Skip to contentSkip to conversation

ACPI blacklisting: move year blacklist into acpi/blacklist.c

PATCH Pavel Machek 3 participants 6 messages Oct 1, 2003 – Oct 8, 2003

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

Hi!

This moves year-based blacklisting to blacklist.c, where it belongs AFAICS. It also adds some externs to include/linux/acpi.h, but I believe *way* more externs are needed. Please apply,

Pavel

usr/src/linux/arch/i386/kernel/dmi_scan.c renamed from usr/src/tmp/linux/arch/i386/kernel/dmi_scan.c +2 -16
@@ -997,24 +998,10 @@
 	int i;
 		
 #ifdef	CONFIG_ACPI_BOOT
-#define	ACPI_BLACKLIST_CUTOFF_YEAR	2001
-
 	if (dmi_ident[DMI_BIOS_DATE]) { 
 		char *s = strrchr(dmi_ident[DMI_BIOS_DATE], '/'); 
-		if (s) { 
-			int year, disable = 0;
-			s++; 
-			year = simple_strtoul(s,NULL,0); 
-			if (year >= 1000) 
-				disable = year < ACPI_BLACKLIST_CUTOFF_YEAR; 
-			else if (year < 1 || (year > 90 && year <= 99))
-				disable = 1; 
-			if (disable && !acpi_force) { 
-				printk(KERN_NOTICE "ACPI disabled because your bios is from %s and too old\n", s);
-				printk(KERN_NOTICE "You can enable it with acpi=force\n");
-				acpi_disabled = 1; 
-			} 
-		}
+		if (s && !acpi_force)
+			acpi_bios_year(s+1);
 	}
 #endif
 
usr/src/linux/drivers/acpi/blacklist.c renamed from usr/src/tmp/linux/drivers/acpi/blacklist.c +21 -0
@@ -83,6 +83,27 @@
 	{""}
 };
 
+#define	ACPI_BLACKLIST_CUTOFF_YEAR	2001
+
+/*
+ * Notice: this is called from dmi_scan.c, which contains second (!) blacklist
+ */
+void __init
+acpi_bios_year(char *s)
+{
+	int year, disable = 0;
+
+	year = simple_strtoul(s,NULL,0); 
+	if (year >= 1000) 
+		disable = year < ACPI_BLACKLIST_CUTOFF_YEAR; 
+	else if (year < 1 || (year > 90 && year <= 99))
+		disable = 1; 
+	if (disable) { 
+		printk(KERN_NOTICE "ACPI disabled because your bios is from %s and too old\n", s);
+		printk(KERN_NOTICE "You can enable it with acpi=force\n");
+		acpi_disabled = 1; 
+	}
+}
 
 int __init
 acpi_blacklisted(void)
usr/src/linux/include/linux/acpi.h renamed from usr/src/tmp/linux/include/linux/acpi.h +8 -7
@@ -403,8 +403,8 @@
 
 struct pci_dev;
 
-int acpi_pci_irq_enable (struct pci_dev *dev);
-int acpi_pci_irq_init (void);
+extern int acpi_pci_irq_enable (struct pci_dev *dev);
+extern int acpi_pci_irq_init (void);
 
 struct acpi_pci_driver {
 	struct acpi_pci_driver *next;
@@ -412,21 +412,22 @@
 	void (*remove)(acpi_handle handle);
 };
 
-int acpi_pci_register_driver(struct acpi_pci_driver *driver);
-void acpi_pci_unregister_driver(struct acpi_pci_driver *driver);
+extern int acpi_pci_register_driver(struct acpi_pci_driver *driver);
+extern void acpi_pci_unregister_driver(struct acpi_pci_driver *driver);
 
 #endif /*CONFIG_ACPI_PCI*/
 
 #ifdef CONFIG_ACPI_EC
 
-int ec_read(u8 addr, u8 *val);
-int ec_write(u8 addr, u8 val);
+extern int ec_read(u8 addr, u8 *val);
+extern int ec_write(u8 addr, u8 val);
 
 #endif /*CONFIG_ACPI_EC*/
 
 #ifdef CONFIG_ACPI
 
-int acpi_blacklisted(void);
+extern int acpi_blacklisted(void);
+extern void acpi_bios_year(char *s);
 
 #else
 
Signature · 2 lines
When do you have a heart between your knees?
[Johanka's followup: and *two* hearts?]
#1 lore raw

usr/src/linux/include/linux/acpi.h line 407

Quoted 13 lines · Pavel Machek

AFAICS. It also adds some externs to include/linux/acpi.h, but I believe *way* more externs are needed. Please apply,

usr/src/linux/include/linux/acpi.h renamed from usr/src/tmp/linux/include/linux/acpi.h +2 -2
@@ -403,8 +403,8 @@
 
 struct pci_dev;
 
-int acpi_pci_irq_enable (struct pci_dev *dev);
-int acpi_pci_irq_init (void);
+extern int acpi_pci_irq_enable (struct pci_dev *dev);
+extern int acpi_pci_irq_init (void);

Why do they need to be externs? The comp.lang.c FAQ suggests they don't have to be.

http://www.eskimo.com/~scs/C-faq/q1.11.html

Signature · 3 lines
"It's not Hollywood.  War is real, war is primarily not about defeat or
victory, it is about death.  I've seen thousands and thousands of dead bodies.
Do you think I want to have an academic debate on this subject?" -- Robert Fisk
#2 lore raw

Hi!

Quoted 17 lines

usr/src/linux/include/linux/acpi.h line 407

Quoted 13 lines

AFAICS. It also adds some externs to include/linux/acpi.h, but I believe *way* more externs are needed. Please apply,

usr/src/linux/include/linux/acpi.h renamed from usr/src/tmp/linux/include/linux/acpi.h +2 -2
@@ -403,8 +403,8 @@
 
 struct pci_dev;
 
-int acpi_pci_irq_enable (struct pci_dev *dev);
-int acpi_pci_irq_init (void);
+extern int acpi_pci_irq_enable (struct pci_dev *dev);
+extern int acpi_pci_irq_init (void);

Why do they need to be externs? The comp.lang.c FAQ suggests they don't have to be.

Quoted 1 line

http://www.eskimo.com/~scs/C-faq/q1.11.html

Well, they don't *have* to be there, but as FAQ says, it is stylistics hint.

Pavel

Signature · 8 lines
Horseback riding is like software...
...vgf orggre jura vgf serr.


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
#3 lore raw
Quoted 7 lines · Pavel Machek
Quoted 4 lines

Why do they need to be externs? The comp.lang.c FAQ suggests they don't have to be.

http://www.eskimo.com/~scs/C-faq/q1.11.html

Well, they don't *have* to be there, but as FAQ says, it is stylistics hint.

I think "being in a .h file" is a sufficient hint ;-)

Signature · 9 lines
"It's not Hollywood.  War is real, war is primarily not about defeat or
victory, it is about death.  I've seen thousands and thousands of dead bodies.
Do you think I want to have an academic debate on this subject?" -- Robert Fisk


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
#4 lore raw
In reply to Pavel Machek

Re: dmi_check_blacklist()

Something like MATCH(DMI_BOARD_NAME, "P2B-DS") boils down to a strstr(a,
b) -- probably something that can be fooled by substring, say "PB2-D"...

But more to the point, there is no way to compare before/after dates, say BEFORE(DMI_BIOS_DATE, "1/3/2002")

This is needed to blacklist things like the Toshiba Tecra 8100 which supplies a DMI date, but no other useful version info.

If we such a date comparison function, we could also use it to make the year-compare code below somewhat cleaner.

Re: externs
looks like using externs in the function protos in include/linux is common practice, so acpi.h should probably do it too -- even if it is just for consistent style.

Re: diffs
I couldn't get patch to digest this format w/o manual intervention --- /usr/src/tmp/linux/arch/...
+++ /usr/src/linux/arch/...

and had to edit it into

--- a/arch/...
+++ b/arch/...

for patch -Np1

Re: acpi_bios_year() definition
note that CONFIG_ACPI_BOOT may be defined when CONFIG_ACPI is not. Ie. in the case where just the ACPI boot code is used to enumerate LAPICs, say for HT, but ACPI_INTERPRETER is not even built in.

In this case, blacklist.c is not built into the kernel, but dmi_scan.c always is.

thanks,
-Len

Quoted 108 lines · Pavel Machek

Hi!

This moves year-based blacklisting to blacklist.c, where it belongs AFAICS. It also adds some externs to include/linux/acpi.h, but I believe *way* more externs are needed. Please apply,

Pavel

usr/src/linux/arch/i386/kernel/dmi_scan.c renamed from usr/src/tmp/linux/arch/i386/kernel/dmi_scan.c +2 -16
@@ -997,24 +998,10 @@
 	int i;
 
 #ifdef	CONFIG_ACPI_BOOT
-#define	ACPI_BLACKLIST_CUTOFF_YEAR	2001
-
 	if (dmi_ident[DMI_BIOS_DATE]) { 
 		char *s = strrchr(dmi_ident[DMI_BIOS_DATE], '/'); 
-		if (s) { 
-			int year, disable = 0;
-			s++; 
-			year = simple_strtoul(s,NULL,0); 
-			if (year >= 1000) 
-				disable = year < ACPI_BLACKLIST_CUTOFF_YEAR; 
-			else if (year < 1 || (year > 90 && year <= 99))
-				disable = 1; 
-			if (disable && !acpi_force) { 
-				printk(KERN_NOTICE "ACPI disabled because your bios is from %s and too old\n", s);
-				printk(KERN_NOTICE "You can enable it with acpi=force\n");
-				acpi_disabled = 1; 
-			} 
-		}
+		if (s && !acpi_force)
+			acpi_bios_year(s+1);
 	}
 #endif
 
usr/src/linux/drivers/acpi/blacklist.c renamed from usr/src/tmp/linux/drivers/acpi/blacklist.c +21 -0
@@ -83,6 +83,27 @@
 	{""}
 };
 
+#define	ACPI_BLACKLIST_CUTOFF_YEAR	2001
+
+/*
+ * Notice: this is called from dmi_scan.c, which contains second (!) blacklist
+ */
+void __init
+acpi_bios_year(char *s)
+{
+	int year, disable = 0;
+
+	year = simple_strtoul(s,NULL,0); 
+	if (year >= 1000) 
+		disable = year < ACPI_BLACKLIST_CUTOFF_YEAR; 
+	else if (year < 1 || (year > 90 && year <= 99))
+		disable = 1; 
+	if (disable) { 
+		printk(KERN_NOTICE "ACPI disabled because your bios is from %s and too old\n", s);
+		printk(KERN_NOTICE "You can enable it with acpi=force\n");
+		acpi_disabled = 1; 
+	}
+}
 
 int __init
 acpi_blacklisted(void)
usr/src/linux/include/linux/acpi.h renamed from usr/src/tmp/linux/include/linux/acpi.h +8 -7
@@ -403,8 +403,8 @@
 
 struct pci_dev;
 
-int acpi_pci_irq_enable (struct pci_dev *dev);
-int acpi_pci_irq_init (void);
+extern int acpi_pci_irq_enable (struct pci_dev *dev);
+extern int acpi_pci_irq_init (void);
 
 struct acpi_pci_driver {
 	struct acpi_pci_driver *next;
@@ -412,21 +412,22 @@
 	void (*remove)(acpi_handle handle);
 };
 
-int acpi_pci_register_driver(struct acpi_pci_driver *driver);
-void acpi_pci_unregister_driver(struct acpi_pci_driver *driver);
+extern int acpi_pci_register_driver(struct acpi_pci_driver *driver);
+extern void acpi_pci_unregister_driver(struct acpi_pci_driver *driver);
 
 #endif /*CONFIG_ACPI_PCI*/
 
 #ifdef CONFIG_ACPI_EC
 
-int ec_read(u8 addr, u8 *val);
-int ec_write(u8 addr, u8 val);
+extern int ec_read(u8 addr, u8 *val);
+extern int ec_write(u8 addr, u8 val);
 
 #endif /*CONFIG_ACPI_EC*/
 
 #ifdef CONFIG_ACPI
 
-int acpi_blacklisted(void);
+extern int acpi_blacklisted(void);
+extern void acpi_bios_year(char *s);
 
 #else
 
#5 lore raw

Hi!

Quoted 5 lines

Re: dmi_check_blacklist()

Something like MATCH(DMI_BOARD_NAME, "P2B-DS") boils down to a strstr(a,
b) -- probably something that can be fooled by substring, say "PB2-D"...

Quoted 8 lines

But more to the point, there is no way to compare before/after dates, say BEFORE(DMI_BIOS_DATE, "1/3/2002")

This is needed to blacklist things like the Toshiba Tecra 8100 which supplies a DMI date, but no other useful version info.

If we such a date comparison function, we could also use it to make the year-compare code below somewhat cleaner.

Well, my main problem is with having two blacklists at diferent places.

Quoted 4 lines

Re: externs
looks like using externs in the function protos in include/linux is common practice, so acpi.h should probably do it too -- even if it is just for consistent style.

Do you want me to submit patch or will you add externs?

Quoted 11 lines

Re: diffs
I couldn't get patch to digest this format w/o manual intervention

usr/src/linux/arch/... renamed from usr/src/tmp/linux/arch/... +0 -0

and had to edit it into

arch/... +0 -0

for patch -Np1

Sorry. (I fixed generating script).

Quoted 7 lines

Re: acpi_bios_year() definition
note that CONFIG_ACPI_BOOT may be defined when CONFIG_ACPI is not. Ie. in the case where just the ACPI boot code is used to enumerate LAPICs, say for HT, but ACPI_INTERPRETER is not even built in.

In this case, blacklist.c is not built into the kernel, but dmi_scan.c always is.

Okay, in such case is it feasible to move blacklist.c functionality into dmi_scan.c? i386 blacklist is probably useless for ia64, anyway... And 2 places seem like bad idea.
Pavel

Signature · 2 lines
When do you have a heart between your knees?
[Johanka's followup: and *two* hearts?]