Re: [PATCH for-4.8 12/12] powerpc/mm/radix: Add a kernel command line to disable radix

Balbir Singh [off-list ref] writes:

On Wed, Jul 13, 2016 at 03:05:31PM +0530, Aneesh Kumar K.V wrote:

quoted

This patch adds the kernel command line disable_radix which disable
the radix MMU mode even if firmware indicates radix support via
ibm,pa-features device tree node.

This helps in testing different MMU mode easily.

Signed-off-by: Aneesh Kumar K.V <redacted>
---
 Documentation/kernel-parameters.txt |  3 +++
 arch/powerpc/kernel/prom.c          | 13 +++++++++++++
 2 files changed, 16 insertions(+)

diff --git a/Documentation/kernel-parameters.txt b/Documentation/kernel-parameters.txt
index 738bae4a5958..bba7ef30d74e 100644
--- a/Documentation/kernel-parameters.txt
+++ b/Documentation/kernel-parameters.txt

@@ -929,6 +929,9 @@ bytes respectively. Such letter suffixes can also be entirely omitted.
 	disable=	[IPV6]
 			See Documentation/networking/ipv6.txt.
 
+	disable_radix	[PPC]
+			Disable RADIX MMU mode on POWER9
+
 	disable_cpu_apicid= [X86,APIC,SMP]
 			Format: <int>
 			The number of initial APIC ID for the

diff --git a/arch/powerpc/kernel/prom.c b/arch/powerpc/kernel/prom.c
index 946e34ffeae9..022540b9366d 100644
--- a/arch/powerpc/kernel/prom.c
+++ b/arch/powerpc/kernel/prom.c

@@ -645,6 +645,14 @@ static void __init early_reserve_mem(void)
 #endif
 }
 
+static bool disable_radix;

__read_mostly?

We actually read it only once.

quoted

+static int __init parse_disable_radix(char *p)
+{
+	disable_radix = true;
+	return 0;
+}
+early_param("disable_radix", parse_disable_radix);
+
 void __init early_init_devtree(void *params)
 {
 	phys_addr_t limit;

@@ -739,6 +747,11 @@ void __init early_init_devtree(void *params)
 	/* Scan and build the list of machine check recoverable ranges */
 	of_scan_flat_dt(early_init_dt_scan_recoverable_ranges, NULL);
 #endif
+	/*
+	 * now fixup radix MMU mode based on kernel command line
+	 */
+	if (disable_radix)
+		cur_cpu_spec->mmu_features &= ~MMU_FTR_RADIX;

Shouldn't this happen before prom_init() and exchanging capabilities?

Didn't get that. We actually set the FTR_RADIX in
check_cpu_pa_features(node) which is called via
of_scan_flat_dt(early_init_dt_scan_cpus, NULL); We clear that
immediately after if the feature is disabled via kernel command line.

-aneesh