| Message ID | 1424889050-1008-1-git-send-email-gregory.clement@free-electrons.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
Dear Gregory CLEMENT, On Wed, 25 Feb 2015 19:30:50 +0100, Gregory CLEMENT wrote: > On Armada 38x SoCs, under heavy I/O load, the system hangs when CPU > Idle is enabled. Waiting for a solution to this issue, this patch > disables the CPU Idle support for this SoC. > > As CPU Hot plug support also uses some of the CPU Idle functions it is > also affected by the same issue. This patch disables it also for the > Armada 38x SoCs. > > Signed-off-by: Gregory CLEMENT <gregory.clement@free-electrons.com> > Cc: <stable@vger.kernel.org> # v3.17 + Tested-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com> I tested on Armada 38x, and indeed CPU hotplug is properly disabled, and it doesn't prevent suspend/resume (not supported yet in mainline, but hopefully soon to be supported). Some minor comments below. > diff --git a/arch/arm/mach-mvebu/pmsu.c b/arch/arm/mach-mvebu/pmsu.c > index 8b9f5e202ccf..2145f192662f 100644 > --- a/arch/arm/mach-mvebu/pmsu.c > +++ b/arch/arm/mach-mvebu/pmsu.c > @@ -415,6 +415,9 @@ static __init int armada_38x_cpuidle_init(void) > void __iomem *mpsoc_base; > u32 reg; > > + pr_warn("CPU Idle is currently broken on Armada 38x: disabling"); Idle -> idle > + return 0; > + > np = of_find_compatible_node(NULL, NULL, > "marvell,armada-380-coherency-fabric"); > if (!np) > @@ -469,13 +472,23 @@ static __init int armada_xp_cpuidle_init(void) > static int __init mvebu_v7_cpu_pm_init(void) > { > struct device_node *np; > - int ret; > + int ret = 0; Not clear why this change is needed. > > np = of_find_matching_node(NULL, of_pmsu_table); > if (!np) > return 0; > of_node_put(np); > > + /* > + * Currently the CPU Idle support for Armada 38x is broken, as > + * the CPU Hotplug uses some of the CPU Idle functions it is > + * broken too, so let's disable it > + */ > + if (of_machine_is_compatible("marvell,armada380")) { > + cpu_hotplug_disable(); > + pr_warn("CPU Hotplug support is currently broken on Armada 38x: disabling"); Hotplug -> hotplug. Thanks, Thomas
Hi Thomas, On 31/03/2015 18:04, Thomas Petazzoni wrote: > Dear Gregory CLEMENT, > > On Wed, 25 Feb 2015 19:30:50 +0100, Gregory CLEMENT wrote: >> On Armada 38x SoCs, under heavy I/O load, the system hangs when CPU >> Idle is enabled. Waiting for a solution to this issue, this patch >> disables the CPU Idle support for this SoC. >> >> As CPU Hot plug support also uses some of the CPU Idle functions it is >> also affected by the same issue. This patch disables it also for the >> Armada 38x SoCs. >> >> Signed-off-by: Gregory CLEMENT <gregory.clement@free-electrons.com> >> Cc: <stable@vger.kernel.org> # v3.17 + > > Tested-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com> > > I tested on Armada 38x, and indeed CPU hotplug is properly disabled, > and it doesn't prevent suspend/resume (not supported yet in mainline, > but hopefully soon to be supported). Great! > > Some minor comments below. > >> diff --git a/arch/arm/mach-mvebu/pmsu.c b/arch/arm/mach-mvebu/pmsu.c >> index 8b9f5e202ccf..2145f192662f 100644 >> --- a/arch/arm/mach-mvebu/pmsu.c >> +++ b/arch/arm/mach-mvebu/pmsu.c >> @@ -415,6 +415,9 @@ static __init int armada_38x_cpuidle_init(void) >> void __iomem *mpsoc_base; >> u32 reg; >> >> + pr_warn("CPU Idle is currently broken on Armada 38x: disabling"); > > Idle -> idle > OK >> + return 0; >> + >> np = of_find_compatible_node(NULL, NULL, >> "marvell,armada-380-coherency-fabric"); >> if (!np) >> @@ -469,13 +472,23 @@ static __init int armada_xp_cpuidle_init(void) >> static int __init mvebu_v7_cpu_pm_init(void) >> { >> struct device_node *np; >> - int ret; >> + int ret = 0; > > Not clear why this change is needed. It was a left over of an older version I will remove it. > >> >> np = of_find_matching_node(NULL, of_pmsu_table); >> if (!np) >> return 0; >> of_node_put(np); >> >> + /* >> + * Currently the CPU Idle support for Armada 38x is broken, as >> + * the CPU Hotplug uses some of the CPU Idle functions it is >> + * broken too, so let's disable it >> + */ >> + if (of_machine_is_compatible("marvell,armada380")) { >> + cpu_hotplug_disable(); >> + pr_warn("CPU Hotplug support is currently broken on Armada 38x: disabling"); > > Hotplug -> hotplug. OK Thanks, Gregory
diff --git a/arch/arm/mach-mvebu/pmsu.c b/arch/arm/mach-mvebu/pmsu.c index 8b9f5e202ccf..2145f192662f 100644 --- a/arch/arm/mach-mvebu/pmsu.c +++ b/arch/arm/mach-mvebu/pmsu.c @@ -415,6 +415,9 @@ static __init int armada_38x_cpuidle_init(void) void __iomem *mpsoc_base; u32 reg; + pr_warn("CPU Idle is currently broken on Armada 38x: disabling"); + return 0; + np = of_find_compatible_node(NULL, NULL, "marvell,armada-380-coherency-fabric"); if (!np) @@ -469,13 +472,23 @@ static __init int armada_xp_cpuidle_init(void) static int __init mvebu_v7_cpu_pm_init(void) { struct device_node *np; - int ret; + int ret = 0; np = of_find_matching_node(NULL, of_pmsu_table); if (!np) return 0; of_node_put(np); + /* + * Currently the CPU Idle support for Armada 38x is broken, as + * the CPU Hotplug uses some of the CPU Idle functions it is + * broken too, so let's disable it + */ + if (of_machine_is_compatible("marvell,armada380")) { + cpu_hotplug_disable(); + pr_warn("CPU Hotplug support is currently broken on Armada 38x: disabling"); + } + if (of_machine_is_compatible("marvell,armadaxp")) ret = armada_xp_cpuidle_init(); else if (of_machine_is_compatible("marvell,armada370")) @@ -489,7 +502,8 @@ static int __init mvebu_v7_cpu_pm_init(void) return ret; mvebu_v7_pmsu_enable_l2_powerdown_onidle(); - platform_device_register(&mvebu_v7_cpuidle_device); + if (mvebu_v7_cpuidle_device.name) + platform_device_register(&mvebu_v7_cpuidle_device); cpu_pm_register_notifier(&mvebu_v7_cpu_pm_notifier); return 0;
On Armada 38x SoCs, under heavy I/O load, the system hangs when CPU Idle is enabled. Waiting for a solution to this issue, this patch disables the CPU Idle support for this SoC. As CPU Hot plug support also uses some of the CPU Idle functions it is also affected by the same issue. This patch disables it also for the Armada 38x SoCs. Signed-off-by: Gregory CLEMENT <gregory.clement@free-electrons.com> Cc: <stable@vger.kernel.org> # v3.17 + --- Hi, The previous code raised a warning about the unused armada_38x_cpuidle_init() function. So I moved the warn message inside the function itself and then I continued to call this function. I also fixed a missing bracket. And finally I only registered the cpuilde platform driver if there was really something to register. Gregory arch/arm/mach-mvebu/pmsu.c | 18 ++++++++++++++++-- 1 file changed, 16 insertions(+), 2 deletions(-)