A 5-level paging capable machine can have memory above 46-bit in the physical address space. This memory is only addressable in the 5-level paging mode: we don't have enough virtual address space to create direct mapping for such memory in the 4-level paging mode.
Currently, we fail boot completely: NULL pointer dereference in subsection_map_init().
Skip creating a memblock for such memory instead and notify user that some memory is not addressable.
Signed-off-by: Kirill A. Shutemov kirill.shutemov@linux.intel.com Reviewed-by: Dave Hansen dave.hansen@intel.com Cc: stable@vger.kernel.org # v4.14 --- arch/x86/kernel/e820.c | 24 ++++++++++++++++++++++-- 1 file changed, 22 insertions(+), 2 deletions(-)
diff --git a/arch/x86/kernel/e820.c b/arch/x86/kernel/e820.c index c5399e80c59c..c10bab121916 100644 --- a/arch/x86/kernel/e820.c +++ b/arch/x86/kernel/e820.c @@ -16,6 +16,7 @@ #include <linux/firmware-map.h> #include <linux/sort.h> #include <linux/memory_hotplug.h> +#include <linux/string_helpers.h>
#include <asm/e820/api.h> #include <asm/setup.h> @@ -1280,8 +1281,8 @@ void __init e820__memory_setup(void)
void __init e820__memblock_setup(void) { + u64 size, end, not_addressable = 0; int i; - u64 end;
/* * The bootstrap memblock region count maximum is 128 entries @@ -1307,7 +1308,26 @@ void __init e820__memblock_setup(void) if (entry->type != E820_TYPE_RAM && entry->type != E820_TYPE_RESERVED_KERN) continue;
- memblock_add(entry->addr, entry->size); + if (entry->addr >= MAXMEM) { + not_addressable += entry->size; + continue; + } + + end = min_t(u64, end, MAXMEM - 1); + size = end - entry->addr; + not_addressable += entry->size - size; + memblock_add(entry->addr, size); + } + + if (not_addressable) { + char tmp[10]; + + string_get_size(not_addressable, 1, STRING_UNITS_2, tmp, sizeof(tmp)); + pr_err("%s of physical memory is not addressable in the %s paging mode\n", + tmp, pgtable_l5_enabled() ? "5-level" : "4-level"); + + if (!pgtable_l5_enabled()) + pr_err("Consider enabling 5-level paging\n"); }
/* Throw away partial pages: */
On Mon, Jun 08, 2020 at 03:54:23PM +0300, Kirill A. Shutemov wrote:
A 5-level paging capable machine can have memory above 46-bit in the physical address space. This memory is only addressable in the 5-level paging mode: we don't have enough virtual address space to create direct mapping for such memory in the 4-level paging mode.
Currently, we fail boot completely: NULL pointer dereference in subsection_map_init().
Skip creating a memblock for such memory instead and notify user that some memory is not addressable.
Signed-off-by: Kirill A. Shutemov kirill.shutemov@linux.intel.com Reviewed-by: Dave Hansen dave.hansen@intel.com Cc: stable@vger.kernel.org # v4.14
Reviewed-by: Mike Rapoport rppt@linux.ibm.com
arch/x86/kernel/e820.c | 24 ++++++++++++++++++++++-- 1 file changed, 22 insertions(+), 2 deletions(-)
diff --git a/arch/x86/kernel/e820.c b/arch/x86/kernel/e820.c index c5399e80c59c..c10bab121916 100644 --- a/arch/x86/kernel/e820.c +++ b/arch/x86/kernel/e820.c @@ -16,6 +16,7 @@ #include <linux/firmware-map.h> #include <linux/sort.h> #include <linux/memory_hotplug.h> +#include <linux/string_helpers.h> #include <asm/e820/api.h> #include <asm/setup.h> @@ -1280,8 +1281,8 @@ void __init e820__memory_setup(void) void __init e820__memblock_setup(void) {
- u64 size, end, not_addressable = 0; int i;
- u64 end;
/* * The bootstrap memblock region count maximum is 128 entries @@ -1307,7 +1308,26 @@ void __init e820__memblock_setup(void) if (entry->type != E820_TYPE_RAM && entry->type != E820_TYPE_RESERVED_KERN) continue;
memblock_add(entry->addr, entry->size);
if (entry->addr >= MAXMEM) {not_addressable += entry->size;continue;}end = min_t(u64, end, MAXMEM - 1);size = end - entry->addr;not_addressable += entry->size - size;memblock_add(entry->addr, size);- }
- if (not_addressable) {
char tmp[10];string_get_size(not_addressable, 1, STRING_UNITS_2, tmp, sizeof(tmp));pr_err("%s of physical memory is not addressable in the %s paging mode\n",tmp, pgtable_l5_enabled() ? "5-level" : "4-level");if (!pgtable_l5_enabled())pr_err("Consider enabling 5-level paging\n");/* Throw away partial pages: */ -- 2.26.2
Greeting,
FYI, we noticed the following commit (built with gcc-9):
commit: c7b2a6d1affb732cdb37677f31cc8c2499fbf57c ("[PATCHv2 1/2] x86/mm: Fix boot with some memory above MAXMEM") url: https://github.com/0day-ci/linux/commits/Kirill-A-Shutemov/x86-Fix-boot-with...
in testcase: boot
on test machine: qemu-system-x86_64 -enable-kvm -cpu SandyBridge -smp 2 -m 16G
caused below changes (please refer to attached dmesg/kmsg for entire log/backtrace):
+---------------------------------------------------------+----------+------------+ | | v5.7-rc7 | c7b2a6d1af | +---------------------------------------------------------+----------+------------+ | boot_successes | 35 | 0 | | boot_failures | 3 | 12 | | BUG:unable_to_handle_page_fault_for_address | 1 | | | Oops:#[##] | 1 | | | invoked_oom-killer:gfp_mask=0x | 2 | | | Mem-Info | 2 | | | Out_of_memory_and_no_killable_processes | 2 | | | Kernel_panic-not_syncing:System_is_deadlocked_on_memory | 2 | | | kernel_BUG_at_arch/x86/mm/physaddr.c | 0 | 12 | | invalid_opcode:#[##] | 0 | 12 | | EIP:__phys_addr | 0 | 12 | | Kernel_panic-not_syncing:Fatal_exception | 0 | 12 | +---------------------------------------------------------+----------+------------+
If you fix the issue, kindly add following tag Reported-by: kernel test robot rong.a.chen@intel.com
[ 0.270304] kernel BUG at arch/x86/mm/physaddr.c:81! [ 0.271167] invalid opcode: 0000 [#1] SMP [ 0.271930] CPU: 0 PID: 0 Comm: swapper/0 Not tainted 5.7.0-rc7-00001-gc7b2a6d1affb7 #1 [ 0.273438] EIP: __phys_addr+0x77/0x170 [ 0.274191] Code: d4 ff cf b4 00 e8 c9 69 00 00 83 05 e8 ff cf b4 01 83 15 ec ff cf b4 00 39 f0 74 3d 83 05 f0 ff cf b4 01 83 15 f4 ff cf b4 00 <0f> 0b 83 05 f8 ff cf b4 01 83 15 fc ff cf b4 00 83 05 00 00 d0 b4 [ 0.277762] EAX: 00000fff EBX: f73fdfff ECX: fffff000 EDX: 00000001 [ 0.279012] ESI: 473fdfff EDI: 00000000 EBP: b44fde34 ESP: b44fde2c [ 0.280216] DS: 007b ES: 007b FS: 00d8 GS: 0000 SS: 0068 EFLAGS: 00210246 [ 0.281576] CR0: 80050033 CR2: ffffffff CR3: 04cc2000 CR4: 00040690 [ 0.282634] Call Trace: [ 0.283117] memtype_kernel_map_sync+0x2e/0x130 [ 0.283995] ? __get_vm_area_node+0x13a/0x230 [ 0.284939] ? get_vm_area_caller+0x58/0x70 [ 0.285757] __ioremap_caller+0x2ea/0x510 [ 0.286734] ? walk_system_ram_range+0xb5/0x110 [ 0.287599] ioremap_cache+0x2f/0x40 [ 0.288288] ? acpi_os_map_iomem+0x34d/0x3c0 [ 0.289103] acpi_os_map_iomem+0x34d/0x3c0 [ 0.289888] acpi_os_map_memory+0x16/0x30 [ 0.290704] acpi_tb_acquire_table+0x4b/0x96 [ 0.291571] acpi_tb_validate_table+0x61/0xb9 [ 0.292482] acpi_tb_validate_temp_table+0x48/0x58 [ 0.293464] acpi_tb_verify_temp_table+0x49/0x4e1 [ 0.294272] ? acpi_ut_acquire_mutex+0x1d8/0x2d9 [ 0.295103] acpi_reallocate_root_table+0x169/0x21a [ 0.296011] acpi_early_init+0x75/0x17f [ 0.296723] start_kernel+0xad5/0xcf3 [ 0.297414] i386_start_kernel+0xc1/0xd1 [ 0.298209] startup_32_smp+0x15f/0x164 [ 0.298991] Modules linked in: [ 0.299653] random: get_random_bytes called from init_oops_id+0x58/0x70 with crng_init=0 [ 0.299660] ---[ end trace 0894cc2a10a079b3 ]---
To reproduce:
# build kernel cd linux cp config-5.7.0-rc7-00001-gc7b2a6d1affb7 .config make HOSTCC=gcc-9 CC=gcc-9 ARCH=i386 olddefconfig prepare modules_prepare bzImage
git clone https://github.com/intel/lkp-tests.git cd lkp-tests bin/lkp qemu -k <bzImage> job-script # job-script is attached in this email
Thanks, Rong Chen
linux-stable-mirror@lists.linaro.org
-
kernel test robot -
Kirill A. Shutemov -
Mike Rapoport