linux-efi.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH v3 0/8] x86, boot: clean up setup_data handling
@ 2015-03-08  0:50 Yinghai Lu
  2015-03-08  0:50 ` [PATCH v3 1/8] x86: Kill E820_RESERVED_KERN Yinghai Lu
  2015-03-08  0:50 ` [PATCH v3 2/8] x86, efi: Copy SETUP_EFI data and access directly Yinghai Lu
  0 siblings, 2 replies; 7+ messages in thread
From: Yinghai Lu @ 2015-03-08  0:50 UTC (permalink / raw)
  To: Matt Fleming, H. Peter Anvin, Ingo Molnar, Borislav Petkov,
	Bjorn Helgaas
  Cc: Thomas Gleixner, Jiri Kosina, Chun-Yi Lee,
	linux-kernel-u79uwXL29TY76Z2rM5mHXA,
	linux-efi-u79uwXL29TY76Z2rM5mHXA,
	linux-pci-u79uwXL29TY76Z2rM5mHXA, Yinghai Lu

Now setup_data is reserved via memblock and e820 and different
handlers have different ways, and it is confusing.
1. SETUP_E820_EXT: is consumed early and will not copy or access again.
        have memory wasted.
2. SETUP_EFI: is accessed via ioremap every time at early stage.
        have memory wasted.
3. SETUP_DTB: is copied locally.
        have memory wasted.
4. SETUP_PCI: is accessed via ioremap for every pci devices, even run-time.
5. SETUP_KASLR: is accessed early, will not copy or access again.
        have memory wasted.

Also setup_data is exported to debugfs for debug purpose.

Here will convert to let every handler to decide how to handle it.
and will not reserve the setup_data generally, so will not
waste memory and also make memblock/e820 keep page aligned.
1. not touch E820 anymore.
2. copy SETUP_EFI to __initdata variable and access it without ioremap.
3. SETUP_DTB: reserver and copy to local and free.
4. SETUP_PCI: reverve localy and convert to list, to avoid keeping ioremap.
5. export SETUP_PCI via sysfs.

Also put them in:
git://git.kernel.org/pub/scm/linux/kernel/git/yinghai/linux-yinghai.git for-x86-4.0-rc2-setup_data

Should be materials for v4.1.

Thanks

Yinghai Lu

-v3: separated with kasl patches, and fix early_ioremap return checking.

Yinghai Lu (8):
  x86: Kill E820_RESERVED_KERN
  x86, efi: Copy SETUP_EFI data and access directly
  x86, of: Let add_dtb reserve setup_data locally
  x86, boot: Add add_pci handler for SETUP_PCI
  x86: Kill not used setup_data handling code
  x86, boot, PCI: Convert SETUP_PCI data to list
  x86, boot, PCI: Copy SETUP_PCI rom to kernel space
  x86, boot, PCI: Export SETUP_PCI data via sysfs

 arch/x86/include/asm/efi.h       |   2 +-
 arch/x86/include/asm/pci.h       |   4 +
 arch/x86/include/asm/prom.h      |   9 +-
 arch/x86/include/uapi/asm/e820.h |   9 --
 arch/x86/kernel/devicetree.c     |  43 +++---
 arch/x86/kernel/e820.c           |   6 +-
 arch/x86/kernel/kdebugfs.c       | 142 ------------------
 arch/x86/kernel/setup.c          |  52 ++-----
 arch/x86/kernel/tboot.c          |   3 +-
 arch/x86/mm/init_64.c            |  11 +-
 arch/x86/pci/common.c            | 316 ++++++++++++++++++++++++++++++++++++---
 arch/x86/platform/efi/efi.c      |  13 +-
 arch/x86/platform/efi/efi_64.c   |  13 +-
 arch/x86/platform/efi/quirks.c   |  23 +--
 14 files changed, 371 insertions(+), 275 deletions(-)

-- 
1.8.4.5

^ permalink raw reply	[flat|nested] 7+ messages in thread

* [PATCH v3 1/8] x86: Kill E820_RESERVED_KERN
  2015-03-08  0:50 [PATCH v3 0/8] x86, boot: clean up setup_data handling Yinghai Lu
@ 2015-03-08  0:50 ` Yinghai Lu
  2015-03-08  0:50 ` [PATCH v3 2/8] x86, efi: Copy SETUP_EFI data and access directly Yinghai Lu
  1 sibling, 0 replies; 7+ messages in thread
From: Yinghai Lu @ 2015-03-08  0:50 UTC (permalink / raw)
  To: Matt Fleming, H. Peter Anvin, Ingo Molnar, Borislav Petkov,
	Bjorn Helgaas
  Cc: Thomas Gleixner, Jiri Kosina, Chun-Yi Lee, linux-kernel,
	linux-efi, linux-pci, Yinghai Lu, stable

Now we are using memblock to do early resource reserver/allocation
instead of using e820 map directly, and setup_data is reserved in
memblock early already.
Also kexec generate setup_data and pass pointer to second kernel,
so second kernel reserve setup_data by their own.
(Now kexec-tools create SETUP_EFI and SETUP_E820_EXT).

We can kill E820_RESERVED_KERN and not touch e820 map at all.

That will fix bug in mark_nonsave_region that can not handle that
case: E820_RAM and E820_RESERVED_KERN ranges are continuous and
boundary is not page aligned.

Bugzilla: https://bugzilla.opensuse.org/show_bug.cgi?id=913885
Reported-by: "Lee, Chun-Yi" <jlee@suse.com>
Tested-by: "Lee, Chun-Yi" <jlee@suse.com>
Cc: "Lee, Chun-Yi" <jlee@suse.com>
Signed-off-by: Yinghai Lu <yinghai@kernel.org>
Cc: stable@vger.kernel.org
---
 arch/x86/include/uapi/asm/e820.h |  9 ---------
 arch/x86/kernel/e820.c           |  6 ++----
 arch/x86/kernel/setup.c          | 26 --------------------------
 arch/x86/kernel/tboot.c          |  3 +--
 arch/x86/mm/init_64.c            | 11 ++++-------
 5 files changed, 7 insertions(+), 48 deletions(-)

diff --git a/arch/x86/include/uapi/asm/e820.h b/arch/x86/include/uapi/asm/e820.h
index d993e33..edc8a71 100644
--- a/arch/x86/include/uapi/asm/e820.h
+++ b/arch/x86/include/uapi/asm/e820.h
@@ -33,15 +33,6 @@
 #define E820_NVS	4
 #define E820_UNUSABLE	5
 
-
-/*
- * reserved RAM used by kernel itself
- * if CONFIG_INTEL_TXT is enabled, memory of this type will be
- * included in the S3 integrity calculation and so should not include
- * any memory that BIOS might alter over the S3 transition
- */
-#define E820_RESERVED_KERN        128
-
 #ifndef __ASSEMBLY__
 #include <linux/types.h>
 struct e820entry {
diff --git a/arch/x86/kernel/e820.c b/arch/x86/kernel/e820.c
index 46201de..2a6bed9 100644
--- a/arch/x86/kernel/e820.c
+++ b/arch/x86/kernel/e820.c
@@ -134,7 +134,6 @@ static void __init e820_print_type(u32 type)
 {
 	switch (type) {
 	case E820_RAM:
-	case E820_RESERVED_KERN:
 		printk(KERN_CONT "usable");
 		break;
 	case E820_RESERVED:
@@ -688,7 +687,7 @@ void __init e820_mark_nosave_regions(unsigned long limit_pfn)
 			register_nosave_region(pfn, PFN_UP(ei->addr));
 
 		pfn = PFN_DOWN(ei->addr + ei->size);
-		if (ei->type != E820_RAM && ei->type != E820_RESERVED_KERN)
+		if (ei->type != E820_RAM)
 			register_nosave_region(PFN_UP(ei->addr), pfn);
 
 		if (pfn >= limit_pfn)
@@ -902,7 +901,6 @@ void __init finish_e820_parsing(void)
 static inline const char *e820_type_to_string(int e820_type)
 {
 	switch (e820_type) {
-	case E820_RESERVED_KERN:
 	case E820_RAM:	return "System RAM";
 	case E820_ACPI:	return "ACPI Tables";
 	case E820_NVS:	return "ACPI Non-volatile Storage";
@@ -1077,7 +1075,7 @@ void __init memblock_x86_fill(void)
 		if (end != (resource_size_t)end)
 			continue;
 
-		if (ei->type != E820_RAM && ei->type != E820_RESERVED_KERN)
+		if (ei->type != E820_RAM)
 			continue;
 
 		memblock_add(ei->addr, ei->size);
diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
index 912f124..aed343b 100644
--- a/arch/x86/kernel/setup.c
+++ b/arch/x86/kernel/setup.c
@@ -478,30 +478,6 @@ static void __init parse_setup_data(void)
 	}
 }
 
-static void __init e820_reserve_setup_data(void)
-{
-	struct setup_data *data;
-	u64 pa_data;
-	int found = 0;
-
-	pa_data = boot_params.hdr.setup_data;
-	while (pa_data) {
-		data = early_memremap(pa_data, sizeof(*data));
-		e820_update_range(pa_data, sizeof(*data)+data->len,
-			 E820_RAM, E820_RESERVED_KERN);
-		found = 1;
-		pa_data = data->next;
-		early_iounmap(data, sizeof(*data));
-	}
-	if (!found)
-		return;
-
-	sanitize_e820_map(e820.map, ARRAY_SIZE(e820.map), &e820.nr_map);
-	memcpy(&e820_saved, &e820, sizeof(struct e820map));
-	printk(KERN_INFO "extended physical RAM map:\n");
-	e820_print_map("reserve setup_data");
-}
-
 static void __init memblock_x86_reserve_range_setup_data(void)
 {
 	struct setup_data *data;
@@ -1037,8 +1013,6 @@ void __init setup_arch(char **cmdline_p)
 		early_dump_pci_devices();
 #endif
 
-	/* update the e820_saved too */
-	e820_reserve_setup_data();
 	finish_e820_parsing();
 
 	if (efi_enabled(EFI_BOOT))
diff --git a/arch/x86/kernel/tboot.c b/arch/x86/kernel/tboot.c
index 91a4496..3c2752a 100644
--- a/arch/x86/kernel/tboot.c
+++ b/arch/x86/kernel/tboot.c
@@ -195,8 +195,7 @@ static int tboot_setup_sleep(void)
 	tboot->num_mac_regions = 0;
 
 	for (i = 0; i < e820.nr_map; i++) {
-		if ((e820.map[i].type != E820_RAM)
-		 && (e820.map[i].type != E820_RESERVED_KERN))
+		if (e820.map[i].type != E820_RAM)
 			continue;
 
 		add_mac_region(e820.map[i].addr, e820.map[i].size);
diff --git a/arch/x86/mm/init_64.c b/arch/x86/mm/init_64.c
index c30efb6..63520ec 100644
--- a/arch/x86/mm/init_64.c
+++ b/arch/x86/mm/init_64.c
@@ -354,8 +354,7 @@ phys_pte_init(pte_t *pte_page, unsigned long addr, unsigned long end,
 		next = (addr & PAGE_MASK) + PAGE_SIZE;
 		if (addr >= end) {
 			if (!after_bootmem &&
-			    !e820_any_mapped(addr & PAGE_MASK, next, E820_RAM) &&
-			    !e820_any_mapped(addr & PAGE_MASK, next, E820_RESERVED_KERN))
+			    !e820_any_mapped(addr & PAGE_MASK, next, E820_RAM))
 				set_pte(pte, __pte(0));
 			continue;
 		}
@@ -401,9 +400,8 @@ phys_pmd_init(pmd_t *pmd_page, unsigned long address, unsigned long end,
 
 		next = (address & PMD_MASK) + PMD_SIZE;
 		if (address >= end) {
-			if (!after_bootmem &&
-			    !e820_any_mapped(address & PMD_MASK, next, E820_RAM) &&
-			    !e820_any_mapped(address & PMD_MASK, next, E820_RESERVED_KERN))
+			if (!after_bootmem && !e820_any_mapped(
+					address & PMD_MASK, next, E820_RAM))
 				set_pmd(pmd, __pmd(0));
 			continue;
 		}
@@ -476,8 +474,7 @@ phys_pud_init(pud_t *pud_page, unsigned long addr, unsigned long end,
 		next = (addr & PUD_MASK) + PUD_SIZE;
 		if (addr >= end) {
 			if (!after_bootmem &&
-			    !e820_any_mapped(addr & PUD_MASK, next, E820_RAM) &&
-			    !e820_any_mapped(addr & PUD_MASK, next, E820_RESERVED_KERN))
+			    !e820_any_mapped(addr & PUD_MASK, next, E820_RAM))
 				set_pud(pud, __pud(0));
 			continue;
 		}
-- 
1.8.4.5

^ permalink raw reply related	[flat|nested] 7+ messages in thread

* [PATCH v3 2/8] x86, efi: Copy SETUP_EFI data and access directly
  2015-03-08  0:50 [PATCH v3 0/8] x86, boot: clean up setup_data handling Yinghai Lu
  2015-03-08  0:50 ` [PATCH v3 1/8] x86: Kill E820_RESERVED_KERN Yinghai Lu
@ 2015-03-08  0:50 ` Yinghai Lu
  1 sibling, 0 replies; 7+ messages in thread
From: Yinghai Lu @ 2015-03-08  0:50 UTC (permalink / raw)
  To: Matt Fleming, H. Peter Anvin, Ingo Molnar, Borislav Petkov,
	Bjorn Helgaas
  Cc: Thomas Gleixner, Jiri Kosina, Chun-Yi Lee, linux-kernel,
	linux-efi, linux-pci, Yinghai Lu

The copy will be in __initdata, and it is small.

We can use pointer to access the setup_data instead of using early_memmap
everywhere.

Cc: Matt Fleming <matt.fleming@intel.com>
Cc: linux-efi@vger.kernel.org
Signed-off-by: Yinghai Lu <yinghai@kernel.org>
---
 arch/x86/include/asm/efi.h     |  2 +-
 arch/x86/platform/efi/efi.c    | 13 ++-----------
 arch/x86/platform/efi/efi_64.c | 13 ++++++++++++-
 arch/x86/platform/efi/quirks.c | 23 ++++++-----------------
 4 files changed, 21 insertions(+), 30 deletions(-)

diff --git a/arch/x86/include/asm/efi.h b/arch/x86/include/asm/efi.h
index 25bce45..edbecd6 100644
--- a/arch/x86/include/asm/efi.h
+++ b/arch/x86/include/asm/efi.h
@@ -114,7 +114,7 @@ struct efi_setup_data {
 	u64 reserved[8];
 };
 
-extern u64 efi_setup;
+extern struct efi_setup_data *efi_setup;
 
 #ifdef CONFIG_EFI
 
diff --git a/arch/x86/platform/efi/efi.c b/arch/x86/platform/efi/efi.c
index dbc8627..1cd38e8 100644
--- a/arch/x86/platform/efi/efi.c
+++ b/arch/x86/platform/efi/efi.c
@@ -68,7 +68,7 @@ static efi_config_table_type_t arch_tables[] __initdata = {
 	{NULL_GUID, NULL, NULL},
 };
 
-u64 efi_setup;		/* efi setup_data physical address */
+struct efi_setup_data *efi_setup __initdata; /* cached efi setup_data pointer */
 
 static int add_efi_memmap __initdata;
 static int __init setup_add_efi_memmap(char *arg)
@@ -225,20 +225,13 @@ static int __init efi_systab_init(void *phys)
 {
 	if (efi_enabled(EFI_64BIT)) {
 		efi_system_table_64_t *systab64;
-		struct efi_setup_data *data = NULL;
+		struct efi_setup_data *data = efi_setup;
 		u64 tmp = 0;
 
-		if (efi_setup) {
-			data = early_memremap(efi_setup, sizeof(*data));
-			if (!data)
-				return -ENOMEM;
-		}
 		systab64 = early_memremap((unsigned long)phys,
 					 sizeof(*systab64));
 		if (systab64 == NULL) {
 			pr_err("Couldn't map the system table!\n");
-			if (data)
-				early_memunmap(data, sizeof(*data));
 			return -ENOMEM;
 		}
 
@@ -271,8 +264,6 @@ static int __init efi_systab_init(void *phys)
 		tmp |= data ? data->tables : systab64->tables;
 
 		early_memunmap(systab64, sizeof(*systab64));
-		if (data)
-			early_memunmap(data, sizeof(*data));
 #ifdef CONFIG_X86_32
 		if (tmp >> 32) {
 			pr_err("EFI data located above 4GB, disabling EFI.\n");
diff --git a/arch/x86/platform/efi/efi_64.c b/arch/x86/platform/efi/efi_64.c
index 17e80d8..a541c6c 100644
--- a/arch/x86/platform/efi/efi_64.c
+++ b/arch/x86/platform/efi/efi_64.c
@@ -292,9 +292,20 @@ void __iomem *__init efi_ioremap(unsigned long phys_addr, unsigned long size,
 	return (void __iomem *)__va(phys_addr);
 }
 
+static struct efi_setup_data efi_setup_data __initdata;
+
 void __init parse_efi_setup(u64 phys_addr, u32 data_len)
 {
-	efi_setup = phys_addr + sizeof(struct setup_data);
+	struct efi_setup_data *data;
+
+	data = early_memremap(phys_addr + sizeof(struct setup_data),
+			      sizeof(*data));
+	if (!data)
+		return;
+
+	efi_setup_data = *data;
+	early_memunmap(data, sizeof(*data));
+	efi_setup = &efi_setup_data;
 }
 
 void __init efi_runtime_mkexec(void)
diff --git a/arch/x86/platform/efi/quirks.c b/arch/x86/platform/efi/quirks.c
index 1c7380d..45fec7d 100644
--- a/arch/x86/platform/efi/quirks.c
+++ b/arch/x86/platform/efi/quirks.c
@@ -203,9 +203,8 @@ void __init efi_free_boot_services(void)
  */
 int __init efi_reuse_config(u64 tables, int nr_tables)
 {
-	int i, sz, ret = 0;
+	int i, sz;
 	void *p, *tablep;
-	struct efi_setup_data *data;
 
 	if (!efi_setup)
 		return 0;
@@ -213,22 +212,15 @@ int __init efi_reuse_config(u64 tables, int nr_tables)
 	if (!efi_enabled(EFI_64BIT))
 		return 0;
 
-	data = early_memremap(efi_setup, sizeof(*data));
-	if (!data) {
-		ret = -ENOMEM;
-		goto out;
-	}
-
-	if (!data->smbios)
-		goto out_memremap;
+	if (!efi_setup->smbios)
+		return 0;
 
 	sz = sizeof(efi_config_table_64_t);
 
 	p = tablep = early_memremap(tables, nr_tables * sz);
 	if (!p) {
 		pr_err("Could not map Configuration table!\n");
-		ret = -ENOMEM;
-		goto out_memremap;
+		return -ENOMEM;
 	}
 
 	for (i = 0; i < efi.systab->nr_tables; i++) {
@@ -237,15 +229,12 @@ int __init efi_reuse_config(u64 tables, int nr_tables)
 		guid = ((efi_config_table_64_t *)p)->guid;
 
 		if (!efi_guidcmp(guid, SMBIOS_TABLE_GUID))
-			((efi_config_table_64_t *)p)->table = data->smbios;
+			((efi_config_table_64_t *)p)->table = efi_setup->smbios;
 		p += sz;
 	}
 	early_memunmap(tablep, nr_tables * sz);
 
-out_memremap:
-	early_memunmap(data, sizeof(*data));
-out:
-	return ret;
+	return 0;
 }
 
 void __init efi_apply_memmap_quirks(void)
-- 
1.8.4.5

^ permalink raw reply related	[flat|nested] 7+ messages in thread

* Re: [PATCH v3 1/8] x86: Kill E820_RESERVED_KERN
       [not found]       ` <alpine.DEB.2.10.1503071748530.5739-X6Q0R45D7oAcqpCFd4KODRPsWskHk0ljAL8bYrjMMd8@public.gmane.org>
  2015-03-08  6:51         ` Yinghai Lu
@ 2015-03-09  0:18         ` joeyli
  1 sibling, 0 replies; 7+ messages in thread
From: joeyli @ 2015-03-09  0:18 UTC (permalink / raw)
  To: David Rientjes
  Cc: Yinghai Lu, Matt Fleming, H. Peter Anvin, Ingo Molnar,
	Borislav Petkov, Bjorn Helgaas, Thomas Gleixner, Jiri Kosina,
	linux-kernel-u79uwXL29TY76Z2rM5mHXA,
	linux-efi-u79uwXL29TY76Z2rM5mHXA,
	linux-pci-u79uwXL29TY76Z2rM5mHXA, stable-u79uwXL29TY76Z2rM5mHXA

On Sat, Mar 07, 2015 at 05:59:14PM -0800, David Rientjes wrote:
> On Sat, 7 Mar 2015, Yinghai Lu wrote:
> 
> > Now we are using memblock to do early resource reserver/allocation
> > instead of using e820 map directly, and setup_data is reserved in
> > memblock early already.
> > Also kexec generate setup_data and pass pointer to second kernel,
> > so second kernel reserve setup_data by their own.
> > (Now kexec-tools create SETUP_EFI and SETUP_E820_EXT).
> > 
> > We can kill E820_RESERVED_KERN and not touch e820 map at all.
> > 
> > That will fix bug in mark_nonsave_region that can not handle that
> > case: E820_RAM and E820_RESERVED_KERN ranges are continuous and
> > boundary is not page aligned.
> > 
> > Bugzilla: https://bugzilla.opensuse.org/show_bug.cgi?id=913885
> 
> Is this the bug referenced in the commit message that is fixed?  If so, 
> it's only a bug for resume, correct?  I'm not sure if that's clear enough 
> just from the commit message, I was looking at this patch for an e820 
> problem I'm currently facing on 3.3.

Yinghai's patches fixed the e820 not page aligned issue that's one of the
issues on bug reporter's machine. I found another issue of the BIOS that
sometimes it doesn't really keep the e820 table unchanging for hibernate
resuming, this BIOS issue causes the total available page number checking
fail. I will file another openSUSE bug to separate those 2 issues.

> 
> > Reported-by: "Lee, Chun-Yi" <jlee-IBi9RG/b67k@public.gmane.org>
> > Tested-by: "Lee, Chun-Yi" <jlee-IBi9RG/b67k@public.gmane.org>
> > Cc: "Lee, Chun-Yi" <jlee-IBi9RG/b67k@public.gmane.org>
> > Signed-off-by: Yinghai Lu <yinghai-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>
> > Cc: stable-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
> 
> Hmm, although the bug is reported for a 3.12 kernel, I assume this is for 
> stable 3.10+?  If so, it should apply fine with the exception of removing 
> e820_reserve_setup_data() from setup_arch() rather than 
> memblock_x86_reserve_range_setup_data().  Or is it for 3.2 as well and 
> needs to be completely rebased for that kernel?

Thanks a lot!
Joey Lee

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [PATCH v3 1/8] x86: Kill E820_RESERVED_KERN
       [not found]       ` <alpine.DEB.2.10.1503071748530.5739-X6Q0R45D7oAcqpCFd4KODRPsWskHk0ljAL8bYrjMMd8@public.gmane.org>
@ 2015-03-08  6:51         ` Yinghai Lu
  2015-03-09  0:18         ` joeyli
  1 sibling, 0 replies; 7+ messages in thread
From: Yinghai Lu @ 2015-03-08  6:51 UTC (permalink / raw)
  To: David Rientjes
  Cc: Matt Fleming, H. Peter Anvin, Ingo Molnar, Borislav Petkov,
	Bjorn Helgaas, Thomas Gleixner, Jiri Kosina, Chun-Yi Lee,
	Linux Kernel Mailing List, linux-efi-u79uwXL29TY76Z2rM5mHXA,
	linux-pci-u79uwXL29TY76Z2rM5mHXA, stable-u79uwXL29TY76Z2rM5mHXA

On Sat, Mar 7, 2015 at 5:59 PM, David Rientjes <rientjes-hpIqsD4AKlfQT0dZR+AlfA@public.gmane.org> wrote:
>
> Hmm, although the bug is reported for a 3.12 kernel, I assume this is for
> stable 3.10+?  If so, it should apply fine with the exception of removing
> e820_reserve_setup_data() from setup_arch() rather than
> memblock_x86_reserve_range_setup_data().  Or is it for 3.2 as well and
> needs to be completely rebased for that kernel?

For 3.10+, we will need to following patches, otherwise will have warning for
SETUP_PCI with ioremap.

for 3.2 that does not SETUP_PCI, should be ok, but will need rebase.

Thanks

Yinghai

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [PATCH v3 1/8] x86: Kill E820_RESERVED_KERN
       [not found]   ` <1425776181-10219-2-git-send-email-yinghai-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>
@ 2015-03-08  1:59     ` David Rientjes
       [not found]       ` <alpine.DEB.2.10.1503071748530.5739-X6Q0R45D7oAcqpCFd4KODRPsWskHk0ljAL8bYrjMMd8@public.gmane.org>
  0 siblings, 1 reply; 7+ messages in thread
From: David Rientjes @ 2015-03-08  1:59 UTC (permalink / raw)
  To: Yinghai Lu
  Cc: Matt Fleming, H. Peter Anvin, Ingo Molnar, Borislav Petkov,
	Bjorn Helgaas, Thomas Gleixner, Jiri Kosina, Chun-Yi Lee,
	linux-kernel-u79uwXL29TY76Z2rM5mHXA,
	linux-efi-u79uwXL29TY76Z2rM5mHXA,
	linux-pci-u79uwXL29TY76Z2rM5mHXA, stable-u79uwXL29TY76Z2rM5mHXA

On Sat, 7 Mar 2015, Yinghai Lu wrote:

> Now we are using memblock to do early resource reserver/allocation
> instead of using e820 map directly, and setup_data is reserved in
> memblock early already.
> Also kexec generate setup_data and pass pointer to second kernel,
> so second kernel reserve setup_data by their own.
> (Now kexec-tools create SETUP_EFI and SETUP_E820_EXT).
> 
> We can kill E820_RESERVED_KERN and not touch e820 map at all.
> 
> That will fix bug in mark_nonsave_region that can not handle that
> case: E820_RAM and E820_RESERVED_KERN ranges are continuous and
> boundary is not page aligned.
> 
> Bugzilla: https://bugzilla.opensuse.org/show_bug.cgi?id=913885

Is this the bug referenced in the commit message that is fixed?  If so, 
it's only a bug for resume, correct?  I'm not sure if that's clear enough 
just from the commit message, I was looking at this patch for an e820 
problem I'm currently facing on 3.3.

> Reported-by: "Lee, Chun-Yi" <jlee-IBi9RG/b67k@public.gmane.org>
> Tested-by: "Lee, Chun-Yi" <jlee-IBi9RG/b67k@public.gmane.org>
> Cc: "Lee, Chun-Yi" <jlee-IBi9RG/b67k@public.gmane.org>
> Signed-off-by: Yinghai Lu <yinghai-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>
> Cc: stable-u79uwXL29TY76Z2rM5mHXA@public.gmane.org

Hmm, although the bug is reported for a 3.12 kernel, I assume this is for 
stable 3.10+?  If so, it should apply fine with the exception of removing 
e820_reserve_setup_data() from setup_arch() rather than 
memblock_x86_reserve_range_setup_data().  Or is it for 3.2 as well and 
needs to be completely rebased for that kernel?

^ permalink raw reply	[flat|nested] 7+ messages in thread

* [PATCH v3 1/8] x86: Kill E820_RESERVED_KERN
  2015-03-08  0:56 [PATCH v3 0/8] x86, boot: clean up setup_data handling Yinghai Lu
@ 2015-03-08  0:56 ` Yinghai Lu
       [not found]   ` <1425776181-10219-2-git-send-email-yinghai-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>
  0 siblings, 1 reply; 7+ messages in thread
From: Yinghai Lu @ 2015-03-08  0:56 UTC (permalink / raw)
  To: Matt Fleming, H. Peter Anvin, Ingo Molnar, Borislav Petkov,
	Bjorn Helgaas
  Cc: Thomas Gleixner, Jiri Kosina, Chun-Yi Lee, linux-kernel,
	linux-efi, linux-pci, Yinghai Lu, stable

Now we are using memblock to do early resource reserver/allocation
instead of using e820 map directly, and setup_data is reserved in
memblock early already.
Also kexec generate setup_data and pass pointer to second kernel,
so second kernel reserve setup_data by their own.
(Now kexec-tools create SETUP_EFI and SETUP_E820_EXT).

We can kill E820_RESERVED_KERN and not touch e820 map at all.

That will fix bug in mark_nonsave_region that can not handle that
case: E820_RAM and E820_RESERVED_KERN ranges are continuous and
boundary is not page aligned.

Bugzilla: https://bugzilla.opensuse.org/show_bug.cgi?id=913885
Reported-by: "Lee, Chun-Yi" <jlee@suse.com>
Tested-by: "Lee, Chun-Yi" <jlee@suse.com>
Cc: "Lee, Chun-Yi" <jlee@suse.com>
Signed-off-by: Yinghai Lu <yinghai@kernel.org>
Cc: stable@vger.kernel.org
---
 arch/x86/include/uapi/asm/e820.h |  9 ---------
 arch/x86/kernel/e820.c           |  6 ++----
 arch/x86/kernel/setup.c          | 26 --------------------------
 arch/x86/kernel/tboot.c          |  3 +--
 arch/x86/mm/init_64.c            | 11 ++++-------
 5 files changed, 7 insertions(+), 48 deletions(-)

diff --git a/arch/x86/include/uapi/asm/e820.h b/arch/x86/include/uapi/asm/e820.h
index d993e33..edc8a71 100644
--- a/arch/x86/include/uapi/asm/e820.h
+++ b/arch/x86/include/uapi/asm/e820.h
@@ -33,15 +33,6 @@
 #define E820_NVS	4
 #define E820_UNUSABLE	5
 
-
-/*
- * reserved RAM used by kernel itself
- * if CONFIG_INTEL_TXT is enabled, memory of this type will be
- * included in the S3 integrity calculation and so should not include
- * any memory that BIOS might alter over the S3 transition
- */
-#define E820_RESERVED_KERN        128
-
 #ifndef __ASSEMBLY__
 #include <linux/types.h>
 struct e820entry {
diff --git a/arch/x86/kernel/e820.c b/arch/x86/kernel/e820.c
index 46201de..2a6bed9 100644
--- a/arch/x86/kernel/e820.c
+++ b/arch/x86/kernel/e820.c
@@ -134,7 +134,6 @@ static void __init e820_print_type(u32 type)
 {
 	switch (type) {
 	case E820_RAM:
-	case E820_RESERVED_KERN:
 		printk(KERN_CONT "usable");
 		break;
 	case E820_RESERVED:
@@ -688,7 +687,7 @@ void __init e820_mark_nosave_regions(unsigned long limit_pfn)
 			register_nosave_region(pfn, PFN_UP(ei->addr));
 
 		pfn = PFN_DOWN(ei->addr + ei->size);
-		if (ei->type != E820_RAM && ei->type != E820_RESERVED_KERN)
+		if (ei->type != E820_RAM)
 			register_nosave_region(PFN_UP(ei->addr), pfn);
 
 		if (pfn >= limit_pfn)
@@ -902,7 +901,6 @@ void __init finish_e820_parsing(void)
 static inline const char *e820_type_to_string(int e820_type)
 {
 	switch (e820_type) {
-	case E820_RESERVED_KERN:
 	case E820_RAM:	return "System RAM";
 	case E820_ACPI:	return "ACPI Tables";
 	case E820_NVS:	return "ACPI Non-volatile Storage";
@@ -1077,7 +1075,7 @@ void __init memblock_x86_fill(void)
 		if (end != (resource_size_t)end)
 			continue;
 
-		if (ei->type != E820_RAM && ei->type != E820_RESERVED_KERN)
+		if (ei->type != E820_RAM)
 			continue;
 
 		memblock_add(ei->addr, ei->size);
diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
index 912f124..aed343b 100644
--- a/arch/x86/kernel/setup.c
+++ b/arch/x86/kernel/setup.c
@@ -478,30 +478,6 @@ static void __init parse_setup_data(void)
 	}
 }
 
-static void __init e820_reserve_setup_data(void)
-{
-	struct setup_data *data;
-	u64 pa_data;
-	int found = 0;
-
-	pa_data = boot_params.hdr.setup_data;
-	while (pa_data) {
-		data = early_memremap(pa_data, sizeof(*data));
-		e820_update_range(pa_data, sizeof(*data)+data->len,
-			 E820_RAM, E820_RESERVED_KERN);
-		found = 1;
-		pa_data = data->next;
-		early_iounmap(data, sizeof(*data));
-	}
-	if (!found)
-		return;
-
-	sanitize_e820_map(e820.map, ARRAY_SIZE(e820.map), &e820.nr_map);
-	memcpy(&e820_saved, &e820, sizeof(struct e820map));
-	printk(KERN_INFO "extended physical RAM map:\n");
-	e820_print_map("reserve setup_data");
-}
-
 static void __init memblock_x86_reserve_range_setup_data(void)
 {
 	struct setup_data *data;
@@ -1037,8 +1013,6 @@ void __init setup_arch(char **cmdline_p)
 		early_dump_pci_devices();
 #endif
 
-	/* update the e820_saved too */
-	e820_reserve_setup_data();
 	finish_e820_parsing();
 
 	if (efi_enabled(EFI_BOOT))
diff --git a/arch/x86/kernel/tboot.c b/arch/x86/kernel/tboot.c
index 91a4496..3c2752a 100644
--- a/arch/x86/kernel/tboot.c
+++ b/arch/x86/kernel/tboot.c
@@ -195,8 +195,7 @@ static int tboot_setup_sleep(void)
 	tboot->num_mac_regions = 0;
 
 	for (i = 0; i < e820.nr_map; i++) {
-		if ((e820.map[i].type != E820_RAM)
-		 && (e820.map[i].type != E820_RESERVED_KERN))
+		if (e820.map[i].type != E820_RAM)
 			continue;
 
 		add_mac_region(e820.map[i].addr, e820.map[i].size);
diff --git a/arch/x86/mm/init_64.c b/arch/x86/mm/init_64.c
index c30efb6..63520ec 100644
--- a/arch/x86/mm/init_64.c
+++ b/arch/x86/mm/init_64.c
@@ -354,8 +354,7 @@ phys_pte_init(pte_t *pte_page, unsigned long addr, unsigned long end,
 		next = (addr & PAGE_MASK) + PAGE_SIZE;
 		if (addr >= end) {
 			if (!after_bootmem &&
-			    !e820_any_mapped(addr & PAGE_MASK, next, E820_RAM) &&
-			    !e820_any_mapped(addr & PAGE_MASK, next, E820_RESERVED_KERN))
+			    !e820_any_mapped(addr & PAGE_MASK, next, E820_RAM))
 				set_pte(pte, __pte(0));
 			continue;
 		}
@@ -401,9 +400,8 @@ phys_pmd_init(pmd_t *pmd_page, unsigned long address, unsigned long end,
 
 		next = (address & PMD_MASK) + PMD_SIZE;
 		if (address >= end) {
-			if (!after_bootmem &&
-			    !e820_any_mapped(address & PMD_MASK, next, E820_RAM) &&
-			    !e820_any_mapped(address & PMD_MASK, next, E820_RESERVED_KERN))
+			if (!after_bootmem && !e820_any_mapped(
+					address & PMD_MASK, next, E820_RAM))
 				set_pmd(pmd, __pmd(0));
 			continue;
 		}
@@ -476,8 +474,7 @@ phys_pud_init(pud_t *pud_page, unsigned long addr, unsigned long end,
 		next = (addr & PUD_MASK) + PUD_SIZE;
 		if (addr >= end) {
 			if (!after_bootmem &&
-			    !e820_any_mapped(addr & PUD_MASK, next, E820_RAM) &&
-			    !e820_any_mapped(addr & PUD_MASK, next, E820_RESERVED_KERN))
+			    !e820_any_mapped(addr & PUD_MASK, next, E820_RAM))
 				set_pud(pud, __pud(0));
 			continue;
 		}
-- 
1.8.4.5

^ permalink raw reply related	[flat|nested] 7+ messages in thread

end of thread, other threads:[~2015-03-09  0:18 UTC | newest]

Thread overview: 7+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2015-03-08  0:50 [PATCH v3 0/8] x86, boot: clean up setup_data handling Yinghai Lu
2015-03-08  0:50 ` [PATCH v3 1/8] x86: Kill E820_RESERVED_KERN Yinghai Lu
2015-03-08  0:50 ` [PATCH v3 2/8] x86, efi: Copy SETUP_EFI data and access directly Yinghai Lu
2015-03-08  0:56 [PATCH v3 0/8] x86, boot: clean up setup_data handling Yinghai Lu
2015-03-08  0:56 ` [PATCH v3 1/8] x86: Kill E820_RESERVED_KERN Yinghai Lu
     [not found]   ` <1425776181-10219-2-git-send-email-yinghai-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>
2015-03-08  1:59     ` David Rientjes
     [not found]       ` <alpine.DEB.2.10.1503071748530.5739-X6Q0R45D7oAcqpCFd4KODRPsWskHk0ljAL8bYrjMMd8@public.gmane.org>
2015-03-08  6:51         ` Yinghai Lu
2015-03-09  0:18         ` joeyli

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).