Skip to contentSkip to conversation

kdump: pass e820 reserved region to 2nd kernel via e820 table or setup_data

PATCH Superseded WANG Chao 2 participants 4 messages Apr 28, 2014 – Apr 29, 2014

Subsystems (2)
lore
Share & tools
lore

WANG Chao

author
#0 lore raw

PATCH 1/1 kdump: pass e820 reserved region to 2nd kernel via e820 table or setup_data

e820 reserved region could be useful in 2nd kernel.

For example, PCI mmconf (extended mode) requires reserved region otherwise it falls back to legacy mode. The following log is from Cliff Wickman <[email protected]>:

PCI: MMCONFIG for domain 1003 [bus 3f-3f] at [mem 0xff0ff00000-0xff0fffffff] (base 0xff0c000000) [Firmware Bug]: PCI: MMCONFIG at [mem 0x80000000-0x80cfffff] not reserved in ACPI motherboard resources
PCI: not using MMCONFIG

PCI devices on segment 1 (>0) can't fall back to legacy mode, thus kernel probing fails and device can't be found.

We don't pass reserved region because these regions could be too much and eat up our very limited kernel command line resource in memmap=exactmap case.

However now we use e820 map and setup_data to pass memory map to 2nd kernel and the number of reserved regions should not be a problem any more.

---
 kexec/arch/i386/crashdump-x86.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)
kexec/arch/i386/crashdump-x86.c +6 -2
index 4a1491b..321ed3d 100644
@@ -300,6 +300,9 @@ static int get_crash_memory_ranges(struct memory_range **range, int *ranges,
 			type = RANGE_ACPI;
 		} else if(memcmp(str,"ACPI Non-volatile Storage\n",26) == 0 ) {
 			type = RANGE_ACPI_NVS;
+		} else if(memcmp(str,"reserved\n",9) == 0 ) {
+			type = RANGE_RANGE;
+		} else if (memcmp(str, "GART\n", 5) == 0) {
 		} else if (memcmp(str, "GART\n", 5) == 0) {
 			gart_start = start;
 			gart_end = end;
@@ -989,14 +992,15 @@ int load_crashdump_segments(struct kexec_info *info, char* mod_cmdline,
 	for (i = 0; i < CRASH_MAX_MEMORY_RANGES; i++) {
 		unsigned long start, end, size, type;
 		if ( !( mem_range[i].type == RANGE_ACPI
-			|| mem_range[i].type == RANGE_ACPI_NVS) )
+			|| mem_range[i].type == RANGE_ACPI_NVS
+			|| mem_range[i].type == RANGE_RESERVED))
 			continue;
 		start = mem_range[i].start;
 		end = mem_range[i].end;
 		type = mem_range[i].type;
 		size = end - start + 1;
 		add_memmap(memmap_p, &nr_memmap, start, size, type);
-		if (arch_options.pass_memmap_cmdline)
+		if (arch_options.pass_memmap_cmdline || type != RANGE_RESERVED)
 			cmdline_add_memmap_acpi(mod_cmdline, start, end);
 	}
 
Signature · 7 lines
1.9.0


_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec
author
#1 lore raw

PATCH RESEND 1/1 kdump: pass e820 reserved region to 2nd kernel via e820 table or setup_data

e820 reserved region could be useful in 2nd kernel.

For example, PCI mmconf (extended mode) requires reserved region otherwise it falls back to legacy mode. The following log is from Cliff Wickman <[email protected]>:

PCI: MMCONFIG for domain 1003 [bus 3f-3f] at [mem 0xff0ff00000-0xff0fffffff] (base 0xff0c000000) [Firmware Bug]: PCI: MMCONFIG at [mem 0x80000000-0x80cfffff] not reserved in ACPI motherboard resources
PCI: not using MMCONFIG

PCI devices on segment 1 (>0) can't fall back to legacy mode, thus kernel probing fails and device can't be found.

We don't pass reserved region because these regions could be too much and eat up our very limited kernel command line resource in memmap=exactmap case.

However now we use e820 map and setup_data to pass memory map to 2nd kernel and the number of reserved regions should not be a problem any more.

---
 kexec/arch/i386/crashdump-x86.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)
kexec/arch/i386/crashdump-x86.c +6 -2
index 4a1491b..cc33347 100644
@@ -300,6 +300,9 @@ static int get_crash_memory_ranges(struct memory_range **range, int *ranges,
 			type = RANGE_ACPI;
 		} else if(memcmp(str,"ACPI Non-volatile Storage\n",26) == 0 ) {
 			type = RANGE_ACPI_NVS;
+		} else if(memcmp(str,"reserved\n",9) == 0 ) {
+			type = RANGE_RESERVED;
+		} else if (memcmp(str, "GART\n", 5) == 0) {
 		} else if (memcmp(str, "GART\n", 5) == 0) {
 			gart_start = start;
 			gart_end = end;
@@ -989,14 +992,15 @@ int load_crashdump_segments(struct kexec_info *info, char* mod_cmdline,
 	for (i = 0; i < CRASH_MAX_MEMORY_RANGES; i++) {
 		unsigned long start, end, size, type;
 		if ( !( mem_range[i].type == RANGE_ACPI
-			|| mem_range[i].type == RANGE_ACPI_NVS) )
+			|| mem_range[i].type == RANGE_ACPI_NVS
+			|| mem_range[i].type == RANGE_RESERVED))
 			continue;
 		start = mem_range[i].start;
 		end = mem_range[i].end;
 		type = mem_range[i].type;
 		size = end - start + 1;
 		add_memmap(memmap_p, &nr_memmap, start, size, type);
-		if (arch_options.pass_memmap_cmdline)
+		if (arch_options.pass_memmap_cmdline || type != RANGE_RESERVED)
 			cmdline_add_memmap_acpi(mod_cmdline, start, end);
 	}
 
Signature · 7 lines
1.9.0


_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec
Applied
#2 lore raw
Quoted 17 lines · WANG Chao

e820 reserved region could be useful in 2nd kernel.

For example, PCI mmconf (extended mode) requires reserved region otherwise it falls back to legacy mode. The following log is from Cliff Wickman <[email protected]>:

PCI: MMCONFIG for domain 1003 [bus 3f-3f] at [mem 0xff0ff00000-0xff0fffffff] (base 0xff0c000000) [Firmware Bug]: PCI: MMCONFIG at [mem 0x80000000-0x80cfffff] not reserved in ACPI motherboard resources
PCI: not using MMCONFIG

PCI devices on segment 1 (>0) can't fall back to legacy mode, thus kernel probing fails and device can't be found.

We don't pass reserved region because these regions could be too much and eat up our very limited kernel command line resource in memmap=exactmap case.

Quoted 5 lines

However now we use e820 map and setup_data to pass memory map to 2nd kernel and the number of reserved regions should not be a problem any more.

Thanks, applied.

_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec

author
#3 lore raw
In reply to WANG Chao
Quoted 31 lines · WANG Chao

e820 reserved region could be useful in 2nd kernel.

For example, PCI mmconf (extended mode) requires reserved region otherwise it falls back to legacy mode. The following log is from Cliff Wickman <[email protected]>:

PCI: MMCONFIG for domain 1003 [bus 3f-3f] at [mem 0xff0ff00000-0xff0fffffff] (base 0xff0c000000) [Firmware Bug]: PCI: MMCONFIG at [mem 0x80000000-0x80cfffff] not reserved in ACPI motherboard resources
PCI: not using MMCONFIG

PCI devices on segment 1 (>0) can't fall back to legacy mode, thus kernel probing fails and device can't be found.

We don't pass reserved region because these regions could be too much and eat up our very limited kernel command line resource in memmap=exactmap case.

However now we use e820 map and setup_data to pass memory map to 2nd kernel and the number of reserved regions should not be a problem any more.

---
 kexec/arch/i386/crashdump-x86.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)
kexec/arch/i386/crashdump-x86.c +0 -0
index 4a1491b..321ed3d 100644

kexec/arch/i386/crashdump-x86.c static int get_crash_memory_ranges(struct memory_range **range, int *ranges, · line 304

Quoted 5 lines
kexec/arch/i386/crashdump-x86.c +2 -0
@@ -300,6 +300,9 @@ static int get_crash_memory_ranges(struct memory_range **range, int *ranges,
 			type = RANGE_ACPI;
 		} else if(memcmp(str,"ACPI Non-volatile Storage\n",26) == 0 ) {
 			type = RANGE_ACPI_NVS;
+		} else if(memcmp(str,"reserved\n",9) == 0 ) {
+			type = RANGE_RANGE;

This is wrong. I've resend the correct one.

_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec