From patchwork series 358597

https://patchwork.ozlabs.org//project/linuxppc-dev/list/?series=358597

Documentation/admin-guide/kernel-parameters.txt

@@ -3157,6 +3157,9 @@
 			Note that even when enabled, there are a few cases where
 			the feature is not effective.
 
+	memory_block_size=size [PPC]
+			 Use this parameter to configure the memory block size value.
+
 	memtest=	[KNL,X86,ARM,M68K,PPC,RISCV] Enable memtest
 			Format: <integer>
 			default : 0 <disable>

arch/powerpc/include/asm/book3s/64/mmu.h

@@ -71,11 +71,6 @@ extern unsigned int mmu_pid_bits;
 /* Base PID to allocate from */
 extern unsigned int mmu_base_pid;
 
-/*
- * memory block size used with radix translation.
- */
-extern unsigned long __ro_after_init radix_mem_block_size;
-
 #define PRTB_SIZE_SHIFT	(mmu_pid_bits + 4)
 #define PRTB_ENTRIES	(1ul << mmu_pid_bits)
 

arch/powerpc/kernel/setup_64.c

@@ -885,13 +885,40 @@ void __init setup_per_cpu_areas(void)
 #endif
 
 #ifdef CONFIG_MEMORY_HOTPLUG
+static unsigned long set_mem_block_size;
 unsigned long memory_block_size_bytes(void)
 {
+	if (set_mem_block_size)
+		return set_mem_block_size;
+
 	if (ppc_md.memory_block_size)
 		return ppc_md.memory_block_size();
 
 	return MIN_MEMORY_BLOCK_SIZE;
 }
+
+/*
+ * Restrict to a power of 2 value for memblock which is larger than
+ * section size
+ */
+static int __init parse_mem_block_size(char *ptr)
+{
+	unsigned int order;
+	unsigned long size = memparse(ptr, NULL);
+
+	order = fls64(size);
+	if (!order)
+		return 0;
+
+	order--;
+	if (order < SECTION_SIZE_BITS)
+		return 0;
+
+	set_mem_block_size = 1UL << order;
+
+	return 0;
+}
+early_param("memory_block_size", parse_mem_block_size);
 #endif
 
 #if defined(CONFIG_PPC_INDIRECT_PIO) || defined(CONFIG_PPC_INDIRECT_MMIO)

arch/powerpc/mm/book3s64/radix_pgtable.c

@@ -37,7 +37,6 @@
 #include <mm/mmu_decl.h>
 
 unsigned int mmu_base_pid;
-unsigned long radix_mem_block_size __ro_after_init;
 
 static __ref void *early_alloc_pgtable(unsigned long size, int nid,
 			unsigned long region_start, unsigned long region_end)
@@ -300,7 +299,7 @@ static int __meminit create_physical_mapping(unsigned long start,
 	bool prev_exec, exec = false;
 	pgprot_t prot;
 	int psize;
-	unsigned long max_mapping_size = radix_mem_block_size;
+	unsigned long max_mapping_size = memory_block_size_bytes();
 
 	if (debug_pagealloc_enabled_or_kfence())
 		max_mapping_size = PAGE_SIZE;
@@ -502,58 +501,6 @@ static int __init radix_dt_scan_page_sizes(unsigned long node,
 	return 1;
 }
 
-#ifdef CONFIG_MEMORY_HOTPLUG
-static int __init probe_memory_block_size(unsigned long node, const char *uname, int
-					  depth, void *data)
-{
-	unsigned long *mem_block_size = (unsigned long *)data;
-	const __be32 *prop;
-	int len;
-
-	if (depth != 1)
-		return 0;
-
-	if (strcmp(uname, "ibm,dynamic-reconfiguration-memory"))
-		return 0;
-
-	prop = of_get_flat_dt_prop(node, "ibm,lmb-size", &len);
-
-	if (!prop || len < dt_root_size_cells * sizeof(__be32))
-		/*
-		 * Nothing in the device tree
-		 */
-		*mem_block_size = MIN_MEMORY_BLOCK_SIZE;
-	else
-		*mem_block_size = of_read_number(prop, dt_root_size_cells);
-	return 1;
-}
-
-static unsigned long __init radix_memory_block_size(void)
-{
-	unsigned long mem_block_size = MIN_MEMORY_BLOCK_SIZE;
-
-	/*
-	 * OPAL firmware feature is set by now. Hence we are ok
-	 * to test OPAL feature.
-	 */
-	if (firmware_has_feature(FW_FEATURE_OPAL))
-		mem_block_size = 1UL * 1024 * 1024 * 1024;
-	else
-		of_scan_flat_dt(probe_memory_block_size, &mem_block_size);
-
-	return mem_block_size;
-}
-
-#else   /* CONFIG_MEMORY_HOTPLUG */
-
-static unsigned long __init radix_memory_block_size(void)
-{
-	return 1UL * 1024 * 1024 * 1024;
-}
-
-#endif /* CONFIG_MEMORY_HOTPLUG */
-
-
 void __init radix__early_init_devtree(void)
 {
 	int rc;
@@ -578,15 +525,6 @@ void __init radix__early_init_devtree(void)
 			psize_to_rpti_pgsize(MMU_PAGE_64K);
 	}
 
-	/*
-	 * Max mapping size used when mapping pages. We don't use
-	 * ppc_md.memory_block_size() here because this get called
-	 * early and we don't have machine probe called yet. Also
-	 * the pseries implementation only check for ibm,lmb-size.
-	 * All hypervisor supporting radix do expose that device
-	 * tree node.
-	 */
-	radix_mem_block_size = radix_memory_block_size();
 	return;
 }
 

arch/powerpc/platforms/powernv/setup.c

@@ -488,9 +488,9 @@ static unsigned long pnv_memory_block_size(void)
 	 * this size.
 	 */
 	if (radix_enabled())
-		return radix_mem_block_size;
+		return 1UL << 30;
 	else
-		return 256UL * 1024 * 1024;
+		return 256UL << 20;
 }
 #endif