Re: [PATCH v4 02/10] ACPI / irqchip: Add self-probe infrastructure to initialize IRQ controller

From: Hanjun Guo
Date: Wed Aug 05 2015 - 09:24:57 EST


Hi Marc,

On 08/04/2015 08:27 PM, Marc Zyngier wrote:
On 29/07/15 11:08, Hanjun Guo wrote:
This self-probe infrastructure works in the similar way as OF,
but there is some different in the mechanism:

For DT, the init fn will be called once it finds compatible strings
in DT, but for ACPI, we init irqchips by static tables, and in
static ACPI tables, there are no compatible strings to indicate
irqchips, but thanks to the GIC version presented in ACPI table,
we can call the corresponding GIC drivers matching the GIC version
with this framework.

This mechanism can also be used for clock declare and may also works
on x86 for some table parsing too.

This patch is based on Tomasz Nowicki <tomasz.nowicki@xxxxxxxxxx>'s
work.

Signed-off-by: Hanjun Guo <hanjun.guo@xxxxxxxxxx>
---
drivers/irqchip/irq-gic-acpi.c | 33 +++++++++++++++++++++++++++++++++
include/asm-generic/vmlinux.lds.h | 13 +++++++++++++
include/linux/acpi.h | 16 ++++++++++++++++
include/linux/irqchip.h | 13 +++++++++++++
include/linux/mod_devicetable.h | 8 ++++++++
5 files changed, 83 insertions(+)

diff --git a/drivers/irqchip/irq-gic-acpi.c b/drivers/irqchip/irq-gic-acpi.c
index 6537b43..011468d 100644
--- a/drivers/irqchip/irq-gic-acpi.c
+++ b/drivers/irqchip/irq-gic-acpi.c
@@ -107,3 +107,36 @@ static int __init acpi_gic_version_init(void)

return 0;
}
+
+/*
+ * This special acpi_table_id is the sentinel at the end of the
+ * acpi_table_id[] array of all irqchips. It is automatically placed at
+ * the end of the array by the linker, thanks to being part of a
+ * special section.
+ */
+static const struct acpi_table_id
+irqchip_acpi_match_end __used __section(__irqchip_acpi_table_end);

What is this thing for? Nobody refers to it (I know
drivers/irqchip/irqchip.c has a similar thing, but that's not enough a
reason...).

If I'm not mistaken, this is for proper section termination, please
see below:

/* scan the irqchip table to match the GIC version and its driver */
for (id = __irqchip_acpi_table; id->id[0]; id++) {
...
}

Then irqchip_acpi_match_end will be the end of the table section,
which will terminate the scanning.


+
+extern struct acpi_table_id __irqchip_acpi_table[];
+
+void __init acpi_irqchip_init(void)
+{
+ struct acpi_table_id *id;
+
+ if (acpi_disabled)
+ return;
+
+ if (acpi_gic_version_init())
+ return;

This is the only place where we need the version, right? So just get
acpi_gic_version_init to return the version number, and loose the global
variable.

The global variable is still needed because we need to get the GIC
version when scanning the MADT tables, we can't return the table value
directly in acpi_table_parse_madt() at now.


+
+ /* scan the irqchip table to match the GIC version and its driver */
+ for (id = __irqchip_acpi_table; id->id[0]; id++) {
+ if (gic_version == (u8)id->driver_data) {
+ acpi_table_parse(id->id,
+ (acpi_tbl_table_handler)id->handler);
+ return;
+ }
+ }
+
+ pr_err("No matched driver GIC version %d\n", gic_version);
+}
diff --git a/include/asm-generic/vmlinux.lds.h b/include/asm-generic/vmlinux.lds.h
index 8bd374d..625776c 100644
--- a/include/asm-generic/vmlinux.lds.h
+++ b/include/asm-generic/vmlinux.lds.h
@@ -181,6 +181,18 @@
#define CPUIDLE_METHOD_OF_TABLES() OF_TABLE(CONFIG_CPU_IDLE, cpuidle_method)
#define EARLYCON_OF_TABLES() OF_TABLE(CONFIG_SERIAL_EARLYCON, earlycon)

+#ifdef CONFIG_ACPI
+#define ACPI_TABLE(name) \
+ . = ALIGN(8); \
+ VMLINUX_SYMBOL(__##name##_acpi_table) = .; \
+ *(__##name##_acpi_table) \
+ *(__##name##_acpi_table_end)
+
+#define IRQCHIP_ACPI_MATCH_TABLE() ACPI_TABLE(irqchip)
+#else
+#define IRQCHIP_ACPI_MATCH_TABLE()
+#endif
+
#define KERNEL_DTB() \
STRUCT_ALIGN(); \
VMLINUX_SYMBOL(__dtb_start) = .; \
@@ -516,6 +528,7 @@
CPUIDLE_METHOD_OF_TABLES() \
KERNEL_DTB() \
IRQCHIP_OF_MATCH_TABLE() \
+ IRQCHIP_ACPI_MATCH_TABLE() \
EARLYCON_TABLE() \
EARLYCON_OF_TABLES()

diff --git a/include/linux/acpi.h b/include/linux/acpi.h
index 0820cb1..04dd0bb 100644
--- a/include/linux/acpi.h
+++ b/include/linux/acpi.h
@@ -829,4 +829,20 @@ static inline struct acpi_device *acpi_get_next_child(struct device *dev,

#endif

+#ifdef CONFIG_ACPI
+#define ACPI_DECLARE(table, name, table_id, data, fn) \
+ static const struct acpi_table_id __acpi_table_##name \
+ __used __section(__##table##_acpi_table) \
+ = { .id = table_id, \
+ .handler = (void *)fn, \
+ .driver_data = data }
+#else
+#define ACPI_DECLARE(table, name, table_id, data, fn) \
+ static const struct acpi_table_id __acpi_table_##name \
+ __attribute__((unused)) \
+ = { .id = table_id, \
+ .handler = (void *)fn, \
+ .driver_data = data }
+#endif
+
#endif /*_LINUX_ACPI_H*/
diff --git a/include/linux/irqchip.h b/include/linux/irqchip.h
index 6388873..6b66d3e 100644
--- a/include/linux/irqchip.h
+++ b/include/linux/irqchip.h
@@ -11,6 +11,7 @@
#ifndef _LINUX_IRQCHIP_H
#define _LINUX_IRQCHIP_H

+#include <linux/acpi.h>
#include <linux/of.h>

/*
@@ -25,6 +26,18 @@
*/
#define IRQCHIP_DECLARE(name, compat, fn) OF_DECLARE_2(irqchip, name, compat, fn)

+/*
+ * This macro must be used by the different ARM GIC drivers to declare
+ * the association between their version and their initialization function.
+ *
+ * @name: name that must be unique accross all IRQCHIP_ACPI_DECLARE of the
+ * same file.
+ * @gic_version: version of GIC
+ * @fn: initialization function
+ */
+#define IRQCHIP_ACPI_DECLARE(name, gic_version, fn) \
+ ACPI_DECLARE(irqchip, name, ACPI_SIG_MADT, gic_version, fn)

I don't think this is the right approach. The MADT table has a *huge*
number of possible subtables, and none of them is matched by the GIC
version.

What you *really* want is MADT -> GICD -> GIC type. Your ACPI_DECLARE
macro doesn't reflect this at all (you've basically cloned OF, and
that's clearly not good enough).

This probably require an intermediate matching function, ending up with
something like:

#define IRQCHIP_ACPI_DECLARE(name, subtable, version, fn) \
ACPI_DECLARE(irqchip, name, ACPI_SIG_MADT, match_madt_subtable,\
subtable, version, fn)

where match_madt_subtable is going to check that a given subtable is
really suitable for a given irqchip. None of that should be GIC
specific, really.

I'm little confused here, in the previous patch, the gic version is
stored, so we only need to match the GIC driver with the gic version.

In the GIC dirver itself, we will match the madt subtable to find
information we need to init the GIC, so I'm not sure why
match_madt_subtable in the ACPI_DECLARE, I think I missed something,
please correct me.

Thanks
Hanjun
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/