> +#endif
>
> #define PLIC_DISABLE_THRESHOLD 0x7
> #define PLIC_ENABLE_THRESHOLD 0
> @@ -358,10 +369,10 @@ static int __init plic_init(struct device_node *node,
> cpumask_set_cpu(cpu, &priv->lmask);
> handler->present = true;
> handler->hart_base =
> - priv->regs + CONTEXT_BASE + i * CONTEXT_PER_HART;
> + priv->regs + CONTEXT_BASE + hartid *
> CONTEXT_PER_HART + CONTEXT_OFFSET;
> raw_spin_lock_init(&handler->enable_lock);
> handler->enable_base =
> - priv->regs + ENABLE_BASE + i * ENABLE_PER_HART;
> + priv->regs + ENABLE_BASE + hartid *
> ENABLE_PER_HART + ENABLE_OFFSET;
> handler->priv = priv;
> done:
> for (hwirq = 1; hwirq <= nr_irqs; hwirq++)
>
This modified patch still breaks SiFive Unleashed board because
it has following PLIC context to HART mappings:
PLIC_context_0 (offset 0x0000) => HART0_M_mode_irq
PLIC_context_1 (offset 0x1000) => HART1_M_mode_irq
PLIC_context_2 (offset 0x2000) => HART1_S_mode_irq
PLIC_context_3 (offset 0x3000) => HART2_M_mode_irq
PLIC_context_4 (offset 0x4000) => HART2_S_mode_irq
PLIC_context_5 (offset 0x5000) => HART3_M_mode_irq
PLIC_context_6 (offset 0x6000) => HART3_S_mode_irq
PLIC_context_7 (offset 0x7000) => HART4_M_mode_irq
PLIC_context_8 (offset 0x8000) => HART4_S_mode_irq
As-per your patch, the Linux S-mode kernel will use
PLIC_context_3 (offset 0x3000) for HART1 which is in-correct
because PLIC_context_3 maps to HART2 M-mode IRQ.
Further with your patch, the Linux M-mode kernel will use
PLIC_context_2 (offset 0x2000) for HART1 which is in-correct
because PLIC_context_2 maps to HART1 S-mode IRQ.
Clearly your modified patch breaks both Linux S-mode kernel
and Linux M-mode kernel on SiFive Unleashed board.
The current driver handles all above cases correctly by parsing
the PLIC context to HART mappings from the interrupts-extended
DT property. I don't see any bug here.
Regards,
Anup
More information about the linux-riscv
mailing list