aboutsummaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authorLinus Torvalds <torvalds@linux-foundation.org>2014-06-03 13:26:41 -0400
committerLinus Torvalds <torvalds@linux-foundation.org>2014-06-03 13:26:41 -0400
commit8f5759aeb88a47448cd92ab55a016d013b154a98 (patch)
treea9c0536e10300a95292b99332171837675af1e16
parente5c4ecdc55b6d824365ba7964bcd3185223f9688 (diff)
parent63aef00b55d37e9fad837a8b38a2c261f0d32041 (diff)
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux into next
Pull first set of s390 updates from Martin Schwidefsky: "The biggest change in this patchset is conversion from the bootmem bitmaps to the memblock code. This conversion requires two common code patches to introduce the 'physmem' memblock list. We experimented with ticket spinlocks but in the end decided against them as they perform poorly on virtualized systems. But the spinlock cleanup and some small improvements are included. The uaccess code got another optimization, the get_user/put_user calls are now inline again for kernel compiles targeted at z10 or newer machines. This makes the text segment shorter and the code gets a little bit faster. And as always some bug fixes" * 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux: (31 commits) s390/lowcore: replace lowcore irb array with a per-cpu variable s390/lowcore: reserve 96 bytes for IRB in lowcore s390/facilities: remove extract-cpu-time facility check s390: require mvcos facility for z10 and newer machines s390/boot: fix boot of compressed kernel built with gcc 4.9 s390/cio: remove weird assignment during argument evaluation s390/time: cast tv_nsec to u64 prior to shift in update_vsyscall s390/oprofile: make return of 0 explicit s390/spinlock: refactor arch_spin_lock_wait[_flags] s390/rwlock: add missing local_irq_restore calls s390/spinlock,rwlock: always to a load-and-test first s390/cio: fix multiple structure definitions s390/spinlock: fix system hang with spin_retry <= 0 s390/appldata: add slab.h for kzalloc/kfree s390/uaccess: provide inline variants of get_user/put_user s390/pci: add some new arch specific pci attributes s390/pci: use pdev->dev.groups for attribute creation s390/pci: use macro for attribute creation s390/pci: improve state check when processing hotplug events s390: split TIF bits into CIF, PIF and TIF bits ...
Diffstat
-rw-r--r--Documentation/s390/zfcpdump.txt73
-rw-r--r--arch/s390/Kconfig14
-rw-r--r--arch/s390/appldata/appldata_mem.c1
-rw-r--r--arch/s390/boot/compressed/Makefile2
-rw-r--r--arch/s390/include/asm/ccwdev.h2
-rw-r--r--arch/s390/include/asm/ccwgroup.h2
-rw-r--r--arch/s390/include/asm/chpid.h11
-rw-r--r--arch/s390/include/asm/futex.h4
-rw-r--r--arch/s390/include/asm/lowcore.h25
-rw-r--r--arch/s390/include/asm/mmu_context.h43
-rw-r--r--arch/s390/include/asm/pci.h6
-rw-r--r--arch/s390/include/asm/pci_clp.h10
-rw-r--r--arch/s390/include/asm/processor.h22
-rw-r--r--arch/s390/include/asm/ptrace.h22
-rw-r--r--arch/s390/include/asm/setup.h16
-rw-r--r--arch/s390/include/asm/smp.h5
-rw-r--r--arch/s390/include/asm/spinlock.h129
-rw-r--r--arch/s390/include/asm/spinlock_types.h6
-rw-r--r--arch/s390/include/asm/switch_to.h1
-rw-r--r--arch/s390/include/asm/syscall.h2
-rw-r--r--arch/s390/include/asm/thread_info.h34
-rw-r--r--arch/s390/include/asm/uaccess.h30
-rw-r--r--arch/s390/kernel/asm-offsets.c3
-rw-r--r--arch/s390/kernel/compat_signal.c2
-rw-r--r--arch/s390/kernel/crash_dump.c83
-rw-r--r--arch/s390/kernel/early.c6
-rw-r--r--arch/s390/kernel/entry.S87
-rw-r--r--arch/s390/kernel/entry64.S82
-rw-r--r--arch/s390/kernel/head.S8
-rw-r--r--arch/s390/kernel/head31.S1
-rw-r--r--arch/s390/kernel/nmi.c8
-rw-r--r--arch/s390/kernel/process.c6
-rw-r--r--arch/s390/kernel/ptrace.c4
-rw-r--r--arch/s390/kernel/setup.c459
-rw-r--r--arch/s390/kernel/signal.c12
-rw-r--r--arch/s390/kernel/smp.c20
-rw-r--r--arch/s390/kernel/time.c2
-rw-r--r--arch/s390/kernel/topology.c4
-rw-r--r--arch/s390/kvm/kvm-s390.c2
-rw-r--r--arch/s390/lib/spinlock.c157
-rw-r--r--arch/s390/lib/uaccess.c10
-rw-r--r--arch/s390/mm/fault.c2
-rw-r--r--arch/s390/mm/mem_detect.c130
-rw-r--r--arch/s390/mm/page-states.c10
-rw-r--r--arch/s390/mm/pgtable.c10
-rw-r--r--arch/s390/mm/vmem.c30
-rw-r--r--arch/s390/oprofile/hwsampler.c14
-rw-r--r--arch/s390/pci/pci_clp.c10
-rw-r--r--arch/s390/pci/pci_event.c5
-rw-r--r--arch/s390/pci/pci_sysfs.c97
-rw-r--r--drivers/s390/char/Makefile2
-rw-r--r--drivers/s390/char/zcore.c44
-rw-r--r--drivers/s390/cio/ccwreq.c4
-rw-r--r--drivers/s390/cio/chp.c2
-rw-r--r--drivers/s390/cio/chp.h2
-rw-r--r--drivers/s390/cio/chsc.h11
-rw-r--r--drivers/s390/cio/chsc_sch.c2
-rw-r--r--drivers/s390/cio/cio.c9
-rw-r--r--drivers/s390/cio/cio.h2
-rw-r--r--drivers/s390/cio/device_fsm.c4
-rw-r--r--drivers/s390/cio/device_ops.c13
-rw-r--r--drivers/s390/cio/eadm_sch.c2
-rw-r--r--drivers/s390/net/qeth_core_main.c12
-rw-r--r--include/linux/memblock.h79
-rw-r--r--mm/Kconfig3
-rw-r--r--mm/memblock.c205
66 files changed, 1048 insertions, 1072 deletions
diff --git a/Documentation/s390/zfcpdump.txt b/Documentation/s390/zfcpdump.txt
index cf45d27c4608..dc929be96016 100644
--- a/Documentation/s390/zfcpdump.txt
+++ b/Documentation/s390/zfcpdump.txt
@@ -1,15 +1,15 @@
1s390 SCSI dump tool (zfcpdump) 1The s390 SCSI dump tool (zfcpdump)
2 2
3System z machines (z900 or higher) provide hardware support for creating system 3System z machines (z900 or higher) provide hardware support for creating system
4dumps on SCSI disks. The dump process is initiated by booting a dump tool, which 4dumps on SCSI disks. The dump process is initiated by booting a dump tool, which
5has to create a dump of the current (probably crashed) Linux image. In order to 5has to create a dump of the current (probably crashed) Linux image. In order to
6not overwrite memory of the crashed Linux with data of the dump tool, the 6not overwrite memory of the crashed Linux with data of the dump tool, the
7hardware saves some memory plus the register sets of the boot cpu before the 7hardware saves some memory plus the register sets of the boot CPU before the
8dump tool is loaded. There exists an SCLP hardware interface to obtain the saved 8dump tool is loaded. There exists an SCLP hardware interface to obtain the saved
9memory afterwards. Currently 32 MB are saved. 9memory afterwards. Currently 32 MB are saved.
10 10
11This zfcpdump implementation consists of a Linux dump kernel together with 11This zfcpdump implementation consists of a Linux dump kernel together with
12a userspace dump tool, which are loaded together into the saved memory region 12a user space dump tool, which are loaded together into the saved memory region
13below 32 MB. zfcpdump is installed on a SCSI disk using zipl (as contained in 13below 32 MB. zfcpdump is installed on a SCSI disk using zipl (as contained in
14the s390-tools package) to make the device bootable. The operator of a Linux 14the s390-tools package) to make the device bootable. The operator of a Linux
15system can then trigger a SCSI dump by booting the SCSI disk, where zfcpdump 15system can then trigger a SCSI dump by booting the SCSI disk, where zfcpdump
@@ -19,68 +19,33 @@ The kernel part of zfcpdump is implemented as a debugfs file under "zcore/mem",
19which exports memory and registers of the crashed Linux in an s390 19which exports memory and registers of the crashed Linux in an s390
20standalone dump format. It can be used in the same way as e.g. /dev/mem. The 20standalone dump format. It can be used in the same way as e.g. /dev/mem. The
21dump format defines a 4K header followed by plain uncompressed memory. The 21dump format defines a 4K header followed by plain uncompressed memory. The
22register sets are stored in the prefix pages of the respective cpus. To build a 22register sets are stored in the prefix pages of the respective CPUs. To build a
23dump enabled kernel with the zcore driver, the kernel config option 23dump enabled kernel with the zcore driver, the kernel config option
24CONFIG_ZFCPDUMP has to be set. When reading from "zcore/mem", the part of 24CONFIG_CRASH_DUMP has to be set. When reading from "zcore/mem", the part of
25memory, which has been saved by hardware is read by the driver via the SCLP 25memory, which has been saved by hardware is read by the driver via the SCLP
26hardware interface. The second part is just copied from the non overwritten real 26hardware interface. The second part is just copied from the non overwritten real
27memory. 27memory.
28 28
29The userspace application of zfcpdump can reside e.g. in an intitramfs or an 29Since kernel version 3.12 also the /proc/vmcore file can also be used to access
30initrd. It reads from zcore/mem and writes the system dump to a file on a 30the dump.
31SCSI disk.
32 31
33To build a zfcpdump kernel use the following settings in your kernel 32To get a valid zfcpdump kernel configuration use "make zfcpdump_defconfig".
34configuration:
35 * CONFIG_ZFCPDUMP=y
36 * Enable ZFCP driver
37 * Enable SCSI driver
38 * Enable ext2 and ext3 filesystems
39 * Disable as many features as possible to keep the kernel small.
40 E.g. network support is not needed at all.
41 33
42To use the zfcpdump userspace application in an initramfs you have to do the 34The s390 zipl tool looks for the zfcpdump kernel and optional initrd/initramfs
43following: 35under the following locations:
44 36
45 * Copy the zfcpdump executable somewhere into your Linux tree. 37* kernel: <zfcpdump directory>/zfcpdump.image
46 E.g. to "arch/s390/boot/zfcpdump. If you do not want to include 38* ramdisk: <zfcpdump directory>/zfcpdump.rd
47 shared libraries, compile the tool with the "-static" gcc option.
48 * If you want to include e2fsck, add it to your source tree, too. The zfcpdump
49 application attempts to start /sbin/e2fsck from the ramdisk.
50 * Use an initramfs config file like the following:
51 39
52 dir /dev 755 0 0 40The zfcpdump directory is defined in the s390-tools package.
53 nod /dev/console 644 0 0 c 5 1
54 nod /dev/null 644 0 0 c 1 3
55 nod /dev/sda1 644 0 0 b 8 1
56 nod /dev/sda2 644 0 0 b 8 2
57 nod /dev/sda3 644 0 0 b 8 3
58 nod /dev/sda4 644 0 0 b 8 4
59 nod /dev/sda5 644 0 0 b 8 5
60 nod /dev/sda6 644 0 0 b 8 6
61 nod /dev/sda7 644 0 0 b 8 7
62 nod /dev/sda8 644 0 0 b 8 8
63 nod /dev/sda9 644 0 0 b 8 9
64 nod /dev/sda10 644 0 0 b 8 10
65 nod /dev/sda11 644 0 0 b 8 11
66 nod /dev/sda12 644 0 0 b 8 12
67 nod /dev/sda13 644 0 0 b 8 13
68 nod /dev/sda14 644 0 0 b 8 14
69 nod /dev/sda15 644 0 0 b 8 15
70 file /init arch/s390/boot/zfcpdump 755 0 0
71 file /sbin/e2fsck arch/s390/boot/e2fsck 755 0 0
72 dir /proc 755 0 0
73 dir /sys 755 0 0
74 dir /mnt 755 0 0
75 dir /sbin 755 0 0
76 41
77 * Issue "make image" to build the zfcpdump image with initramfs. 42The user space application of zfcpdump can reside in an intitramfs or an
43initrd. It can also be included in a built-in kernel initramfs. The application
44reads from /proc/vmcore or zcore/mem and writes the system dump to a SCSI disk.
78 45
79In a Linux distribution the zfcpdump enabled kernel image must be copied to 46The s390-tools package version 1.24.0 and above builds an external zfcpdump
80/usr/share/zfcpdump/zfcpdump.image, where the s390 zipl tool is looking for the 47initramfs with a user space application that writes the dump to a SCSI
81dump kernel when preparing a SCSI dump disk. 48partition.
82
83If you use a ramdisk copy it to "/usr/share/zfcpdump/zfcpdump.rd".
84 49
85For more information on how to use zfcpdump refer to the s390 'Using the Dump 50For more information on how to use zfcpdump refer to the s390 'Using the Dump
86Tools book', which is available from 51Tools book', which is available from
diff --git a/arch/s390/Kconfig b/arch/s390/Kconfig
index d68fe34799b0..bb63499fc5d3 100644
--- a/arch/s390/Kconfig
+++ b/arch/s390/Kconfig
@@ -60,7 +60,6 @@ config PCI_QUIRKS
60 60
61config S390 61config S390
62 def_bool y 62 def_bool y
63 select ARCH_DISCARD_MEMBLOCK
64 select ARCH_HAS_ATOMIC64_DEC_IF_POSITIVE 63 select ARCH_HAS_ATOMIC64_DEC_IF_POSITIVE
65 select ARCH_HAS_DEBUG_STRICT_USER_COPY_CHECKS 64 select ARCH_HAS_DEBUG_STRICT_USER_COPY_CHECKS
66 select ARCH_HAVE_NMI_SAFE_CMPXCHG 65 select ARCH_HAVE_NMI_SAFE_CMPXCHG
@@ -130,6 +129,7 @@ config S390
130 select HAVE_KVM if 64BIT 129 select HAVE_KVM if 64BIT
131 select HAVE_MEMBLOCK 130 select HAVE_MEMBLOCK
132 select HAVE_MEMBLOCK_NODE_MAP 131 select HAVE_MEMBLOCK_NODE_MAP
132 select HAVE_MEMBLOCK_PHYS_MAP
133 select HAVE_MOD_ARCH_SPECIFIC 133 select HAVE_MOD_ARCH_SPECIFIC
134 select HAVE_OPROFILE 134 select HAVE_OPROFILE
135 select HAVE_PERF_EVENTS 135 select HAVE_PERF_EVENTS
@@ -139,6 +139,7 @@ config S390
139 select HAVE_VIRT_CPU_ACCOUNTING 139 select HAVE_VIRT_CPU_ACCOUNTING
140 select KTIME_SCALAR if 32BIT 140 select KTIME_SCALAR if 32BIT
141 select MODULES_USE_ELF_RELA 141 select MODULES_USE_ELF_RELA
142 select NO_BOOTMEM
142 select OLD_SIGACTION 143 select OLD_SIGACTION
143 select OLD_SIGSUSPEND3 144 select OLD_SIGSUSPEND3
144 select SYSCTL_EXCEPTION_TRACE 145 select SYSCTL_EXCEPTION_TRACE
@@ -592,21 +593,14 @@ config CRASH_DUMP
592 bool "kernel crash dumps" 593 bool "kernel crash dumps"
593 depends on 64BIT && SMP 594 depends on 64BIT && SMP
594 select KEXEC 595 select KEXEC
595 select ZFCPDUMP
596 help 596 help
597 Generate crash dump after being started by kexec. 597 Generate crash dump after being started by kexec.
598 Crash dump kernels are loaded in the main kernel with kexec-tools 598 Crash dump kernels are loaded in the main kernel with kexec-tools
599 into a specially reserved region and then later executed after 599 into a specially reserved region and then later executed after
600 a crash by kdump/kexec. 600 a crash by kdump/kexec.
601 For more details see Documentation/kdump/kdump.txt
602
603config ZFCPDUMP
604 def_bool n
605 prompt "zfcpdump support"
606 depends on 64BIT && SMP
607 help
608 Select this option if you want to build an zfcpdump enabled kernel.
609 Refer to <file:Documentation/s390/zfcpdump.txt> for more details on this. 601 Refer to <file:Documentation/s390/zfcpdump.txt> for more details on this.
602 This option also enables s390 zfcpdump.
603 See also <file:Documentation/s390/zfcpdump.txt>
610 604
611endmenu 605endmenu
612 606
diff --git a/arch/s390/appldata/appldata_mem.c b/arch/s390/appldata/appldata_mem.c
index 42be53743133..edcf2a706942 100644
--- a/arch/s390/appldata/appldata_mem.c
+++ b/arch/s390/appldata/appldata_mem.c
@@ -13,6 +13,7 @@
13#include <linux/kernel_stat.h> 13#include <linux/kernel_stat.h>
14#include <linux/pagemap.h> 14#include <linux/pagemap.h>
15#include <linux/swap.h> 15#include <linux/swap.h>
16#include <linux/slab.h>
16#include <asm/io.h> 17#include <asm/io.h>
17 18
18#include "appldata.h" 19#include "appldata.h"
diff --git a/arch/s390/boot/compressed/Makefile b/arch/s390/boot/compressed/Makefile
index 866ecbe670e4..f90d1fc6d603 100644
--- a/arch/s390/boot/compressed/Makefile
+++ b/arch/s390/boot/compressed/Makefile
@@ -12,7 +12,7 @@ targets += misc.o piggy.o sizes.h head$(BITS).o
12 12
13KBUILD_CFLAGS := -m$(BITS) -D__KERNEL__ $(LINUX_INCLUDE) -O2 13KBUILD_CFLAGS := -m$(BITS) -D__KERNEL__ $(LINUX_INCLUDE) -O2
14KBUILD_CFLAGS += -DDISABLE_BRANCH_PROFILING 14KBUILD_CFLAGS += -DDISABLE_BRANCH_PROFILING
15KBUILD_CFLAGS += $(cflags-y) 15KBUILD_CFLAGS += $(cflags-y) -fno-delete-null-pointer-checks
16KBUILD_CFLAGS += $(call cc-option,-mpacked-stack) 16KBUILD_CFLAGS += $(call cc-option,-mpacked-stack)
17KBUILD_CFLAGS += $(call cc-option,-ffreestanding) 17KBUILD_CFLAGS += $(call cc-option,-ffreestanding)
18 18
diff --git a/arch/s390/include/asm/ccwdev.h b/arch/s390/include/asm/ccwdev.h
index a9c2c0686177..b80e456d6428 100644
--- a/arch/s390/include/asm/ccwdev.h
+++ b/arch/s390/include/asm/ccwdev.h
@@ -229,5 +229,5 @@ int ccw_device_siosl(struct ccw_device *);
229 229
230extern void ccw_device_get_schid(struct ccw_device *, struct subchannel_id *); 230extern void ccw_device_get_schid(struct ccw_device *, struct subchannel_id *);
231 231
232extern void *ccw_device_get_chp_desc(struct ccw_device *, int); 232struct channel_path_desc *ccw_device_get_chp_desc(struct ccw_device *, int);
233#endif /* _S390_CCWDEV_H_ */ 233#endif /* _S390_CCWDEV_H_ */
diff --git a/arch/s390/include/asm/ccwgroup.h b/arch/s390/include/asm/ccwgroup.h
index ebc2913f9ee0..057ce0ca6377 100644
--- a/arch/s390/include/asm/ccwgroup.h
+++ b/arch/s390/include/asm/ccwgroup.h
@@ -10,6 +10,8 @@ struct ccw_driver;
10 * @count: number of attached slave devices 10 * @count: number of attached slave devices
11 * @dev: embedded device structure 11 * @dev: embedded device structure
12 * @cdev: variable number of slave devices, allocated as needed 12 * @cdev: variable number of slave devices, allocated as needed
13 * @ungroup_work: work to be done when a ccwgroup notifier has action
14 * type %BUS_NOTIFY_UNBIND_DRIVER
13 */ 15 */
14struct ccwgroup_device { 16struct ccwgroup_device {
15 enum { 17 enum {
diff --git a/arch/s390/include/asm/chpid.h b/arch/s390/include/asm/chpid.h
index 38c405ef89ce..7298eec98541 100644
--- a/arch/s390/include/asm/chpid.h
+++ b/arch/s390/include/asm/chpid.h
@@ -8,6 +8,17 @@
8#include <uapi/asm/chpid.h> 8#include <uapi/asm/chpid.h>
9#include <asm/cio.h> 9#include <asm/cio.h>
10 10
11struct channel_path_desc {
12 u8 flags;
13 u8 lsn;
14 u8 desc;
15 u8 chpid;
16 u8 swla;
17 u8 zeroes;
18 u8 chla;
19 u8 chpp;
20} __packed;
21
11static inline void chp_id_init(struct chp_id *chpid) 22static inline void chp_id_init(struct chp_id *chpid)
12{ 23{
13 memset(chpid, 0, sizeof(struct chp_id)); 24 memset(chpid, 0, sizeof(struct chp_id));
diff --git a/arch/s390/include/asm/futex.h b/arch/s390/include/asm/futex.h
index 69cf5b5eddc9..a4811aa0304d 100644
--- a/arch/s390/include/asm/futex.h
+++ b/arch/s390/include/asm/futex.h
@@ -29,7 +29,7 @@ static inline int futex_atomic_op_inuser(int encoded_op, u32 __user *uaddr)
29 int cmparg = (encoded_op << 20) >> 20; 29 int cmparg = (encoded_op << 20) >> 20;
30 int oldval = 0, newval, ret; 30 int oldval = 0, newval, ret;
31 31
32 update_primary_asce(current); 32 load_kernel_asce();
33 if (encoded_op & (FUTEX_OP_OPARG_SHIFT << 28)) 33 if (encoded_op & (FUTEX_OP_OPARG_SHIFT << 28))
34 oparg = 1 << oparg; 34 oparg = 1 << oparg;
35 35
@@ -79,7 +79,7 @@ static inline int futex_atomic_cmpxchg_inatomic(u32 *uval, u32 __user *uaddr,
79{ 79{
80 int ret; 80 int ret;
81 81
82 update_primary_asce(current); 82 load_kernel_asce();
83 asm volatile( 83 asm volatile(
84 " sacf 256\n" 84 " sacf 256\n"
85 "0: cs %1,%4,0(%5)\n" 85 "0: cs %1,%4,0(%5)\n"
diff --git a/arch/s390/include/asm/lowcore.h b/arch/s390/include/asm/lowcore.h
index bbf8141408cd..2070cad80e9e 100644
--- a/arch/s390/include/asm/lowcore.h
+++ b/arch/s390/include/asm/lowcore.h
@@ -93,7 +93,9 @@ struct _lowcore {
93 __u32 save_area_sync[8]; /* 0x0200 */ 93 __u32 save_area_sync[8]; /* 0x0200 */
94 __u32 save_area_async[8]; /* 0x0220 */ 94 __u32 save_area_async[8]; /* 0x0220 */
95 __u32 save_area_restart[1]; /* 0x0240 */ 95 __u32 save_area_restart[1]; /* 0x0240 */
96 __u8 pad_0x0244[0x0248-0x0244]; /* 0x0244 */ 96
97 /* CPU flags. */
98 __u32 cpu_flags; /* 0x0244 */
97 99
98 /* Return psws. */ 100 /* Return psws. */
99 psw_t return_psw; /* 0x0248 */ 101 psw_t return_psw; /* 0x0248 */
@@ -139,12 +141,9 @@ struct _lowcore {
139 __u32 percpu_offset; /* 0x02f0 */ 141 __u32 percpu_offset; /* 0x02f0 */
140 __u32 machine_flags; /* 0x02f4 */ 142 __u32 machine_flags; /* 0x02f4 */
141 __u32 ftrace_func; /* 0x02f8 */ 143 __u32 ftrace_func; /* 0x02f8 */
142 __u8 pad_0x02fc[0x0300-0x02fc]; /* 0x02fc */ 144 __u32 spinlock_lockval; /* 0x02fc */
143
144 /* Interrupt response block */
145 __u8 irb[64]; /* 0x0300 */
146 145
147 __u8 pad_0x0340[0x0e00-0x0340]; /* 0x0340 */ 146 __u8 pad_0x0300[0x0e00-0x0300]; /* 0x0300 */
148 147
149 /* 148 /*
150 * 0xe00 contains the address of the IPL Parameter Information 149 * 0xe00 contains the address of the IPL Parameter Information
@@ -237,7 +236,9 @@ struct _lowcore {
237 __u64 save_area_sync[8]; /* 0x0200 */ 236 __u64 save_area_sync[8]; /* 0x0200 */
238 __u64 save_area_async[8]; /* 0x0240 */ 237 __u64 save_area_async[8]; /* 0x0240 */
239 __u64 save_area_restart[1]; /* 0x0280 */ 238 __u64 save_area_restart[1]; /* 0x0280 */
240 __u8 pad_0x0288[0x0290-0x0288]; /* 0x0288 */ 239
240 /* CPU flags. */
241 __u64 cpu_flags; /* 0x0288 */
241 242
242 /* Return psws. */ 243 /* Return psws. */
243 psw_t return_psw; /* 0x0290 */ 244 psw_t return_psw; /* 0x0290 */
@@ -285,15 +286,13 @@ struct _lowcore {
285 __u64 machine_flags; /* 0x0388 */ 286 __u64 machine_flags; /* 0x0388 */
286 __u64 ftrace_func; /* 0x0390 */ 287 __u64 ftrace_func; /* 0x0390 */
287 __u64 gmap; /* 0x0398 */ 288 __u64 gmap; /* 0x0398 */
288 __u8 pad_0x03a0[0x0400-0x03a0]; /* 0x03a0 */ 289 __u32 spinlock_lockval; /* 0x03a0 */
289 290 __u8 pad_0x03a0[0x0400-0x03a4]; /* 0x03a4 */
290 /* Interrupt response block. */
291 __u8 irb[64]; /* 0x0400 */
292 291
293 /* Per cpu primary space access list */ 292 /* Per cpu primary space access list */
294 __u32 paste[16]; /* 0x0440 */ 293 __u32 paste[16]; /* 0x0400 */
295 294
296 __u8 pad_0x0480[0x0e00-0x0480]; /* 0x0480 */ 295 __u8 pad_0x04c0[0x0e00-0x0440]; /* 0x0440 */
297 296
298 /* 297 /*
299 * 0xe00 contains the address of the IPL Parameter Information 298 * 0xe00 contains the address of the IPL Parameter Information
diff --git a/arch/s390/include/asm/mmu_context.h b/arch/s390/include/asm/mmu_context.h
index 71be346d0e3c..056d7eff2a16 100644
--- a/arch/s390/include/asm/mmu_context.h
+++ b/arch/s390/include/asm/mmu_context.h
@@ -30,33 +30,31 @@ static inline int init_new_context(struct task_struct *tsk,
30 30
31#define destroy_context(mm) do { } while (0) 31#define destroy_context(mm) do { } while (0)
32 32
33static inline void update_user_asce(struct mm_struct *mm, int load_primary) 33static inline void set_user_asce(struct mm_struct *mm)
34{ 34{
35 pgd_t *pgd = mm->pgd; 35 pgd_t *pgd = mm->pgd;
36 36
37 S390_lowcore.user_asce = mm->context.asce_bits | __pa(pgd); 37 S390_lowcore.user_asce = mm->context.asce_bits | __pa(pgd);
38 if (load_primary)
39 __ctl_load(S390_lowcore.user_asce, 1, 1);
40 set_fs(current->thread.mm_segment); 38 set_fs(current->thread.mm_segment);
39 set_cpu_flag(CIF_ASCE);
41} 40}
42 41
43static inline void clear_user_asce(struct mm_struct *mm, int load_primary) 42static inline void clear_user_asce(void)
44{ 43{
45 S390_lowcore.user_asce = S390_lowcore.kernel_asce; 44 S390_lowcore.user_asce = S390_lowcore.kernel_asce;
46 45
47 if (load_primary) 46 __ctl_load(S390_lowcore.user_asce, 1, 1);
48 __ctl_load(S390_lowcore.user_asce, 1, 1);
49 __ctl_load(S390_lowcore.user_asce, 7, 7); 47 __ctl_load(S390_lowcore.user_asce, 7, 7);
50} 48}
51 49
52static inline void update_primary_asce(struct task_struct *tsk) 50static inline void load_kernel_asce(void)
53{ 51{
54 unsigned long asce; 52 unsigned long asce;
55 53
56 __ctl_store(asce, 1, 1); 54 __ctl_store(asce, 1, 1);
57 if (asce != S390_lowcore.kernel_asce) 55 if (asce != S390_lowcore.kernel_asce)
58 __ctl_load(S390_lowcore.kernel_asce, 1, 1); 56 __ctl_load(S390_lowcore.kernel_asce, 1, 1);
59 set_tsk_thread_flag(tsk, TIF_ASCE); 57 set_cpu_flag(CIF_ASCE);
60} 58}
61 59
62static inline void switch_mm(struct mm_struct *prev, struct mm_struct *next, 60static inline void switch_mm(struct mm_struct *prev, struct mm_struct *next,
@@ -64,25 +62,17 @@ static inline void switch_mm(struct mm_struct *prev, struct mm_struct *next,
64{ 62{
65 int cpu = smp_processor_id(); 63 int cpu = smp_processor_id();
66 64
67 update_primary_asce(tsk);
68 if (prev == next) 65 if (prev == next)
69 return; 66 return;
70 if (MACHINE_HAS_TLB_LC) 67 if (MACHINE_HAS_TLB_LC)
71 cpumask_set_cpu(cpu, &next->context.cpu_attach_mask); 68 cpumask_set_cpu(cpu, &next->context.cpu_attach_mask);
72 if (atomic_inc_return(&next->context.attach_count) >> 16) { 69 /* Clear old ASCE by loading the kernel ASCE. */
73 /* Delay update_user_asce until all TLB flushes are done. */ 70 __ctl_load(S390_lowcore.kernel_asce, 1, 1);
74 set_tsk_thread_flag(tsk, TIF_TLB_WAIT); 71 __ctl_load(S390_lowcore.kernel_asce, 7, 7);
75 /* Clear old ASCE by loading the kernel ASCE. */ 72 /* Delay loading of the new ASCE to control registers CR1 & CR7 */
76 clear_user_asce(next, 0); 73 set_cpu_flag(CIF_ASCE);
77 } else { 74 atomic_inc(&next->context.attach_count);
78 cpumask_set_cpu(cpu, mm_cpumask(next));
79 update_user_asce(next, 0);
80 if (next->context.flush_mm)
81 /* Flush pending TLBs */
82 __tlb_flush_mm(next);
83 }
84 atomic_dec(&prev->context.attach_count); 75 atomic_dec(&prev->context.attach_count);
85 WARN_ON(atomic_read(&prev->context.attach_count) < 0);
86 if (MACHINE_HAS_TLB_LC) 76 if (MACHINE_HAS_TLB_LC)
87 cpumask_clear_cpu(cpu, &prev->context.cpu_attach_mask); 77 cpumask_clear_cpu(cpu, &prev->context.cpu_attach_mask);
88} 78}
@@ -93,15 +83,14 @@ static inline void finish_arch_post_lock_switch(void)
93 struct task_struct *tsk = current; 83 struct task_struct *tsk = current;
94 struct mm_struct *mm = tsk->mm; 84 struct mm_struct *mm = tsk->mm;
95 85
96 if (!test_tsk_thread_flag(tsk, TIF_TLB_WAIT)) 86 if (!mm)
97 return; 87 return;
98 preempt_disable(); 88 preempt_disable();
99 clear_tsk_thread_flag(tsk, TIF_TLB_WAIT);
100 while (atomic_read(&mm->context.attach_count) >> 16) 89 while (atomic_read(&mm->context.attach_count) >> 16)
101 cpu_relax(); 90 cpu_relax();
102 91
103 cpumask_set_cpu(smp_processor_id(), mm_cpumask(mm)); 92 cpumask_set_cpu(smp_processor_id(), mm_cpumask(mm));
104 update_user_asce(mm, 0); 93 set_user_asce(mm);
105 if (mm->context.flush_mm) 94 if (mm->context.flush_mm)
106 __tlb_flush_mm(mm); 95 __tlb_flush_mm(mm);
107 preempt_enable(); 96 preempt_enable();
@@ -113,7 +102,9 @@ static inline void finish_arch_post_lock_switch(void)
113static inline void activate_mm(struct mm_struct *prev, 102static inline void activate_mm(struct mm_struct *prev,
114 struct mm_struct *next) 103 struct mm_struct *next)
115{ 104{
116 switch_mm(prev, next, current); 105 switch_mm(prev, next, current);
106 cpumask_set_cpu(smp_processor_id(), mm_cpumask(next));
107 set_user_asce(next);
117} 108}
118 109
119static inline void arch_dup_mmap(struct mm_struct *oldmm, 110static inline void arch_dup_mmap(struct mm_struct *oldmm,
diff --git a/arch/s390/include/asm/pci.h b/arch/s390/include/asm/pci.h
index 79b5f0783a30..c030900320e0 100644
--- a/arch/s390/include/asm/pci.h
+++ b/arch/s390/include/asm/pci.h
@@ -78,10 +78,16 @@ struct zpci_dev {
78 enum zpci_state state; 78 enum zpci_state state;
79 u32 fid; /* function ID, used by sclp */ 79 u32 fid; /* function ID, used by sclp */
80 u32 fh; /* function handle, used by insn's */ 80 u32 fh; /* function handle, used by insn's */
81 u16 vfn; /* virtual function number */
81 u16 pchid; /* physical channel ID */ 82 u16 pchid; /* physical channel ID */
82 u8 pfgid; /* function group ID */ 83 u8 pfgid; /* function group ID */
84 u8 pft; /* pci function type */
83 u16 domain; 85 u16 domain;
84 86
87 u8 pfip[CLP_PFIP_NR_SEGMENTS]; /* pci function internal path */
88 u32 uid; /* user defined id */
89 u8 util_str[CLP_UTIL_STR_LEN]; /* utility string */
90
85 /* IRQ stuff */ 91 /* IRQ stuff */
86 u64 msi_addr; /* MSI address */ 92 u64 msi_addr; /* MSI address */
87 struct airq_iv *aibv; /* adapter interrupt bit vector */ 93 struct airq_iv *aibv; /* adapter interrupt bit vector */
diff --git a/arch/s390/include/asm/pci_clp.h b/arch/s390/include/asm/pci_clp.h
index d31d739f8689..dd78f92f1cce 100644
--- a/arch/s390/include/asm/pci_clp.h
+++ b/arch/s390/include/asm/pci_clp.h
@@ -44,6 +44,7 @@ struct clp_fh_list_entry {
44#define CLP_SET_DISABLE_PCI_FN 1 /* Yes, 1 disables it */ 44#define CLP_SET_DISABLE_PCI_FN 1 /* Yes, 1 disables it */
45 45
46#define CLP_UTIL_STR_LEN 64 46#define CLP_UTIL_STR_LEN 64
47#define CLP_PFIP_NR_SEGMENTS 4
47 48
48/* List PCI functions request */ 49/* List PCI functions request */
49struct clp_req_list_pci { 50struct clp_req_list_pci {
@@ -85,7 +86,7 @@ struct clp_rsp_query_pci {
85 struct clp_rsp_hdr hdr; 86 struct clp_rsp_hdr hdr;
86 u32 fmt : 4; /* cmd request block format */ 87 u32 fmt : 4; /* cmd request block format */
87 u32 : 28; 88 u32 : 28;
88 u64 reserved1; 89 u64 : 64;
89 u16 vfn; /* virtual fn number */ 90 u16 vfn; /* virtual fn number */
90 u16 : 7; 91 u16 : 7;
91 u16 util_str_avail : 1; /* utility string available? */ 92 u16 util_str_avail : 1; /* utility string available? */
@@ -94,10 +95,13 @@ struct clp_rsp_query_pci {
94 u8 bar_size[PCI_BAR_COUNT]; 95 u8 bar_size[PCI_BAR_COUNT];
95 u16 pchid; 96 u16 pchid;
96 u32 bar[PCI_BAR_COUNT]; 97 u32 bar[PCI_BAR_COUNT];
97 u64 reserved2; 98 u8 pfip[CLP_PFIP_NR_SEGMENTS]; /* pci function internal path */
99 u32 : 24;
100 u8 pft; /* pci function type */
98 u64 sdma; /* start dma as */ 101 u64 sdma; /* start dma as */
99 u64 edma; /* end dma as */ 102 u64 edma; /* end dma as */
100 u64 reserved3[6]; 103 u32 reserved[11];
104 u32 uid; /* user defined id */
101 u8 util_str[CLP_UTIL_STR_LEN]; /* utility string */ 105 u8 util_str[CLP_UTIL_STR_LEN]; /* utility string */
102} __packed; 106} __packed;
103 107
diff --git a/arch/s390/include/asm/processor.h b/arch/s390/include/asm/processor.h
index dc5fc4f90e52..6f02d452bbee 100644
--- a/arch/s390/include/asm/processor.h
+++ b/arch/s390/include/asm/processor.h
@@ -11,6 +11,13 @@
11#ifndef __ASM_S390_PROCESSOR_H 11#ifndef __ASM_S390_PROCESSOR_H
12#define __ASM_S390_PROCESSOR_H 12#define __ASM_S390_PROCESSOR_H
13 13
14#define CIF_MCCK_PENDING 0 /* machine check handling is pending */
15#define CIF_ASCE 1 /* user asce needs fixup / uaccess */
16
17#define _CIF_MCCK_PENDING (1<<CIF_MCCK_PENDING)
18#define _CIF_ASCE (1<<CIF_ASCE)
19
20
14#ifndef __ASSEMBLY__ 21#ifndef __ASSEMBLY__
15 22
16#include <linux/linkage.h> 23#include <linux/linkage.h>
@@ -21,6 +28,21 @@
21#include <asm/setup.h> 28#include <asm/setup.h>
22#include <asm/runtime_instr.h> 29#include <asm/runtime_instr.h>
23 30
31static inline void set_cpu_flag(int flag)
32{
33 S390_lowcore.cpu_flags |= (1U << flag);
34}
35
36static inline void clear_cpu_flag(int flag)
37{
38 S390_lowcore.cpu_flags &= ~(1U << flag);
39}
40
41static inline int test_cpu_flag(int flag)
42{
43 return !!(S390_lowcore.cpu_flags & (1U << flag));
44}
45
24/* 46/*
25 * Default implementation of macro that returns current 47 * Default implementation of macro that returns current
26 * instruction pointer ("program counter"). 48 * instruction pointer ("program counter").
diff --git a/arch/s390/include/asm/ptrace.h b/arch/s390/include/asm/ptrace.h
index f4783c0b7b43..1b5300cd6d22 100644
--- a/arch/s390/include/asm/ptrace.h
+++ b/arch/s390/include/asm/ptrace.h
@@ -8,6 +8,12 @@
8 8
9#include <uapi/asm/ptrace.h> 9#include <uapi/asm/ptrace.h>
10 10
11#define PIF_SYSCALL 0 /* inside a system call */
12#define PIF_PER_TRAP 1 /* deliver sigtrap on return to user */
13
14#define _PIF_SYSCALL (1<<PIF_SYSCALL)
15#define _PIF_PER_TRAP (1<<PIF_PER_TRAP)
16
11#ifndef __ASSEMBLY__ 17#ifndef __ASSEMBLY__
12 18
13#define PSW_KERNEL_BITS (PSW_DEFAULT_KEY | PSW_MASK_BASE | PSW_ASC_HOME | \ 19#define PSW_KERNEL_BITS (PSW_DEFAULT_KEY | PSW_MASK_BASE | PSW_ASC_HOME | \
@@ -29,6 +35,7 @@ struct pt_regs
29 unsigned int int_code; 35 unsigned int int_code;
30 unsigned int int_parm; 36 unsigned int int_parm;
31 unsigned long int_parm_long; 37 unsigned long int_parm_long;
38 unsigned long flags;
32}; 39};
33 40
34/* 41/*
@@ -79,6 +86,21 @@ struct per_struct_kernel {
79#define PER_CONTROL_SUSPENSION 0x00400000UL 86#define PER_CONTROL_SUSPENSION 0x00400000UL
80#define PER_CONTROL_ALTERATION 0x00200000UL 87#define PER_CONTROL_ALTERATION 0x00200000UL
81 88
89static inline void set_pt_regs_flag(struct pt_regs *regs, int flag)
90{
91 regs->flags |= (1U << flag);
92}
93
94static inline void clear_pt_regs_flag(struct pt_regs *regs, int flag)
95{
96 regs->flags &= ~(1U << flag);
97}
98
99static inline int test_pt_regs_flag(struct pt_regs *regs, int flag)
100{
101 return !!(regs->flags & (1U << flag));
102}
103
82/* 104/*
83 * These are defined as per linux/ptrace.h, which see. 105 * These are defined as per linux/ptrace.h, which see.
84 */ 106 */
diff --git a/arch/s390/include/asm/setup.h b/arch/s390/include/asm/setup.h
index b31b22dba948..089a49814c50 100644
--- a/arch/s390/include/asm/setup.h
+++ b/arch/s390/include/asm/setup.h
@@ -9,7 +9,6 @@
9 9
10 10
11#define PARMAREA 0x10400 11#define PARMAREA 0x10400
12#define MEMORY_CHUNKS 256
13 12
14#ifndef __ASSEMBLY__ 13#ifndef __ASSEMBLY__
15 14
@@ -31,22 +30,11 @@
31#endif /* CONFIG_64BIT */ 30#endif /* CONFIG_64BIT */
32#define COMMAND_LINE ((char *) (0x10480)) 31#define COMMAND_LINE ((char *) (0x10480))
33 32
34#define CHUNK_READ_WRITE 0
35#define CHUNK_READ_ONLY 1
36
37struct mem_chunk {
38 unsigned long addr;
39 unsigned long size;
40 int type;
41};
42
43extern struct mem_chunk memory_chunk[];
44extern int memory_end_set; 33extern int memory_end_set;
45extern unsigned long memory_end; 34extern unsigned long memory_end;
35extern unsigned long max_physmem_end;
46 36
47void detect_memory_layout(struct mem_chunk chunk[], unsigned long maxsize); 37extern void detect_memory_memblock(void);
48void create_mem_hole(struct mem_chunk mem_chunk[], unsigned long addr,
49 unsigned long size);
50 38
51/* 39/*
52 * Machine features detected in head.S 40 * Machine features detected in head.S
diff --git a/arch/s390/include/asm/smp.h b/arch/s390/include/asm/smp.h
index 21703f85b48d..4f1307962a95 100644
--- a/arch/s390/include/asm/smp.h
+++ b/arch/s390/include/asm/smp.h
@@ -30,7 +30,6 @@ extern int smp_store_status(int cpu);
30extern int smp_vcpu_scheduled(int cpu); 30extern int smp_vcpu_scheduled(int cpu);
31extern void smp_yield_cpu(int cpu); 31extern void smp_yield_cpu(int cpu);
32extern void smp_yield(void); 32extern void smp_yield(void);
33extern void smp_stop_cpu(void);
34extern void smp_cpu_set_polarization(int cpu, int val); 33extern void smp_cpu_set_polarization(int cpu, int val);
35extern int smp_cpu_get_polarization(int cpu); 34extern int smp_cpu_get_polarization(int cpu);
36extern void smp_fill_possible_mask(void); 35extern void smp_fill_possible_mask(void);
@@ -54,6 +53,8 @@ static inline void smp_yield_cpu(int cpu) { }
54static inline void smp_yield(void) { } 53static inline void smp_yield(void) { }
55static inline void smp_fill_possible_mask(void) { } 54static inline void smp_fill_possible_mask(void) { }
56 55
56#endif /* CONFIG_SMP */
57
57static inline void smp_stop_cpu(void) 58static inline void smp_stop_cpu(void)
58{ 59{
59 u16 pcpu = stap(); 60 u16 pcpu = stap();
@@ -64,8 +65,6 @@ static inline void smp_stop_cpu(void)
64 } 65 }
65} 66}
66 67
67#endif /* CONFIG_SMP */
68
69#ifdef CONFIG_HOTPLUG_CPU 68#ifdef CONFIG_HOTPLUG_CPU
70extern int smp_rescan_cpus(void); 69extern int smp_rescan_cpus(void);
71extern void __noreturn cpu_die(void); 70extern void __noreturn cpu_die(void);
diff --git a/arch/s390/include/asm/spinlock.h b/arch/s390/include/asm/spinlock.h
index 83e5d216105e..96879f7ad6da 100644
--- a/arch/s390/include/asm/spinlock.h
+++ b/arch/s390/include/asm/spinlock.h
@@ -11,18 +11,21 @@
11 11
12#include <linux/smp.h> 12#include <linux/smp.h>
13 13
14#define SPINLOCK_LOCKVAL (S390_lowcore.spinlock_lockval)
15
14extern int spin_retry; 16extern int spin_retry;
15 17
16static inline int 18static inline int
17_raw_compare_and_swap(volatile unsigned int *lock, 19_raw_compare_and_swap(unsigned int *lock, unsigned int old, unsigned int new)
18 unsigned int old, unsigned int new)
19{ 20{
21 unsigned int old_expected = old;
22
20 asm volatile( 23 asm volatile(
21 " cs %0,%3,%1" 24 " cs %0,%3,%1"
22 : "=d" (old), "=Q" (*lock) 25 : "=d" (old), "=Q" (*lock)
23 : "0" (old), "d" (new), "Q" (*lock) 26 : "0" (old), "d" (new), "Q" (*lock)
24 : "cc", "memory" ); 27 : "cc", "memory" );
25 return old; 28 return old == old_expected;
26} 29}
27 30
28/* 31/*
@@ -34,57 +37,69 @@ _raw_compare_and_swap(volatile unsigned int *lock,
34 * (the type definitions are in asm/spinlock_types.h) 37 * (the type definitions are in asm/spinlock_types.h)
35 */ 38 */
36 39
37#define arch_spin_is_locked(x) ((x)->owner_cpu != 0) 40void arch_spin_lock_wait(arch_spinlock_t *);
38#define arch_spin_unlock_wait(lock) \ 41int arch_spin_trylock_retry(arch_spinlock_t *);
39 do { while (arch_spin_is_locked(lock)) \ 42void arch_spin_relax(arch_spinlock_t *);
40 arch_spin_relax(lock); } while (0) 43void arch_spin_lock_wait_flags(arch_spinlock_t *, unsigned long flags);
41 44
42extern void arch_spin_lock_wait(arch_spinlock_t *); 45static inline u32 arch_spin_lockval(int cpu)
43extern void arch_spin_lock_wait_flags(arch_spinlock_t *, unsigned long flags); 46{
44extern int arch_spin_trylock_retry(arch_spinlock_t *); 47 return ~cpu;
45extern void arch_spin_relax(arch_spinlock_t *lock); 48}
46 49
47static inline int arch_spin_value_unlocked(arch_spinlock_t lock) 50static inline int arch_spin_value_unlocked(arch_spinlock_t lock)
48{ 51{
49 return lock.owner_cpu == 0; 52 return lock.lock == 0;
50} 53}
51 54
52static inline void arch_spin_lock(arch_spinlock_t *lp) 55static inline int arch_spin_is_locked(arch_spinlock_t *lp)
53{ 56{
54 int old; 57 return ACCESS_ONCE(lp->lock) != 0;
58}
55 59
56 old = _raw_compare_and_swap(&lp->owner_cpu, 0, ~smp_processor_id()); 60static inline int arch_spin_trylock_once(arch_spinlock_t *lp)
57 if (likely(old == 0)) 61{
58 return; 62 barrier();
59 arch_spin_lock_wait(lp); 63 return likely(arch_spin_value_unlocked(*lp) &&
64 _raw_compare_and_swap(&lp->lock, 0, SPINLOCK_LOCKVAL));
60} 65}
61 66
62static inline void arch_spin_lock_flags(arch_spinlock_t *lp, 67static inline int arch_spin_tryrelease_once(arch_spinlock_t *lp)
63 unsigned long flags)
64{ 68{
65 int old; 69 return _raw_compare_and_swap(&lp->lock, SPINLOCK_LOCKVAL, 0);
70}
66 71
67 old = _raw_compare_and_swap(&lp->owner_cpu, 0, ~smp_processor_id()); 72static inline void arch_spin_lock(arch_spinlock_t *lp)
68 if (likely(old == 0)) 73{
69 return; 74 if (!arch_spin_trylock_once(lp))
70 arch_spin_lock_wait_flags(lp, flags); 75 arch_spin_lock_wait(lp);
71} 76}
72 77
73static inline int arch_spin_trylock(arch_spinlock_t *lp) 78static inline void arch_spin_lock_flags(arch_spinlock_t *lp,
79 unsigned long flags)
74{ 80{
75 int old; 81 if (!arch_spin_trylock_once(lp))
82 arch_spin_lock_wait_flags(lp, flags);
83}
76 84
77 old = _raw_compare_and_swap(&lp->owner_cpu, 0, ~smp_processor_id()); 85static inline int arch_spin_trylock(arch_spinlock_t *lp)
78 if (likely(old == 0)) 86{
79 return 1; 87 if (!arch_spin_trylock_once(lp))
80 return arch_spin_trylock_retry(lp); 88 return arch_spin_trylock_retry(lp);
89 return 1;
81} 90}
82 91
83static inline void arch_spin_unlock(arch_spinlock_t *lp) 92static inline void arch_spin_unlock(arch_spinlock_t *lp)
84{ 93{
85 _raw_compare_and_swap(&lp->owner_cpu, lp->owner_cpu, 0); 94 arch_spin_tryrelease_once(lp);
95}
96
97static inline void arch_spin_unlock_wait(arch_spinlock_t *lock)
98{
99 while (arch_spin_is_locked(lock))
100 arch_spin_relax(lock);
86} 101}
87 102
88/* 103/*
89 * Read-write spinlocks, allowing multiple readers 104 * Read-write spinlocks, allowing multiple readers
90 * but only one writer. 105 * but only one writer.
@@ -115,42 +130,50 @@ extern void _raw_write_lock_wait(arch_rwlock_t *lp);
115extern void _raw_write_lock_wait_flags(arch_rwlock_t *lp, unsigned long flags); 130extern void _raw_write_lock_wait_flags(arch_rwlock_t *lp, unsigned long flags);
116extern int _raw_write_trylock_retry(arch_rwlock_t *lp); 131extern int _raw_write_trylock_retry(arch_rwlock_t *lp);
117 132
133static inline int arch_read_trylock_once(arch_rwlock_t *rw)
134{
135 unsigned int old = ACCESS_ONCE(rw->lock);
136 return likely((int) old >= 0 &&
137 _raw_compare_and_swap(&rw->lock, old, old + 1));
138}
139
140static inline int arch_write_trylock_once(arch_rwlock_t *rw)
141{
142 unsigned int old = ACCESS_ONCE(rw->lock);
143 return likely(old == 0 &&
144 _raw_compare_and_swap(&rw->lock, 0, 0x80000000));
145}
146
118static inline void arch_read_lock(arch_rwlock_t *rw) 147static inline void arch_read_lock(arch_rwlock_t *rw)
119{ 148{
120 unsigned int old; 149 if (!arch_read_trylock_once(rw))
121 old = rw->lock & 0x7fffffffU;
122 if (_raw_compare_and_swap(&rw->lock, old, old + 1) != old)
123 _raw_read_lock_wait(rw); 150 _raw_read_lock_wait(rw);
124} 151}
125 152
126static inline void arch_read_lock_flags(arch_rwlock_t *rw, unsigned long flags) 153static inline void arch_read_lock_flags(arch_rwlock_t *rw, unsigned long flags)
127{ 154{
128 unsigned int old; 155 if (!arch_read_trylock_once(rw))
129 old = rw->lock & 0x7fffffffU;
130 if (_raw_compare_and_swap(&rw->lock, old, old + 1) != old)
131 _raw_read_lock_wait_flags(rw, flags); 156 _raw_read_lock_wait_flags(rw, flags);
132} 157}
133 158
134static inline void arch_read_unlock(arch_rwlock_t *rw) 159static inline void arch_read_unlock(arch_rwlock_t *rw)
135{ 160{
136 unsigned int old, cmp; 161 unsigned int old;
137 162
138 old = rw->lock;
139 do { 163 do {
140 cmp = old; 164 old = ACCESS_ONCE(rw->lock);
141 old = _raw_compare_and_swap(&rw->lock, old, old - 1); 165 } while (!_raw_compare_and_swap(&rw->lock, old, old - 1));
142 } while (cmp != old);
143} 166}
144 167
145static inline void arch_write_lock(arch_rwlock_t *rw) 168static inline void arch_write_lock(arch_rwlock_t *rw)
146{ 169{
147 if (unlikely(_raw_compare_and_swap(&rw->lock, 0, 0x80000000) != 0)) 170 if (!arch_write_trylock_once(rw))
148 _raw_write_lock_wait(rw); 171 _raw_write_lock_wait(rw);
149} 172}
150 173
151static inline void arch_write_lock_flags(arch_rwlock_t *rw, unsigned long flags) 174static inline void arch_write_lock_flags(arch_rwlock_t *rw, unsigned long flags)
152{ 175{
153 if (unlikely(_raw_compare_and_swap(&rw->lock, 0, 0x80000000) != 0)) 176 if (!arch_write_trylock_once(rw))
154 _raw_write_lock_wait_flags(rw, flags); 177 _raw_write_lock_wait_flags(rw, flags);
155} 178}
156 179
@@ -161,18 +184,16 @@ static inline void arch_write_unlock(arch_rwlock_t *rw)
161 184
162static inline int arch_read_trylock(arch_rwlock_t *rw) 185static inline int arch_read_trylock(arch_rwlock_t *rw)
163{ 186{
164 unsigned int old; 187 if (!arch_read_trylock_once(rw))
165 old = rw->lock & 0x7fffffffU; 188 return _raw_read_trylock_retry(rw);
166 if (likely(_raw_compare_and_swap(&rw->lock, old, old + 1) == old)) 189 return 1;
167 return 1;
168 return _raw_read_trylock_retry(rw);
169} 190}
170 191
171static inline int arch_write_trylock(arch_rwlock_t *rw) 192static inline int arch_write_trylock(arch_rwlock_t *rw)
172{ 193{
173 if (likely(_raw_compare_and_swap(&rw->lock, 0, 0x80000000) == 0)) 194 if (!arch_write_trylock_once(rw))
174 return 1; 195 return _raw_write_trylock_retry(rw);
175 return _raw_write_trylock_retry(rw); 196 return 1;
176} 197}
177 198
178#define arch_read_relax(lock) cpu_relax() 199#define arch_read_relax(lock) cpu_relax()
diff --git a/arch/s390/include/asm/spinlock_types.h b/arch/s390/include/asm/spinlock_types.h
index 9c76656a0af0..b2cd6ff7c2c5 100644
--- a/arch/s390/include/asm/spinlock_types.h
+++ b/arch/s390/include/asm/spinlock_types.h
@@ -6,13 +6,13 @@
6#endif 6#endif
7 7
8typedef struct { 8typedef struct {
9 volatile unsigned int owner_cpu; 9 unsigned int lock;
10} __attribute__ ((aligned (4))) arch_spinlock_t; 10} __attribute__ ((aligned (4))) arch_spinlock_t;
11 11
12#define __ARCH_SPIN_LOCK_UNLOCKED { 0 } 12#define __ARCH_SPIN_LOCK_UNLOCKED { .lock = 0, }
13 13
14typedef struct { 14typedef struct {
15 volatile unsigned int lock; 15 unsigned int lock;
16} arch_rwlock_t; 16} arch_rwlock_t;
17 17
18#define __ARCH_RW_LOCK_UNLOCKED { 0 } 18#define __ARCH_RW_LOCK_UNLOCKED { 0 }
diff --git a/arch/s390/include/asm/switch_to.h b/arch/s390/include/asm/switch_to.h
index e759181357fc..29c81f82705e 100644
--- a/arch/s390/include/asm/switch_to.h
+++ b/arch/s390/include/asm/switch_to.h
@@ -132,7 +132,6 @@ static inline void restore_access_regs(unsigned int *acrs)
132 update_cr_regs(next); \ 132 update_cr_regs(next); \
133 } \ 133 } \
134 prev = __switch_to(prev,next); \ 134 prev = __switch_to(prev,next); \
135 update_primary_asce(current); \
136} while (0) 135} while (0)
137 136
138#define finish_arch_switch(prev) do { \ 137#define finish_arch_switch(prev) do { \
diff --git a/arch/s390/include/asm/syscall.h b/arch/s390/include/asm/syscall.h
index 777687055e7b..abad78d5b10c 100644
--- a/arch/s390/include/asm/syscall.h
+++ b/arch/s390/include/asm/syscall.h
@@ -28,7 +28,7 @@ extern const unsigned int sys_call_table_emu[];
28static inline long syscall_get_nr(struct task_struct *task, 28static inline long syscall_get_nr(struct task_struct *task,
29 struct pt_regs *regs) 29 struct pt_regs *regs)
30{ 30{
31 return test_tsk_thread_flag(task, TIF_SYSCALL) ? 31 return test_pt_regs_flag(regs, PIF_SYSCALL) ?
32 (regs->int_code & 0xffff) : -1; 32 (regs->int_code & 0xffff) : -1;
33} 33}
34 34
diff --git a/arch/s390/include/asm/thread_info.h b/arch/s390/include/asm/thread_info.h
index 50630e6a35de..b833e9c0bfbf 100644
--- a/arch/s390/include/asm/thread_info.h
+++ b/arch/s390/include/asm/thread_info.h
@@ -77,32 +77,22 @@ static inline struct thread_info *current_thread_info(void)
77/* 77/*
78 * thread information flags bit numbers 78 * thread information flags bit numbers
79 */ 79 */
80#define TIF_SYSCALL 0 /* inside a system call */ 80#define TIF_NOTIFY_RESUME 0 /* callback before returning to user */
81#define TIF_NOTIFY_RESUME 1 /* callback before returning to user */ 81#define TIF_SIGPENDING 1 /* signal pending */
82#define TIF_SIGPENDING 2 /* signal pending */ 82#define TIF_NEED_RESCHED 2 /* rescheduling necessary */
83#define TIF_NEED_RESCHED 3 /* rescheduling necessary */ 83#define TIF_SYSCALL_TRACE 3 /* syscall trace active */
84#define TIF_TLB_WAIT 4 /* wait for TLB flush completion */ 84#define TIF_SYSCALL_AUDIT 4 /* syscall auditing active */
85#define TIF_ASCE 5 /* primary asce needs fixup / uaccess */ 85#define TIF_SECCOMP 5 /* secure computing */
86#define TIF_PER_TRAP 6 /* deliver sigtrap on return to user */ 86#define TIF_SYSCALL_TRACEPOINT 6 /* syscall tracepoint instrumentation */
87#define TIF_MCCK_PENDING 7 /* machine check handling is pending */ 87#define TIF_31BIT 16 /* 32bit process */
88#define TIF_SYSCALL_TRACE 8 /* syscall trace active */ 88#define TIF_MEMDIE 17 /* is terminating due to OOM killer */
89#define TIF_SYSCALL_AUDIT 9 /* syscall auditing active */ 89#define TIF_RESTORE_SIGMASK 18 /* restore signal mask in do_signal() */
90#define TIF_SECCOMP 10 /* secure computing */ 90#define TIF_SINGLE_STEP 19 /* This task is single stepped */
91#define TIF_SYSCALL_TRACEPOINT 11 /* syscall tracepoint instrumentation */ 91#define TIF_BLOCK_STEP 20 /* This task is block stepped */
92#define TIF_31BIT 17 /* 32bit process */
93#define TIF_MEMDIE 18 /* is terminating due to OOM killer */
94#define TIF_RESTORE_SIGMASK 19 /* restore signal mask in do_signal() */
95#define TIF_SINGLE_STEP 20 /* This task is single stepped */
96#define TIF_BLOCK_STEP 21 /* This task is block stepped */
97 92
98#define _TIF_SYSCALL (1<<TIF_SYSCALL)
99#define _TIF_NOTIFY_RESUME (1<<TIF_NOTIFY_RESUME) 93#define _TIF_NOTIFY_RESUME (1<<TIF_NOTIFY_RESUME)
100#define _TIF_SIGPENDING (1<<TIF_SIGPENDING) 94#define _TIF_SIGPENDING (1<<TIF_SIGPENDING)
101#define _TIF_NEED_RESCHED (1<<TIF_NEED_RESCHED) 95#define _TIF_NEED_RESCHED (1<<TIF_NEED_RESCHED)
102#define _TIF_TLB_WAIT (1<<TIF_TLB_WAIT)
103#define _TIF_ASCE (1<<TIF_ASCE)
104#define _TIF_PER_TRAP (1<<TIF_PER_TRAP)
105#define _TIF_MCCK_PENDING (1<<TIF_MCCK_PENDING)
106#define _TIF_SYSCALL_TRACE (1<<TIF_SYSCALL_TRACE) 96#define _TIF_SYSCALL_TRACE (1<<TIF_SYSCALL_TRACE)
107#define _TIF_SYSCALL_AUDIT (1<<TIF_SYSCALL_AUDIT) 97#define _TIF_SYSCALL_AUDIT (1<<TIF_SYSCALL_AUDIT)
108#define _TIF_SECCOMP (1<<TIF_SECCOMP) 98#define _TIF_SECCOMP (1<<TIF_SECCOMP)
diff --git a/arch/s390/include/asm/uaccess.h b/arch/s390/include/asm/uaccess.h
index 1be64a1506d0..cd4c68e0398d 100644
--- a/arch/s390/include/asm/uaccess.h
+++ b/arch/s390/include/asm/uaccess.h
@@ -132,6 +132,34 @@ unsigned long __must_check __copy_to_user(void __user *to, const void *from,
132#define __copy_to_user_inatomic __copy_to_user 132#define __copy_to_user_inatomic __copy_to_user
133#define __copy_from_user_inatomic __copy_from_user 133#define __copy_from_user_inatomic __copy_from_user
134 134
135#ifdef CONFIG_HAVE_MARCH_Z10_FEATURES
136
137#define __put_get_user_asm(to, from, size, spec) \
138({ \
139 register unsigned long __reg0 asm("0") = spec; \
140 int __rc; \
141 \
142 asm volatile( \
143 "0: mvcos %1,%3,%2\n" \
144 "1: xr %0,%0\n" \
145 "2:\n" \
146 ".pushsection .fixup, \"ax\"\n" \
147 "3: lhi %0,%5\n" \
148 " jg 2b\n" \
149 ".popsection\n" \
150 EX_TABLE(0b,3b) EX_TABLE(1b,3b) \
151 : "=d" (__rc), "=Q" (*(to)) \
152 : "d" (size), "Q" (*(from)), \
153 "d" (__reg0), "K" (-EFAULT) \
154 : "cc"); \
155 __rc; \
156})
157
158#define __put_user_fn(x, ptr, size) __put_get_user_asm(ptr, x, size, 0x810000UL)
159#define __get_user_fn(x, ptr, size) __put_get_user_asm(x, ptr, size, 0x81UL)
160
161#else /* CONFIG_HAVE_MARCH_Z10_FEATURES */
162
135static inline int __put_user_fn(void *x, void __user *ptr, unsigned long size) 163static inline int __put_user_fn(void *x, void __user *ptr, unsigned long size)
136{ 164{
137 size = __copy_to_user(ptr, x, size); 165 size = __copy_to_user(ptr, x, size);
@@ -144,6 +172,8 @@ static inline int __get_user_fn(void *x, const void __user *ptr, unsigned long s
144 return size ? -EFAULT : 0; 172 return size ? -EFAULT : 0;
145} 173}
146 174
175#endif /* CONFIG_HAVE_MARCH_Z10_FEATURES */
176
147/* 177/*
148 * These are the main single-value transfer routines. They automatically 178 * These are the main single-value transfer routines. They automatically
149 * use the right size if we just have the right pointer type. 179 * use the right size if we just have the right pointer type.
diff --git a/arch/s390/kernel/asm-offsets.c b/arch/s390/kernel/asm-offsets.c
index cc10cdd4d6a2..0c070c44cde2 100644
--- a/arch/s390/kernel/asm-offsets.c
+++ b/arch/s390/kernel/asm-offsets.c
@@ -50,6 +50,7 @@ int main(void)
50 DEFINE(__PT_INT_CODE, offsetof(struct pt_regs, int_code)); 50 DEFINE(__PT_INT_CODE, offsetof(struct pt_regs, int_code));
51 DEFINE(__PT_INT_PARM, offsetof(struct pt_regs, int_parm)); 51 DEFINE(__PT_INT_PARM, offsetof(struct pt_regs, int_parm));
52 DEFINE(__PT_INT_PARM_LONG, offsetof(struct pt_regs, int_parm_long)); 52 DEFINE(__PT_INT_PARM_LONG, offsetof(struct pt_regs, int_parm_long));
53 DEFINE(__PT_FLAGS, offsetof(struct pt_regs, flags));
53 DEFINE(__PT_SIZE, sizeof(struct pt_regs)); 54 DEFINE(__PT_SIZE, sizeof(struct pt_regs));
54 BLANK(); 55 BLANK();
55 DEFINE(__SF_BACKCHAIN, offsetof(struct stack_frame, back_chain)); 56 DEFINE(__SF_BACKCHAIN, offsetof(struct stack_frame, back_chain));
@@ -115,6 +116,7 @@ int main(void)
115 DEFINE(__LC_SAVE_AREA_SYNC, offsetof(struct _lowcore, save_area_sync)); 116 DEFINE(__LC_SAVE_AREA_SYNC, offsetof(struct _lowcore, save_area_sync));
116 DEFINE(__LC_SAVE_AREA_ASYNC, offsetof(struct _lowcore, save_area_async)); 117 DEFINE(__LC_SAVE_AREA_ASYNC, offsetof(struct _lowcore, save_area_async));
117 DEFINE(__LC_SAVE_AREA_RESTART, offsetof(struct _lowcore, save_area_restart)); 118 DEFINE(__LC_SAVE_AREA_RESTART, offsetof(struct _lowcore, save_area_restart));
119 DEFINE(__LC_CPU_FLAGS, offsetof(struct _lowcore, cpu_flags));
118 DEFINE(__LC_RETURN_PSW, offsetof(struct _lowcore, return_psw)); 120 DEFINE(__LC_RETURN_PSW, offsetof(struct _lowcore, return_psw));
119 DEFINE(__LC_RETURN_MCCK_PSW, offsetof(struct _lowcore, return_mcck_psw)); 121 DEFINE(__LC_RETURN_MCCK_PSW, offsetof(struct _lowcore, return_mcck_psw));
120 DEFINE(__LC_SYNC_ENTER_TIMER, offsetof(struct _lowcore, sync_enter_timer)); 122 DEFINE(__LC_SYNC_ENTER_TIMER, offsetof(struct _lowcore, sync_enter_timer));
@@ -142,7 +144,6 @@ int main(void)
142 DEFINE(__LC_MCCK_CLOCK, offsetof(struct _lowcore, mcck_clock)); 144 DEFINE(__LC_MCCK_CLOCK, offsetof(struct _lowcore, mcck_clock));
143 DEFINE(__LC_MACHINE_FLAGS, offsetof(struct _lowcore, machine_flags)); 145 DEFINE(__LC_MACHINE_FLAGS, offsetof(struct _lowcore, machine_flags));
144 DEFINE(__LC_FTRACE_FUNC, offsetof(struct _lowcore, ftrace_func)); 146 DEFINE(__LC_FTRACE_FUNC, offsetof(struct _lowcore, ftrace_func));
145 DEFINE(__LC_IRB, offsetof(struct _lowcore, irb));
146 DEFINE(__LC_DUMP_REIPL, offsetof(struct _lowcore, ipib)); 147 DEFINE(__LC_DUMP_REIPL, offsetof(struct _lowcore, ipib));
147 BLANK(); 148 BLANK();
148 DEFINE(__LC_CPU_TIMER_SAVE_AREA, offsetof(struct _lowcore, cpu_timer_save_area)); 149 DEFINE(__LC_CPU_TIMER_SAVE_AREA, offsetof(struct _lowcore, cpu_timer_save_area));
diff --git a/arch/s390/kernel/compat_signal.c b/arch/s390/kernel/compat_signal.c
index 7df5ed9f44d7..f204d6920368 100644
--- a/arch/s390/kernel/compat_signal.c
+++ b/arch/s390/kernel/compat_signal.c
@@ -213,7 +213,7 @@ static int restore_sigregs32(struct pt_regs *regs,_sigregs32 __user *sregs)
213 sizeof(current->thread.fp_regs)); 213 sizeof(current->thread.fp_regs));
214 214
215 restore_fp_regs(current->thread.fp_regs.fprs); 215 restore_fp_regs(current->thread.fp_regs.fprs);
216 clear_thread_flag(TIF_SYSCALL); /* No longer in a system call */ 216 clear_pt_regs_flag(regs, PIF_SYSCALL); /* No longer in a system call */
217 return 0; 217 return 0;
218} 218}
219 219
diff --git a/arch/s390/kernel/crash_dump.c b/arch/s390/kernel/crash_dump.c
index d7658c4b2ed5..a3b9150e6802 100644
--- a/arch/s390/kernel/crash_dump.c
+++ b/arch/s390/kernel/crash_dump.c
@@ -13,6 +13,7 @@
13#include <linux/slab.h> 13#include <linux/slab.h>
14#include <linux/bootmem.h> 14#include <linux/bootmem.h>
15#include <linux/elf.h> 15#include <linux/elf.h>
16#include <linux/memblock.h>
16#include <asm/os_info.h> 17#include <asm/os_info.h>
17#include <asm/elf.h> 18#include <asm/elf.h>
18#include <asm/ipl.h> 19#include <asm/ipl.h>
@@ -22,6 +23,24 @@
22#define PTR_SUB(x, y) (((char *) (x)) - ((unsigned long) (y))) 23#define PTR_SUB(x, y) (((char *) (x)) - ((unsigned long) (y)))
23#define PTR_DIFF(x, y) ((unsigned long)(((char *) (x)) - ((unsigned long) (y)))) 24#define PTR_DIFF(x, y) ((unsigned long)(((char *) (x)) - ((unsigned long) (y))))
24 25
26static struct memblock_region oldmem_region;
27
28static struct memblock_type oldmem_type = {
29 .cnt = 1,
30 .max = 1,
31 .total_size = 0,
32 .regions = &oldmem_region,
33};
34
35#define for_each_dump_mem_range(i, nid, p_start, p_end, p_nid) \
36 for (i = 0, __next_mem_range(&i, nid, &memblock.physmem, \
37 &oldmem_type, p_start, \
38 p_end, p_nid); \
39 i != (u64)ULLONG_MAX; \
40 __next_mem_range(&i, nid, &memblock.physmem, \
41 &oldmem_type, \
42 p_start, p_end, p_nid))
43
25struct dump_save_areas dump_save_areas; 44struct dump_save_areas dump_save_areas;
26 45
27/* 46/*
@@ -264,19 +283,6 @@ static void *kzalloc_panic(int len)
264} 283}
265 284
266/* 285/*
267 * Get memory layout and create hole for oldmem
268 */
269static struct mem_chunk *get_memory_layout(void)
270{
271 struct mem_chunk *chunk_array;
272
273 chunk_array = kzalloc_panic(MEMORY_CHUNKS * sizeof(struct mem_chunk));
274 detect_memory_layout(chunk_array, 0);
275 create_mem_hole(chunk_array, OLDMEM_BASE, OLDMEM_SIZE);
276 return chunk_array;
277}
278
279/*
280 * Initialize ELF note 286 * Initialize ELF note
281 */ 287 */
282static void *nt_init(void *buf, Elf64_Word type, void *desc, int d_len, 288static void *nt_init(void *buf, Elf64_Word type, void *desc, int d_len,
@@ -490,52 +496,33 @@ static int get_cpu_cnt(void)
490 */ 496 */
491static int get_mem_chunk_cnt(void) 497static int get_mem_chunk_cnt(void)
492{ 498{
493 struct mem_chunk *chunk_array, *mem_chunk; 499 int cnt = 0;
494 int i, cnt = 0; 500 u64 idx;
495 501
496 chunk_array = get_memory_layout(); 502 for_each_dump_mem_range(idx, NUMA_NO_NODE, NULL, NULL, NULL)
497 for (i = 0; i < MEMORY_CHUNKS; i++) {
498 mem_chunk = &chunk_array[i];
499 if (chunk_array[i].type != CHUNK_READ_WRITE &&
500 chunk_array[i].type != CHUNK_READ_ONLY)
501 continue;
502 if (mem_chunk->size == 0)
503 continue;
504 cnt++; 503 cnt++;
505 }
506 kfree(chunk_array);
507 return cnt; 504 return cnt;
508} 505}
509 506
510/* 507/*
511 * Initialize ELF loads (new kernel) 508 * Initialize ELF loads (new kernel)
512 */ 509 */
513static int loads_init(Elf64_Phdr *phdr, u64 loads_offset) 510static void loads_init(Elf64_Phdr *phdr, u64 loads_offset)
514{ 511{
515 struct mem_chunk *chunk_array, *mem_chunk; 512 phys_addr_t start, end;
516 int i; 513 u64 idx;
517 514
518 chunk_array = get_memory_layout(); 515 for_each_dump_mem_range(idx, NUMA_NO_NODE, &start, &end, NULL) {
519 for (i = 0; i < MEMORY_CHUNKS; i++) { 516 phdr->p_filesz = end - start;
520 mem_chunk = &chunk_array[i];
521 if (mem_chunk->size == 0)
522 continue;
523 if (chunk_array[i].type != CHUNK_READ_WRITE &&
524 chunk_array[i].type != CHUNK_READ_ONLY)
525 continue;
526 else
527 phdr->p_filesz = mem_chunk->size;
528 phdr->p_type = PT_LOAD; 517 phdr->p_type = PT_LOAD;
529 phdr->p_offset = mem_chunk->addr; 518 phdr->p_offset = start;
530 phdr->p_vaddr = mem_chunk->addr; 519 phdr->p_vaddr = start;
531 phdr->p_paddr = mem_chunk->addr; 520 phdr->p_paddr = start;
532 phdr->p_memsz = mem_chunk->size; 521 phdr->p_memsz = end - start;
533 phdr->p_flags = PF_R | PF_W | PF_X; 522 phdr->p_flags = PF_R | PF_W | PF_X;
534 phdr->p_align = PAGE_SIZE; 523 phdr->p_align = PAGE_SIZE;
535 phdr++; 524 phdr++;
536 } 525 }
537 kfree(chunk_array);
538 return i;
539} 526}
540 527
541/* 528/*
@@ -584,6 +571,14 @@ int elfcorehdr_alloc(unsigned long long *addr, unsigned long long *size)
584 /* If we cannot get HSA size for zfcpdump return error */ 571 /* If we cannot get HSA size for zfcpdump return error */
585 if (ipl_info.type == IPL_TYPE_FCP_DUMP && !sclp_get_hsa_size()) 572 if (ipl_info.type == IPL_TYPE_FCP_DUMP && !sclp_get_hsa_size())
586 return -ENODEV; 573 return -ENODEV;
574
575 /* For kdump, exclude previous crashkernel memory */
576 if (OLDMEM_BASE) {
577 oldmem_region.base = OLDMEM_BASE;
578 oldmem_region.size = OLDMEM_SIZE;
579 oldmem_type.total_size = OLDMEM_SIZE;
580 }
581
587 mem_chunk_cnt = get_mem_chunk_cnt(); 582 mem_chunk_cnt = get_mem_chunk_cnt();
588 583
589 alloc_size = 0x1000 + get_cpu_cnt() * 0x300 + 584 alloc_size = 0x1000 + get_cpu_cnt() * 0x300 +
diff --git a/arch/s390/kernel/early.c b/arch/s390/kernel/early.c
index a734f3585ceb..0dff972a169c 100644
--- a/arch/s390/kernel/early.c
+++ b/arch/s390/kernel/early.c
@@ -258,13 +258,19 @@ static __init void setup_topology(void)
258static void early_pgm_check_handler(void) 258static void early_pgm_check_handler(void)
259{ 259{
260 const struct exception_table_entry *fixup; 260 const struct exception_table_entry *fixup;
261 unsigned long cr0, cr0_new;
261 unsigned long addr; 262 unsigned long addr;
262 263
263 addr = S390_lowcore.program_old_psw.addr; 264 addr = S390_lowcore.program_old_psw.addr;
264 fixup = search_exception_tables(addr & PSW_ADDR_INSN); 265 fixup = search_exception_tables(addr & PSW_ADDR_INSN);
265 if (!fixup) 266 if (!fixup)
266 disabled_wait(0); 267 disabled_wait(0);
268 /* Disable low address protection before storing into lowcore. */
269 __ctl_store(cr0, 0, 0);
270 cr0_new = cr0 & ~(1UL << 28);
271 __ctl_load(cr0_new, 0, 0);
267 S390_lowcore.program_old_psw.addr = extable_fixup(fixup)|PSW_ADDR_AMODE; 272 S390_lowcore.program_old_psw.addr = extable_fixup(fixup)|PSW_ADDR_AMODE;
273 __ctl_load(cr0, 0, 0);
268} 274}
269 275
270static noinline __init void setup_lowcore_early(void) 276static noinline __init void setup_lowcore_early(void)
diff --git a/arch/s390/kernel/entry.S b/arch/s390/kernel/entry.S
index 1662038516c0..18e5af848f9a 100644
--- a/arch/s390/kernel/entry.S
+++ b/arch/s390/kernel/entry.S
@@ -10,6 +10,7 @@
10 10
11#include <linux/init.h> 11#include <linux/init.h>
12#include <linux/linkage.h> 12#include <linux/linkage.h>
13#include <asm/processor.h>
13#include <asm/cache.h> 14#include <asm/cache.h>
14#include <asm/errno.h> 15#include <asm/errno.h>
15#include <asm/ptrace.h> 16#include <asm/ptrace.h>
@@ -37,18 +38,16 @@ __PT_R13 = __PT_GPRS + 524
37__PT_R14 = __PT_GPRS + 56 38__PT_R14 = __PT_GPRS + 56
38__PT_R15 = __PT_GPRS + 60 39__PT_R15 = __PT_GPRS + 60
39 40
40_TIF_WORK_SVC = (_TIF_SIGPENDING | _TIF_NOTIFY_RESUME | _TIF_NEED_RESCHED | \
41 _TIF_MCCK_PENDING | _TIF_PER_TRAP | _TIF_ASCE)
42_TIF_WORK_INT = (_TIF_SIGPENDING | _TIF_NOTIFY_RESUME | _TIF_NEED_RESCHED | \
43 _TIF_MCCK_PENDING | _TIF_ASCE)
44_TIF_TRACE = (_TIF_SYSCALL_TRACE | _TIF_SYSCALL_AUDIT | _TIF_SECCOMP | \
45 _TIF_SYSCALL_TRACEPOINT)
46_TIF_TRANSFER = (_TIF_MCCK_PENDING | _TIF_TLB_WAIT)
47
48STACK_SHIFT = PAGE_SHIFT + THREAD_ORDER 41STACK_SHIFT = PAGE_SHIFT + THREAD_ORDER
49STACK_SIZE = 1 << STACK_SHIFT 42STACK_SIZE = 1 << STACK_SHIFT
50STACK_INIT = STACK_SIZE - STACK_FRAME_OVERHEAD - __PT_SIZE 43STACK_INIT = STACK_SIZE - STACK_FRAME_OVERHEAD - __PT_SIZE
51 44
45_TIF_WORK = (_TIF_SIGPENDING | _TIF_NOTIFY_RESUME | _TIF_NEED_RESCHED)
46_TIF_TRACE = (_TIF_SYSCALL_TRACE | _TIF_SYSCALL_AUDIT | _TIF_SECCOMP | \
47 _TIF_SYSCALL_TRACEPOINT)
48_CIF_WORK = (_CIF_MCCK_PENDING | _CIF_ASCE)
49_PIF_WORK = (_PIF_PER_TRAP)
50
52#define BASED(name) name-system_call(%r13) 51#define BASED(name) name-system_call(%r13)
53 52
54 .macro TRACE_IRQS_ON 53 .macro TRACE_IRQS_ON
@@ -160,13 +159,7 @@ ENTRY(__switch_to)
160 lctl %c4,%c4,__TASK_pid(%r3) # load pid to control reg. 4 159 lctl %c4,%c4,__TASK_pid(%r3) # load pid to control reg. 4
161 mvc __LC_CURRENT_PID(4,%r0),__TASK_pid(%r3) # store pid of next 160 mvc __LC_CURRENT_PID(4,%r0),__TASK_pid(%r3) # store pid of next
162 l %r15,__THREAD_ksp(%r3) # load kernel stack of next 161 l %r15,__THREAD_ksp(%r3) # load kernel stack of next
163 lhi %r6,_TIF_TRANSFER # transfer TIF bits 162 lm %r6,%r15,__SF_GPRS(%r15) # load gprs of next task
164 n %r6,__TI_flags(%r4) # isolate TIF bits
165 jz 0f
166 o %r6,__TI_flags(%r5) # set TIF bits of next
167 st %r6,__TI_flags(%r5)
168 ni __TI_flags+3(%r4),255-_TIF_TRANSFER # clear TIF bits of prev
1690: lm %r6,%r15,__SF_GPRS(%r15) # load gprs of next task
170 br %r14 163 br %r14
171 164
172__critical_start: 165__critical_start:
@@ -181,6 +174,7 @@ sysc_stm:
181 stm %r8,%r15,__LC_SAVE_AREA_SYNC 174 stm %r8,%r15,__LC_SAVE_AREA_SYNC
182 l %r12,__LC_THREAD_INFO 175 l %r12,__LC_THREAD_INFO
183 l %r13,__LC_SVC_NEW_PSW+4 176 l %r13,__LC_SVC_NEW_PSW+4
177 lhi %r14,_PIF_SYSCALL
184sysc_per: 178sysc_per:
185 l %r15,__LC_KERNEL_STACK 179 l %r15,__LC_KERNEL_STACK
186 la %r11,STACK_FRAME_OVERHEAD(%r15) # pointer to pt_regs 180 la %r11,STACK_FRAME_OVERHEAD(%r15) # pointer to pt_regs
@@ -190,8 +184,8 @@ sysc_vtime:
190 mvc __PT_R8(32,%r11),__LC_SAVE_AREA_SYNC 184 mvc __PT_R8(32,%r11),__LC_SAVE_AREA_SYNC
191 mvc __PT_PSW(8,%r11),__LC_SVC_OLD_PSW 185 mvc __PT_PSW(8,%r11),__LC_SVC_OLD_PSW
192 mvc __PT_INT_CODE(4,%r11),__LC_SVC_ILC 186 mvc __PT_INT_CODE(4,%r11),__LC_SVC_ILC
187 st %r14,__PT_FLAGS(%r11)
193sysc_do_svc: 188sysc_do_svc:
194 oi __TI_flags+3(%r12),_TIF_SYSCALL
195 l %r10,__TI_sysc_table(%r12) # 31 bit system call table 189 l %r10,__TI_sysc_table(%r12) # 31 bit system call table
196 lh %r8,__PT_INT_CODE+2(%r11) 190 lh %r8,__PT_INT_CODE+2(%r11)
197 sla %r8,2 # shift and test for svc0 191 sla %r8,2 # shift and test for svc0
@@ -207,7 +201,7 @@ sysc_nr_ok:
207 st %r2,__PT_ORIG_GPR2(%r11) 201 st %r2,__PT_ORIG_GPR2(%r11)
208 st %r7,STACK_FRAME_OVERHEAD(%r15) 202 st %r7,STACK_FRAME_OVERHEAD(%r15)
209 l %r9,0(%r8,%r10) # get system call addr. 203 l %r9,0(%r8,%r10) # get system call addr.
210 tm __TI_flags+2(%r12),_TIF_TRACE >> 8 204 tm __TI_flags+3(%r12),_TIF_TRACE
211 jnz sysc_tracesys 205 jnz sysc_tracesys
212 basr %r14,%r9 # call sys_xxxx 206 basr %r14,%r9 # call sys_xxxx
213 st %r2,__PT_R2(%r11) # store return value 207 st %r2,__PT_R2(%r11) # store return value
@@ -217,9 +211,12 @@ sysc_return:
217sysc_tif: 211sysc_tif:
218 tm __PT_PSW+1(%r11),0x01 # returning to user ? 212 tm __PT_PSW+1(%r11),0x01 # returning to user ?
219 jno sysc_restore 213 jno sysc_restore
220 tm __TI_flags+3(%r12),_TIF_WORK_SVC 214 tm __PT_FLAGS+3(%r11),_PIF_WORK
221 jnz sysc_work # check for work 215 jnz sysc_work
222 ni __TI_flags+3(%r12),255-_TIF_SYSCALL 216 tm __TI_flags+3(%r12),_TIF_WORK
217 jnz sysc_work # check for thread work
218 tm __LC_CPU_FLAGS+3,_CIF_WORK
219 jnz sysc_work
223sysc_restore: 220sysc_restore:
224 mvc __LC_RETURN_PSW(8),__PT_PSW(%r11) 221 mvc __LC_RETURN_PSW(8),__PT_PSW(%r11)
225 stpt __LC_EXIT_TIMER 222 stpt __LC_EXIT_TIMER
@@ -231,17 +228,17 @@ sysc_done:
231# One of the work bits is on. Find out which one. 228# One of the work bits is on. Find out which one.
232# 229#
233sysc_work: 230sysc_work:
234 tm __TI_flags+3(%r12),_TIF_MCCK_PENDING 231 tm __LC_CPU_FLAGS+3,_CIF_MCCK_PENDING
235 jo sysc_mcck_pending 232 jo sysc_mcck_pending
236 tm __TI_flags+3(%r12),_TIF_NEED_RESCHED 233 tm __TI_flags+3(%r12),_TIF_NEED_RESCHED
237 jo sysc_reschedule 234 jo sysc_reschedule
238 tm __TI_flags+3(%r12),_TIF_PER_TRAP 235 tm __PT_FLAGS+3(%r11),_PIF_PER_TRAP
239 jo sysc_singlestep 236 jo sysc_singlestep
240 tm __TI_flags+3(%r12),_TIF_SIGPENDING 237 tm __TI_flags+3(%r12),_TIF_SIGPENDING
241 jo sysc_sigpending 238 jo sysc_sigpending
242 tm __TI_flags+3(%r12),_TIF_NOTIFY_RESUME 239 tm __TI_flags+3(%r12),_TIF_NOTIFY_RESUME
243 jo sysc_notify_resume 240 jo sysc_notify_resume
244 tm __TI_flags+3(%r12),_TIF_ASCE 241 tm __LC_CPU_FLAGS+3,_CIF_ASCE
245 jo sysc_uaccess 242 jo sysc_uaccess
246 j sysc_return # beware of critical section cleanup 243 j sysc_return # beware of critical section cleanup
247 244
@@ -254,7 +251,7 @@ sysc_reschedule:
254 br %r1 # call schedule 251 br %r1 # call schedule
255 252
256# 253#
257# _TIF_MCCK_PENDING is set, call handler 254# _CIF_MCCK_PENDING is set, call handler
258# 255#
259sysc_mcck_pending: 256sysc_mcck_pending:
260 l %r1,BASED(.Lhandle_mcck) 257 l %r1,BASED(.Lhandle_mcck)
@@ -262,10 +259,10 @@ sysc_mcck_pending:
262 br %r1 # TIF bit will be cleared by handler 259 br %r1 # TIF bit will be cleared by handler
263 260
264# 261#
265# _TIF_ASCE is set, load user space asce 262# _CIF_ASCE is set, load user space asce
266# 263#
267sysc_uaccess: 264sysc_uaccess:
268 ni __TI_flags+3(%r12),255-_TIF_ASCE 265 ni __LC_CPU_FLAGS+3,255-_CIF_ASCE
269 lctl %c1,%c1,__LC_USER_ASCE # load primary asce 266 lctl %c1,%c1,__LC_USER_ASCE # load primary asce
270 j sysc_return 267 j sysc_return
271 268
@@ -276,7 +273,7 @@ sysc_sigpending:
276 lr %r2,%r11 # pass pointer to pt_regs 273 lr %r2,%r11 # pass pointer to pt_regs
277 l %r1,BASED(.Ldo_signal) 274 l %r1,BASED(.Ldo_signal)
278 basr %r14,%r1 # call do_signal 275 basr %r14,%r1 # call do_signal
279 tm __TI_flags+3(%r12),_TIF_SYSCALL 276 tm __PT_FLAGS+3(%r11),_PIF_SYSCALL
280 jno sysc_return 277 jno sysc_return
281 lm %r2,%r7,__PT_R2(%r11) # load svc arguments 278 lm %r2,%r7,__PT_R2(%r11) # load svc arguments
282 l %r10,__TI_sysc_table(%r12) # 31 bit system call table 279 l %r10,__TI_sysc_table(%r12) # 31 bit system call table
@@ -297,10 +294,10 @@ sysc_notify_resume:
297 br %r1 # call do_notify_resume 294 br %r1 # call do_notify_resume
298 295
299# 296#
300# _TIF_PER_TRAP is set, call do_per_trap 297# _PIF_PER_TRAP is set, call do_per_trap
301# 298#
302sysc_singlestep: 299sysc_singlestep:
303 ni __TI_flags+3(%r12),255-_TIF_PER_TRAP 300 ni __PT_FLAGS+3(%r11),255-_PIF_PER_TRAP
304 lr %r2,%r11 # pass pointer to pt_regs 301 lr %r2,%r11 # pass pointer to pt_regs
305 l %r1,BASED(.Ldo_per_trap) 302 l %r1,BASED(.Ldo_per_trap)
306 la %r14,BASED(sysc_return) 303 la %r14,BASED(sysc_return)
@@ -330,7 +327,7 @@ sysc_tracego:
330 basr %r14,%r9 # call sys_xxx 327 basr %r14,%r9 # call sys_xxx
331 st %r2,__PT_R2(%r11) # store return value 328 st %r2,__PT_R2(%r11) # store return value
332sysc_tracenogo: 329sysc_tracenogo:
333 tm __TI_flags+2(%r12),_TIF_TRACE >> 8 330 tm __TI_flags+3(%r12),_TIF_TRACE
334 jz sysc_return 331 jz sysc_return
335 l %r1,BASED(.Ltrace_exit) 332 l %r1,BASED(.Ltrace_exit)
336 lr %r2,%r11 # pass pointer to pt_regs 333 lr %r2,%r11 # pass pointer to pt_regs
@@ -384,12 +381,13 @@ ENTRY(pgm_check_handler)
384 stm %r8,%r9,__PT_PSW(%r11) 381 stm %r8,%r9,__PT_PSW(%r11)
385 mvc __PT_INT_CODE(4,%r11),__LC_PGM_ILC 382 mvc __PT_INT_CODE(4,%r11),__LC_PGM_ILC
386 mvc __PT_INT_PARM_LONG(4,%r11),__LC_TRANS_EXC_CODE 383 mvc __PT_INT_PARM_LONG(4,%r11),__LC_TRANS_EXC_CODE
384 xc __PT_FLAGS(4,%r11),__PT_FLAGS(%r11)
387 tm __LC_PGM_ILC+3,0x80 # check for per exception 385 tm __LC_PGM_ILC+3,0x80 # check for per exception
388 jz 0f 386 jz 0f
389 l %r1,__TI_task(%r12) 387 l %r1,__TI_task(%r12)
390 tmh %r8,0x0001 # kernel per event ? 388 tmh %r8,0x0001 # kernel per event ?
391 jz pgm_kprobe 389 jz pgm_kprobe
392 oi __TI_flags+3(%r12),_TIF_PER_TRAP 390 oi __PT_FLAGS+3(%r11),_PIF_PER_TRAP
393 mvc __THREAD_per_address(4,%r1),__LC_PER_ADDRESS 391 mvc __THREAD_per_address(4,%r1),__LC_PER_ADDRESS
394 mvc __THREAD_per_cause(2,%r1),__LC_PER_CAUSE 392 mvc __THREAD_per_cause(2,%r1),__LC_PER_CAUSE
395 mvc __THREAD_per_paid(1,%r1),__LC_PER_PAID 393 mvc __THREAD_per_paid(1,%r1),__LC_PER_PAID
@@ -420,9 +418,9 @@ pgm_kprobe:
420# single stepped system call 418# single stepped system call
421# 419#
422pgm_svcper: 420pgm_svcper:
423 oi __TI_flags+3(%r12),_TIF_PER_TRAP
424 mvc __LC_RETURN_PSW(4),__LC_SVC_NEW_PSW 421 mvc __LC_RETURN_PSW(4),__LC_SVC_NEW_PSW
425 mvc __LC_RETURN_PSW+4(4),BASED(.Lsysc_per) 422 mvc __LC_RETURN_PSW+4(4),BASED(.Lsysc_per)
423 lhi %r14,_PIF_SYSCALL | _PIF_PER_TRAP
426 lpsw __LC_RETURN_PSW # branch to sysc_per and enable irqs 424 lpsw __LC_RETURN_PSW # branch to sysc_per and enable irqs
427 425
428/* 426/*
@@ -445,6 +443,7 @@ io_skip:
445 mvc __PT_R8(32,%r11),__LC_SAVE_AREA_ASYNC 443 mvc __PT_R8(32,%r11),__LC_SAVE_AREA_ASYNC
446 stm %r8,%r9,__PT_PSW(%r11) 444 stm %r8,%r9,__PT_PSW(%r11)
447 mvc __PT_INT_CODE(12,%r11),__LC_SUBCHANNEL_ID 445 mvc __PT_INT_CODE(12,%r11),__LC_SUBCHANNEL_ID
446 xc __PT_FLAGS(4,%r11),__PT_FLAGS(%r11)
448 TRACE_IRQS_OFF 447 TRACE_IRQS_OFF
449 xc __SF_BACKCHAIN(4,%r15),__SF_BACKCHAIN(%r15) 448 xc __SF_BACKCHAIN(4,%r15),__SF_BACKCHAIN(%r15)
450io_loop: 449io_loop:
@@ -466,8 +465,10 @@ io_return:
466 LOCKDEP_SYS_EXIT 465 LOCKDEP_SYS_EXIT
467 TRACE_IRQS_ON 466 TRACE_IRQS_ON
468io_tif: 467io_tif:
469 tm __TI_flags+3(%r12),_TIF_WORK_INT 468 tm __TI_flags+3(%r12),_TIF_WORK
470 jnz io_work # there is work to do (signals etc.) 469 jnz io_work # there is work to do (signals etc.)
470 tm __LC_CPU_FLAGS+3,_CIF_WORK
471 jnz io_work
471io_restore: 472io_restore:
472 mvc __LC_RETURN_PSW(8),__PT_PSW(%r11) 473 mvc __LC_RETURN_PSW(8),__PT_PSW(%r11)
473 stpt __LC_EXIT_TIMER 474 stpt __LC_EXIT_TIMER
@@ -477,7 +478,7 @@ io_done:
477 478
478# 479#
479# There is work todo, find out in which context we have been interrupted: 480# There is work todo, find out in which context we have been interrupted:
480# 1) if we return to user space we can do all _TIF_WORK_INT work 481# 1) if we return to user space we can do all _TIF_WORK work
481# 2) if we return to kernel code and preemptive scheduling is enabled check 482# 2) if we return to kernel code and preemptive scheduling is enabled check
482# the preemption counter and if it is zero call preempt_schedule_irq 483# the preemption counter and if it is zero call preempt_schedule_irq
483# Before any work can be done, a switch to the kernel stack is required. 484# Before any work can be done, a switch to the kernel stack is required.
@@ -520,11 +521,9 @@ io_work_user:
520 521
521# 522#
522# One of the work bits is on. Find out which one. 523# One of the work bits is on. Find out which one.
523# Checked are: _TIF_SIGPENDING, _TIF_NOTIFY_RESUME, _TIF_NEED_RESCHED
524# and _TIF_MCCK_PENDING
525# 524#
526io_work_tif: 525io_work_tif:
527 tm __TI_flags+3(%r12),_TIF_MCCK_PENDING 526 tm __LC_CPU_FLAGS+3(%r12),_CIF_MCCK_PENDING
528 jo io_mcck_pending 527 jo io_mcck_pending
529 tm __TI_flags+3(%r12),_TIF_NEED_RESCHED 528 tm __TI_flags+3(%r12),_TIF_NEED_RESCHED
530 jo io_reschedule 529 jo io_reschedule
@@ -532,12 +531,12 @@ io_work_tif:
532 jo io_sigpending 531 jo io_sigpending
533 tm __TI_flags+3(%r12),_TIF_NOTIFY_RESUME 532 tm __TI_flags+3(%r12),_TIF_NOTIFY_RESUME
534 jo io_notify_resume 533 jo io_notify_resume
535 tm __TI_flags+3(%r12),_TIF_ASCE 534 tm __LC_CPU_FLAGS+3,_CIF_ASCE
536 jo io_uaccess 535 jo io_uaccess
537 j io_return # beware of critical section cleanup 536 j io_return # beware of critical section cleanup
538 537
539# 538#
540# _TIF_MCCK_PENDING is set, call handler 539# _CIF_MCCK_PENDING is set, call handler
541# 540#
542io_mcck_pending: 541io_mcck_pending:
543 # TRACE_IRQS_ON already done at io_return 542 # TRACE_IRQS_ON already done at io_return
@@ -547,10 +546,10 @@ io_mcck_pending:
547 j io_return 546 j io_return
548 547
549# 548#
550# _TIF_ASCE is set, load user space asce 549# _CIF_ASCE is set, load user space asce
551# 550#
552io_uaccess: 551io_uaccess:
553 ni __TI_flags+3(%r12),255-_TIF_ASCE 552 ni __LC_CPU_FLAGS+3,255-_CIF_ASCE
554 lctl %c1,%c1,__LC_USER_ASCE # load primary asce 553 lctl %c1,%c1,__LC_USER_ASCE # load primary asce
555 j io_return 554 j io_return
556 555
@@ -613,6 +612,7 @@ ext_skip:
613 stm %r8,%r9,__PT_PSW(%r11) 612 stm %r8,%r9,__PT_PSW(%r11)
614 mvc __PT_INT_CODE(4,%r11),__LC_EXT_CPU_ADDR 613 mvc __PT_INT_CODE(4,%r11),__LC_EXT_CPU_ADDR
615 mvc __PT_INT_PARM(4,%r11),__LC_EXT_PARAMS 614 mvc __PT_INT_PARM(4,%r11),__LC_EXT_PARAMS
615 xc __PT_FLAGS(4,%r11),__PT_FLAGS(%r11)
616 TRACE_IRQS_OFF 616 TRACE_IRQS_OFF
617 l %r1,BASED(.Ldo_IRQ) 617 l %r1,BASED(.Ldo_IRQ)
618 lr %r2,%r11 # pass pointer to pt_regs 618 lr %r2,%r11 # pass pointer to pt_regs
@@ -677,6 +677,7 @@ mcck_skip:
677 stm %r0,%r7,__PT_R0(%r11) 677 stm %r0,%r7,__PT_R0(%r11)
678 mvc __PT_R8(32,%r11),__LC_GPREGS_SAVE_AREA+32 678 mvc __PT_R8(32,%r11),__LC_GPREGS_SAVE_AREA+32
679 stm %r8,%r9,__PT_PSW(%r11) 679 stm %r8,%r9,__PT_PSW(%r11)
680 xc __PT_FLAGS(4,%r11),__PT_FLAGS(%r11)
680 xc __SF_BACKCHAIN(4,%r15),__SF_BACKCHAIN(%r15) 681 xc __SF_BACKCHAIN(4,%r15),__SF_BACKCHAIN(%r15)
681 l %r1,BASED(.Ldo_machine_check) 682 l %r1,BASED(.Ldo_machine_check)
682 lr %r2,%r11 # pass pointer to pt_regs 683 lr %r2,%r11 # pass pointer to pt_regs
@@ -689,7 +690,7 @@ mcck_skip:
689 la %r11,STACK_FRAME_OVERHEAD(%r15) 690 la %r11,STACK_FRAME_OVERHEAD(%r15)
690 lr %r15,%r1 691 lr %r15,%r1
691 ssm __LC_PGM_NEW_PSW # turn dat on, keep irqs off 692 ssm __LC_PGM_NEW_PSW # turn dat on, keep irqs off
692 tm __TI_flags+3(%r12),_TIF_MCCK_PENDING 693 tm __LC_CPU_FLAGS+3,_CIF_MCCK_PENDING
693 jno mcck_return 694 jno mcck_return
694 TRACE_IRQS_OFF 695 TRACE_IRQS_OFF
695 l %r1,BASED(.Lhandle_mcck) 696 l %r1,BASED(.Lhandle_mcck)
@@ -842,6 +843,8 @@ cleanup_system_call:
842 stm %r0,%r7,__PT_R0(%r9) 843 stm %r0,%r7,__PT_R0(%r9)
843 mvc __PT_PSW(8,%r9),__LC_SVC_OLD_PSW 844 mvc __PT_PSW(8,%r9),__LC_SVC_OLD_PSW
844 mvc __PT_INT_CODE(4,%r9),__LC_SVC_ILC 845 mvc __PT_INT_CODE(4,%r9),__LC_SVC_ILC
846 xc __PT_FLAGS(4,%r9),__PT_FLAGS(%r9)
847 mvi __PT_FLAGS+3(%r9),_PIF_SYSCALL
845 # setup saved register 15 848 # setup saved register 15
846 st %r15,28(%r11) # r15 stack pointer 849 st %r15,28(%r11) # r15 stack pointer
847 # set new psw address and exit 850 # set new psw address and exit
diff --git a/arch/s390/kernel/entry64.S b/arch/s390/kernel/entry64.S
index 5963e43618bb..c41f3f906720 100644
--- a/arch/s390/kernel/entry64.S
+++ b/arch/s390/kernel/entry64.S
@@ -42,13 +42,11 @@ STACK_SHIFT = PAGE_SHIFT + THREAD_ORDER
42STACK_SIZE = 1 << STACK_SHIFT 42STACK_SIZE = 1 << STACK_SHIFT
43STACK_INIT = STACK_SIZE - STACK_FRAME_OVERHEAD - __PT_SIZE 43STACK_INIT = STACK_SIZE - STACK_FRAME_OVERHEAD - __PT_SIZE
44 44
45_TIF_WORK_SVC = (_TIF_SIGPENDING | _TIF_NOTIFY_RESUME | _TIF_NEED_RESCHED | \ 45_TIF_WORK = (_TIF_SIGPENDING | _TIF_NOTIFY_RESUME | _TIF_NEED_RESCHED)
46 _TIF_MCCK_PENDING | _TIF_PER_TRAP | _TIF_ASCE) 46_TIF_TRACE = (_TIF_SYSCALL_TRACE | _TIF_SYSCALL_AUDIT | _TIF_SECCOMP | \
47_TIF_WORK_INT = (_TIF_SIGPENDING | _TIF_NOTIFY_RESUME | _TIF_NEED_RESCHED | \ 47 _TIF_SYSCALL_TRACEPOINT)
48 _TIF_MCCK_PENDING | _TIF_ASCE) 48_CIF_WORK = (_CIF_MCCK_PENDING | _CIF_ASCE)
49_TIF_TRACE = (_TIF_SYSCALL_TRACE | _TIF_SYSCALL_AUDIT | _TIF_SECCOMP | \ 49_PIF_WORK = (_PIF_PER_TRAP)
50 _TIF_SYSCALL_TRACEPOINT)
51_TIF_TRANSFER = (_TIF_MCCK_PENDING | _TIF_TLB_WAIT)
52 50
53#define BASED(name) name-system_call(%r13) 51#define BASED(name) name-system_call(%r13)
54 52
@@ -190,13 +188,7 @@ ENTRY(__switch_to)
190 lctl %c4,%c4,__TASK_pid(%r3) # load pid to control reg. 4 188 lctl %c4,%c4,__TASK_pid(%r3) # load pid to control reg. 4
191 mvc __LC_CURRENT_PID+4(4,%r0),__TASK_pid(%r3) # store pid of next 189 mvc __LC_CURRENT_PID+4(4,%r0),__TASK_pid(%r3) # store pid of next
192 lg %r15,__THREAD_ksp(%r3) # load kernel stack of next 190 lg %r15,__THREAD_ksp(%r3) # load kernel stack of next
193 llill %r6,_TIF_TRANSFER # transfer TIF bits 191 lmg %r6,%r15,__SF_GPRS(%r15) # load gprs of next task
194 ng %r6,__TI_flags(%r4) # isolate TIF bits
195 jz 0f
196 og %r6,__TI_flags(%r5) # set TIF bits of next
197 stg %r6,__TI_flags(%r5)
198 ni __TI_flags+7(%r4),255-_TIF_TRANSFER # clear TIF bits of prev
1990: lmg %r6,%r15,__SF_GPRS(%r15) # load gprs of next task
200 br %r14 192 br %r14
201 193
202__critical_start: 194__critical_start:
@@ -211,6 +203,7 @@ sysc_stmg:
211 stmg %r8,%r15,__LC_SAVE_AREA_SYNC 203 stmg %r8,%r15,__LC_SAVE_AREA_SYNC
212 lg %r10,__LC_LAST_BREAK 204 lg %r10,__LC_LAST_BREAK
213 lg %r12,__LC_THREAD_INFO 205 lg %r12,__LC_THREAD_INFO
206 lghi %r14,_PIF_SYSCALL
214sysc_per: 207sysc_per:
215 lg %r15,__LC_KERNEL_STACK 208 lg %r15,__LC_KERNEL_STACK
216 la %r11,STACK_FRAME_OVERHEAD(%r15) # pointer to pt_regs 209 la %r11,STACK_FRAME_OVERHEAD(%r15) # pointer to pt_regs
@@ -221,8 +214,8 @@ sysc_vtime:
221 mvc __PT_R8(64,%r11),__LC_SAVE_AREA_SYNC 214 mvc __PT_R8(64,%r11),__LC_SAVE_AREA_SYNC
222 mvc __PT_PSW(16,%r11),__LC_SVC_OLD_PSW 215 mvc __PT_PSW(16,%r11),__LC_SVC_OLD_PSW
223 mvc __PT_INT_CODE(4,%r11),__LC_SVC_ILC 216 mvc __PT_INT_CODE(4,%r11),__LC_SVC_ILC
217 stg %r14,__PT_FLAGS(%r11)
224sysc_do_svc: 218sysc_do_svc:
225 oi __TI_flags+7(%r12),_TIF_SYSCALL
226 lg %r10,__TI_sysc_table(%r12) # address of system call table 219 lg %r10,__TI_sysc_table(%r12) # address of system call table
227 llgh %r8,__PT_INT_CODE+2(%r11) 220 llgh %r8,__PT_INT_CODE+2(%r11)
228 slag %r8,%r8,2 # shift and test for svc 0 221 slag %r8,%r8,2 # shift and test for svc 0
@@ -238,7 +231,7 @@ sysc_nr_ok:
238 stg %r2,__PT_ORIG_GPR2(%r11) 231 stg %r2,__PT_ORIG_GPR2(%r11)
239 stg %r7,STACK_FRAME_OVERHEAD(%r15) 232 stg %r7,STACK_FRAME_OVERHEAD(%r15)
240 lgf %r9,0(%r8,%r10) # get system call add. 233 lgf %r9,0(%r8,%r10) # get system call add.
241 tm __TI_flags+6(%r12),_TIF_TRACE >> 8 234 tm __TI_flags+7(%r12),_TIF_TRACE
242 jnz sysc_tracesys 235 jnz sysc_tracesys
243 basr %r14,%r9 # call sys_xxxx 236 basr %r14,%r9 # call sys_xxxx
244 stg %r2,__PT_R2(%r11) # store return value 237 stg %r2,__PT_R2(%r11) # store return value
@@ -248,9 +241,12 @@ sysc_return:
248sysc_tif: 241sysc_tif:
249 tm __PT_PSW+1(%r11),0x01 # returning to user ? 242 tm __PT_PSW+1(%r11),0x01 # returning to user ?
250 jno sysc_restore 243 jno sysc_restore
251 tm __TI_flags+7(%r12),_TIF_WORK_SVC 244 tm __PT_FLAGS+7(%r11),_PIF_WORK
245 jnz sysc_work
246 tm __TI_flags+7(%r12),_TIF_WORK
252 jnz sysc_work # check for work 247 jnz sysc_work # check for work
253 ni __TI_flags+7(%r12),255-_TIF_SYSCALL 248 tm __LC_CPU_FLAGS+7,_CIF_WORK
249 jnz sysc_work
254sysc_restore: 250sysc_restore:
255 lg %r14,__LC_VDSO_PER_CPU 251 lg %r14,__LC_VDSO_PER_CPU
256 lmg %r0,%r10,__PT_R0(%r11) 252 lmg %r0,%r10,__PT_R0(%r11)
@@ -265,17 +261,17 @@ sysc_done:
265# One of the work bits is on. Find out which one. 261# One of the work bits is on. Find out which one.
266# 262#
267sysc_work: 263sysc_work:
268 tm __TI_flags+7(%r12),_TIF_MCCK_PENDING 264 tm __LC_CPU_FLAGS+7,_CIF_MCCK_PENDING
269 jo sysc_mcck_pending 265 jo sysc_mcck_pending
270 tm __TI_flags+7(%r12),_TIF_NEED_RESCHED 266 tm __TI_flags+7(%r12),_TIF_NEED_RESCHED
271 jo sysc_reschedule 267 jo sysc_reschedule
272 tm __TI_flags+7(%r12),_TIF_PER_TRAP 268 tm __PT_FLAGS+7(%r11),_PIF_PER_TRAP
273 jo sysc_singlestep 269 jo sysc_singlestep
274 tm __TI_flags+7(%r12),_TIF_SIGPENDING 270 tm __TI_flags+7(%r12),_TIF_SIGPENDING
275 jo sysc_sigpending 271 jo sysc_sigpending
276 tm __TI_flags+7(%r12),_TIF_NOTIFY_RESUME 272 tm __TI_flags+7(%r12),_TIF_NOTIFY_RESUME
277 jo sysc_notify_resume 273 jo sysc_notify_resume
278 tm __TI_flags+7(%r12),_TIF_ASCE 274 tm __LC_CPU_FLAGS+7,_CIF_ASCE
279 jo sysc_uaccess 275 jo sysc_uaccess
280 j sysc_return # beware of critical section cleanup 276 j sysc_return # beware of critical section cleanup
281 277
@@ -287,17 +283,17 @@ sysc_reschedule:
287 jg schedule 283 jg schedule
288 284
289# 285#
290# _TIF_MCCK_PENDING is set, call handler 286# _CIF_MCCK_PENDING is set, call handler
291# 287#
292sysc_mcck_pending: 288sysc_mcck_pending:
293 larl %r14,sysc_return 289 larl %r14,sysc_return
294 jg s390_handle_mcck # TIF bit will be cleared by handler 290 jg s390_handle_mcck # TIF bit will be cleared by handler
295 291
296# 292#
297# _TIF_ASCE is set, load user space asce 293# _CIF_ASCE is set, load user space asce
298# 294#
299sysc_uaccess: 295sysc_uaccess:
300 ni __TI_flags+7(%r12),255-_TIF_ASCE 296 ni __LC_CPU_FLAGS+7,255-_CIF_ASCE
301 lctlg %c1,%c1,__LC_USER_ASCE # load primary asce 297 lctlg %c1,%c1,__LC_USER_ASCE # load primary asce
302 j sysc_return 298 j sysc_return
303 299
@@ -307,7 +303,7 @@ sysc_uaccess:
307sysc_sigpending: 303sysc_sigpending:
308 lgr %r2,%r11 # pass pointer to pt_regs 304 lgr %r2,%r11 # pass pointer to pt_regs
309 brasl %r14,do_signal 305 brasl %r14,do_signal
310 tm __TI_flags+7(%r12),_TIF_SYSCALL 306 tm __PT_FLAGS+7(%r11),_PIF_SYSCALL
311 jno sysc_return 307 jno sysc_return
312 lmg %r2,%r7,__PT_R2(%r11) # load svc arguments 308 lmg %r2,%r7,__PT_R2(%r11) # load svc arguments
313 lg %r10,__TI_sysc_table(%r12) # address of system call table 309 lg %r10,__TI_sysc_table(%r12) # address of system call table
@@ -327,10 +323,10 @@ sysc_notify_resume:
327 jg do_notify_resume 323 jg do_notify_resume
328 324
329# 325#
330# _TIF_PER_TRAP is set, call do_per_trap 326# _PIF_PER_TRAP is set, call do_per_trap
331# 327#
332sysc_singlestep: 328sysc_singlestep:
333 ni __TI_flags+7(%r12),255-_TIF_PER_TRAP 329 ni __PT_FLAGS+7(%r11),255-_PIF_PER_TRAP
334 lgr %r2,%r11 # pass pointer to pt_regs 330 lgr %r2,%r11 # pass pointer to pt_regs
335 larl %r14,sysc_return 331 larl %r14,sysc_return
336 jg do_per_trap 332 jg do_per_trap
@@ -357,7 +353,7 @@ sysc_tracego:
357 basr %r14,%r9 # call sys_xxx 353 basr %r14,%r9 # call sys_xxx
358 stg %r2,__PT_R2(%r11) # store return value 354 stg %r2,__PT_R2(%r11) # store return value
359sysc_tracenogo: 355sysc_tracenogo:
360 tm __TI_flags+6(%r12),_TIF_TRACE >> 8 356 tm __TI_flags+7(%r12),_TIF_TRACE
361 jz sysc_return 357 jz sysc_return
362 lgr %r2,%r11 # pass pointer to pt_regs 358 lgr %r2,%r11 # pass pointer to pt_regs
363 larl %r14,sysc_return 359 larl %r14,sysc_return
@@ -416,12 +412,13 @@ ENTRY(pgm_check_handler)
416 stmg %r8,%r9,__PT_PSW(%r11) 412 stmg %r8,%r9,__PT_PSW(%r11)
417 mvc __PT_INT_CODE(4,%r11),__LC_PGM_ILC 413 mvc __PT_INT_CODE(4,%r11),__LC_PGM_ILC
418 mvc __PT_INT_PARM_LONG(8,%r11),__LC_TRANS_EXC_CODE 414 mvc __PT_INT_PARM_LONG(8,%r11),__LC_TRANS_EXC_CODE
415 xc __PT_FLAGS(8,%r11),__PT_FLAGS(%r11)
419 stg %r10,__PT_ARGS(%r11) 416 stg %r10,__PT_ARGS(%r11)
420 tm __LC_PGM_ILC+3,0x80 # check for per exception 417 tm __LC_PGM_ILC+3,0x80 # check for per exception
421 jz 0f 418 jz 0f
422 tmhh %r8,0x0001 # kernel per event ? 419 tmhh %r8,0x0001 # kernel per event ?
423 jz pgm_kprobe 420 jz pgm_kprobe
424 oi __TI_flags+7(%r12),_TIF_PER_TRAP 421 oi __PT_FLAGS+7(%r11),_PIF_PER_TRAP
425 mvc __THREAD_per_address(8,%r14),__LC_PER_ADDRESS 422 mvc __THREAD_per_address(8,%r14),__LC_PER_ADDRESS
426 mvc __THREAD_per_cause(2,%r14),__LC_PER_CAUSE 423 mvc __THREAD_per_cause(2,%r14),__LC_PER_CAUSE
427 mvc __THREAD_per_paid(1,%r14),__LC_PER_PAID 424 mvc __THREAD_per_paid(1,%r14),__LC_PER_PAID
@@ -451,10 +448,10 @@ pgm_kprobe:
451# single stepped system call 448# single stepped system call
452# 449#
453pgm_svcper: 450pgm_svcper:
454 oi __TI_flags+7(%r12),_TIF_PER_TRAP
455 mvc __LC_RETURN_PSW(8),__LC_SVC_NEW_PSW 451 mvc __LC_RETURN_PSW(8),__LC_SVC_NEW_PSW
456 larl %r14,sysc_per 452 larl %r14,sysc_per
457 stg %r14,__LC_RETURN_PSW+8 453 stg %r14,__LC_RETURN_PSW+8
454 lghi %r14,_PIF_SYSCALL | _PIF_PER_TRAP
458 lpswe __LC_RETURN_PSW # branch to sysc_per and enable irqs 455 lpswe __LC_RETURN_PSW # branch to sysc_per and enable irqs
459 456
460/* 457/*
@@ -479,6 +476,7 @@ io_skip:
479 mvc __PT_R8(64,%r11),__LC_SAVE_AREA_ASYNC 476 mvc __PT_R8(64,%r11),__LC_SAVE_AREA_ASYNC
480 stmg %r8,%r9,__PT_PSW(%r11) 477 stmg %r8,%r9,__PT_PSW(%r11)
481 mvc __PT_INT_CODE(12,%r11),__LC_SUBCHANNEL_ID 478 mvc __PT_INT_CODE(12,%r11),__LC_SUBCHANNEL_ID
479 xc __PT_FLAGS(8,%r11),__PT_FLAGS(%r11)
482 TRACE_IRQS_OFF 480 TRACE_IRQS_OFF
483 xc __SF_BACKCHAIN(8,%r15),__SF_BACKCHAIN(%r15) 481 xc __SF_BACKCHAIN(8,%r15),__SF_BACKCHAIN(%r15)
484io_loop: 482io_loop:
@@ -499,8 +497,10 @@ io_return:
499 LOCKDEP_SYS_EXIT 497 LOCKDEP_SYS_EXIT
500 TRACE_IRQS_ON 498 TRACE_IRQS_ON
501io_tif: 499io_tif:
502 tm __TI_flags+7(%r12),_TIF_WORK_INT 500 tm __TI_flags+7(%r12),_TIF_WORK
503 jnz io_work # there is work to do (signals etc.) 501 jnz io_work # there is work to do (signals etc.)
502 tm __LC_CPU_FLAGS+7,_CIF_WORK
503 jnz io_work
504io_restore: 504io_restore:
505 lg %r14,__LC_VDSO_PER_CPU 505 lg %r14,__LC_VDSO_PER_CPU
506 lmg %r0,%r10,__PT_R0(%r11) 506 lmg %r0,%r10,__PT_R0(%r11)
@@ -513,7 +513,7 @@ io_done:
513 513
514# 514#
515# There is work todo, find out in which context we have been interrupted: 515# There is work todo, find out in which context we have been interrupted:
516# 1) if we return to user space we can do all _TIF_WORK_INT work 516# 1) if we return to user space we can do all _TIF_WORK work
517# 2) if we return to kernel code and kvm is enabled check if we need to 517# 2) if we return to kernel code and kvm is enabled check if we need to
518# modify the psw to leave SIE 518# modify the psw to leave SIE
519# 3) if we return to kernel code and preemptive scheduling is enabled check 519# 3) if we return to kernel code and preemptive scheduling is enabled check
@@ -557,11 +557,9 @@ io_work_user:
557 557
558# 558#
559# One of the work bits is on. Find out which one. 559# One of the work bits is on. Find out which one.
560# Checked are: _TIF_SIGPENDING, _TIF_NOTIFY_RESUME, _TIF_NEED_RESCHED
561# and _TIF_MCCK_PENDING
562# 560#
563io_work_tif: 561io_work_tif:
564 tm __TI_flags+7(%r12),_TIF_MCCK_PENDING 562 tm __LC_CPU_FLAGS+7,_CIF_MCCK_PENDING
565 jo io_mcck_pending 563 jo io_mcck_pending
566 tm __TI_flags+7(%r12),_TIF_NEED_RESCHED 564 tm __TI_flags+7(%r12),_TIF_NEED_RESCHED
567 jo io_reschedule 565 jo io_reschedule
@@ -569,12 +567,12 @@ io_work_tif:
569 jo io_sigpending 567 jo io_sigpending
570 tm __TI_flags+7(%r12),_TIF_NOTIFY_RESUME 568 tm __TI_flags+7(%r12),_TIF_NOTIFY_RESUME
571 jo io_notify_resume 569 jo io_notify_resume
572 tm __TI_flags+7(%r12),_TIF_ASCE 570 tm __LC_CPU_FLAGS+7,_CIF_ASCE
573 jo io_uaccess 571 jo io_uaccess
574 j io_return # beware of critical section cleanup 572 j io_return # beware of critical section cleanup
575 573
576# 574#
577# _TIF_MCCK_PENDING is set, call handler 575# _CIF_MCCK_PENDING is set, call handler
578# 576#
579io_mcck_pending: 577io_mcck_pending:
580 # TRACE_IRQS_ON already done at io_return 578 # TRACE_IRQS_ON already done at io_return
@@ -583,10 +581,10 @@ io_mcck_pending:
583 j io_return 581 j io_return
584 582
585# 583#
586# _TIF_ASCE is set, load user space asce 584# _CIF_ASCE is set, load user space asce
587# 585#
588io_uaccess: 586io_uaccess:
589 ni __TI_flags+7(%r12),255-_TIF_ASCE 587 ni __LC_CPU_FLAGS+7,255-_CIF_ASCE
590 lctlg %c1,%c1,__LC_USER_ASCE # load primary asce 588 lctlg %c1,%c1,__LC_USER_ASCE # load primary asce
591 j io_return 589 j io_return
592 590
@@ -650,6 +648,7 @@ ext_skip:
650 mvc __PT_INT_CODE(4,%r11),__LC_EXT_CPU_ADDR 648 mvc __PT_INT_CODE(4,%r11),__LC_EXT_CPU_ADDR
651 mvc __PT_INT_PARM(4,%r11),__LC_EXT_PARAMS 649 mvc __PT_INT_PARM(4,%r11),__LC_EXT_PARAMS
652 mvc __PT_INT_PARM_LONG(8,%r11),0(%r1) 650 mvc __PT_INT_PARM_LONG(8,%r11),0(%r1)
651 xc __PT_FLAGS(8,%r11),__PT_FLAGS(%r11)
653 TRACE_IRQS_OFF 652 TRACE_IRQS_OFF
654 xc __SF_BACKCHAIN(8,%r15),__SF_BACKCHAIN(%r15) 653 xc __SF_BACKCHAIN(8,%r15),__SF_BACKCHAIN(%r15)
655 lgr %r2,%r11 # pass pointer to pt_regs 654 lgr %r2,%r11 # pass pointer to pt_regs
@@ -716,6 +715,7 @@ mcck_skip:
716 stmg %r0,%r7,__PT_R0(%r11) 715 stmg %r0,%r7,__PT_R0(%r11)
717 mvc __PT_R8(64,%r11),0(%r14) 716 mvc __PT_R8(64,%r11),0(%r14)
718 stmg %r8,%r9,__PT_PSW(%r11) 717 stmg %r8,%r9,__PT_PSW(%r11)
718 xc __PT_FLAGS(8,%r11),__PT_FLAGS(%r11)
719 xc __SF_BACKCHAIN(8,%r15),__SF_BACKCHAIN(%r15) 719 xc __SF_BACKCHAIN(8,%r15),__SF_BACKCHAIN(%r15)
720 lgr %r2,%r11 # pass pointer to pt_regs 720 lgr %r2,%r11 # pass pointer to pt_regs
721 brasl %r14,s390_do_machine_check 721 brasl %r14,s390_do_machine_check
@@ -727,7 +727,7 @@ mcck_skip:
727 la %r11,STACK_FRAME_OVERHEAD(%r1) 727 la %r11,STACK_FRAME_OVERHEAD(%r1)
728 lgr %r15,%r1 728 lgr %r15,%r1
729 ssm __LC_PGM_NEW_PSW # turn dat on, keep irqs off 729 ssm __LC_PGM_NEW_PSW # turn dat on, keep irqs off
730 tm __TI_flags+7(%r12),_TIF_MCCK_PENDING 730 tm __LC_CPU_FLAGS+7,_CIF_MCCK_PENDING
731 jno mcck_return 731 jno mcck_return
732 TRACE_IRQS_OFF 732 TRACE_IRQS_OFF
733 brasl %r14,s390_handle_mcck 733 brasl %r14,s390_handle_mcck
@@ -884,6 +884,8 @@ cleanup_system_call:
884 stmg %r0,%r7,__PT_R0(%r9) 884 stmg %r0,%r7,__PT_R0(%r9)
885 mvc __PT_PSW(16,%r9),__LC_SVC_OLD_PSW 885 mvc __PT_PSW(16,%r9),__LC_SVC_OLD_PSW
886 mvc __PT_INT_CODE(4,%r9),__LC_SVC_ILC 886 mvc __PT_INT_CODE(4,%r9),__LC_SVC_ILC
887 xc __PT_FLAGS(8,%r9),__PT_FLAGS(%r9)
888 mvi __PT_FLAGS+7(%r9),_PIF_SYSCALL
887 # setup saved register r15 889 # setup saved register r15
888 stg %r15,56(%r11) # r15 stack pointer 890 stg %r15,56(%r11) # r15 stack pointer
889 # set new psw address and exit 891 # set new psw address and exit
diff --git a/arch/s390/kernel/head.S b/arch/s390/kernel/head.S
index 429afcc480cb..7ba7d6784510 100644
--- a/arch/s390/kernel/head.S
+++ b/arch/s390/kernel/head.S
@@ -437,13 +437,13 @@ ENTRY(startup_kdump)
437 437
438#if defined(CONFIG_64BIT) 438#if defined(CONFIG_64BIT)
439#if defined(CONFIG_MARCH_ZEC12) 439#if defined(CONFIG_MARCH_ZEC12)
440 .long 3, 0xc100efe3, 0xf46ce800, 0x00400000 440 .long 3, 0xc100efea, 0xf46ce800, 0x00400000
441#elif defined(CONFIG_MARCH_Z196) 441#elif defined(CONFIG_MARCH_Z196)
442 .long 2, 0xc100efe3, 0xf46c0000 442 .long 2, 0xc100efea, 0xf46c0000
443#elif defined(CONFIG_MARCH_Z10) 443#elif defined(CONFIG_MARCH_Z10)
444 .long 2, 0xc100efe3, 0xf0680000 444 .long 2, 0xc100efea, 0xf0680000
445#elif defined(CONFIG_MARCH_Z9_109) 445#elif defined(CONFIG_MARCH_Z9_109)
446 .long 1, 0xc100efc3 446 .long 1, 0xc100efc2
447#elif defined(CONFIG_MARCH_Z990) 447#elif defined(CONFIG_MARCH_Z990)
448 .long 1, 0xc0002000 448 .long 1, 0xc0002000
449#elif defined(CONFIG_MARCH_Z900) 449#elif defined(CONFIG_MARCH_Z900)
diff --git a/arch/s390/kernel/head31.S b/arch/s390/kernel/head31.S
index 9a99856df1c9..6dbe80983a24 100644
--- a/arch/s390/kernel/head31.S
+++ b/arch/s390/kernel/head31.S
@@ -59,7 +59,6 @@ ENTRY(startup_continue)
59 .long 0 # cr13: home space segment table 59 .long 0 # cr13: home space segment table
60 .long 0xc0000000 # cr14: machine check handling off 60 .long 0xc0000000 # cr14: machine check handling off
61 .long 0 # cr15: linkage stack operations 61 .long 0 # cr15: linkage stack operations
62.Lmchunk:.long memory_chunk
63.Lbss_bgn: .long __bss_start 62.Lbss_bgn: .long __bss_start
64.Lbss_end: .long _end 63.Lbss_end: .long _end
65.Lparmaddr: .long PARMAREA 64.Lparmaddr: .long PARMAREA
diff --git a/arch/s390/kernel/nmi.c b/arch/s390/kernel/nmi.c
index c4c033819879..210e1285f75a 100644
--- a/arch/s390/kernel/nmi.c
+++ b/arch/s390/kernel/nmi.c
@@ -55,7 +55,7 @@ void s390_handle_mcck(void)
55 local_mcck_disable(); 55 local_mcck_disable();
56 mcck = __get_cpu_var(cpu_mcck); 56 mcck = __get_cpu_var(cpu_mcck);
57 memset(&__get_cpu_var(cpu_mcck), 0, sizeof(struct mcck_struct)); 57 memset(&__get_cpu_var(cpu_mcck), 0, sizeof(struct mcck_struct));
58 clear_thread_flag(TIF_MCCK_PENDING); 58 clear_cpu_flag(CIF_MCCK_PENDING);
59 local_mcck_enable(); 59 local_mcck_enable();
60 local_irq_restore(flags); 60 local_irq_restore(flags);
61 61
@@ -313,7 +313,7 @@ void notrace s390_do_machine_check(struct pt_regs *regs)
313 */ 313 */
314 mcck->kill_task = 1; 314 mcck->kill_task = 1;
315 mcck->mcck_code = *(unsigned long long *) mci; 315 mcck->mcck_code = *(unsigned long long *) mci;
316 set_thread_flag(TIF_MCCK_PENDING); 316 set_cpu_flag(CIF_MCCK_PENDING);
317 } else { 317 } else {
318 /* 318 /*
319 * Couldn't restore all register contents while in 319 * Couldn't restore all register contents while in
@@ -352,12 +352,12 @@ void notrace s390_do_machine_check(struct pt_regs *regs)
352 if (mci->cp) { 352 if (mci->cp) {
353 /* Channel report word pending */ 353 /* Channel report word pending */
354 mcck->channel_report = 1; 354 mcck->channel_report = 1;
355 set_thread_flag(TIF_MCCK_PENDING); 355 set_cpu_flag(CIF_MCCK_PENDING);
356 } 356 }
357 if (mci->w) { 357 if (mci->w) {
358 /* Warning pending */ 358 /* Warning pending */
359 mcck->warning = 1; 359 mcck->warning = 1;
360 set_thread_flag(TIF_MCCK_PENDING); 360 set_cpu_flag(CIF_MCCK_PENDING);
361 } 361 }
362 nmi_exit(); 362 nmi_exit();
363} 363}
diff --git a/arch/s390/kernel/process.c b/arch/s390/kernel/process.c
index dd145321d215..93b9ca42e5c0 100644
--- a/arch/s390/kernel/process.c
+++ b/arch/s390/kernel/process.c
@@ -64,7 +64,7 @@ unsigned long thread_saved_pc(struct task_struct *tsk)
64void arch_cpu_idle(void) 64void arch_cpu_idle(void)
65{ 65{
66 local_mcck_disable(); 66 local_mcck_disable();
67 if (test_thread_flag(TIF_MCCK_PENDING)) { 67 if (test_cpu_flag(CIF_MCCK_PENDING)) {
68 local_mcck_enable(); 68 local_mcck_enable();
69 local_irq_enable(); 69 local_irq_enable();
70 return; 70 return;
@@ -76,7 +76,7 @@ void arch_cpu_idle(void)
76 76
77void arch_cpu_idle_exit(void) 77void arch_cpu_idle_exit(void)
78{ 78{
79 if (test_thread_flag(TIF_MCCK_PENDING)) 79 if (test_cpu_flag(CIF_MCCK_PENDING))
80 s390_handle_mcck(); 80 s390_handle_mcck();
81} 81}
82 82
@@ -123,7 +123,6 @@ int copy_thread(unsigned long clone_flags, unsigned long new_stackp,
123 memset(&p->thread.per_user, 0, sizeof(p->thread.per_user)); 123 memset(&p->thread.per_user, 0, sizeof(p->thread.per_user));
124 memset(&p->thread.per_event, 0, sizeof(p->thread.per_event)); 124 memset(&p->thread.per_event, 0, sizeof(p->thread.per_event));
125 clear_tsk_thread_flag(p, TIF_SINGLE_STEP); 125 clear_tsk_thread_flag(p, TIF_SINGLE_STEP);
126 clear_tsk_thread_flag(p, TIF_PER_TRAP);
127 /* Initialize per thread user and system timer values */ 126 /* Initialize per thread user and system timer values */
128 ti = task_thread_info(p); 127 ti = task_thread_info(p);
129 ti->user_timer = 0; 128 ti->user_timer = 0;
@@ -152,6 +151,7 @@ int copy_thread(unsigned long clone_flags, unsigned long new_stackp,
152 } 151 }
153 frame->childregs = *current_pt_regs(); 152 frame->childregs = *current_pt_regs();
154 frame->childregs.gprs[2] = 0; /* child returns 0 on fork. */ 153 frame->childregs.gprs[2] = 0; /* child returns 0 on fork. */
154 frame->childregs.flags = 0;
155 if (new_stackp) 155 if (new_stackp)
156 frame->childregs.gprs[15] = new_stackp; 156 frame->childregs.gprs[15] = new_stackp;
157 157
diff --git a/arch/s390/kernel/ptrace.c b/arch/s390/kernel/ptrace.c
index 1c82619eb4f7..2d716734b5b1 100644
--- a/arch/s390/kernel/ptrace.c
+++ b/arch/s390/kernel/ptrace.c
@@ -136,7 +136,7 @@ void ptrace_disable(struct task_struct *task)
136 memset(&task->thread.per_user, 0, sizeof(task->thread.per_user)); 136 memset(&task->thread.per_user, 0, sizeof(task->thread.per_user));
137 memset(&task->thread.per_event, 0, sizeof(task->thread.per_event)); 137 memset(&task->thread.per_event, 0, sizeof(task->thread.per_event));
138 clear_tsk_thread_flag(task, TIF_SINGLE_STEP); 138 clear_tsk_thread_flag(task, TIF_SINGLE_STEP);
139 clear_tsk_thread_flag(task, TIF_PER_TRAP); 139 clear_pt_regs_flag(task_pt_regs(task), PIF_PER_TRAP);
140 task->thread.per_flags = 0; 140 task->thread.per_flags = 0;
141} 141}
142 142
@@ -813,7 +813,7 @@ asmlinkage long do_syscall_trace_enter(struct pt_regs *regs)
813 * debugger stored an invalid system call number. Skip 813 * debugger stored an invalid system call number. Skip
814 * the system call and the system call restart handling. 814 * the system call and the system call restart handling.
815 */ 815 */
816 clear_thread_flag(TIF_SYSCALL); 816 clear_pt_regs_flag(regs, PIF_SYSCALL);
817 ret = -1; 817 ret = -1;
818 } 818 }
819 819
diff --git a/arch/s390/kernel/setup.c b/arch/s390/kernel/setup.c
index 88d1ca81e2dd..1e2264b46e4c 100644
--- a/arch/s390/kernel/setup.c
+++ b/arch/s390/kernel/setup.c
@@ -78,10 +78,9 @@ EXPORT_SYMBOL(console_irq);
78unsigned long elf_hwcap = 0; 78unsigned long elf_hwcap = 0;
79char elf_platform[ELF_PLATFORM_SIZE]; 79char elf_platform[ELF_PLATFORM_SIZE];
80 80
81struct mem_chunk __initdata memory_chunk[MEMORY_CHUNKS];
82
83int __initdata memory_end_set; 81int __initdata memory_end_set;
84unsigned long __initdata memory_end; 82unsigned long __initdata memory_end;
83unsigned long __initdata max_physmem_end;
85 84
86unsigned long VMALLOC_START; 85unsigned long VMALLOC_START;
87EXPORT_SYMBOL(VMALLOC_START); 86EXPORT_SYMBOL(VMALLOC_START);
@@ -212,7 +211,7 @@ static void __init conmode_default(void)
212 } 211 }
213} 212}
214 213
215#ifdef CONFIG_ZFCPDUMP 214#ifdef CONFIG_CRASH_DUMP
216static void __init setup_zfcpdump(void) 215static void __init setup_zfcpdump(void)
217{ 216{
218 if (ipl_info.type != IPL_TYPE_FCP_DUMP) 217 if (ipl_info.type != IPL_TYPE_FCP_DUMP)
@@ -224,7 +223,7 @@ static void __init setup_zfcpdump(void)
224} 223}
225#else 224#else
226static inline void setup_zfcpdump(void) {} 225static inline void setup_zfcpdump(void) {}
227#endif /* CONFIG_ZFCPDUMP */ 226#endif /* CONFIG_CRASH_DUMP */
228 227
229 /* 228 /*
230 * Reboot, halt and power_off stubs. They just call _machine_restart, 229 * Reboot, halt and power_off stubs. They just call _machine_restart,
@@ -273,6 +272,7 @@ EXPORT_SYMBOL_GPL(pm_power_off);
273static int __init early_parse_mem(char *p) 272static int __init early_parse_mem(char *p)
274{ 273{
275 memory_end = memparse(p, &p); 274 memory_end = memparse(p, &p);
275 memory_end &= PAGE_MASK;
276 memory_end_set = 1; 276 memory_end_set = 1;
277 return 0; 277 return 0;
278} 278}
@@ -373,6 +373,10 @@ static void __init setup_lowcore(void)
373 mem_assign_absolute(S390_lowcore.restart_source, lc->restart_source); 373 mem_assign_absolute(S390_lowcore.restart_source, lc->restart_source);
374 mem_assign_absolute(S390_lowcore.restart_psw, lc->restart_psw); 374 mem_assign_absolute(S390_lowcore.restart_psw, lc->restart_psw);
375 375
376#ifdef CONFIG_SMP
377 lc->spinlock_lockval = arch_spin_lockval(0);
378#endif
379
376 set_prefix((u32)(unsigned long) lc); 380 set_prefix((u32)(unsigned long) lc);
377 lowcore_ptr[0] = lc; 381 lowcore_ptr[0] = lc;
378} 382}
@@ -401,7 +405,8 @@ static struct resource __initdata *standard_resources[] = {
401static void __init setup_resources(void) 405static void __init setup_resources(void)
402{ 406{
403 struct resource *res, *std_res, *sub_res; 407 struct resource *res, *std_res, *sub_res;
404 int i, j; 408 struct memblock_region *reg;
409 int j;
405 410
406 code_resource.start = (unsigned long) &_text; 411 code_resource.start = (unsigned long) &_text;
407 code_resource.end = (unsigned long) &_etext - 1; 412 code_resource.end = (unsigned long) &_etext - 1;
@@ -410,24 +415,13 @@ static void __init setup_resources(void)
410 bss_resource.start = (unsigned long) &__bss_start; 415 bss_resource.start = (unsigned long) &__bss_start;
411 bss_resource.end = (unsigned long) &__bss_stop - 1; 416 bss_resource.end = (unsigned long) &__bss_stop - 1;
412 417
413 for (i = 0; i < MEMORY_CHUNKS; i++) { 418 for_each_memblock(memory, reg) {
414 if (!memory_chunk[i].size)
415 continue;
416 res = alloc_bootmem_low(sizeof(*res)); 419 res = alloc_bootmem_low(sizeof(*res));
417 res->flags = IORESOURCE_BUSY | IORESOURCE_MEM; 420 res->flags = IORESOURCE_BUSY | IORESOURCE_MEM;
418 switch (memory_chunk[i].type) { 421
419 case CHUNK_READ_WRITE: 422 res->name = "System RAM";
420 res->name = "System RAM"; 423 res->start = reg->base;
421 break; 424 res->end = reg->base + reg->size - 1;
422 case CHUNK_READ_ONLY:
423 res->name = "System ROM";
424 res->flags |= IORESOURCE_READONLY;
425 break;
426 default:
427 res->name = "reserved";
428 }
429 res->start = memory_chunk[i].addr;
430 res->end = res->start + memory_chunk[i].size - 1;
431 request_resource(&iomem_resource, res); 425 request_resource(&iomem_resource, res);
432 426
433 for (j = 0; j < ARRAY_SIZE(standard_resources); j++) { 427 for (j = 0; j < ARRAY_SIZE(standard_resources); j++) {
@@ -451,48 +445,11 @@ static void __init setup_resources(void)
451static void __init setup_memory_end(void) 445static void __init setup_memory_end(void)
452{ 446{
453 unsigned long vmax, vmalloc_size, tmp; 447 unsigned long vmax, vmalloc_size, tmp;
454 unsigned long real_memory_size = 0;
455 int i;
456
457
458#ifdef CONFIG_ZFCPDUMP
459 if (ipl_info.type == IPL_TYPE_FCP_DUMP &&
460 !OLDMEM_BASE && sclp_get_hsa_size()) {
461 memory_end = sclp_get_hsa_size();
462 memory_end_set = 1;
463 }
464#endif
465 memory_end &= PAGE_MASK;
466
467 /*
468 * Make sure all chunks are MAX_ORDER aligned so we don't need the
469 * extra checks that HOLES_IN_ZONE would require.
470 */
471 for (i = 0; i < MEMORY_CHUNKS; i++) {
472 unsigned long start, end;
473 struct mem_chunk *chunk;
474 unsigned long align;
475
476 chunk = &memory_chunk[i];
477 if (!chunk->size)
478 continue;
479 align = 1UL << (MAX_ORDER + PAGE_SHIFT - 1);
480 start = (chunk->addr + align - 1) & ~(align - 1);
481 end = (chunk->addr + chunk->size) & ~(align - 1);
482 if (start >= end)
483 memset(chunk, 0, sizeof(*chunk));
484 else {
485 chunk->addr = start;
486 chunk->size = end - start;
487 }
488 real_memory_size = max(real_memory_size,
489 chunk->addr + chunk->size);
490 }
491 448
492 /* Choose kernel address space layout: 2, 3, or 4 levels. */ 449 /* Choose kernel address space layout: 2, 3, or 4 levels. */
493#ifdef CONFIG_64BIT 450#ifdef CONFIG_64BIT
494 vmalloc_size = VMALLOC_END ?: (128UL << 30) - MODULES_LEN; 451 vmalloc_size = VMALLOC_END ?: (128UL << 30) - MODULES_LEN;
495 tmp = (memory_end ?: real_memory_size) / PAGE_SIZE; 452 tmp = (memory_end ?: max_physmem_end) / PAGE_SIZE;
496 tmp = tmp * (sizeof(struct page) + PAGE_SIZE) + vmalloc_size; 453 tmp = tmp * (sizeof(struct page) + PAGE_SIZE) + vmalloc_size;
497 if (tmp <= (1UL << 42)) 454 if (tmp <= (1UL << 42))
498 vmax = 1UL << 42; /* 3-level kernel page table */ 455 vmax = 1UL << 42; /* 3-level kernel page table */
@@ -520,21 +477,11 @@ static void __init setup_memory_end(void)
520 vmemmap = (struct page *) tmp; 477 vmemmap = (struct page *) tmp;
521 478
522 /* Take care that memory_end is set and <= vmemmap */ 479 /* Take care that memory_end is set and <= vmemmap */
523 memory_end = min(memory_end ?: real_memory_size, tmp); 480 memory_end = min(memory_end ?: max_physmem_end, tmp);
524 481 max_pfn = max_low_pfn = PFN_DOWN(memory_end);
525 /* Fixup memory chunk array to fit into 0..memory_end */ 482 memblock_remove(memory_end, ULONG_MAX);
526 for (i = 0; i < MEMORY_CHUNKS; i++) {
527 struct mem_chunk *chunk = &memory_chunk[i];
528 483
529 if (!chunk->size) 484 pr_notice("Max memory size: %luMB\n", memory_end >> 20);
530 continue;
531 if (chunk->addr >= memory_end) {
532 memset(chunk, 0, sizeof(*chunk));
533 continue;
534 }
535 if (chunk->addr + chunk->size > memory_end)
536 chunk->size = memory_end - chunk->addr;
537 }
538} 485}
539 486
540static void __init setup_vmcoreinfo(void) 487static void __init setup_vmcoreinfo(void)
@@ -545,89 +492,6 @@ static void __init setup_vmcoreinfo(void)
545#ifdef CONFIG_CRASH_DUMP 492#ifdef CONFIG_CRASH_DUMP
546 493
547/* 494/*
548 * Find suitable location for crashkernel memory
549 */
550static unsigned long __init find_crash_base(unsigned long crash_size,
551 char **msg)
552{
553 unsigned long crash_base;
554 struct mem_chunk *chunk;
555 int i;
556
557 if (memory_chunk[0].size < crash_size) {
558 *msg = "first memory chunk must be at least crashkernel size";
559 return 0;
560 }
561 if (OLDMEM_BASE && crash_size == OLDMEM_SIZE)
562 return OLDMEM_BASE;
563
564 for (i = MEMORY_CHUNKS - 1; i >= 0; i--) {
565 chunk = &memory_chunk[i];
566 if (chunk->size == 0)
567 continue;
568 if (chunk->type != CHUNK_READ_WRITE)
569 continue;
570 if (chunk->size < crash_size)
571 continue;
572 crash_base = (chunk->addr + chunk->size) - crash_size;
573 if (crash_base < crash_size)
574 continue;
575 if (crash_base < sclp_get_hsa_size())
576 continue;
577 if (crash_base < (unsigned long) INITRD_START + INITRD_SIZE)
578 continue;
579 return crash_base;
580 }
581 *msg = "no suitable area found";
582 return 0;
583}
584
585/*
586 * Check if crash_base and crash_size is valid
587 */
588static int __init verify_crash_base(unsigned long crash_base,
589 unsigned long crash_size,
590 char **msg)
591{
592 struct mem_chunk *chunk;
593 int i;
594
595 /*
596 * Because we do the swap to zero, we must have at least 'crash_size'
597 * bytes free space before crash_base
598 */
599 if (crash_size > crash_base) {
600 *msg = "crashkernel offset must be greater than size";
601 return -EINVAL;
602 }
603
604 /* First memory chunk must be at least crash_size */
605 if (memory_chunk[0].size < crash_size) {
606 *msg = "first memory chunk must be at least crashkernel size";
607 return -EINVAL;
608 }
609 /* Check if we fit into the respective memory chunk */
610 for (i = 0; i < MEMORY_CHUNKS; i++) {
611 chunk = &memory_chunk[i];
612 if (chunk->size == 0)
613 continue;
614 if (crash_base < chunk->addr)
615 continue;
616 if (crash_base >= chunk->addr + chunk->size)
617 continue;
618 /* we have found the memory chunk */
619 if (crash_base + crash_size > chunk->addr + chunk->size) {
620 *msg = "selected memory chunk is too small for "
621 "crashkernel memory";
622 return -EINVAL;
623 }
624 return 0;
625 }
626 *msg = "invalid memory range specified";
627 return -EINVAL;
628}
629
630/*
631 * When kdump is enabled, we have to ensure that no memory from 495 * When kdump is enabled, we have to ensure that no memory from
632 * the area [0 - crashkernel memory size] and 496 * the area [0 - crashkernel memory size] and
633 * [crashk_res.start - crashk_res.end] is set offline. 497 * [crashk_res.start - crashk_res.end] is set offline.
@@ -653,23 +517,44 @@ static struct notifier_block kdump_mem_nb = {
653#endif 517#endif
654 518
655/* 519/*
520 * Make sure that the area behind memory_end is protected
521 */
522static void reserve_memory_end(void)
523{
524#ifdef CONFIG_CRASH_DUMP
525 if (ipl_info.type == IPL_TYPE_FCP_DUMP &&
526 !OLDMEM_BASE && sclp_get_hsa_size()) {
527 memory_end = sclp_get_hsa_size();
528 memory_end &= PAGE_MASK;
529 memory_end_set = 1;
530 }
531#endif
532 if (!memory_end_set)
533 return;
534 memblock_reserve(memory_end, ULONG_MAX);
535}
536
537/*
656 * Make sure that oldmem, where the dump is stored, is protected 538 * Make sure that oldmem, where the dump is stored, is protected
657 */ 539 */
658static void reserve_oldmem(void) 540static void reserve_oldmem(void)
659{ 541{
660#ifdef CONFIG_CRASH_DUMP 542#ifdef CONFIG_CRASH_DUMP
661 unsigned long real_size = 0; 543 if (OLDMEM_BASE)
662 int i; 544 /* Forget all memory above the running kdump system */
663 545 memblock_reserve(OLDMEM_SIZE, (phys_addr_t)ULONG_MAX);
664 if (!OLDMEM_BASE) 546#endif
665 return; 547}
666 for (i = 0; i < MEMORY_CHUNKS; i++) {
667 struct mem_chunk *chunk = &memory_chunk[i];
668 548
669 real_size = max(real_size, chunk->addr + chunk->size); 549/*
670 } 550 * Make sure that oldmem, where the dump is stored, is protected
671 create_mem_hole(memory_chunk, OLDMEM_BASE, OLDMEM_SIZE); 551 */
672 create_mem_hole(memory_chunk, OLDMEM_SIZE, real_size - OLDMEM_SIZE); 552static void remove_oldmem(void)
553{
554#ifdef CONFIG_CRASH_DUMP
555 if (OLDMEM_BASE)
556 /* Forget all memory above the running kdump system */
557 memblock_remove(OLDMEM_SIZE, (phys_addr_t)ULONG_MAX);
673#endif 558#endif
674} 559}
675 560
@@ -680,167 +565,132 @@ static void __init reserve_crashkernel(void)
680{ 565{
681#ifdef CONFIG_CRASH_DUMP 566#ifdef CONFIG_CRASH_DUMP
682 unsigned long long crash_base, crash_size; 567 unsigned long long crash_base, crash_size;
683 char *msg = NULL; 568 phys_addr_t low, high;
684 int rc; 569 int rc;
685 570
686 rc = parse_crashkernel(boot_command_line, memory_end, &crash_size, 571 rc = parse_crashkernel(boot_command_line, memory_end, &crash_size,
687 &crash_base); 572 &crash_base);
688 if (rc || crash_size == 0) 573
689 return;
690 crash_base = ALIGN(crash_base, KEXEC_CRASH_MEM_ALIGN); 574 crash_base = ALIGN(crash_base, KEXEC_CRASH_MEM_ALIGN);
691 crash_size = ALIGN(crash_size, KEXEC_CRASH_MEM_ALIGN); 575 crash_size = ALIGN(crash_size, KEXEC_CRASH_MEM_ALIGN);
692 if (register_memory_notifier(&kdump_mem_nb)) 576 if (rc || crash_size == 0)
693 return; 577 return;
694 if (!crash_base) 578
695 crash_base = find_crash_base(crash_size, &msg); 579 if (memblock.memory.regions[0].size < crash_size) {
696 if (!crash_base) { 580 pr_info("crashkernel reservation failed: %s\n",
697 pr_info("crashkernel reservation failed: %s\n", msg); 581 "first memory chunk must be at least crashkernel size");
698 unregister_memory_notifier(&kdump_mem_nb);
699 return; 582 return;
700 } 583 }
701 if (verify_crash_base(crash_base, crash_size, &msg)) { 584
702 pr_info("crashkernel reservation failed: %s\n", msg); 585 low = crash_base ?: OLDMEM_BASE;
703 unregister_memory_notifier(&kdump_mem_nb); 586 high = low + crash_size;
587 if (low >= OLDMEM_BASE && high <= OLDMEM_BASE + OLDMEM_SIZE) {
588 /* The crashkernel fits into OLDMEM, reuse OLDMEM */
589 crash_base = low;
590 } else {
591 /* Find suitable area in free memory */
592 low = max_t(unsigned long, crash_size, sclp_get_hsa_size());
593 high = crash_base ? crash_base + crash_size : ULONG_MAX;
594
595 if (crash_base && crash_base < low) {
596 pr_info("crashkernel reservation failed: %s\n",
597 "crash_base too low");
598 return;
599 }
600 low = crash_base ?: low;
601 crash_base = memblock_find_in_range(low, high, crash_size,
602 KEXEC_CRASH_MEM_ALIGN);
603 }
604
605 if (!crash_base) {
606 pr_info("crashkernel reservation failed: %s\n",
607 "no suitable area found");
704 return; 608 return;
705 } 609 }
610
611 if (register_memory_notifier(&kdump_mem_nb))
612 return;
613
706 if (!OLDMEM_BASE && MACHINE_IS_VM) 614 if (!OLDMEM_BASE && MACHINE_IS_VM)
707 diag10_range(PFN_DOWN(crash_base), PFN_DOWN(crash_size)); 615 diag10_range(PFN_DOWN(crash_base), PFN_DOWN(crash_size));
708 crashk_res.start = crash_base; 616 crashk_res.start = crash_base;
709 crashk_res.end = crash_base + crash_size - 1; 617 crashk_res.end = crash_base + crash_size - 1;
710 insert_resource(&iomem_resource, &crashk_res); 618 insert_resource(&iomem_resource, &crashk_res);
711 create_mem_hole(memory_chunk, crash_base, crash_size); 619 memblock_remove(crash_base, crash_size);
712 pr_info("Reserving %lluMB of memory at %lluMB " 620 pr_info("Reserving %lluMB of memory at %lluMB "
713 "for crashkernel (System RAM: %luMB)\n", 621 "for crashkernel (System RAM: %luMB)\n",
714 crash_size >> 20, crash_base >> 20, memory_end >> 20); 622 crash_size >> 20, crash_base >> 20,
623 (unsigned long)memblock.memory.total_size >> 20);
715 os_info_crashkernel_add(crash_base, crash_size); 624 os_info_crashkernel_add(crash_base, crash_size);
716#endif 625#endif
717} 626}
718 627
719static void __init setup_memory(void) 628/*
629 * Reserve the initrd from being used by memblock
630 */
631static void __init reserve_initrd(void)
720{ 632{
721 unsigned long bootmap_size; 633#ifdef CONFIG_BLK_DEV_INITRD
722 unsigned long start_pfn, end_pfn; 634 initrd_start = INITRD_START;
723 int i; 635 initrd_end = initrd_start + INITRD_SIZE;
636 memblock_reserve(INITRD_START, INITRD_SIZE);
637#endif
638}
724 639
725 /* 640/*
726 * partially used pages are not usable - thus 641 * Check for initrd being in usable memory
727 * we are rounding upwards: 642 */
728 */ 643static void __init check_initrd(void)
644{
645#ifdef CONFIG_BLK_DEV_INITRD
646 if (INITRD_START && INITRD_SIZE &&
647 !memblock_is_region_memory(INITRD_START, INITRD_SIZE)) {
648 pr_err("initrd does not fit memory.\n");
649 memblock_free(INITRD_START, INITRD_SIZE);
650 initrd_start = initrd_end = 0;
651 }
652#endif
653}
654
655/*
656 * Reserve all kernel text
657 */
658static void __init reserve_kernel(void)
659{
660 unsigned long start_pfn;
729 start_pfn = PFN_UP(__pa(&_end)); 661 start_pfn = PFN_UP(__pa(&_end));
730 end_pfn = max_pfn = PFN_DOWN(memory_end);
731 662
732#ifdef CONFIG_BLK_DEV_INITRD
733 /* 663 /*
734 * Move the initrd in case the bitmap of the bootmem allocater 664 * Reserve memory used for lowcore/command line/kernel image.
735 * would overwrite it.
736 */ 665 */
666 memblock_reserve(0, (unsigned long)_ehead);
667 memblock_reserve((unsigned long)_stext, PFN_PHYS(start_pfn)
668 - (unsigned long)_stext);
669}
737 670
738 if (INITRD_START && INITRD_SIZE) { 671static void __init reserve_elfcorehdr(void)
739 unsigned long bmap_size; 672{
740 unsigned long start;
741
742 bmap_size = bootmem_bootmap_pages(end_pfn - start_pfn + 1);
743 bmap_size = PFN_PHYS(bmap_size);
744
745 if (PFN_PHYS(start_pfn) + bmap_size > INITRD_START) {
746 start = PFN_PHYS(start_pfn) + bmap_size + PAGE_SIZE;
747
748#ifdef CONFIG_CRASH_DUMP 673#ifdef CONFIG_CRASH_DUMP
749 if (OLDMEM_BASE) { 674 if (is_kdump_kernel())
750 /* Move initrd behind kdump oldmem */ 675 memblock_reserve(elfcorehdr_addr - OLDMEM_BASE,
751 if (start + INITRD_SIZE > OLDMEM_BASE && 676 PAGE_ALIGN(elfcorehdr_size));
752 start < OLDMEM_BASE + OLDMEM_SIZE)
753 start = OLDMEM_BASE + OLDMEM_SIZE;
754 }
755#endif
756 if (start + INITRD_SIZE > memory_end) {
757 pr_err("initrd extends beyond end of "
758 "memory (0x%08lx > 0x%08lx) "
759 "disabling initrd\n",
760 start + INITRD_SIZE, memory_end);
761 INITRD_START = INITRD_SIZE = 0;
762 } else {
763 pr_info("Moving initrd (0x%08lx -> "
764 "0x%08lx, size: %ld)\n",
765 INITRD_START, start, INITRD_SIZE);
766 memmove((void *) start, (void *) INITRD_START,
767 INITRD_SIZE);
768 INITRD_START = start;
769 }
770 }
771 }
772#endif 677#endif
678}
773 679
774 /* 680static void __init setup_memory(void)
775 * Initialize the boot-time allocator 681{
776 */ 682 struct memblock_region *reg;
777 bootmap_size = init_bootmem(start_pfn, end_pfn);
778 683
779 /* 684 /*
780 * Register RAM areas with the bootmem allocator. 685 * Init storage key for present memory
781 */ 686 */
782 687 for_each_memblock(memory, reg) {
783 for (i = 0; i < MEMORY_CHUNKS; i++) { 688 storage_key_init_range(reg->base, reg->base + reg->size);
784 unsigned long start_chunk, end_chunk, pfn;
785
786 if (!memory_chunk[i].size)
787 continue;
788 start_chunk = PFN_DOWN(memory_chunk[i].addr);
789 end_chunk = start_chunk + PFN_DOWN(memory_chunk[i].size);
790 end_chunk = min(end_chunk, end_pfn);
791 if (start_chunk >= end_chunk)
792 continue;
793 memblock_add_node(PFN_PHYS(start_chunk),
794 PFN_PHYS(end_chunk - start_chunk), 0);
795 pfn = max(start_chunk, start_pfn);
796 storage_key_init_range(PFN_PHYS(pfn), PFN_PHYS(end_chunk));
797 } 689 }
798
799 psw_set_key(PAGE_DEFAULT_KEY); 690 psw_set_key(PAGE_DEFAULT_KEY);
800 691
801 free_bootmem_with_active_regions(0, max_pfn); 692 /* Only cosmetics */
802 693 memblock_enforce_memory_limit(memblock_end_of_DRAM());
803 /*
804 * Reserve memory used for lowcore/command line/kernel image.
805 */
806 reserve_bootmem(0, (unsigned long)_ehead, BOOTMEM_DEFAULT);
807 reserve_bootmem((unsigned long)_stext,
808 PFN_PHYS(start_pfn) - (unsigned long)_stext,
809 BOOTMEM_DEFAULT);
810 /*
811 * Reserve the bootmem bitmap itself as well. We do this in two
812 * steps (first step was init_bootmem()) because this catches
813 * the (very unlikely) case of us accidentally initializing the
814 * bootmem allocator with an invalid RAM area.
815 */
816 reserve_bootmem(start_pfn << PAGE_SHIFT, bootmap_size,
817 BOOTMEM_DEFAULT);
818
819#ifdef CONFIG_CRASH_DUMP
820 if (crashk_res.start)
821 reserve_bootmem(crashk_res.start,
822 crashk_res.end - crashk_res.start + 1,
823 BOOTMEM_DEFAULT);
824 if (is_kdump_kernel())
825 reserve_bootmem(elfcorehdr_addr - OLDMEM_BASE,
826 PAGE_ALIGN(elfcorehdr_size), BOOTMEM_DEFAULT);
827#endif
828#ifdef CONFIG_BLK_DEV_INITRD
829 if (INITRD_START && INITRD_SIZE) {
830 if (INITRD_START + INITRD_SIZE <= memory_end) {
831 reserve_bootmem(INITRD_START, INITRD_SIZE,
832 BOOTMEM_DEFAULT);
833 initrd_start = INITRD_START;
834 initrd_end = initrd_start + INITRD_SIZE;
835 } else {
836 pr_err("initrd extends beyond end of "
837 "memory (0x%08lx > 0x%08lx) "
838 "disabling initrd\n",
839 initrd_start + INITRD_SIZE, memory_end);
840 initrd_start = initrd_end = 0;
841 }
842 }
843#endif
844} 694}
845 695
846/* 696/*
@@ -989,23 +839,46 @@ void __init setup_arch(char **cmdline_p)
989 839
990 ROOT_DEV = Root_RAM0; 840 ROOT_DEV = Root_RAM0;
991 841
842 /* Is init_mm really needed? */
992 init_mm.start_code = PAGE_OFFSET; 843 init_mm.start_code = PAGE_OFFSET;
993 init_mm.end_code = (unsigned long) &_etext; 844 init_mm.end_code = (unsigned long) &_etext;
994 init_mm.end_data = (unsigned long) &_edata; 845 init_mm.end_data = (unsigned long) &_edata;
995 init_mm.brk = (unsigned long) &_end; 846 init_mm.brk = (unsigned long) &_end;
996 847
997 parse_early_param(); 848 parse_early_param();
998 detect_memory_layout(memory_chunk, memory_end);
999 os_info_init(); 849 os_info_init();
1000 setup_ipl(); 850 setup_ipl();
851
852 /* Do some memory reservations *before* memory is added to memblock */
853 reserve_memory_end();
1001 reserve_oldmem(); 854 reserve_oldmem();
855 reserve_kernel();
856 reserve_initrd();
857 reserve_elfcorehdr();
858 memblock_allow_resize();
859
860 /* Get information about *all* installed memory */
861 detect_memory_memblock();
862
863 remove_oldmem();
864
865 /*
866 * Make sure all chunks are MAX_ORDER aligned so we don't need the
867 * extra checks that HOLES_IN_ZONE would require.
868 *
869 * Is this still required?
870 */
871 memblock_trim_memory(1UL << (MAX_ORDER - 1 + PAGE_SHIFT));
872
1002 setup_memory_end(); 873 setup_memory_end();
1003 reserve_crashkernel();
1004 setup_memory(); 874 setup_memory();
875
876 check_initrd();
877 reserve_crashkernel();
878
1005 setup_resources(); 879 setup_resources();
1006 setup_vmcoreinfo(); 880 setup_vmcoreinfo();
1007 setup_lowcore(); 881 setup_lowcore();
1008
1009 smp_fill_possible_mask(); 882 smp_fill_possible_mask();
1010 cpu_init(); 883 cpu_init();
1011 s390_init_cpu_topology(); 884 s390_init_cpu_topology();
diff --git a/arch/s390/kernel/signal.c b/arch/s390/kernel/signal.c
index d8fd508ccd1e..42b49f9e19bf 100644
--- a/arch/s390/kernel/signal.c
+++ b/arch/s390/kernel/signal.c
@@ -113,7 +113,7 @@ static int restore_sigregs(struct pt_regs *regs, _sigregs __user *sregs)
113 sizeof(current->thread.fp_regs)); 113 sizeof(current->thread.fp_regs));
114 114
115 restore_fp_regs(current->thread.fp_regs.fprs); 115 restore_fp_regs(current->thread.fp_regs.fprs);
116 clear_thread_flag(TIF_SYSCALL); /* No longer in a system call */ 116 clear_pt_regs_flag(regs, PIF_SYSCALL); /* No longer in a system call */
117 return 0; 117 return 0;
118} 118}
119 119
@@ -356,7 +356,7 @@ void do_signal(struct pt_regs *regs)
356 * call information. 356 * call information.
357 */ 357 */
358 current_thread_info()->system_call = 358 current_thread_info()->system_call =
359 test_thread_flag(TIF_SYSCALL) ? regs->int_code : 0; 359 test_pt_regs_flag(regs, PIF_SYSCALL) ? regs->int_code : 0;
360 signr = get_signal_to_deliver(&info, &ka, regs, NULL); 360 signr = get_signal_to_deliver(&info, &ka, regs, NULL);
361 361
362 if (signr > 0) { 362 if (signr > 0) {
@@ -384,7 +384,7 @@ void do_signal(struct pt_regs *regs)
384 } 384 }
385 } 385 }
386 /* No longer in a system call */ 386 /* No longer in a system call */
387 clear_thread_flag(TIF_SYSCALL); 387 clear_pt_regs_flag(regs, PIF_SYSCALL);
388 388
389 if (is_compat_task()) 389 if (is_compat_task())
390 handle_signal32(signr, &ka, &info, oldset, regs); 390 handle_signal32(signr, &ka, &info, oldset, regs);
@@ -394,7 +394,7 @@ void do_signal(struct pt_regs *regs)
394 } 394 }
395 395
396 /* No handlers present - check for system call restart */ 396 /* No handlers present - check for system call restart */
397 clear_thread_flag(TIF_SYSCALL); 397 clear_pt_regs_flag(regs, PIF_SYSCALL);
398 if (current_thread_info()->system_call) { 398 if (current_thread_info()->system_call) {
399 regs->int_code = current_thread_info()->system_call; 399 regs->int_code = current_thread_info()->system_call;
400 switch (regs->gprs[2]) { 400 switch (regs->gprs[2]) {
@@ -407,9 +407,9 @@ void do_signal(struct pt_regs *regs)
407 case -ERESTARTNOINTR: 407 case -ERESTARTNOINTR:
408 /* Restart system call with magic TIF bit. */ 408 /* Restart system call with magic TIF bit. */
409 regs->gprs[2] = regs->orig_gpr2; 409 regs->gprs[2] = regs->orig_gpr2;
410 set_thread_flag(TIF_SYSCALL); 410 set_pt_regs_flag(regs, PIF_SYSCALL);
411 if (test_thread_flag(TIF_SINGLE_STEP)) 411 if (test_thread_flag(TIF_SINGLE_STEP))
412 set_thread_flag(TIF_PER_TRAP); 412 clear_pt_regs_flag(regs, PIF_PER_TRAP);
413 break; 413 break;
414 } 414 }
415 } 415 }
diff --git a/arch/s390/kernel/smp.c b/arch/s390/kernel/smp.c
index 86e65ec3422b..243c7e512600 100644
--- a/arch/s390/kernel/smp.c
+++ b/arch/s390/kernel/smp.c
@@ -170,6 +170,7 @@ static int pcpu_alloc_lowcore(struct pcpu *pcpu, int cpu)
170 lc->panic_stack = pcpu->panic_stack + PAGE_SIZE 170 lc->panic_stack = pcpu->panic_stack + PAGE_SIZE
171 - STACK_FRAME_OVERHEAD - sizeof(struct pt_regs); 171 - STACK_FRAME_OVERHEAD - sizeof(struct pt_regs);
172 lc->cpu_nr = cpu; 172 lc->cpu_nr = cpu;
173 lc->spinlock_lockval = arch_spin_lockval(cpu);
173#ifndef CONFIG_64BIT 174#ifndef CONFIG_64BIT
174 if (MACHINE_HAS_IEEE) { 175 if (MACHINE_HAS_IEEE) {
175 lc->extended_save_area_addr = get_zeroed_page(GFP_KERNEL); 176 lc->extended_save_area_addr = get_zeroed_page(GFP_KERNEL);
@@ -226,6 +227,7 @@ static void pcpu_prepare_secondary(struct pcpu *pcpu, int cpu)
226 cpumask_set_cpu(cpu, mm_cpumask(&init_mm)); 227 cpumask_set_cpu(cpu, mm_cpumask(&init_mm));
227 atomic_inc(&init_mm.context.attach_count); 228 atomic_inc(&init_mm.context.attach_count);
228 lc->cpu_nr = cpu; 229 lc->cpu_nr = cpu;
230 lc->spinlock_lockval = arch_spin_lockval(cpu);
229 lc->percpu_offset = __per_cpu_offset[cpu]; 231 lc->percpu_offset = __per_cpu_offset[cpu];
230 lc->kernel_asce = S390_lowcore.kernel_asce; 232 lc->kernel_asce = S390_lowcore.kernel_asce;
231 lc->machine_flags = S390_lowcore.machine_flags; 233 lc->machine_flags = S390_lowcore.machine_flags;
@@ -403,15 +405,6 @@ void smp_send_stop(void)
403} 405}
404 406
405/* 407/*
406 * Stop the current cpu.
407 */
408void smp_stop_cpu(void)
409{
410 pcpu_sigp_retry(pcpu_devices + smp_processor_id(), SIGP_STOP, 0);
411 for (;;) ;
412}
413
414/*
415 * This is the main routine where commands issued by other 408 * This is the main routine where commands issued by other
416 * cpus are handled. 409 * cpus are handled.
417 */ 410 */
@@ -519,7 +512,7 @@ void smp_ctl_clear_bit(int cr, int bit)
519} 512}
520EXPORT_SYMBOL(smp_ctl_clear_bit); 513EXPORT_SYMBOL(smp_ctl_clear_bit);
521 514
522#if defined(CONFIG_ZFCPDUMP) || defined(CONFIG_CRASH_DUMP) 515#ifdef CONFIG_CRASH_DUMP
523 516
524static void __init smp_get_save_area(int cpu, u16 address) 517static void __init smp_get_save_area(int cpu, u16 address)
525{ 518{
@@ -534,14 +527,12 @@ static void __init smp_get_save_area(int cpu, u16 address)
534 save_area = dump_save_area_create(cpu); 527 save_area = dump_save_area_create(cpu);
535 if (!save_area) 528 if (!save_area)
536 panic("could not allocate memory for save area\n"); 529 panic("could not allocate memory for save area\n");
537#ifdef CONFIG_CRASH_DUMP
538 if (address == boot_cpu_address) { 530 if (address == boot_cpu_address) {
539 /* Copy the registers of the boot cpu. */ 531 /* Copy the registers of the boot cpu. */
540 copy_oldmem_page(1, (void *) save_area, sizeof(*save_area), 532 copy_oldmem_page(1, (void *) save_area, sizeof(*save_area),
541 SAVE_AREA_BASE - PAGE_SIZE, 0); 533 SAVE_AREA_BASE - PAGE_SIZE, 0);
542 return; 534 return;
543 } 535 }
544#endif
545 /* Get the registers of a non-boot cpu. */ 536 /* Get the registers of a non-boot cpu. */
546 __pcpu_sigp_relax(address, SIGP_STOP_AND_STORE_STATUS, 0, NULL); 537 __pcpu_sigp_relax(address, SIGP_STOP_AND_STORE_STATUS, 0, NULL);
547 memcpy_real(save_area, lc + SAVE_AREA_BASE, sizeof(*save_area)); 538 memcpy_real(save_area, lc + SAVE_AREA_BASE, sizeof(*save_area));
@@ -558,11 +549,11 @@ int smp_store_status(int cpu)
558 return 0; 549 return 0;
559} 550}
560 551
561#else /* CONFIG_ZFCPDUMP || CONFIG_CRASH_DUMP */ 552#else /* CONFIG_CRASH_DUMP */
562 553
563static inline void smp_get_save_area(int cpu, u16 address) { } 554static inline void smp_get_save_area(int cpu, u16 address) { }
564 555
565#endif /* CONFIG_ZFCPDUMP || CONFIG_CRASH_DUMP */ 556#endif /* CONFIG_CRASH_DUMP */
566 557
567void smp_cpu_set_polarization(int cpu, int val) 558void smp_cpu_set_polarization(int cpu, int val)
568{ 559{
@@ -809,6 +800,7 @@ void __init smp_cpus_done(unsigned int max_cpus)
809void __init smp_setup_processor_id(void) 800void __init smp_setup_processor_id(void)
810{ 801{
811 S390_lowcore.cpu_nr = 0; 802 S390_lowcore.cpu_nr = 0;
803 S390_lowcore.spinlock_lockval = arch_spin_lockval(0);
812} 804}
813 805
814/* 806/*
diff --git a/arch/s390/kernel/time.c b/arch/s390/kernel/time.c
index 386d37a228bb..0931b110c826 100644
--- a/arch/s390/kernel/time.c
+++ b/arch/s390/kernel/time.c
@@ -226,7 +226,7 @@ void update_vsyscall(struct timekeeper *tk)
226 vdso_data->wtom_clock_sec = 226 vdso_data->wtom_clock_sec =
227 tk->xtime_sec + tk->wall_to_monotonic.tv_sec; 227 tk->xtime_sec + tk->wall_to_monotonic.tv_sec;
228 vdso_data->wtom_clock_nsec = tk->xtime_nsec + 228 vdso_data->wtom_clock_nsec = tk->xtime_nsec +
229 + (tk->wall_to_monotonic.tv_nsec << tk->shift); 229 + ((u64) tk->wall_to_monotonic.tv_nsec << tk->shift);
230 nsecps = (u64) NSEC_PER_SEC << tk->shift; 230 nsecps = (u64) NSEC_PER_SEC << tk->shift;
231 while (vdso_data->wtom_clock_nsec >= nsecps) { 231 while (vdso_data->wtom_clock_nsec >= nsecps) {
232 vdso_data->wtom_clock_nsec -= nsecps; 232 vdso_data->wtom_clock_nsec -= nsecps;
diff --git a/arch/s390/kernel/topology.c b/arch/s390/kernel/topology.c
index 6298fed11ced..fa3b8cdaadac 100644
--- a/arch/s390/kernel/topology.c
+++ b/arch/s390/kernel/topology.c
@@ -333,7 +333,9 @@ static void __init alloc_masks(struct sysinfo_15_1_x *info,
333 nr_masks *= info->mag[TOPOLOGY_NR_MAG - offset - 1 - i]; 333 nr_masks *= info->mag[TOPOLOGY_NR_MAG - offset - 1 - i];
334 nr_masks = max(nr_masks, 1); 334 nr_masks = max(nr_masks, 1);
335 for (i = 0; i < nr_masks; i++) { 335 for (i = 0; i < nr_masks; i++) {
336 mask->next = alloc_bootmem(sizeof(struct mask_info)); 336 mask->next = alloc_bootmem_align(
337 roundup_pow_of_two(sizeof(struct mask_info)),
338 roundup_pow_of_two(sizeof(struct mask_info)));
337 mask = mask->next; 339 mask = mask->next;
338 } 340 }
339} 341}
diff --git a/arch/s390/kvm/kvm-s390.c b/arch/s390/kvm/kvm-s390.c
index 9ae6664ff08c..825fe7bf95a6 100644
--- a/arch/s390/kvm/kvm-s390.c
+++ b/arch/s390/kvm/kvm-s390.c
@@ -907,7 +907,7 @@ static int vcpu_pre_run(struct kvm_vcpu *vcpu)
907 if (need_resched()) 907 if (need_resched())
908 schedule(); 908 schedule();
909 909
910 if (test_thread_flag(TIF_MCCK_PENDING)) 910 if (test_cpu_flag(CIF_MCCK_PENDING))
911 s390_handle_mcck(); 911 s390_handle_mcck();
912 912
913 if (!kvm_is_ucontrol(vcpu->kvm)) 913 if (!kvm_is_ucontrol(vcpu->kvm))
diff --git a/arch/s390/lib/spinlock.c b/arch/s390/lib/spinlock.c
index f709983f41f8..5b0e445bc3f3 100644
--- a/arch/s390/lib/spinlock.c
+++ b/arch/s390/lib/spinlock.c
@@ -26,83 +26,81 @@ __setup("spin_retry=", spin_retry_setup);
26 26
27void arch_spin_lock_wait(arch_spinlock_t *lp) 27void arch_spin_lock_wait(arch_spinlock_t *lp)
28{ 28{
29 int count = spin_retry; 29 unsigned int cpu = SPINLOCK_LOCKVAL;
30 unsigned int cpu = ~smp_processor_id();
31 unsigned int owner; 30 unsigned int owner;
31 int count;
32 32
33 while (1) { 33 while (1) {
34 owner = lp->owner_cpu; 34 owner = ACCESS_ONCE(lp->lock);
35 if (!owner || smp_vcpu_scheduled(~owner)) { 35 /* Try to get the lock if it is free. */
36 for (count = spin_retry; count > 0; count--) { 36 if (!owner) {
37 if (arch_spin_is_locked(lp)) 37 if (_raw_compare_and_swap(&lp->lock, 0, cpu))
38 continue; 38 return;
39 if (_raw_compare_and_swap(&lp->owner_cpu, 0, 39 continue;
40 cpu) == 0)
41 return;
42 }
43 if (MACHINE_IS_LPAR)
44 continue;
45 } 40 }
46 owner = lp->owner_cpu; 41 /* Check if the lock owner is running. */
47 if (owner) 42 if (!smp_vcpu_scheduled(~owner)) {
43 smp_yield_cpu(~owner);
44 continue;
45 }
46 /* Loop for a while on the lock value. */
47 count = spin_retry;
48 do {
49 owner = ACCESS_ONCE(lp->lock);
50 } while (owner && count-- > 0);
51 if (!owner)
52 continue;
53 /*
54 * For multiple layers of hypervisors, e.g. z/VM + LPAR
55 * yield the CPU if the lock is still unavailable.
56 */
57 if (!MACHINE_IS_LPAR)
48 smp_yield_cpu(~owner); 58 smp_yield_cpu(~owner);
49 if (_raw_compare_and_swap(&lp->owner_cpu, 0, cpu) == 0)
50 return;
51 } 59 }
52} 60}
53EXPORT_SYMBOL(arch_spin_lock_wait); 61EXPORT_SYMBOL(arch_spin_lock_wait);
54 62
55void arch_spin_lock_wait_flags(arch_spinlock_t *lp, unsigned long flags) 63void arch_spin_lock_wait_flags(arch_spinlock_t *lp, unsigned long flags)
56{ 64{
57 int count = spin_retry; 65 unsigned int cpu = SPINLOCK_LOCKVAL;
58 unsigned int cpu = ~smp_processor_id();
59 unsigned int owner; 66 unsigned int owner;
67 int count;
60 68
61 local_irq_restore(flags); 69 local_irq_restore(flags);
62 while (1) { 70 while (1) {
63 owner = lp->owner_cpu; 71 owner = ACCESS_ONCE(lp->lock);
64 if (!owner || smp_vcpu_scheduled(~owner)) { 72 /* Try to get the lock if it is free. */
65 for (count = spin_retry; count > 0; count--) { 73 if (!owner) {
66 if (arch_spin_is_locked(lp)) 74 local_irq_disable();
67 continue; 75 if (_raw_compare_and_swap(&lp->lock, 0, cpu))
68 local_irq_disable(); 76 return;
69 if (_raw_compare_and_swap(&lp->owner_cpu, 0, 77 local_irq_restore(flags);
70 cpu) == 0)
71 return;
72 local_irq_restore(flags);
73 }
74 if (MACHINE_IS_LPAR)
75 continue;
76 } 78 }
77 owner = lp->owner_cpu; 79 /* Check if the lock owner is running. */
78 if (owner) 80 if (!smp_vcpu_scheduled(~owner)) {
79 smp_yield_cpu(~owner); 81 smp_yield_cpu(~owner);
80 local_irq_disable();
81 if (_raw_compare_and_swap(&lp->owner_cpu, 0, cpu) == 0)
82 return;
83 local_irq_restore(flags);
84 }
85}
86EXPORT_SYMBOL(arch_spin_lock_wait_flags);
87
88int arch_spin_trylock_retry(arch_spinlock_t *lp)
89{
90 unsigned int cpu = ~smp_processor_id();
91 int count;
92
93 for (count = spin_retry; count > 0; count--) {
94 if (arch_spin_is_locked(lp))
95 continue; 82 continue;
96 if (_raw_compare_and_swap(&lp->owner_cpu, 0, cpu) == 0) 83 }
97 return 1; 84 /* Loop for a while on the lock value. */
85 count = spin_retry;
86 do {
87 owner = ACCESS_ONCE(lp->lock);
88 } while (owner && count-- > 0);
89 if (!owner)
90 continue;
91 /*
92 * For multiple layers of hypervisors, e.g. z/VM + LPAR
93 * yield the CPU if the lock is still unavailable.
94 */
95 if (!MACHINE_IS_LPAR)
96 smp_yield_cpu(~owner);
98 } 97 }
99 return 0;
100} 98}
101EXPORT_SYMBOL(arch_spin_trylock_retry); 99EXPORT_SYMBOL(arch_spin_lock_wait_flags);
102 100
103void arch_spin_relax(arch_spinlock_t *lock) 101void arch_spin_relax(arch_spinlock_t *lp)
104{ 102{
105 unsigned int cpu = lock->owner_cpu; 103 unsigned int cpu = lp->lock;
106 if (cpu != 0) { 104 if (cpu != 0) {
107 if (MACHINE_IS_VM || MACHINE_IS_KVM || 105 if (MACHINE_IS_VM || MACHINE_IS_KVM ||
108 !smp_vcpu_scheduled(~cpu)) 106 !smp_vcpu_scheduled(~cpu))
@@ -111,6 +109,17 @@ void arch_spin_relax(arch_spinlock_t *lock)
111} 109}
112EXPORT_SYMBOL(arch_spin_relax); 110EXPORT_SYMBOL(arch_spin_relax);
113 111
112int arch_spin_trylock_retry(arch_spinlock_t *lp)
113{
114 int count;
115
116 for (count = spin_retry; count > 0; count--)
117 if (arch_spin_trylock_once(lp))
118 return 1;
119 return 0;
120}
121EXPORT_SYMBOL(arch_spin_trylock_retry);
122
114void _raw_read_lock_wait(arch_rwlock_t *rw) 123void _raw_read_lock_wait(arch_rwlock_t *rw)
115{ 124{
116 unsigned int old; 125 unsigned int old;
@@ -121,10 +130,10 @@ void _raw_read_lock_wait(arch_rwlock_t *rw)
121 smp_yield(); 130 smp_yield();
122 count = spin_retry; 131 count = spin_retry;
123 } 132 }
124 if (!arch_read_can_lock(rw)) 133 old = ACCESS_ONCE(rw->lock);
134 if ((int) old < 0)
125 continue; 135 continue;
126 old = rw->lock & 0x7fffffffU; 136 if (_raw_compare_and_swap(&rw->lock, old, old + 1))
127 if (_raw_compare_and_swap(&rw->lock, old, old + 1) == old)
128 return; 137 return;
129 } 138 }
130} 139}
@@ -141,12 +150,13 @@ void _raw_read_lock_wait_flags(arch_rwlock_t *rw, unsigned long flags)
141 smp_yield(); 150 smp_yield();
142 count = spin_retry; 151 count = spin_retry;
143 } 152 }
144 if (!arch_read_can_lock(rw)) 153 old = ACCESS_ONCE(rw->lock);
154 if ((int) old < 0)
145 continue; 155 continue;
146 old = rw->lock & 0x7fffffffU;
147 local_irq_disable(); 156 local_irq_disable();
148 if (_raw_compare_and_swap(&rw->lock, old, old + 1) == old) 157 if (_raw_compare_and_swap(&rw->lock, old, old + 1))
149 return; 158 return;
159 local_irq_restore(flags);
150 } 160 }
151} 161}
152EXPORT_SYMBOL(_raw_read_lock_wait_flags); 162EXPORT_SYMBOL(_raw_read_lock_wait_flags);
@@ -157,10 +167,10 @@ int _raw_read_trylock_retry(arch_rwlock_t *rw)
157 int count = spin_retry; 167 int count = spin_retry;
158 168
159 while (count-- > 0) { 169 while (count-- > 0) {
160 if (!arch_read_can_lock(rw)) 170 old = ACCESS_ONCE(rw->lock);
171 if ((int) old < 0)
161 continue; 172 continue;
162 old = rw->lock & 0x7fffffffU; 173 if (_raw_compare_and_swap(&rw->lock, old, old + 1))
163 if (_raw_compare_and_swap(&rw->lock, old, old + 1) == old)
164 return 1; 174 return 1;
165 } 175 }
166 return 0; 176 return 0;
@@ -169,6 +179,7 @@ EXPORT_SYMBOL(_raw_read_trylock_retry);
169 179
170void _raw_write_lock_wait(arch_rwlock_t *rw) 180void _raw_write_lock_wait(arch_rwlock_t *rw)
171{ 181{
182 unsigned int old;
172 int count = spin_retry; 183 int count = spin_retry;
173 184
174 while (1) { 185 while (1) {
@@ -176,9 +187,10 @@ void _raw_write_lock_wait(arch_rwlock_t *rw)
176 smp_yield(); 187 smp_yield();
177 count = spin_retry; 188 count = spin_retry;
178 } 189 }
179 if (!arch_write_can_lock(rw)) 190 old = ACCESS_ONCE(rw->lock);
191 if (old)
180 continue; 192 continue;
181 if (_raw_compare_and_swap(&rw->lock, 0, 0x80000000) == 0) 193 if (_raw_compare_and_swap(&rw->lock, 0, 0x80000000))
182 return; 194 return;
183 } 195 }
184} 196}
@@ -186,6 +198,7 @@ EXPORT_SYMBOL(_raw_write_lock_wait);
186 198
187void _raw_write_lock_wait_flags(arch_rwlock_t *rw, unsigned long flags) 199void _raw_write_lock_wait_flags(arch_rwlock_t *rw, unsigned long flags)
188{ 200{
201 unsigned int old;
189 int count = spin_retry; 202 int count = spin_retry;
190 203
191 local_irq_restore(flags); 204 local_irq_restore(flags);
@@ -194,23 +207,27 @@ void _raw_write_lock_wait_flags(arch_rwlock_t *rw, unsigned long flags)
194 smp_yield(); 207 smp_yield();
195 count = spin_retry; 208 count = spin_retry;
196 } 209 }
197 if (!arch_write_can_lock(rw)) 210 old = ACCESS_ONCE(rw->lock);
211 if (old)
198 continue; 212 continue;
199 local_irq_disable(); 213 local_irq_disable();
200 if (_raw_compare_and_swap(&rw->lock, 0, 0x80000000) == 0) 214 if (_raw_compare_and_swap(&rw->lock, 0, 0x80000000))
201 return; 215 return;
216 local_irq_restore(flags);
202 } 217 }
203} 218}
204EXPORT_SYMBOL(_raw_write_lock_wait_flags); 219EXPORT_SYMBOL(_raw_write_lock_wait_flags);
205 220
206int _raw_write_trylock_retry(arch_rwlock_t *rw) 221int _raw_write_trylock_retry(arch_rwlock_t *rw)
207{ 222{
223 unsigned int old;
208 int count = spin_retry; 224 int count = spin_retry;
209 225
210 while (count-- > 0) { 226 while (count-- > 0) {
211 if (!arch_write_can_lock(rw)) 227 old = ACCESS_ONCE(rw->lock);
228 if (old)
212 continue; 229 continue;
213 if (_raw_compare_and_swap(&rw->lock, 0, 0x80000000) == 0) 230 if (_raw_compare_and_swap(&rw->lock, 0, 0x80000000))
214 return 1; 231 return 1;
215 } 232 }
216 return 0; 233 return 0;
diff --git a/arch/s390/lib/uaccess.c b/arch/s390/lib/uaccess.c
index 7416efe8eae4..53dd5d7a0c96 100644
--- a/arch/s390/lib/uaccess.c
+++ b/arch/s390/lib/uaccess.c
@@ -76,7 +76,7 @@ static inline unsigned long copy_from_user_mvcp(void *x, const void __user *ptr,
76{ 76{
77 unsigned long tmp1, tmp2; 77 unsigned long tmp1, tmp2;
78 78
79 update_primary_asce(current); 79 load_kernel_asce();
80 tmp1 = -256UL; 80 tmp1 = -256UL;
81 asm volatile( 81 asm volatile(
82 " sacf 0\n" 82 " sacf 0\n"
@@ -159,7 +159,7 @@ static inline unsigned long copy_to_user_mvcs(void __user *ptr, const void *x,
159{ 159{
160 unsigned long tmp1, tmp2; 160 unsigned long tmp1, tmp2;
161 161
162 update_primary_asce(current); 162 load_kernel_asce();
163 tmp1 = -256UL; 163 tmp1 = -256UL;
164 asm volatile( 164 asm volatile(
165 " sacf 0\n" 165 " sacf 0\n"
@@ -225,7 +225,7 @@ static inline unsigned long copy_in_user_mvc(void __user *to, const void __user
225{ 225{
226 unsigned long tmp1; 226 unsigned long tmp1;
227 227
228 update_primary_asce(current); 228 load_kernel_asce();
229 asm volatile( 229 asm volatile(
230 " sacf 256\n" 230 " sacf 256\n"
231 " "AHI" %0,-1\n" 231 " "AHI" %0,-1\n"
@@ -292,7 +292,7 @@ static inline unsigned long clear_user_xc(void __user *to, unsigned long size)
292{ 292{
293 unsigned long tmp1, tmp2; 293 unsigned long tmp1, tmp2;
294 294
295 update_primary_asce(current); 295 load_kernel_asce();
296 asm volatile( 296 asm volatile(
297 " sacf 256\n" 297 " sacf 256\n"
298 " "AHI" %0,-1\n" 298 " "AHI" %0,-1\n"
@@ -358,7 +358,7 @@ unsigned long __strnlen_user(const char __user *src, unsigned long size)
358{ 358{
359 if (unlikely(!size)) 359 if (unlikely(!size))
360 return 0; 360 return 0;
361 update_primary_asce(current); 361 load_kernel_asce();
362 return strnlen_user_srst(src, size); 362 return strnlen_user_srst(src, size);
363} 363}
364EXPORT_SYMBOL(__strnlen_user); 364EXPORT_SYMBOL(__strnlen_user);
diff --git a/arch/s390/mm/fault.c b/arch/s390/mm/fault.c
index 2f51a998a67e..3f3b35403d0a 100644
--- a/arch/s390/mm/fault.c
+++ b/arch/s390/mm/fault.c
@@ -415,7 +415,7 @@ static inline int do_exception(struct pt_regs *regs, int access)
415 * The instruction that caused the program check has 415 * The instruction that caused the program check has
416 * been nullified. Don't signal single step via SIGTRAP. 416 * been nullified. Don't signal single step via SIGTRAP.
417 */ 417 */
418 clear_tsk_thread_flag(tsk, TIF_PER_TRAP); 418 clear_pt_regs_flag(regs, PIF_PER_TRAP);
419 419
420 if (notify_page_fault(regs)) 420 if (notify_page_fault(regs))
421 return 0; 421 return 0;
diff --git a/arch/s390/mm/mem_detect.c b/arch/s390/mm/mem_detect.c
index cca388253a39..5535cfe0ee11 100644
--- a/arch/s390/mm/mem_detect.c
+++ b/arch/s390/mm/mem_detect.c
@@ -6,130 +6,60 @@
6 6
7#include <linux/kernel.h> 7#include <linux/kernel.h>
8#include <linux/module.h> 8#include <linux/module.h>
9#include <linux/memblock.h>
10#include <linux/init.h>
11#include <linux/debugfs.h>
12#include <linux/seq_file.h>
9#include <asm/ipl.h> 13#include <asm/ipl.h>
10#include <asm/sclp.h> 14#include <asm/sclp.h>
11#include <asm/setup.h> 15#include <asm/setup.h>
12 16
13#define ADDR2G (1ULL << 31) 17#define ADDR2G (1ULL << 31)
14 18
15static void find_memory_chunks(struct mem_chunk chunk[], unsigned long maxsize) 19#define CHUNK_READ_WRITE 0
20#define CHUNK_READ_ONLY 1
21
22static inline void memblock_physmem_add(phys_addr_t start, phys_addr_t size)
23{
24 memblock_add_range(&memblock.memory, start, size, 0, 0);
25 memblock_add_range(&memblock.physmem, start, size, 0, 0);
26}
27
28void __init detect_memory_memblock(void)
16{ 29{
17 unsigned long long memsize, rnmax, rzm; 30 unsigned long long memsize, rnmax, rzm;
18 unsigned long addr = 0, size; 31 unsigned long addr, size;
19 int i = 0, type; 32 int type;
20 33
21 rzm = sclp_get_rzm(); 34 rzm = sclp_get_rzm();
22 rnmax = sclp_get_rnmax(); 35 rnmax = sclp_get_rnmax();
23 memsize = rzm * rnmax; 36 memsize = rzm * rnmax;
24 if (!rzm) 37 if (!rzm)
25 rzm = 1ULL << 17; 38 rzm = 1ULL << 17;
26 if (sizeof(long) == 4) { 39 if (IS_ENABLED(CONFIG_32BIT)) {
27 rzm = min(ADDR2G, rzm); 40 rzm = min(ADDR2G, rzm);
28 memsize = memsize ? min(ADDR2G, memsize) : ADDR2G; 41 memsize = min(ADDR2G, memsize);
29 } 42 }
30 if (maxsize) 43 max_physmem_end = memsize;
31 memsize = memsize ? min((unsigned long)memsize, maxsize) : maxsize; 44 addr = 0;
45 /* keep memblock lists close to the kernel */
46 memblock_set_bottom_up(true);
32 do { 47 do {
33 size = 0; 48 size = 0;
34 type = tprot(addr); 49 type = tprot(addr);
35 do { 50 do {
36 size += rzm; 51 size += rzm;
37 if (memsize && addr + size >= memsize) 52 if (max_physmem_end && addr + size >= max_physmem_end)
38 break; 53 break;
39 } while (type == tprot(addr + size)); 54 } while (type == tprot(addr + size));
40 if (type == CHUNK_READ_WRITE || type == CHUNK_READ_ONLY) { 55 if (type == CHUNK_READ_WRITE || type == CHUNK_READ_ONLY) {
41 if (memsize && (addr + size > memsize)) 56 if (max_physmem_end && (addr + size > max_physmem_end))
42 size = memsize - addr; 57 size = max_physmem_end - addr;
43 chunk[i].addr = addr; 58 memblock_physmem_add(addr, size);
44 chunk[i].size = size;
45 chunk[i].type = type;
46 i++;
47 } 59 }
48 addr += size; 60 addr += size;
49 } while (addr < memsize && i < MEMORY_CHUNKS); 61 } while (addr < max_physmem_end);
50} 62 memblock_set_bottom_up(false);
51 63 if (!max_physmem_end)
52/** 64 max_physmem_end = memblock_end_of_DRAM();
53 * detect_memory_layout - fill mem_chunk array with memory layout data
54 * @chunk: mem_chunk array to be filled
55 * @maxsize: maximum address where memory detection should stop
56 *
57 * Fills the passed in memory chunk array with the memory layout of the
58 * machine. The array must have a size of at least MEMORY_CHUNKS and will
59 * be fully initialized afterwards.
60 * If the maxsize paramater has a value > 0 memory detection will stop at
61 * that address. It is guaranteed that all chunks have an ending address
62 * that is smaller than maxsize.
63 * If maxsize is 0 all memory will be detected.
64 */
65void detect_memory_layout(struct mem_chunk chunk[], unsigned long maxsize)
66{
67 unsigned long flags, flags_dat, cr0;
68
69 memset(chunk, 0, MEMORY_CHUNKS * sizeof(struct mem_chunk));
70 /*
71 * Disable IRQs, DAT and low address protection so tprot does the
72 * right thing and we don't get scheduled away with low address
73 * protection disabled.
74 */
75 local_irq_save(flags);
76 flags_dat = __arch_local_irq_stnsm(0xfb);
77 /*
78 * In case DAT was enabled, make sure chunk doesn't reside in vmalloc
79 * space. We have disabled DAT and any access to vmalloc area will
80 * cause an exception.
81 * If DAT was disabled we are called from early ipl code.
82 */
83 if (test_bit(5, &flags_dat)) {
84 if (WARN_ON_ONCE(is_vmalloc_or_module_addr(chunk)))
85 goto out;
86 }
87 __ctl_store(cr0, 0, 0);
88 __ctl_clear_bit(0, 28);
89 find_memory_chunks(chunk, maxsize);
90 __ctl_load(cr0, 0, 0);
91out:
92 __arch_local_irq_ssm(flags_dat);
93 local_irq_restore(flags);
94}
95EXPORT_SYMBOL(detect_memory_layout);
96
97/*
98 * Create memory hole with given address and size.
99 */
100void create_mem_hole(struct mem_chunk mem_chunk[], unsigned long addr,
101 unsigned long size)
102{
103 int i;
104
105 for (i = 0; i < MEMORY_CHUNKS; i++) {
106 struct mem_chunk *chunk = &mem_chunk[i];
107
108 if (chunk->size == 0)
109 continue;
110 if (addr > chunk->addr + chunk->size)
111 continue;
112 if (addr + size <= chunk->addr)
113 continue;
114 /* Split */
115 if ((addr > chunk->addr) &&
116 (addr + size < chunk->addr + chunk->size)) {
117 struct mem_chunk *new = chunk + 1;
118
119 memmove(new, chunk, (MEMORY_CHUNKS-i-1) * sizeof(*new));
120 new->addr = addr + size;
121 new->size = chunk->addr + chunk->size - new->addr;
122 chunk->size = addr - chunk->addr;
123 continue;
124 } else if ((addr <= chunk->addr) &&
125 (addr + size >= chunk->addr + chunk->size)) {
126 memmove(chunk, chunk + 1, (MEMORY_CHUNKS-i-1) * sizeof(*chunk));
127 memset(&mem_chunk[MEMORY_CHUNKS-1], 0, sizeof(*chunk));
128 } else if (addr + size < chunk->addr + chunk->size) {
129 chunk->size = chunk->addr + chunk->size - addr - size;
130 chunk->addr = addr + size;
131 } else if (addr > chunk->addr) {
132 chunk->size = addr - chunk->addr;
133 }
134 }
135} 65}
diff --git a/arch/s390/mm/page-states.c b/arch/s390/mm/page-states.c
index 27c50f4d90cb..a90d45e9dfb0 100644
--- a/arch/s390/mm/page-states.c
+++ b/arch/s390/mm/page-states.c
@@ -12,8 +12,6 @@
12#include <linux/mm.h> 12#include <linux/mm.h>
13#include <linux/gfp.h> 13#include <linux/gfp.h>
14#include <linux/init.h> 14#include <linux/init.h>
15#include <asm/setup.h>
16#include <asm/ipl.h>
17 15
18#define ESSA_SET_STABLE 1 16#define ESSA_SET_STABLE 1
19#define ESSA_SET_UNUSED 2 17#define ESSA_SET_UNUSED 2
@@ -43,14 +41,6 @@ void __init cmma_init(void)
43 41
44 if (!cmma_flag) 42 if (!cmma_flag)
45 return; 43 return;
46 /*
47 * Disable CMM for dump, otherwise the tprot based memory
48 * detection can fail because of unstable pages.
49 */
50 if (OLDMEM_BASE || ipl_info.type == IPL_TYPE_FCP_DUMP) {
51 cmma_flag = 0;
52 return;
53 }
54 asm volatile( 44 asm volatile(
55 " .insn rrf,0xb9ab0000,%1,%1,0,0\n" 45 " .insn rrf,0xb9ab0000,%1,%1,0,0\n"
56 "0: la %0,0\n" 46 "0: la %0,0\n"
diff --git a/arch/s390/mm/pgtable.c b/arch/s390/mm/pgtable.c
index d7cfd57815fb..7881d4eb8b6b 100644
--- a/arch/s390/mm/pgtable.c
+++ b/arch/s390/mm/pgtable.c
@@ -53,8 +53,10 @@ static void __crst_table_upgrade(void *arg)
53{ 53{
54 struct mm_struct *mm = arg; 54 struct mm_struct *mm = arg;
55 55
56 if (current->active_mm == mm) 56 if (current->active_mm == mm) {
57 update_user_asce(mm, 1); 57 clear_user_asce();
58 set_user_asce(mm);
59 }
58 __tlb_flush_local(); 60 __tlb_flush_local();
59} 61}
60 62
@@ -108,7 +110,7 @@ void crst_table_downgrade(struct mm_struct *mm, unsigned long limit)
108 pgd_t *pgd; 110 pgd_t *pgd;
109 111
110 if (current->active_mm == mm) { 112 if (current->active_mm == mm) {
111 clear_user_asce(mm, 1); 113 clear_user_asce();
112 __tlb_flush_mm(mm); 114 __tlb_flush_mm(mm);
113 } 115 }
114 while (mm->context.asce_limit > limit) { 116 while (mm->context.asce_limit > limit) {
@@ -134,7 +136,7 @@ void crst_table_downgrade(struct mm_struct *mm, unsigned long limit)
134 crst_table_free(mm, (unsigned long *) pgd); 136 crst_table_free(mm, (unsigned long *) pgd);
135 } 137 }
136 if (current->active_mm == mm) 138 if (current->active_mm == mm)
137 update_user_asce(mm, 1); 139 set_user_asce(mm);
138} 140}
139#endif 141#endif
140 142
diff --git a/arch/s390/mm/vmem.c b/arch/s390/mm/vmem.c
index 72b04de18283..fe9012a49aa5 100644
--- a/arch/s390/mm/vmem.c
+++ b/arch/s390/mm/vmem.c
@@ -10,6 +10,7 @@
10#include <linux/list.h> 10#include <linux/list.h>
11#include <linux/hugetlb.h> 11#include <linux/hugetlb.h>
12#include <linux/slab.h> 12#include <linux/slab.h>
13#include <linux/memblock.h>
13#include <asm/pgalloc.h> 14#include <asm/pgalloc.h>
14#include <asm/pgtable.h> 15#include <asm/pgtable.h>
15#include <asm/setup.h> 16#include <asm/setup.h>
@@ -66,7 +67,8 @@ static pte_t __ref *vmem_pte_alloc(unsigned long address)
66 if (slab_is_available()) 67 if (slab_is_available())
67 pte = (pte_t *) page_table_alloc(&init_mm, address); 68 pte = (pte_t *) page_table_alloc(&init_mm, address);
68 else 69 else
69 pte = alloc_bootmem(PTRS_PER_PTE * sizeof(pte_t)); 70 pte = alloc_bootmem_align(PTRS_PER_PTE * sizeof(pte_t),
71 PTRS_PER_PTE * sizeof(pte_t));
70 if (!pte) 72 if (!pte)
71 return NULL; 73 return NULL;
72 clear_table((unsigned long *) pte, _PAGE_INVALID, 74 clear_table((unsigned long *) pte, _PAGE_INVALID,
@@ -371,16 +373,14 @@ out:
371void __init vmem_map_init(void) 373void __init vmem_map_init(void)
372{ 374{
373 unsigned long ro_start, ro_end; 375 unsigned long ro_start, ro_end;
374 unsigned long start, end; 376 struct memblock_region *reg;
375 int i; 377 phys_addr_t start, end;
376 378
377 ro_start = PFN_ALIGN((unsigned long)&_stext); 379 ro_start = PFN_ALIGN((unsigned long)&_stext);
378 ro_end = (unsigned long)&_eshared & PAGE_MASK; 380 ro_end = (unsigned long)&_eshared & PAGE_MASK;
379 for (i = 0; i < MEMORY_CHUNKS; i++) { 381 for_each_memblock(memory, reg) {
380 if (!memory_chunk[i].size) 382 start = reg->base;
381 continue; 383 end = reg->base + reg->size - 1;
382 start = memory_chunk[i].addr;
383 end = memory_chunk[i].addr + memory_chunk[i].size;
384 if (start >= ro_end || end <= ro_start) 384 if (start >= ro_end || end <= ro_start)
385 vmem_add_mem(start, end - start, 0); 385 vmem_add_mem(start, end - start, 0);
386 else if (start >= ro_start && end <= ro_end) 386 else if (start >= ro_start && end <= ro_end)
@@ -400,23 +400,21 @@ void __init vmem_map_init(void)
400} 400}
401 401
402/* 402/*
403 * Convert memory chunk array to a memory segment list so there is a single 403 * Convert memblock.memory to a memory segment list so there is a single
404 * list that contains both r/w memory and shared memory segments. 404 * list that contains all memory segments.
405 */ 405 */
406static int __init vmem_convert_memory_chunk(void) 406static int __init vmem_convert_memory_chunk(void)
407{ 407{
408 struct memblock_region *reg;
408 struct memory_segment *seg; 409 struct memory_segment *seg;
409 int i;
410 410
411 mutex_lock(&vmem_mutex); 411 mutex_lock(&vmem_mutex);
412 for (i = 0; i < MEMORY_CHUNKS; i++) { 412 for_each_memblock(memory, reg) {
413 if (!memory_chunk[i].size)
414 continue;
415 seg = kzalloc(sizeof(*seg), GFP_KERNEL); 413 seg = kzalloc(sizeof(*seg), GFP_KERNEL);
416 if (!seg) 414 if (!seg)
417 panic("Out of memory...\n"); 415 panic("Out of memory...\n");
418 seg->start = memory_chunk[i].addr; 416 seg->start = reg->base;
419 seg->size = memory_chunk[i].size; 417 seg->size = reg->size;
420 insert_memory_segment(seg); 418 insert_memory_segment(seg);
421 } 419 }
422 mutex_unlock(&vmem_mutex); 420 mutex_unlock(&vmem_mutex);
diff --git a/arch/s390/oprofile/hwsampler.c b/arch/s390/oprofile/hwsampler.c
index 276f2e26c761..e53c6f268807 100644
--- a/arch/s390/oprofile/hwsampler.c
+++ b/arch/s390/oprofile/hwsampler.c
@@ -209,13 +209,11 @@ static void init_all_cpu_buffers(void)
209 } 209 }
210} 210}
211 211
212static int prepare_cpu_buffers(void) 212static void prepare_cpu_buffers(void)
213{ 213{
214 int cpu;
215 int rc;
216 struct hws_cpu_buffer *cb; 214 struct hws_cpu_buffer *cb;
215 int cpu;
217 216
218 rc = 0;
219 for_each_online_cpu(cpu) { 217 for_each_online_cpu(cpu) {
220 cb = &per_cpu(sampler_cpu_buffer, cpu); 218 cb = &per_cpu(sampler_cpu_buffer, cpu);
221 atomic_set(&cb->ext_params, 0); 219 atomic_set(&cb->ext_params, 0);
@@ -230,8 +228,6 @@ static int prepare_cpu_buffers(void)
230 cb->oom = 0; 228 cb->oom = 0;
231 cb->stop_mode = 0; 229 cb->stop_mode = 0;
232 } 230 }
233
234 return rc;
235} 231}
236 232
237/* 233/*
@@ -1107,9 +1103,7 @@ int hwsampler_start_all(unsigned long rate)
1107 if (rc) 1103 if (rc)
1108 goto start_all_exit; 1104 goto start_all_exit;
1109 1105
1110 rc = prepare_cpu_buffers(); 1106 prepare_cpu_buffers();
1111 if (rc)
1112 goto start_all_exit;
1113 1107
1114 for_each_online_cpu(cpu) { 1108 for_each_online_cpu(cpu) {
1115 rc = start_sampling(cpu); 1109 rc = start_sampling(cpu);
@@ -1156,7 +1150,7 @@ int hwsampler_stop_all(void)
1156 rc = 0; 1150 rc = 0;
1157 if (hws_state == HWS_INIT) { 1151 if (hws_state == HWS_INIT) {
1158 mutex_unlock(&hws_sem); 1152 mutex_unlock(&hws_sem);
1159 return rc; 1153 return 0;
1160 } 1154 }
1161 hws_state = HWS_STOPPING; 1155 hws_state = HWS_STOPPING;
1162 mutex_unlock(&hws_sem); 1156 mutex_unlock(&hws_sem);
diff --git a/arch/s390/pci/pci_clp.c b/arch/s390/pci/pci_clp.c
index c747394029ee..96545d7659fd 100644
--- a/arch/s390/pci/pci_clp.c
+++ b/arch/s390/pci/pci_clp.c
@@ -114,6 +114,16 @@ static int clp_store_query_pci_fn(struct zpci_dev *zdev,
114 zdev->end_dma = response->edma; 114 zdev->end_dma = response->edma;
115 zdev->pchid = response->pchid; 115 zdev->pchid = response->pchid;
116 zdev->pfgid = response->pfgid; 116 zdev->pfgid = response->pfgid;
117 zdev->pft = response->pft;
118 zdev->vfn = response->vfn;
119 zdev->uid = response->uid;
120
121 memcpy(zdev->pfip, response->pfip, sizeof(zdev->pfip));
122 if (response->util_str_avail) {
123 memcpy(zdev->util_str, response->util_str,
124 sizeof(zdev->util_str));
125 }
126
117 return 0; 127 return 0;
118} 128}
119 129
diff --git a/arch/s390/pci/pci_event.c b/arch/s390/pci/pci_event.c
index 01e251b1da0c..6d7f5a3016ca 100644
--- a/arch/s390/pci/pci_event.c
+++ b/arch/s390/pci/pci_event.c
@@ -76,7 +76,7 @@ static void __zpci_event_availability(struct zpci_ccdf_avail *ccdf)
76 76
77 switch (ccdf->pec) { 77 switch (ccdf->pec) {
78 case 0x0301: /* Standby -> Configured */ 78 case 0x0301: /* Standby -> Configured */
79 if (!zdev || zdev->state == ZPCI_FN_STATE_CONFIGURED) 79 if (!zdev || zdev->state != ZPCI_FN_STATE_STANDBY)
80 break; 80 break;
81 zdev->state = ZPCI_FN_STATE_CONFIGURED; 81 zdev->state = ZPCI_FN_STATE_CONFIGURED;
82 zdev->fh = ccdf->fh; 82 zdev->fh = ccdf->fh;
@@ -86,7 +86,8 @@ static void __zpci_event_availability(struct zpci_ccdf_avail *ccdf)
86 pci_rescan_bus(zdev->bus); 86 pci_rescan_bus(zdev->bus);
87 break; 87 break;
88 case 0x0302: /* Reserved -> Standby */ 88 case 0x0302: /* Reserved -> Standby */
89 clp_add_pci_device(ccdf->fid, ccdf->fh, 0); 89 if (!zdev)
90 clp_add_pci_device(ccdf->fid, ccdf->fh, 0);
90 break; 91 break;
91 case 0x0303: /* Deconfiguration requested */ 92 case 0x0303: /* Deconfiguration requested */
92 if (pdev) 93 if (pdev)
diff --git a/arch/s390/pci/pci_sysfs.c b/arch/s390/pci/pci_sysfs.c
index b56a3958f1a7..9190214b8702 100644
--- a/arch/s390/pci/pci_sysfs.c
+++ b/arch/s390/pci/pci_sysfs.c
@@ -12,43 +12,29 @@
12#include <linux/stat.h> 12#include <linux/stat.h>
13#include <linux/pci.h> 13#include <linux/pci.h>
14 14
15static ssize_t show_fid(struct device *dev, struct device_attribute *attr, 15#define zpci_attr(name, fmt, member) \
16 char *buf) 16static ssize_t name##_show(struct device *dev, \
17{ 17 struct device_attribute *attr, char *buf) \
18 struct zpci_dev *zdev = get_zdev(to_pci_dev(dev)); 18{ \
19 19 struct zpci_dev *zdev = get_zdev(to_pci_dev(dev)); \
20 return sprintf(buf, "0x%08x\n", zdev->fid); 20 \
21} 21 return sprintf(buf, fmt, zdev->member); \
22static DEVICE_ATTR(function_id, S_IRUGO, show_fid, NULL); 22} \
23 23static DEVICE_ATTR_RO(name)
24static ssize_t show_fh(struct device *dev, struct device_attribute *attr, 24
25 char *buf) 25zpci_attr(function_id, "0x%08x\n", fid);
26{ 26zpci_attr(function_handle, "0x%08x\n", fh);
27 struct zpci_dev *zdev = get_zdev(to_pci_dev(dev)); 27zpci_attr(pchid, "0x%04x\n", pchid);
28 28zpci_attr(pfgid, "0x%02x\n", pfgid);
29 return sprintf(buf, "0x%08x\n", zdev->fh); 29zpci_attr(vfn, "0x%04x\n", vfn);
30} 30zpci_attr(pft, "0x%02x\n", pft);
31static DEVICE_ATTR(function_handle, S_IRUGO, show_fh, NULL); 31zpci_attr(uid, "0x%x\n", uid);
32 32zpci_attr(segment0, "0x%02x\n", pfip[0]);
33static ssize_t show_pchid(struct device *dev, struct device_attribute *attr, 33zpci_attr(segment1, "0x%02x\n", pfip[1]);
34 char *buf) 34zpci_attr(segment2, "0x%02x\n", pfip[2]);
35{ 35zpci_attr(segment3, "0x%02x\n", pfip[3]);
36 struct zpci_dev *zdev = get_zdev(to_pci_dev(dev)); 36
37 37static ssize_t recover_store(struct device *dev, struct device_attribute *attr,
38 return sprintf(buf, "0x%04x\n", zdev->pchid);
39}
40static DEVICE_ATTR(pchid, S_IRUGO, show_pchid, NULL);
41
42static ssize_t show_pfgid(struct device *dev, struct device_attribute *attr,
43 char *buf)
44{
45 struct zpci_dev *zdev = get_zdev(to_pci_dev(dev));
46
47 return sprintf(buf, "0x%02x\n", zdev->pfgid);
48}
49static DEVICE_ATTR(pfgid, S_IRUGO, show_pfgid, NULL);
50
51static ssize_t store_recover(struct device *dev, struct device_attribute *attr,
52 const char *buf, size_t count) 38 const char *buf, size_t count)
53{ 39{
54 struct pci_dev *pdev = to_pci_dev(dev); 40 struct pci_dev *pdev = to_pci_dev(dev);
@@ -70,20 +56,55 @@ static ssize_t store_recover(struct device *dev, struct device_attribute *attr,
70 pci_rescan_bus(zdev->bus); 56 pci_rescan_bus(zdev->bus);
71 return count; 57 return count;
72} 58}
73static DEVICE_ATTR(recover, S_IWUSR, NULL, store_recover); 59static DEVICE_ATTR_WO(recover);
60
61static ssize_t util_string_read(struct file *filp, struct kobject *kobj,
62 struct bin_attribute *attr, char *buf,
63 loff_t off, size_t count)
64{
65 struct device *dev = kobj_to_dev(kobj);
66 struct pci_dev *pdev = to_pci_dev(dev);
67 struct zpci_dev *zdev = get_zdev(pdev);
68
69 return memory_read_from_buffer(buf, count, &off, zdev->util_str,
70 sizeof(zdev->util_str));
71}
72static BIN_ATTR_RO(util_string, CLP_UTIL_STR_LEN);
73static struct bin_attribute *zpci_bin_attrs[] = {
74 &bin_attr_util_string,
75 NULL,
76};
74 77
75static struct attribute *zpci_dev_attrs[] = { 78static struct attribute *zpci_dev_attrs[] = {
76 &dev_attr_function_id.attr, 79 &dev_attr_function_id.attr,
77 &dev_attr_function_handle.attr, 80 &dev_attr_function_handle.attr,
78 &dev_attr_pchid.attr, 81 &dev_attr_pchid.attr,
79 &dev_attr_pfgid.attr, 82 &dev_attr_pfgid.attr,
83 &dev_attr_pft.attr,
84 &dev_attr_vfn.attr,
85 &dev_attr_uid.attr,
80 &dev_attr_recover.attr, 86 &dev_attr_recover.attr,
81 NULL, 87 NULL,
82}; 88};
83static struct attribute_group zpci_attr_group = { 89static struct attribute_group zpci_attr_group = {
84 .attrs = zpci_dev_attrs, 90 .attrs = zpci_dev_attrs,
91 .bin_attrs = zpci_bin_attrs,
85}; 92};
93
94static struct attribute *pfip_attrs[] = {
95 &dev_attr_segment0.attr,
96 &dev_attr_segment1.attr,
97 &dev_attr_segment2.attr,
98 &dev_attr_segment3.attr,
99 NULL,
100};
101static struct attribute_group pfip_attr_group = {
102 .name = "pfip",
103 .attrs = pfip_attrs,
104};
105
86const struct attribute_group *zpci_attr_groups[] = { 106const struct attribute_group *zpci_attr_groups[] = {
87 &zpci_attr_group, 107 &zpci_attr_group,
108 &pfip_attr_group,
88 NULL, 109 NULL,
89}; 110};
diff --git a/drivers/s390/char/Makefile b/drivers/s390/char/Makefile
index b69ab17f13fa..629fcc275e92 100644
--- a/drivers/s390/char/Makefile
+++ b/drivers/s390/char/Makefile
@@ -33,4 +33,4 @@ obj-$(CONFIG_MONWRITER) += monwriter.o
33obj-$(CONFIG_S390_VMUR) += vmur.o 33obj-$(CONFIG_S390_VMUR) += vmur.o
34 34
35zcore_mod-objs := sclp_sdias.o zcore.o 35zcore_mod-objs := sclp_sdias.o zcore.o
36obj-$(CONFIG_ZFCPDUMP) += zcore_mod.o 36obj-$(CONFIG_CRASH_DUMP) += zcore_mod.o
diff --git a/drivers/s390/char/zcore.c b/drivers/s390/char/zcore.c
index 3d8e4d63f514..1884653e4472 100644
--- a/drivers/s390/char/zcore.c
+++ b/drivers/s390/char/zcore.c
@@ -17,6 +17,8 @@
17#include <linux/miscdevice.h> 17#include <linux/miscdevice.h>
18#include <linux/debugfs.h> 18#include <linux/debugfs.h>
19#include <linux/module.h> 19#include <linux/module.h>
20#include <linux/memblock.h>
21
20#include <asm/asm-offsets.h> 22#include <asm/asm-offsets.h>
21#include <asm/ipl.h> 23#include <asm/ipl.h>
22#include <asm/sclp.h> 24#include <asm/sclp.h>
@@ -411,33 +413,24 @@ static ssize_t zcore_memmap_read(struct file *filp, char __user *buf,
411 size_t count, loff_t *ppos) 413 size_t count, loff_t *ppos)
412{ 414{
413 return simple_read_from_buffer(buf, count, ppos, filp->private_data, 415 return simple_read_from_buffer(buf, count, ppos, filp->private_data,
414 MEMORY_CHUNKS * CHUNK_INFO_SIZE); 416 memblock.memory.cnt * CHUNK_INFO_SIZE);
415} 417}
416 418
417static int zcore_memmap_open(struct inode *inode, struct file *filp) 419static int zcore_memmap_open(struct inode *inode, struct file *filp)
418{ 420{
419 int i; 421 struct memblock_region *reg;
420 char *buf; 422 char *buf;
421 struct mem_chunk *chunk_array; 423 int i = 0;
422 424
423 chunk_array = kzalloc(MEMORY_CHUNKS * sizeof(struct mem_chunk), 425 buf = kzalloc(memblock.memory.cnt * CHUNK_INFO_SIZE, GFP_KERNEL);
424 GFP_KERNEL);
425 if (!chunk_array)
426 return -ENOMEM;
427 detect_memory_layout(chunk_array, 0);
428 buf = kzalloc(MEMORY_CHUNKS * CHUNK_INFO_SIZE, GFP_KERNEL);
429 if (!buf) { 426 if (!buf) {
430 kfree(chunk_array);
431 return -ENOMEM; 427 return -ENOMEM;
432 } 428 }
433 for (i = 0; i < MEMORY_CHUNKS; i++) { 429 for_each_memblock(memory, reg) {
434 sprintf(buf + (i * CHUNK_INFO_SIZE), "%016llx %016llx ", 430 sprintf(buf + (i++ * CHUNK_INFO_SIZE), "%016llx %016llx ",
435 (unsigned long long) chunk_array[i].addr, 431 (unsigned long long) reg->base,
436 (unsigned long long) chunk_array[i].size); 432 (unsigned long long) reg->size);
437 if (chunk_array[i].size == 0)
438 break;
439 } 433 }
440 kfree(chunk_array);
441 filp->private_data = buf; 434 filp->private_data = buf;
442 return nonseekable_open(inode, filp); 435 return nonseekable_open(inode, filp);
443} 436}
@@ -593,21 +586,12 @@ static int __init check_sdias(void)
593 586
594static int __init get_mem_info(unsigned long *mem, unsigned long *end) 587static int __init get_mem_info(unsigned long *mem, unsigned long *end)
595{ 588{
596 int i; 589 struct memblock_region *reg;
597 struct mem_chunk *chunk_array;
598 590
599 chunk_array = kzalloc(MEMORY_CHUNKS * sizeof(struct mem_chunk), 591 for_each_memblock(memory, reg) {
600 GFP_KERNEL); 592 *mem += reg->size;
601 if (!chunk_array) 593 *end = max_t(unsigned long, *end, reg->base + reg->size);
602 return -ENOMEM;
603 detect_memory_layout(chunk_array, 0);
604 for (i = 0; i < MEMORY_CHUNKS; i++) {
605 if (chunk_array[i].size == 0)
606 break;
607 *mem += chunk_array[i].size;
608 *end = max(*end, chunk_array[i].addr + chunk_array[i].size);
609 } 594 }
610 kfree(chunk_array);
611 return 0; 595 return 0;
612} 596}
613 597
diff --git a/drivers/s390/cio/ccwreq.c b/drivers/s390/cio/ccwreq.c
index 5156264d0c74..07676c22d514 100644
--- a/drivers/s390/cio/ccwreq.c
+++ b/drivers/s390/cio/ccwreq.c
@@ -46,7 +46,7 @@ static u16 ccwreq_next_path(struct ccw_device *cdev)
46 goto out; 46 goto out;
47 } 47 }
48 req->retries = req->maxretries; 48 req->retries = req->maxretries;
49 req->mask = lpm_adjust(req->mask >>= 1, req->lpm); 49 req->mask = lpm_adjust(req->mask >> 1, req->lpm);
50out: 50out:
51 return req->mask; 51 return req->mask;
52} 52}
@@ -252,7 +252,7 @@ static void ccwreq_log_status(struct ccw_device *cdev, enum io_status status)
252 */ 252 */
253void ccw_request_handler(struct ccw_device *cdev) 253void ccw_request_handler(struct ccw_device *cdev)
254{ 254{
255 struct irb *irb = (struct irb *)&S390_lowcore.irb; 255 struct irb *irb = &__get_cpu_var(cio_irb);
256 struct ccw_request *req = &cdev->private->req; 256 struct ccw_request *req = &cdev->private->req;
257 enum io_status status; 257 enum io_status status;
258 int rc = -EOPNOTSUPP; 258 int rc = -EOPNOTSUPP;
diff --git a/drivers/s390/cio/chp.c b/drivers/s390/cio/chp.c
index 6c440d4349d4..d497aa05a72f 100644
--- a/drivers/s390/cio/chp.c
+++ b/drivers/s390/cio/chp.c
@@ -509,7 +509,7 @@ out:
509 * On success return a newly allocated copy of the channel-path description 509 * On success return a newly allocated copy of the channel-path description
510 * data associated with the given channel-path ID. Return %NULL on error. 510 * data associated with the given channel-path ID. Return %NULL on error.
511 */ 511 */
512void *chp_get_chp_desc(struct chp_id chpid) 512struct channel_path_desc *chp_get_chp_desc(struct chp_id chpid)
513{ 513{
514 struct channel_path *chp; 514 struct channel_path *chp;
515 struct channel_path_desc *desc; 515 struct channel_path_desc *desc;
diff --git a/drivers/s390/cio/chp.h b/drivers/s390/cio/chp.h
index 9284b785a06f..4efd5b867cc3 100644
--- a/drivers/s390/cio/chp.h
+++ b/drivers/s390/cio/chp.h
@@ -60,7 +60,7 @@ static inline struct channel_path *chpid_to_chp(struct chp_id chpid)
60int chp_get_status(struct chp_id chpid); 60int chp_get_status(struct chp_id chpid);
61u8 chp_get_sch_opm(struct subchannel *sch); 61u8 chp_get_sch_opm(struct subchannel *sch);
62int chp_is_registered(struct chp_id chpid); 62int chp_is_registered(struct chp_id chpid);
63void *chp_get_chp_desc(struct chp_id chpid); 63struct channel_path_desc *chp_get_chp_desc(struct chp_id chpid);
64void chp_remove_cmg_attr(struct channel_path *chp); 64void chp_remove_cmg_attr(struct channel_path *chp);
65int chp_add_cmg_attr(struct channel_path *chp); 65int chp_add_cmg_attr(struct channel_path *chp);
66int chp_update_desc(struct channel_path *chp); 66int chp_update_desc(struct channel_path *chp);
diff --git a/drivers/s390/cio/chsc.h b/drivers/s390/cio/chsc.h
index 7e53a9c8b0b9..76c9b50700b2 100644
--- a/drivers/s390/cio/chsc.h
+++ b/drivers/s390/cio/chsc.h
@@ -21,17 +21,6 @@ struct cmg_entry {
21 u32 values[NR_MEASUREMENT_ENTRIES]; 21 u32 values[NR_MEASUREMENT_ENTRIES];
22} __attribute__ ((packed)); 22} __attribute__ ((packed));
23 23
24struct channel_path_desc {
25 u8 flags;
26 u8 lsn;
27 u8 desc;
28 u8 chpid;
29 u8 swla;
30 u8 zeroes;
31 u8 chla;
32 u8 chpp;
33} __attribute__ ((packed));
34
35struct channel_path_desc_fmt1 { 24struct channel_path_desc_fmt1 {
36 u8 flags; 25 u8 flags;
37 u8 lsn; 26 u8 lsn;
diff --git a/drivers/s390/cio/chsc_sch.c b/drivers/s390/cio/chsc_sch.c
index 1d3661af7bd8..3d22d2a4ce14 100644
--- a/drivers/s390/cio/chsc_sch.c
+++ b/drivers/s390/cio/chsc_sch.c
@@ -58,7 +58,7 @@ static void chsc_subchannel_irq(struct subchannel *sch)
58{ 58{
59 struct chsc_private *private = dev_get_drvdata(&sch->dev); 59 struct chsc_private *private = dev_get_drvdata(&sch->dev);
60 struct chsc_request *request = private->request; 60 struct chsc_request *request = private->request;
61 struct irb *irb = (struct irb *)&S390_lowcore.irb; 61 struct irb *irb = &__get_cpu_var(cio_irb);
62 62
63 CHSC_LOG(4, "irb"); 63 CHSC_LOG(4, "irb");
64 CHSC_LOG_HEX(4, irb, sizeof(*irb)); 64 CHSC_LOG_HEX(4, irb, sizeof(*irb));
diff --git a/drivers/s390/cio/cio.c b/drivers/s390/cio/cio.c
index 9e058c4657a3..77f9c92df4b9 100644
--- a/drivers/s390/cio/cio.c
+++ b/drivers/s390/cio/cio.c
@@ -46,6 +46,9 @@ debug_info_t *cio_debug_msg_id;
46debug_info_t *cio_debug_trace_id; 46debug_info_t *cio_debug_trace_id;
47debug_info_t *cio_debug_crw_id; 47debug_info_t *cio_debug_crw_id;
48 48
49DEFINE_PER_CPU_ALIGNED(struct irb, cio_irb);
50EXPORT_PER_CPU_SYMBOL(cio_irb);
51
49/* 52/*
50 * Function: cio_debug_init 53 * Function: cio_debug_init
51 * Initializes three debug logs for common I/O: 54 * Initializes three debug logs for common I/O:
@@ -560,7 +563,7 @@ static irqreturn_t do_cio_interrupt(int irq, void *dummy)
560 563
561 __this_cpu_write(s390_idle.nohz_delay, 1); 564 __this_cpu_write(s390_idle.nohz_delay, 1);
562 tpi_info = (struct tpi_info *) &get_irq_regs()->int_code; 565 tpi_info = (struct tpi_info *) &get_irq_regs()->int_code;
563 irb = (struct irb *) &S390_lowcore.irb; 566 irb = &__get_cpu_var(cio_irb);
564 sch = (struct subchannel *)(unsigned long) tpi_info->intparm; 567 sch = (struct subchannel *)(unsigned long) tpi_info->intparm;
565 if (!sch) { 568 if (!sch) {
566 /* Clear pending interrupt condition. */ 569 /* Clear pending interrupt condition. */
@@ -609,7 +612,7 @@ void cio_tsch(struct subchannel *sch)
609 struct irb *irb; 612 struct irb *irb;
610 int irq_context; 613 int irq_context;
611 614
612 irb = (struct irb *)&S390_lowcore.irb; 615 irb = &__get_cpu_var(cio_irb);
613 /* Store interrupt response block to lowcore. */ 616 /* Store interrupt response block to lowcore. */
614 if (tsch(sch->schid, irb) != 0) 617 if (tsch(sch->schid, irb) != 0)
615 /* Not status pending or not operational. */ 618 /* Not status pending or not operational. */
@@ -746,7 +749,7 @@ __clear_io_subchannel_easy(struct subchannel_id schid)
746 struct tpi_info ti; 749 struct tpi_info ti;
747 750
748 if (tpi(&ti)) { 751 if (tpi(&ti)) {
749 tsch(ti.schid, (struct irb *)&S390_lowcore.irb); 752 tsch(ti.schid, &__get_cpu_var(cio_irb));
750 if (schid_equal(&ti.schid, &schid)) 753 if (schid_equal(&ti.schid, &schid))
751 return 0; 754 return 0;
752 } 755 }
diff --git a/drivers/s390/cio/cio.h b/drivers/s390/cio/cio.h
index d42f67412bd8..a01376ae1749 100644
--- a/drivers/s390/cio/cio.h
+++ b/drivers/s390/cio/cio.h
@@ -102,6 +102,8 @@ struct subchannel {
102 struct schib_config config; 102 struct schib_config config;
103} __attribute__ ((aligned(8))); 103} __attribute__ ((aligned(8)));
104 104
105DECLARE_PER_CPU(struct irb, cio_irb);
106
105#define to_subchannel(n) container_of(n, struct subchannel, dev) 107#define to_subchannel(n) container_of(n, struct subchannel, dev)
106 108
107extern int cio_validate_subchannel (struct subchannel *, struct subchannel_id); 109extern int cio_validate_subchannel (struct subchannel *, struct subchannel_id);
diff --git a/drivers/s390/cio/device_fsm.c b/drivers/s390/cio/device_fsm.c
index c7638c543250..0bc902b3cd84 100644
--- a/drivers/s390/cio/device_fsm.c
+++ b/drivers/s390/cio/device_fsm.c
@@ -739,7 +739,7 @@ ccw_device_irq(struct ccw_device *cdev, enum dev_event dev_event)
739 struct irb *irb; 739 struct irb *irb;
740 int is_cmd; 740 int is_cmd;
741 741
742 irb = (struct irb *)&S390_lowcore.irb; 742 irb = &__get_cpu_var(cio_irb);
743 is_cmd = !scsw_is_tm(&irb->scsw); 743 is_cmd = !scsw_is_tm(&irb->scsw);
744 /* Check for unsolicited interrupt. */ 744 /* Check for unsolicited interrupt. */
745 if (!scsw_is_solicited(&irb->scsw)) { 745 if (!scsw_is_solicited(&irb->scsw)) {
@@ -805,7 +805,7 @@ ccw_device_w4sense(struct ccw_device *cdev, enum dev_event dev_event)
805{ 805{
806 struct irb *irb; 806 struct irb *irb;
807 807
808 irb = (struct irb *)&S390_lowcore.irb; 808 irb = &__get_cpu_var(cio_irb);
809 /* Check for unsolicited interrupt. */ 809 /* Check for unsolicited interrupt. */
810 if (scsw_stctl(&irb->scsw) == 810 if (scsw_stctl(&irb->scsw) ==
811 (SCSW_STCTL_STATUS_PEND | SCSW_STCTL_ALERT_STATUS)) { 811 (SCSW_STCTL_STATUS_PEND | SCSW_STCTL_ALERT_STATUS)) {
diff --git a/drivers/s390/cio/device_ops.c b/drivers/s390/cio/device_ops.c
index 4845d64f2842..f3c417943dad 100644
--- a/drivers/s390/cio/device_ops.c
+++ b/drivers/s390/cio/device_ops.c
@@ -563,14 +563,23 @@ out_unlock:
563 return rc; 563 return rc;
564} 564}
565 565
566void *ccw_device_get_chp_desc(struct ccw_device *cdev, int chp_no) 566/**
567 * chp_get_chp_desc - return newly allocated channel-path descriptor
568 * @cdev: device to obtain the descriptor for
569 * @chp_idx: index of the channel path
570 *
571 * On success return a newly allocated copy of the channel-path description
572 * data associated with the given channel path. Return %NULL on error.
573 */
574struct channel_path_desc *ccw_device_get_chp_desc(struct ccw_device *cdev,
575 int chp_idx)
567{ 576{
568 struct subchannel *sch; 577 struct subchannel *sch;
569 struct chp_id chpid; 578 struct chp_id chpid;
570 579
571 sch = to_subchannel(cdev->dev.parent); 580 sch = to_subchannel(cdev->dev.parent);
572 chp_id_init(&chpid); 581 chp_id_init(&chpid);
573 chpid.id = sch->schib.pmcw.chpid[chp_no]; 582 chpid.id = sch->schib.pmcw.chpid[chp_idx];
574 return chp_get_chp_desc(chpid); 583 return chp_get_chp_desc(chpid);
575} 584}
576 585
diff --git a/drivers/s390/cio/eadm_sch.c b/drivers/s390/cio/eadm_sch.c
index 3a2ee4a740b4..c4f7bf3e24c2 100644
--- a/drivers/s390/cio/eadm_sch.c
+++ b/drivers/s390/cio/eadm_sch.c
@@ -134,7 +134,7 @@ static void eadm_subchannel_irq(struct subchannel *sch)
134{ 134{
135 struct eadm_private *private = get_eadm_private(sch); 135 struct eadm_private *private = get_eadm_private(sch);
136 struct eadm_scsw *scsw = &sch->schib.scsw.eadm; 136 struct eadm_scsw *scsw = &sch->schib.scsw.eadm;
137 struct irb *irb = (struct irb *)&S390_lowcore.irb; 137 struct irb *irb = &__get_cpu_var(cio_irb);
138 int error = 0; 138 int error = 0;
139 139
140 EADM_LOG(6, "irq"); 140 EADM_LOG(6, "irq");
diff --git a/drivers/s390/net/qeth_core_main.c b/drivers/s390/net/qeth_core_main.c
index 22470a3b182f..e89f38c31176 100644
--- a/drivers/s390/net/qeth_core_main.c
+++ b/drivers/s390/net/qeth_core_main.c
@@ -22,6 +22,7 @@
22#include <net/iucv/af_iucv.h> 22#include <net/iucv/af_iucv.h>
23 23
24#include <asm/ebcdic.h> 24#include <asm/ebcdic.h>
25#include <asm/chpid.h>
25#include <asm/io.h> 26#include <asm/io.h>
26#include <asm/sysinfo.h> 27#include <asm/sysinfo.h>
27#include <asm/compat.h> 28#include <asm/compat.h>
@@ -1344,16 +1345,7 @@ static void qeth_set_multiple_write_queues(struct qeth_card *card)
1344static void qeth_update_from_chp_desc(struct qeth_card *card) 1345static void qeth_update_from_chp_desc(struct qeth_card *card)
1345{ 1346{
1346 struct ccw_device *ccwdev; 1347 struct ccw_device *ccwdev;
1347 struct channelPath_dsc { 1348 struct channel_path_desc *chp_dsc;
1348 u8 flags;
1349 u8 lsn;
1350 u8 desc;
1351 u8 chpid;
1352 u8 swla;
1353 u8 zeroes;
1354 u8 chla;
1355 u8 chpp;
1356 } *chp_dsc;
1357 1349
1358 QETH_DBF_TEXT(SETUP, 2, "chp_desc"); 1350 QETH_DBF_TEXT(SETUP, 2, "chp_desc");
1359 1351
diff --git a/include/linux/memblock.h b/include/linux/memblock.h
index 8a20a51ed42d..73dc382e72d8 100644
--- a/include/linux/memblock.h
+++ b/include/linux/memblock.h
@@ -18,6 +18,7 @@
18#include <linux/mm.h> 18#include <linux/mm.h>
19 19
20#define INIT_MEMBLOCK_REGIONS 128 20#define INIT_MEMBLOCK_REGIONS 128
21#define INIT_PHYSMEM_REGIONS 4
21 22
22/* Definition of memblock flags. */ 23/* Definition of memblock flags. */
23#define MEMBLOCK_HOTPLUG 0x1 /* hotpluggable region */ 24#define MEMBLOCK_HOTPLUG 0x1 /* hotpluggable region */
@@ -43,6 +44,9 @@ struct memblock {
43 phys_addr_t current_limit; 44 phys_addr_t current_limit;
44 struct memblock_type memory; 45 struct memblock_type memory;
45 struct memblock_type reserved; 46 struct memblock_type reserved;
47#ifdef CONFIG_HAVE_MEMBLOCK_PHYS_MAP
48 struct memblock_type physmem;
49#endif
46}; 50};
47 51
48extern struct memblock memblock; 52extern struct memblock memblock;
@@ -71,6 +75,63 @@ int memblock_reserve(phys_addr_t base, phys_addr_t size);
71void memblock_trim_memory(phys_addr_t align); 75void memblock_trim_memory(phys_addr_t align);
72int memblock_mark_hotplug(phys_addr_t base, phys_addr_t size); 76int memblock_mark_hotplug(phys_addr_t base, phys_addr_t size);
73int memblock_clear_hotplug(phys_addr_t base, phys_addr_t size); 77int memblock_clear_hotplug(phys_addr_t base, phys_addr_t size);
78
79/* Low level functions */
80int memblock_add_range(struct memblock_type *type,
81 phys_addr_t base, phys_addr_t size,
82 int nid, unsigned long flags);
83
84int memblock_remove_range(struct memblock_type *type,
85 phys_addr_t base,
86 phys_addr_t size);
87
88void __next_mem_range(u64 *idx, int nid, struct memblock_type *type_a,
89 struct memblock_type *type_b, phys_addr_t *out_start,
90 phys_addr_t *out_end, int *out_nid);
91
92void __next_mem_range_rev(u64 *idx, int nid, struct memblock_type *type_a,
93 struct memblock_type *type_b, phys_addr_t *out_start,
94 phys_addr_t *out_end, int *out_nid);
95
96/**
97 * for_each_mem_range - iterate through memblock areas from type_a and not
98 * included in type_b. Or just type_a if type_b is NULL.
99 * @i: u64 used as loop variable
100 * @type_a: ptr to memblock_type to iterate
101 * @type_b: ptr to memblock_type which excludes from the iteration
102 * @nid: node selector, %NUMA_NO_NODE for all nodes
103 * @p_start: ptr to phys_addr_t for start address of the range, can be %NULL
104 * @p_end: ptr to phys_addr_t for end address of the range, can be %NULL
105 * @p_nid: ptr to int for nid of the range, can be %NULL
106 */
107#define for_each_mem_range(i, type_a, type_b, nid, \
108 p_start, p_end, p_nid) \
109 for (i = 0, __next_mem_range(&i, nid, type_a, type_b, \
110 p_start, p_end, p_nid); \
111 i != (u64)ULLONG_MAX; \
112 __next_mem_range(&i, nid, type_a, type_b, \
113 p_start, p_end, p_nid))
114
115/**
116 * for_each_mem_range_rev - reverse iterate through memblock areas from
117 * type_a and not included in type_b. Or just type_a if type_b is NULL.
118 * @i: u64 used as loop variable
119 * @type_a: ptr to memblock_type to iterate
120 * @type_b: ptr to memblock_type which excludes from the iteration
121 * @nid: node selector, %NUMA_NO_NODE for all nodes
122 * @p_start: ptr to phys_addr_t for start address of the range, can be %NULL
123 * @p_end: ptr to phys_addr_t for end address of the range, can be %NULL
124 * @p_nid: ptr to int for nid of the range, can be %NULL
125 */
126#define for_each_mem_range_rev(i, type_a, type_b, nid, \
127 p_start, p_end, p_nid) \
128 for (i = (u64)ULLONG_MAX, \
129 __next_mem_range_rev(&i, nid, type_a, type_b, \
130 p_start, p_end, p_nid); \
131 i != (u64)ULLONG_MAX; \
132 __next_mem_range_rev(&i, nid, type_a, type_b, \
133 p_start, p_end, p_nid))
134
74#ifdef CONFIG_MOVABLE_NODE 135#ifdef CONFIG_MOVABLE_NODE
75static inline bool memblock_is_hotpluggable(struct memblock_region *m) 136static inline bool memblock_is_hotpluggable(struct memblock_region *m)
76{ 137{
@@ -113,9 +174,6 @@ void __next_mem_pfn_range(int *idx, int nid, unsigned long *out_start_pfn,
113 i >= 0; __next_mem_pfn_range(&i, nid, p_start, p_end, p_nid)) 174 i >= 0; __next_mem_pfn_range(&i, nid, p_start, p_end, p_nid))
114#endif /* CONFIG_HAVE_MEMBLOCK_NODE_MAP */ 175#endif /* CONFIG_HAVE_MEMBLOCK_NODE_MAP */
115 176
116void __next_free_mem_range(u64 *idx, int nid, phys_addr_t *out_start,
117 phys_addr_t *out_end, int *out_nid);
118
119/** 177/**
120 * for_each_free_mem_range - iterate through free memblock areas 178 * for_each_free_mem_range - iterate through free memblock areas
121 * @i: u64 used as loop variable 179 * @i: u64 used as loop variable
@@ -128,13 +186,8 @@ void __next_free_mem_range(u64 *idx, int nid, phys_addr_t *out_start,
128 * soon as memblock is initialized. 186 * soon as memblock is initialized.
129 */ 187 */
130#define for_each_free_mem_range(i, nid, p_start, p_end, p_nid) \ 188#define for_each_free_mem_range(i, nid, p_start, p_end, p_nid) \
131 for (i = 0, \ 189 for_each_mem_range(i, &memblock.memory, &memblock.reserved, \
132 __next_free_mem_range(&i, nid, p_start, p_end, p_nid); \ 190 nid, p_start, p_end, p_nid)
133 i != (u64)ULLONG_MAX; \
134 __next_free_mem_range(&i, nid, p_start, p_end, p_nid))
135
136void __next_free_mem_range_rev(u64 *idx, int nid, phys_addr_t *out_start,
137 phys_addr_t *out_end, int *out_nid);
138 191
139/** 192/**
140 * for_each_free_mem_range_reverse - rev-iterate through free memblock areas 193 * for_each_free_mem_range_reverse - rev-iterate through free memblock areas
@@ -148,10 +201,8 @@ void __next_free_mem_range_rev(u64 *idx, int nid, phys_addr_t *out_start,
148 * order. Available as soon as memblock is initialized. 201 * order. Available as soon as memblock is initialized.
149 */ 202 */
150#define for_each_free_mem_range_reverse(i, nid, p_start, p_end, p_nid) \ 203#define for_each_free_mem_range_reverse(i, nid, p_start, p_end, p_nid) \
151 for (i = (u64)ULLONG_MAX, \ 204 for_each_mem_range_rev(i, &memblock.memory, &memblock.reserved, \
152 __next_free_mem_range_rev(&i, nid, p_start, p_end, p_nid); \ 205 nid, p_start, p_end, p_nid)
153 i != (u64)ULLONG_MAX; \
154 __next_free_mem_range_rev(&i, nid, p_start, p_end, p_nid))
155 206
156static inline void memblock_set_region_flags(struct memblock_region *r, 207static inline void memblock_set_region_flags(struct memblock_region *r,
157 unsigned long flags) 208 unsigned long flags)
diff --git a/mm/Kconfig b/mm/Kconfig
index 1b5a95f0fa01..28cec518f4d4 100644
--- a/mm/Kconfig
+++ b/mm/Kconfig
@@ -134,6 +134,9 @@ config HAVE_MEMBLOCK
134config HAVE_MEMBLOCK_NODE_MAP 134config HAVE_MEMBLOCK_NODE_MAP
135 boolean 135 boolean
136 136
137config HAVE_MEMBLOCK_PHYS_MAP
138 boolean
139
137config ARCH_DISCARD_MEMBLOCK 140config ARCH_DISCARD_MEMBLOCK
138 boolean 141 boolean
139 142
diff --git a/mm/memblock.c b/mm/memblock.c
index e9d6ca9a01a9..a810ba923cdd 100644
--- a/mm/memblock.c
+++ b/mm/memblock.c
@@ -27,6 +27,9 @@
27 27
28static struct memblock_region memblock_memory_init_regions[INIT_MEMBLOCK_REGIONS] __initdata_memblock; 28static struct memblock_region memblock_memory_init_regions[INIT_MEMBLOCK_REGIONS] __initdata_memblock;
29static struct memblock_region memblock_reserved_init_regions[INIT_MEMBLOCK_REGIONS] __initdata_memblock; 29static struct memblock_region memblock_reserved_init_regions[INIT_MEMBLOCK_REGIONS] __initdata_memblock;
30#ifdef CONFIG_HAVE_MEMBLOCK_PHYS_MAP
31static struct memblock_region memblock_physmem_init_regions[INIT_PHYSMEM_REGIONS] __initdata_memblock;
32#endif
30 33
31struct memblock memblock __initdata_memblock = { 34struct memblock memblock __initdata_memblock = {
32 .memory.regions = memblock_memory_init_regions, 35 .memory.regions = memblock_memory_init_regions,
@@ -37,6 +40,12 @@ struct memblock memblock __initdata_memblock = {
37 .reserved.cnt = 1, /* empty dummy entry */ 40 .reserved.cnt = 1, /* empty dummy entry */
38 .reserved.max = INIT_MEMBLOCK_REGIONS, 41 .reserved.max = INIT_MEMBLOCK_REGIONS,
39 42
43#ifdef CONFIG_HAVE_MEMBLOCK_PHYS_MAP
44 .physmem.regions = memblock_physmem_init_regions,
45 .physmem.cnt = 1, /* empty dummy entry */
46 .physmem.max = INIT_PHYSMEM_REGIONS,
47#endif
48
40 .bottom_up = false, 49 .bottom_up = false,
41 .current_limit = MEMBLOCK_ALLOC_ANYWHERE, 50 .current_limit = MEMBLOCK_ALLOC_ANYWHERE,
42}; 51};
@@ -472,7 +481,7 @@ static void __init_memblock memblock_insert_region(struct memblock_type *type,
472} 481}
473 482
474/** 483/**
475 * memblock_add_region - add new memblock region 484 * memblock_add_range - add new memblock region
476 * @type: memblock type to add new region into 485 * @type: memblock type to add new region into
477 * @base: base address of the new region 486 * @base: base address of the new region
478 * @size: size of the new region 487 * @size: size of the new region
@@ -487,7 +496,7 @@ static void __init_memblock memblock_insert_region(struct memblock_type *type,
487 * RETURNS: 496 * RETURNS:
488 * 0 on success, -errno on failure. 497 * 0 on success, -errno on failure.
489 */ 498 */
490static int __init_memblock memblock_add_region(struct memblock_type *type, 499int __init_memblock memblock_add_range(struct memblock_type *type,
491 phys_addr_t base, phys_addr_t size, 500 phys_addr_t base, phys_addr_t size,
492 int nid, unsigned long flags) 501 int nid, unsigned long flags)
493{ 502{
@@ -569,12 +578,12 @@ repeat:
569int __init_memblock memblock_add_node(phys_addr_t base, phys_addr_t size, 578int __init_memblock memblock_add_node(phys_addr_t base, phys_addr_t size,
570 int nid) 579 int nid)
571{ 580{
572 return memblock_add_region(&memblock.memory, base, size, nid, 0); 581 return memblock_add_range(&memblock.memory, base, size, nid, 0);
573} 582}
574 583
575int __init_memblock memblock_add(phys_addr_t base, phys_addr_t size) 584int __init_memblock memblock_add(phys_addr_t base, phys_addr_t size)
576{ 585{
577 return memblock_add_region(&memblock.memory, base, size, 586 return memblock_add_range(&memblock.memory, base, size,
578 MAX_NUMNODES, 0); 587 MAX_NUMNODES, 0);
579} 588}
580 589
@@ -654,8 +663,8 @@ static int __init_memblock memblock_isolate_range(struct memblock_type *type,
654 return 0; 663 return 0;
655} 664}
656 665
657static int __init_memblock __memblock_remove(struct memblock_type *type, 666int __init_memblock memblock_remove_range(struct memblock_type *type,
658 phys_addr_t base, phys_addr_t size) 667 phys_addr_t base, phys_addr_t size)
659{ 668{
660 int start_rgn, end_rgn; 669 int start_rgn, end_rgn;
661 int i, ret; 670 int i, ret;
@@ -671,9 +680,10 @@ static int __init_memblock __memblock_remove(struct memblock_type *type,
671 680
672int __init_memblock memblock_remove(phys_addr_t base, phys_addr_t size) 681int __init_memblock memblock_remove(phys_addr_t base, phys_addr_t size)
673{ 682{
674 return __memblock_remove(&memblock.memory, base, size); 683 return memblock_remove_range(&memblock.memory, base, size);
675} 684}
676 685
686
677int __init_memblock memblock_free(phys_addr_t base, phys_addr_t size) 687int __init_memblock memblock_free(phys_addr_t base, phys_addr_t size)
678{ 688{
679 memblock_dbg(" memblock_free: [%#016llx-%#016llx] %pF\n", 689 memblock_dbg(" memblock_free: [%#016llx-%#016llx] %pF\n",
@@ -681,7 +691,7 @@ int __init_memblock memblock_free(phys_addr_t base, phys_addr_t size)
681 (unsigned long long)base + size - 1, 691 (unsigned long long)base + size - 1,
682 (void *)_RET_IP_); 692 (void *)_RET_IP_);
683 693
684 return __memblock_remove(&memblock.reserved, base, size); 694 return memblock_remove_range(&memblock.reserved, base, size);
685} 695}
686 696
687static int __init_memblock memblock_reserve_region(phys_addr_t base, 697static int __init_memblock memblock_reserve_region(phys_addr_t base,
@@ -696,7 +706,7 @@ static int __init_memblock memblock_reserve_region(phys_addr_t base,
696 (unsigned long long)base + size - 1, 706 (unsigned long long)base + size - 1,
697 flags, (void *)_RET_IP_); 707 flags, (void *)_RET_IP_);
698 708
699 return memblock_add_region(_rgn, base, size, nid, flags); 709 return memblock_add_range(_rgn, base, size, nid, flags);
700} 710}
701 711
702int __init_memblock memblock_reserve(phys_addr_t base, phys_addr_t size) 712int __init_memblock memblock_reserve(phys_addr_t base, phys_addr_t size)
@@ -758,17 +768,19 @@ int __init_memblock memblock_clear_hotplug(phys_addr_t base, phys_addr_t size)
758} 768}
759 769
760/** 770/**
761 * __next_free_mem_range - next function for for_each_free_mem_range() 771 * __next__mem_range - next function for for_each_free_mem_range() etc.
762 * @idx: pointer to u64 loop variable 772 * @idx: pointer to u64 loop variable
763 * @nid: node selector, %NUMA_NO_NODE for all nodes 773 * @nid: node selector, %NUMA_NO_NODE for all nodes
774 * @type_a: pointer to memblock_type from where the range is taken
775 * @type_b: pointer to memblock_type which excludes memory from being taken
764 * @out_start: ptr to phys_addr_t for start address of the range, can be %NULL 776 * @out_start: ptr to phys_addr_t for start address of the range, can be %NULL
765 * @out_end: ptr to phys_addr_t for end address of the range, can be %NULL 777 * @out_end: ptr to phys_addr_t for end address of the range, can be %NULL
766 * @out_nid: ptr to int for nid of the range, can be %NULL 778 * @out_nid: ptr to int for nid of the range, can be %NULL
767 * 779 *
768 * Find the first free area from *@idx which matches @nid, fill the out 780 * Find the first area from *@idx which matches @nid, fill the out
769 * parameters, and update *@idx for the next iteration. The lower 32bit of 781 * parameters, and update *@idx for the next iteration. The lower 32bit of
770 * *@idx contains index into memory region and the upper 32bit indexes the 782 * *@idx contains index into type_a and the upper 32bit indexes the
771 * areas before each reserved region. For example, if reserved regions 783 * areas before each region in type_b. For example, if type_b regions
772 * look like the following, 784 * look like the following,
773 * 785 *
774 * 0:[0-16), 1:[32-48), 2:[128-130) 786 * 0:[0-16), 1:[32-48), 2:[128-130)
@@ -780,53 +792,77 @@ int __init_memblock memblock_clear_hotplug(phys_addr_t base, phys_addr_t size)
780 * As both region arrays are sorted, the function advances the two indices 792 * As both region arrays are sorted, the function advances the two indices
781 * in lockstep and returns each intersection. 793 * in lockstep and returns each intersection.
782 */ 794 */
783void __init_memblock __next_free_mem_range(u64 *idx, int nid, 795void __init_memblock __next_mem_range(u64 *idx, int nid,
784 phys_addr_t *out_start, 796 struct memblock_type *type_a,
785 phys_addr_t *out_end, int *out_nid) 797 struct memblock_type *type_b,
798 phys_addr_t *out_start,
799 phys_addr_t *out_end, int *out_nid)
786{ 800{
787 struct memblock_type *mem = &memblock.memory; 801 int idx_a = *idx & 0xffffffff;
788 struct memblock_type *rsv = &memblock.reserved; 802 int idx_b = *idx >> 32;
789 int mi = *idx & 0xffffffff;
790 int ri = *idx >> 32;
791 803
792 if (WARN_ONCE(nid == MAX_NUMNODES, "Usage of MAX_NUMNODES is deprecated. Use NUMA_NO_NODE instead\n")) 804 if (WARN_ONCE(nid == MAX_NUMNODES,
805 "Usage of MAX_NUMNODES is deprecated. Use NUMA_NO_NODE instead\n"))
793 nid = NUMA_NO_NODE; 806 nid = NUMA_NO_NODE;
794 807
795 for ( ; mi < mem->cnt; mi++) { 808 for (; idx_a < type_a->cnt; idx_a++) {
796 struct memblock_region *m = &mem->regions[mi]; 809 struct memblock_region *m = &type_a->regions[idx_a];
810
797 phys_addr_t m_start = m->base; 811 phys_addr_t m_start = m->base;
798 phys_addr_t m_end = m->base + m->size; 812 phys_addr_t m_end = m->base + m->size;
813 int m_nid = memblock_get_region_node(m);
799 814
800 /* only memory regions are associated with nodes, check it */ 815 /* only memory regions are associated with nodes, check it */
801 if (nid != NUMA_NO_NODE && nid != memblock_get_region_node(m)) 816 if (nid != NUMA_NO_NODE && nid != m_nid)
802 continue; 817 continue;
803 818
804 /* scan areas before each reservation for intersection */ 819 if (!type_b) {
805 for ( ; ri < rsv->cnt + 1; ri++) { 820 if (out_start)
806 struct memblock_region *r = &rsv->regions[ri]; 821 *out_start = m_start;
807 phys_addr_t r_start = ri ? r[-1].base + r[-1].size : 0; 822 if (out_end)
808 phys_addr_t r_end = ri < rsv->cnt ? r->base : ULLONG_MAX; 823 *out_end = m_end;
824 if (out_nid)
825 *out_nid = m_nid;
826 idx_a++;
827 *idx = (u32)idx_a | (u64)idx_b << 32;
828 return;
829 }
809 830
810 /* if ri advanced past mi, break out to advance mi */ 831 /* scan areas before each reservation */
832 for (; idx_b < type_b->cnt + 1; idx_b++) {
833 struct memblock_region *r;
834 phys_addr_t r_start;
835 phys_addr_t r_end;
836
837 r = &type_b->regions[idx_b];
838 r_start = idx_b ? r[-1].base + r[-1].size : 0;
839 r_end = idx_b < type_b->cnt ?
840 r->base : ULLONG_MAX;
841
842 /*
843 * if idx_b advanced past idx_a,
844 * break out to advance idx_a
845 */
811 if (r_start >= m_end) 846 if (r_start >= m_end)
812 break; 847 break;
813 /* if the two regions intersect, we're done */ 848 /* if the two regions intersect, we're done */
814 if (m_start < r_end) { 849 if (m_start < r_end) {
815 if (out_start) 850 if (out_start)
816 *out_start = max(m_start, r_start); 851 *out_start =
852 max(m_start, r_start);
817 if (out_end) 853 if (out_end)
818 *out_end = min(m_end, r_end); 854 *out_end = min(m_end, r_end);
819 if (out_nid) 855 if (out_nid)
820 *out_nid = memblock_get_region_node(m); 856 *out_nid = m_nid;
821 /* 857 /*
822 * The region which ends first is advanced 858 * The region which ends first is
823 * for the next iteration. 859 * advanced for the next iteration.
824 */ 860 */
825 if (m_end <= r_end) 861 if (m_end <= r_end)
826 mi++; 862 idx_a++;
827 else 863 else
828 ri++; 864 idx_b++;
829 *idx = (u32)mi | (u64)ri << 32; 865 *idx = (u32)idx_a | (u64)idx_b << 32;
830 return; 866 return;
831 } 867 }
832 } 868 }
@@ -837,57 +873,80 @@ void __init_memblock __next_free_mem_range(u64 *idx, int nid,
837} 873}
838 874
839/** 875/**
840 * __next_free_mem_range_rev - next function for for_each_free_mem_range_reverse() 876 * __next_mem_range_rev - generic next function for for_each_*_range_rev()
877 *
878 * Finds the next range from type_a which is not marked as unsuitable
879 * in type_b.
880 *
841 * @idx: pointer to u64 loop variable 881 * @idx: pointer to u64 loop variable
842 * @nid: nid: node selector, %NUMA_NO_NODE for all nodes 882 * @nid: nid: node selector, %NUMA_NO_NODE for all nodes
883 * @type_a: pointer to memblock_type from where the range is taken
884 * @type_b: pointer to memblock_type which excludes memory from being taken
843 * @out_start: ptr to phys_addr_t for start address of the range, can be %NULL 885 * @out_start: ptr to phys_addr_t for start address of the range, can be %NULL
844 * @out_end: ptr to phys_addr_t for end address of the range, can be %NULL 886 * @out_end: ptr to phys_addr_t for end address of the range, can be %NULL
845 * @out_nid: ptr to int for nid of the range, can be %NULL 887 * @out_nid: ptr to int for nid of the range, can be %NULL
846 * 888 *
847 * Reverse of __next_free_mem_range(). 889 * Reverse of __next_mem_range().
848 *
849 * Linux kernel cannot migrate pages used by itself. Memory hotplug users won't
850 * be able to hot-remove hotpluggable memory used by the kernel. So this
851 * function skip hotpluggable regions if needed when allocating memory for the
852 * kernel.
853 */ 890 */
854void __init_memblock __next_free_mem_range_rev(u64 *idx, int nid, 891void __init_memblock __next_mem_range_rev(u64 *idx, int nid,
855 phys_addr_t *out_start, 892 struct memblock_type *type_a,
856 phys_addr_t *out_end, int *out_nid) 893 struct memblock_type *type_b,
894 phys_addr_t *out_start,
895 phys_addr_t *out_end, int *out_nid)
857{ 896{
858 struct memblock_type *mem = &memblock.memory; 897 int idx_a = *idx & 0xffffffff;
859 struct memblock_type *rsv = &memblock.reserved; 898 int idx_b = *idx >> 32;
860 int mi = *idx & 0xffffffff;
861 int ri = *idx >> 32;
862 899
863 if (WARN_ONCE(nid == MAX_NUMNODES, "Usage of MAX_NUMNODES is deprecated. Use NUMA_NO_NODE instead\n")) 900 if (WARN_ONCE(nid == MAX_NUMNODES, "Usage of MAX_NUMNODES is deprecated. Use NUMA_NO_NODE instead\n"))
864 nid = NUMA_NO_NODE; 901 nid = NUMA_NO_NODE;
865 902
866 if (*idx == (u64)ULLONG_MAX) { 903 if (*idx == (u64)ULLONG_MAX) {
867 mi = mem->cnt - 1; 904 idx_a = type_a->cnt - 1;
868 ri = rsv->cnt; 905 idx_b = type_b->cnt;
869 } 906 }
870 907
871 for ( ; mi >= 0; mi--) { 908 for (; idx_a >= 0; idx_a--) {
872 struct memblock_region *m = &mem->regions[mi]; 909 struct memblock_region *m = &type_a->regions[idx_a];
910
873 phys_addr_t m_start = m->base; 911 phys_addr_t m_start = m->base;
874 phys_addr_t m_end = m->base + m->size; 912 phys_addr_t m_end = m->base + m->size;
913 int m_nid = memblock_get_region_node(m);
875 914
876 /* only memory regions are associated with nodes, check it */ 915 /* only memory regions are associated with nodes, check it */
877 if (nid != NUMA_NO_NODE && nid != memblock_get_region_node(m)) 916 if (nid != NUMA_NO_NODE && nid != m_nid)
878 continue; 917 continue;
879 918
880 /* skip hotpluggable memory regions if needed */ 919 /* skip hotpluggable memory regions if needed */
881 if (movable_node_is_enabled() && memblock_is_hotpluggable(m)) 920 if (movable_node_is_enabled() && memblock_is_hotpluggable(m))
882 continue; 921 continue;
883 922
884 /* scan areas before each reservation for intersection */ 923 if (!type_b) {
885 for ( ; ri >= 0; ri--) { 924 if (out_start)
886 struct memblock_region *r = &rsv->regions[ri]; 925 *out_start = m_start;
887 phys_addr_t r_start = ri ? r[-1].base + r[-1].size : 0; 926 if (out_end)
888 phys_addr_t r_end = ri < rsv->cnt ? r->base : ULLONG_MAX; 927 *out_end = m_end;
928 if (out_nid)
929 *out_nid = m_nid;
930 idx_a++;
931 *idx = (u32)idx_a | (u64)idx_b << 32;
932 return;
933 }
934
935 /* scan areas before each reservation */
936 for (; idx_b >= 0; idx_b--) {
937 struct memblock_region *r;
938 phys_addr_t r_start;
939 phys_addr_t r_end;
940
941 r = &type_b->regions[idx_b];
942 r_start = idx_b ? r[-1].base + r[-1].size : 0;
943 r_end = idx_b < type_b->cnt ?
944 r->base : ULLONG_MAX;
945 /*
946 * if idx_b advanced past idx_a,
947 * break out to advance idx_a
948 */
889 949
890 /* if ri advanced past mi, break out to advance mi */
891 if (r_end <= m_start) 950 if (r_end <= m_start)
892 break; 951 break;
893 /* if the two regions intersect, we're done */ 952 /* if the two regions intersect, we're done */
@@ -897,18 +956,17 @@ void __init_memblock __next_free_mem_range_rev(u64 *idx, int nid,
897 if (out_end) 956 if (out_end)
898 *out_end = min(m_end, r_end); 957 *out_end = min(m_end, r_end);
899 if (out_nid) 958 if (out_nid)
900 *out_nid = memblock_get_region_node(m); 959 *out_nid = m_nid;
901
902 if (m_start >= r_start) 960 if (m_start >= r_start)
903 mi--; 961 idx_a--;
904 else 962 else
905 ri--; 963 idx_b--;
906 *idx = (u32)mi | (u64)ri << 32; 964 *idx = (u32)idx_a | (u64)idx_b << 32;
907 return; 965 return;
908 } 966 }
909 } 967 }
910 } 968 }
911 969 /* signal end of iteration */
912 *idx = ULLONG_MAX; 970 *idx = ULLONG_MAX;
913} 971}
914 972
@@ -1201,7 +1259,7 @@ void __init __memblock_free_early(phys_addr_t base, phys_addr_t size)
1201 __func__, (u64)base, (u64)base + size - 1, 1259 __func__, (u64)base, (u64)base + size - 1,
1202 (void *)_RET_IP_); 1260 (void *)_RET_IP_);
1203 kmemleak_free_part(__va(base), size); 1261 kmemleak_free_part(__va(base), size);
1204 __memblock_remove(&memblock.reserved, base, size); 1262 memblock_remove_range(&memblock.reserved, base, size);
1205} 1263}
1206 1264
1207/* 1265/*
@@ -1287,8 +1345,10 @@ void __init memblock_enforce_memory_limit(phys_addr_t limit)
1287 } 1345 }
1288 1346
1289 /* truncate both memory and reserved regions */ 1347 /* truncate both memory and reserved regions */
1290 __memblock_remove(&memblock.memory, max_addr, (phys_addr_t)ULLONG_MAX); 1348 memblock_remove_range(&memblock.memory, max_addr,
1291 __memblock_remove(&memblock.reserved, max_addr, (phys_addr_t)ULLONG_MAX); 1349 (phys_addr_t)ULLONG_MAX);
1350 memblock_remove_range(&memblock.reserved, max_addr,
1351 (phys_addr_t)ULLONG_MAX);
1292} 1352}
1293 1353
1294static int __init_memblock memblock_search(struct memblock_type *type, phys_addr_t addr) 1354static int __init_memblock memblock_search(struct memblock_type *type, phys_addr_t addr)
@@ -1502,6 +1562,9 @@ static int __init memblock_init_debugfs(void)
1502 return -ENXIO; 1562 return -ENXIO;
1503 debugfs_create_file("memory", S_IRUGO, root, &memblock.memory, &memblock_debug_fops); 1563 debugfs_create_file("memory", S_IRUGO, root, &memblock.memory, &memblock_debug_fops);
1504 debugfs_create_file("reserved", S_IRUGO, root, &memblock.reserved, &memblock_debug_fops); 1564 debugfs_create_file("reserved", S_IRUGO, root, &memblock.reserved, &memblock_debug_fops);
1565#ifdef CONFIG_HAVE_MEMBLOCK_PHYS_MAP
1566 debugfs_create_file("physmem", S_IRUGO, root, &memblock.physmem, &memblock_debug_fops);
1567#endif
1505 1568
1506 return 0; 1569 return 0;
1507} 1570}
generated by cgit v1.2.2 (git 2.25.0) at 2025-08-22 10:08:58 -0400