Regression with legacy IRQ numbers caused by 9a1091ef0017
Marc Zyngier
marc.zyngier at arm.com
Thu Jan 15 05:42:57 PST 2015
On Wed, Jan 14 2015 at 10:14:08 pm GMT, Tony Lindgren <tony at atomide.com> wrote:
> Hi all,
>
> Looks like the legacy IRQ numbers are now all wrong at least for omap4
> since commit 9a1091ef0017 ("irqchip: gic: Support hierarchy irq domain.").
>
> Instead of this:
>
> # cat /proc/interrupts
> CPU0 CPU1
> 29: 1124 981 GIC 29 twd
> 39: 0 0 GIC 39 TWL6030-PIH
> 41: 0 0 GIC 41 l3-dbg-irq
> 42: 0 0 GIC 42 l3-app-irq
> 44: 0 0 GIC 44 DMA
> 45: 7854 0 GIC 45 omap-dma-engine
> 52: 0 0 GIC 52 gpmc
> ...
>
>
> We now have:
>
> # cat /proc/interrupts
> CPU0 CPU1
> 16: 343 0 GIC 69 gp_timer
> 17: 1160 1017 GIC 29 twd
> 18: 0 0 GIC 41 l3-dbg-irq
> 19: 1 0 GIC 42 l3-app-irq
> 22: 7850 0 GIC 45 omap-dma-engine
> 44: 0 0 4a310000.gpio 18 DMA
> 61: 2730 0 48055000.gpio 2 eth0
> 223: 0 0 GIC 52 gpmc
> ...
>
> So the DMA interrupt using the legacy mapping with something like
> irq = 12 + OMAP44XX_IRQ_GIC_START now is wrong and unfortunately
> at least omaps still have a bunch of the legacy interrupts still
> around.
Holy crap. How much of this do we have hanging around?
> And that naturally produces all kinds of strange errors like:
>
> WARNING: CPU: 0 PID: 1 at drivers/bus/omap_l3_noc.c:147 l3_interrupt_handler+0x214/0x340()
> 44000000.ocp:L3 Custom Error: MASTER MPU TARGET L4CFG (Idle): Data Access in Supervisor mode during Functional access
> ...
> [<c05f21e4>] (__irq_svc) from [<c05f1974>] (_raw_spin_unlock_irqrestore+0x34/0x44)
> [<c05f1974>] (_raw_spin_unlock_irqrestore) from [<c00914a8>] (__setup_irq+0x244/0x530)
> [<c00914a8>] (__setup_irq) from [<c00917d4>] (setup_irq+0x40/0x8c)
> [<c00917d4>] (setup_irq) from [<c0039c8c>] (omap_system_dma_probe+0x1d4/0x2b4)
> [<c0039c8c>] (omap_system_dma_probe) from [<c03b2200>] (platform_drv_probe+0x44/0xa4)
> ...
>
> Looks like the logic changed from:
>
> if (of_property_read_u32(node, "arm,routable-irqs", &nr_routable_irqs))
>
> to just
>
> if (node)
>
> Which now causes irq_domain_add_linear() to be called instead of
> irq_domain_add_legacy(), which causes the breakage.
>
> Anybody got a sane fix in mind for the -rc series, or should we just
> revert it for now?
Reverting it is going to kill other platforms, and I'd rather have a
workaround, short of fixing it for good (which seems ambitious at -rc4).
How about something along these lines:
diff --git a/arch/arm/mach-omap2/common.h b/arch/arm/mach-omap2/common.h
index 377eea8..b664494 100644
--- a/arch/arm/mach-omap2/common.h
+++ b/arch/arm/mach-omap2/common.h
@@ -211,6 +211,7 @@ extern struct device *omap2_get_iva_device(void);
extern struct device *omap2_get_l3_device(void);
extern struct device *omap4_get_dsp_device(void);
+unsigned int omap4_xlate_irq(unsigned int hwirq);
void omap_gic_of_init(void);
#ifdef CONFIG_CACHE_L2X0
diff --git a/arch/arm/mach-omap2/omap4-common.c b/arch/arm/mach-omap2/omap4-common.c
index b7cb44a..cc30e49 100644
--- a/arch/arm/mach-omap2/omap4-common.c
+++ b/arch/arm/mach-omap2/omap4-common.c
@@ -256,6 +256,38 @@ static int __init omap4_sar_ram_init(void)
}
omap_early_initcall(omap4_sar_ram_init);
+static struct of_device_id gic_match[] = {
+ { .compatible = "arm,cortex-a9-gic", },
+ { .compatible = "arm,cortex-a15-gic", },
+ { },
+};
+
+static struct device_node *gic_node;
+
+unsigned int omap4_xlate_irq(unsigned int hwirq)
+{
+ struct of_phandle_args irq_data;
+ unsigned int irq;
+
+ if (!gic_node)
+ gic_node = of_find_matching_node(NULL, gic_match);
+
+ if (WARN_ON(!gic_node))
+ return hwirq;
+
+ irq_data.np = gic_node;
+ irq_data.args_count = 3;
+ irq_data.args[0] = 0;
+ irq_data.args[1] = hwirq - OMAP44XX_IRQ_GIC_START;
+ irq_data.args[2] = IRQ_TYPE_LEVEL_HIGH;
+
+ irq = irq_create_of_mapping(&irq_data);
+ if (WARN_ON(!irq))
+ irq = hwirq;
+
+ return irq;
+}
+
void __init omap_gic_of_init(void)
{
struct device_node *np;
diff --git a/arch/arm/mach-omap2/omap_hwmod.c b/arch/arm/mach-omap2/omap_hwmod.c
index cbb908d..9025fff 100644
--- a/arch/arm/mach-omap2/omap_hwmod.c
+++ b/arch/arm/mach-omap2/omap_hwmod.c
@@ -3534,9 +3534,15 @@ int omap_hwmod_fill_resources(struct omap_hwmod *oh, struct resource *res)
mpu_irqs_cnt = _count_mpu_irqs(oh);
for (i = 0; i < mpu_irqs_cnt; i++) {
+ unsigned int irq;
+
+ if (oh->xlate_irq)
+ irq = oh->xlate_irq((oh->mpu_irqs + i)->irq);
+ else
+ irq = (oh->mpu_irqs + i)->irq;
(res + r)->name = (oh->mpu_irqs + i)->name;
- (res + r)->start = (oh->mpu_irqs + i)->irq;
- (res + r)->end = (oh->mpu_irqs + i)->irq;
+ (res + r)->start = irq;
+ (res + r)->end = irq;
(res + r)->flags = IORESOURCE_IRQ;
r++;
}
diff --git a/arch/arm/mach-omap2/omap_hwmod.h b/arch/arm/mach-omap2/omap_hwmod.h
index 35ca6ef..5b42faf 100644
--- a/arch/arm/mach-omap2/omap_hwmod.h
+++ b/arch/arm/mach-omap2/omap_hwmod.h
@@ -676,6 +676,7 @@ struct omap_hwmod {
spinlock_t _lock;
struct list_head node;
struct omap_hwmod_ocp_if *_mpu_port;
+ unsigned int (*xlate_irq)(unsigned int);
u16 flags;
u8 mpu_rt_idx;
u8 response_lat;
diff --git a/arch/arm/mach-omap2/omap_hwmod_44xx_data.c b/arch/arm/mach-omap2/omap_hwmod_44xx_data.c
index c314b3c..f5e68a7 100644
--- a/arch/arm/mach-omap2/omap_hwmod_44xx_data.c
+++ b/arch/arm/mach-omap2/omap_hwmod_44xx_data.c
@@ -479,6 +479,7 @@ static struct omap_hwmod omap44xx_dma_system_hwmod = {
.class = &omap44xx_dma_hwmod_class,
.clkdm_name = "l3_dma_clkdm",
.mpu_irqs = omap44xx_dma_system_irqs,
+ .xlate_irq = omap4_xlate_irq,
.main_clk = "l3_div_ck",
.prcm = {
.omap4 = {
@@ -640,6 +641,7 @@ static struct omap_hwmod omap44xx_dss_dispc_hwmod = {
.class = &omap44xx_dispc_hwmod_class,
.clkdm_name = "l3_dss_clkdm",
.mpu_irqs = omap44xx_dss_dispc_irqs,
+ .xlate_irq = omap4_xlate_irq,
.sdma_reqs = omap44xx_dss_dispc_sdma_reqs,
.main_clk = "dss_dss_clk",
.prcm = {
@@ -693,6 +695,7 @@ static struct omap_hwmod omap44xx_dss_dsi1_hwmod = {
.class = &omap44xx_dsi_hwmod_class,
.clkdm_name = "l3_dss_clkdm",
.mpu_irqs = omap44xx_dss_dsi1_irqs,
+ .xlate_irq = omap4_xlate_irq,
.sdma_reqs = omap44xx_dss_dsi1_sdma_reqs,
.main_clk = "dss_dss_clk",
.prcm = {
@@ -726,6 +729,7 @@ static struct omap_hwmod omap44xx_dss_dsi2_hwmod = {
.class = &omap44xx_dsi_hwmod_class,
.clkdm_name = "l3_dss_clkdm",
.mpu_irqs = omap44xx_dss_dsi2_irqs,
+ .xlate_irq = omap4_xlate_irq,
.sdma_reqs = omap44xx_dss_dsi2_sdma_reqs,
.main_clk = "dss_dss_clk",
.prcm = {
@@ -784,6 +788,7 @@ static struct omap_hwmod omap44xx_dss_hdmi_hwmod = {
*/
.flags = HWMOD_SWSUP_SIDLE,
.mpu_irqs = omap44xx_dss_hdmi_irqs,
+ .xlate_irq = omap4_xlate_irq,
.sdma_reqs = omap44xx_dss_hdmi_sdma_reqs,
.main_clk = "dss_48mhz_clk",
.prcm = {
It seems to make my Panda-ES happy enough, but there is of course some
more to fix (prm44xx.c and twl_common.c). OMAP5 can be worked around the
same way.
Of course, this is in no way a proper fix, but I suppose the OMAP DT is
still missing a few bits...
Thoughts?
M.
--
Jazz is not dead. It just smells funny.
More information about the linux-arm-kernel
mailing list