s390/kernel: introduce .dma sections

With a relocatable kernel that could reside at any place in memory, code
and data that has to stay below 2 GB needs special handling.

This patch introduces .dma sections for such text, data and ex_table.
The sections will be part of the decompressor kernel, so they will not
be relocated and stay below 2 GB. Their location is passed over to the
decompressed / relocated kernel via the .boot.preserved.data section.

The duald and aste for control register setup also need to stay below
2 GB, so move the setup code from arch/s390/kernel/head64.S to
arch/s390/boot/head.S. The duct and linkage_stack could reside above
2 GB, but their content has to be preserved for the decompresed kernel,
so they are also moved into the .dma section.

The start and end address of the .dma sections is added to vmcoreinfo,
for crash support, to help debugging in case the kernel crashed there.

Signed-off-by: Gerald Schaefer <gerald.schaefer@de.ibm.com>
Reviewed-by: Philipp Rudo <prudo@linux.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>

5 files changed, 259 insertions, 2 deletions

diff --git a/arch/s390/boot/Makefile b/arch/s390/boot/Makefile
index 4df43e83363a..88932c25ad26 100644
--- a/arch/s390/boot/Makefile
+++ b/arch/s390/boot/Makefile
@@ -30,7 +30,7 @@ CFLAGS_sclp_early_core.o += -I$(srctree)/drivers/s390/char
 
 obj-y	:= head.o als.o startup.o mem_detect.o ipl_parm.o ipl_report.o
 obj-y	+= string.o ebcdic.o sclp_early_core.o mem.o ipl_vmparm.o cmdline.o
-obj-y	+= ctype.o
+obj-y	+= ctype.o text_dma.o
 obj-$(CONFIG_PROTECTED_VIRTUALIZATION_GUEST)	+= uv.o
 obj-$(CONFIG_RELOCATABLE)	+= machine_kexec_reloc.o
 targets	:= bzImage startup.a section_cmp.boot.data section_cmp.boot.preserved.data $(obj-y)
diff --git a/arch/s390/boot/compressed/vmlinux.lds.S b/arch/s390/boot/compressed/vmlinux.lds.S
index 1c8ef393c656..112b8d9f1e4c 100644
--- a/arch/s390/boot/compressed/vmlinux.lds.S
+++ b/arch/s390/boot/compressed/vmlinux.lds.S
@@ -33,6 +33,27 @@ SECTIONS
 		*(.data.*)
 		_edata = . ;
 	}
+	/*
+	* .dma section for code, data, ex_table that need to stay below 2 GB,
+	* even when the kernel is relocate: above 2 GB.
+	*/
+	_sdma = .;
+	.dma.text : {
+		. = ALIGN(PAGE_SIZE);
+		_stext_dma = .;
+		*(.dma.text)
+		. = ALIGN(PAGE_SIZE);
+		_etext_dma = .;
+	}
+	. = ALIGN(16);
+	.dma.ex_table : {
+		_start_dma_ex_table = .;
+		KEEP(*(.dma.ex_table))
+		_stop_dma_ex_table = .;
+	}
+	.dma.data : { *(.dma.data) }
+	_edma = .;
+
 	BOOT_DATA
 	BOOT_DATA_PRESERVED
 
diff --git a/arch/s390/boot/head.S b/arch/s390/boot/head.S
index c6b4b4c5dd58..028aab03a9e7 100644
--- a/arch/s390/boot/head.S
+++ b/arch/s390/boot/head.S
@@ -305,7 +305,7 @@ ENTRY(startup_kdump)
 	xc	0x300(256),0x300
 	xc	0xe00(256),0xe00
 	xc	0xf00(256),0xf00
-	lctlg	%c0,%c15,0x200(%r0)	# initialize control registers
+	lctlg	%c0,%c15,.Lctl-.LPG0(%r13)	# load control registers
 	stcke	__LC_BOOT_CLOCK
 	mvc	__LC_LAST_UPDATE_CLOCK(8),__LC_BOOT_CLOCK+1
 	spt	6f-.LPG0(%r13)
@@ -319,6 +319,36 @@ ENTRY(startup_kdump)
 	.align	8
 6:	.long	0x7fffffff,0xffffffff
 
+.Lctl:	.quad	0x04040000		# cr0: AFP registers & secondary space
+	.quad	0			# cr1: primary space segment table
+	.quad	.Lduct			# cr2: dispatchable unit control table
+	.quad	0			# cr3: instruction authorization
+	.quad	0xffff			# cr4: instruction authorization
+	.quad	.Lduct			# cr5: primary-aste origin
+	.quad	0			# cr6:	I/O interrupts
+	.quad	0			# cr7:	secondary space segment table
+	.quad	0			# cr8:	access registers translation
+	.quad	0			# cr9:	tracing off
+	.quad	0			# cr10: tracing off
+	.quad	0			# cr11: tracing off
+	.quad	0			# cr12: tracing off
+	.quad	0			# cr13: home space segment table
+	.quad	0xc0000000		# cr14: machine check handling off
+	.quad	.Llinkage_stack		# cr15: linkage stack operations
+
+	.section .dma.data,"aw",@progbits
+.Lduct: .long	0,.Laste,.Laste,0,.Lduald,0,0,0
+	.long	0,0,0,0,0,0,0,0
+.Llinkage_stack:
+	.long	0,0,0x89000000,0,0,0,0x8a000000,0
+	.align 64
+.Laste:	.quad	0,0xffffffffffffffff,0,0,0,0,0,0
+	.align	128
+.Lduald:.rept	8
+	.long	0x80000000,0,0,0	# invalid access-list entries
+	.endr
+	.previous
+
 #include "head_kdump.S"
 
 #
diff --git a/arch/s390/boot/startup.c b/arch/s390/boot/startup.c
index b7d6a76cb5e9..e3f339d248ce 100644
--- a/arch/s390/boot/startup.c
+++ b/arch/s390/boot/startup.c
@@ -1,9 +1,11 @@
 // SPDX-License-Identifier: GPL-2.0
 #include <linux/string.h>
 #include <linux/elf.h>
+#include <asm/sections.h>
 #include <asm/setup.h>
 #include <asm/kexec.h>
 #include <asm/sclp.h>
+#include <asm/diag.h>
 #include <asm/uv.h>
 #include "compressed/decompressor.h"
 #include "boot.h"
@@ -11,6 +13,43 @@
 extern char __boot_data_start[], __boot_data_end[];
 extern char __boot_data_preserved_start[], __boot_data_preserved_end[];
 
+/*
+ * Some code and data needs to stay below 2 GB, even when the kernel would be
+ * relocated above 2 GB, because it has to use 31 bit addresses.
+ * Such code and data is part of the .dma section, and its location is passed
+ * over to the decompressed / relocated kernel via the .boot.preserved.data
+ * section.
+ */
+extern char _sdma[], _edma[];
+extern char _stext_dma[], _etext_dma[];
+extern struct exception_table_entry _start_dma_ex_table[];
+extern struct exception_table_entry _stop_dma_ex_table[];
+unsigned long __bootdata_preserved(__sdma) = __pa(&_sdma);
+unsigned long __bootdata_preserved(__edma) = __pa(&_edma);
+unsigned long __bootdata_preserved(__stext_dma) = __pa(&_stext_dma);
+unsigned long __bootdata_preserved(__etext_dma) = __pa(&_etext_dma);
+struct exception_table_entry *
+	__bootdata_preserved(__start_dma_ex_table) = _start_dma_ex_table;
+struct exception_table_entry *
+	__bootdata_preserved(__stop_dma_ex_table) = _stop_dma_ex_table;
+
+int _diag210_dma(struct diag210 *addr);
+int _diag26c_dma(void *req, void *resp, enum diag26c_sc subcode);
+int _diag14_dma(unsigned long rx, unsigned long ry1, unsigned long subcode);
+void _diag0c_dma(struct hypfs_diag0c_entry *entry);
+void _diag308_reset_dma(void);
+struct diag_ops __bootdata_preserved(diag_dma_ops) = {
+	.diag210 = _diag210_dma,
+	.diag26c = _diag26c_dma,
+	.diag14 = _diag14_dma,
+	.diag0c = _diag0c_dma,
+	.diag308_reset = _diag308_reset_dma
+};
+static struct diag210 _diag210_tmp_dma __section(".dma.data");
+struct diag210 *__bootdata_preserved(__diag210_tmp_dma) = &_diag210_tmp_dma;
+void _swsusp_reset_dma(void);
+unsigned long __bootdata_preserved(__swsusp_reset_dma) = __pa(_swsusp_reset_dma);
+
 void error(char *x)
 {
 	sclp_early_printk("\n\n");
diff --git a/arch/s390/boot/text_dma.S b/arch/s390/boot/text_dma.S
new file mode 100644
index 000000000000..2414360ff22d
--- /dev/null
+++ b/arch/s390/boot/text_dma.S
@@ -0,0 +1,167 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Code that needs to run below 2 GB.
+ *
+ * Copyright IBM Corp. 2019
+ */
+
+#include <linux/linkage.h>
+#include <asm/errno.h>
+#include <asm/sigp.h>
+
+	.section .dma.text,"ax"
+/*
+ * Simplified version of expoline thunk. The normal thunks can not be used here,
+ * because they might be more than 2 GB away, and not reachable by the relative
+ * branch. No comdat, exrl, etc. optimizations used here, because it only
+ * affects a few functions that are not performance-relevant.
+ */
+	.macro BR_EX_DMA_r14
+	larl	%r1,0f
+	ex	0,0(%r1)
+	j	.
+0:	br	%r14
+	.endm
+
+/*
+ * int _diag14_dma(unsigned long rx, unsigned long ry1, unsigned long subcode)
+ */
+ENTRY(_diag14_dma)
+	lgr	%r1,%r2
+	lgr	%r2,%r3
+	lgr	%r3,%r4
+	lhi	%r5,-EIO
+	sam31
+	diag	%r1,%r2,0x14
+.Ldiag14_ex:
+	ipm	%r5
+	srl	%r5,28
+.Ldiag14_fault:
+	sam64
+	lgfr	%r2,%r5
+	BR_EX_DMA_r14
+	EX_TABLE_DMA(.Ldiag14_ex, .Ldiag14_fault)
+
+/*
+ * int _diag210_dma(struct diag210 *addr)
+ */
+ENTRY(_diag210_dma)
+	lgr	%r1,%r2
+	lhi	%r2,-1
+	sam31
+	diag	%r1,%r0,0x210
+.Ldiag210_ex:
+	ipm	%r2
+	srl	%r2,28
+.Ldiag210_fault:
+	sam64
+	lgfr	%r2,%r2
+	BR_EX_DMA_r14
+	EX_TABLE_DMA(.Ldiag210_ex, .Ldiag210_fault)
+
+/*
+ * int _diag26c_dma(void *req, void *resp, enum diag26c_sc subcode)
+ */
+ENTRY(_diag26c_dma)
+	lghi	%r5,-EOPNOTSUPP
+	sam31
+	diag	%r2,%r4,0x26c
+.Ldiag26c_ex:
+	sam64
+	lgfr	%r2,%r5
+	BR_EX_DMA_r14
+	EX_TABLE_DMA(.Ldiag26c_ex, .Ldiag26c_ex)
+
+/*
+ * void _diag0c_dma(struct hypfs_diag0c_entry *entry)
+ */
+ENTRY(_diag0c_dma)
+	sam31
+	diag	%r2,%r2,0x0c
+	sam64
+	BR_EX_DMA_r14
+
+/*
+ * void _swsusp_reset_dma(void)
+ */
+ENTRY(_swsusp_reset_dma)
+	larl	%r1,restart_entry
+	larl	%r2,.Lrestart_diag308_psw
+	og	%r1,0(%r2)
+	stg	%r1,0(%r0)
+	lghi	%r0,0
+	diag	%r0,%r0,0x308
+restart_entry:
+	lhi	%r1,1
+	sigp	%r1,%r0,SIGP_SET_ARCHITECTURE
+	sam64
+	BR_EX_DMA_r14
+
+/*
+ * void _diag308_reset_dma(void)
+ *
+ * Calls diag 308 subcode 1 and continues execution
+ */
+ENTRY(_diag308_reset_dma)
+	larl	%r4,.Lctlregs		# Save control registers
+	stctg	%c0,%c15,0(%r4)
+	lg	%r2,0(%r4)		# Disable lowcore protection
+	nilh	%r2,0xefff
+	larl	%r4,.Lctlreg0
+	stg	%r2,0(%r4)
+	lctlg	%c0,%c0,0(%r4)
+	larl	%r4,.Lfpctl		# Floating point control register
+	stfpc	0(%r4)
+	larl	%r4,.Lprefix		# Save prefix register
+	stpx	0(%r4)
+	larl	%r4,.Lprefix_zero	# Set prefix register to 0
+	spx	0(%r4)
+	larl	%r4,.Lcontinue_psw	# Save PSW flags
+	epsw	%r2,%r3
+	stm	%r2,%r3,0(%r4)
+	larl	%r4,restart_part2	# Setup restart PSW at absolute 0
+	larl	%r3,.Lrestart_diag308_psw
+	og	%r4,0(%r3)		# Save PSW
+	lghi	%r3,0
+	sturg	%r4,%r3			# Use sturg, because of large pages
+	lghi	%r1,1
+	lghi	%r0,0
+	diag	%r0,%r1,0x308
+restart_part2:
+	lhi	%r0,0			# Load r0 with zero
+	lhi	%r1,2			# Use mode 2 = ESAME (dump)
+	sigp	%r1,%r0,SIGP_SET_ARCHITECTURE	# Switch to ESAME mode
+	sam64				# Switch to 64 bit addressing mode
+	larl	%r4,.Lctlregs		# Restore control registers
+	lctlg	%c0,%c15,0(%r4)
+	larl	%r4,.Lfpctl		# Restore floating point ctl register
+	lfpc	0(%r4)
+	larl	%r4,.Lprefix		# Restore prefix register
+	spx	0(%r4)
+	larl	%r4,.Lcontinue_psw	# Restore PSW flags
+	lpswe	0(%r4)
+.Lcontinue:
+	BR_EX_DMA_r14
+
+	.section .dma.data,"aw",@progbits
+.align	8
+.Lrestart_diag308_psw:
+	.long	0x00080000,0x80000000
+
+.align 8
+.Lcontinue_psw:
+	.quad	0,.Lcontinue
+
+.align 8
+.Lctlreg0:
+	.quad	0
+.Lctlregs:
+	.rept	16
+	.quad	0
+	.endr
+.Lfpctl:
+	.long	0
+.Lprefix:
+	.long	0
+.Lprefix_zero:
+	.long	0