Andrew Morton's on June 6, 2019 7:22 am: > On Thu, 6 Jun 2019 00:48:13 +1000 Nicholas Piggin <[email protected]> wrote: > >> The kernel currently clamps large system hashes to MAX_ORDER when >> hashdist is not set, which is rather arbitrary. >> >> vmalloc space is limited on 32-bit machines, but this shouldn't >> result in much more used because of small physical memory limiting >> system hash sizes. >> >> Include "vmalloc" or "linear" in the kernel log message. >> >> Signed-off-by: Nicholas Piggin <[email protected]> >> --- >> >> This is a better solution than the previous one for the case of !NUMA >> systems running on CONFIG_NUMA kernels, we can clear the default >> hashdist early and have everything allocated out of the linear map. >> >> The hugepage vmap series I will post later, but it's quite >> independent from this improvement. >> >> ... >> >> --- a/mm/page_alloc.c >> +++ b/mm/page_alloc.c >> @@ -7966,6 +7966,7 @@ void *__init alloc_large_system_hash(const char >> *tablename, >> unsigned long log2qty, size; >> void *table = NULL; >> gfp_t gfp_flags; >> + bool virt; >> >> /* allow the kernel cmdline to have a say */ >> if (!numentries) { >> @@ -8022,6 +8023,7 @@ void *__init alloc_large_system_hash(const char >> *tablename, >> >> gfp_flags = (flags & HASH_ZERO) ? GFP_ATOMIC | __GFP_ZERO : GFP_ATOMIC; >> do { >> + virt = false; >> size = bucketsize << log2qty; >> if (flags & HASH_EARLY) { >> if (flags & HASH_ZERO) >> @@ -8029,26 +8031,26 @@ void *__init alloc_large_system_hash(const char >> *tablename, >> else >> table = memblock_alloc_raw(size, >> SMP_CACHE_BYTES); >> - } else if (hashdist) { >> + } else if (get_order(size) >= MAX_ORDER || hashdist) { >> table = __vmalloc(size, gfp_flags, PAGE_KERNEL); >> + virt = true; >> } else { >> /* >> * If bucketsize is not a power-of-two, we may free >> * some pages at the end of hash table which >> * alloc_pages_exact() automatically does >> */ >> - if (get_order(size) < MAX_ORDER) { >> - table = alloc_pages_exact(size, gfp_flags); >> - kmemleak_alloc(table, size, 1, gfp_flags); >> - } >> + table = alloc_pages_exact(size, gfp_flags); >> + kmemleak_alloc(table, size, 1, gfp_flags); >> } >> } while (!table && size > PAGE_SIZE && --log2qty); >> >> if (!table) >> panic("Failed to allocate %s hash table\n", tablename); >> >> - pr_info("%s hash table entries: %ld (order: %d, %lu bytes)\n", >> - tablename, 1UL << log2qty, ilog2(size) - PAGE_SHIFT, size); >> + pr_info("%s hash table entries: %ld (order: %d, %lu bytes, %s)\n", >> + tablename, 1UL << log2qty, ilog2(size) - PAGE_SHIFT, size, >> + virt ? "vmalloc" : "linear"); > > Could remove `bool virt' and use is_vmalloc_addr() in the printk? >
It can run before mem_init() and it looks like some archs set VMALLOC_START/END (high_memory) there (e.g., x86-32, ppc32). Thanks, Nick
