Optional Beeping During Resume From Suspend To Ram.
Subsystems (4)
PATCH 1/1 Optional Beeping During Resume From Suspend To Ram.
Hi all
Here's what I have after today's work.
I haven't yet been able to test on x86, but can confirm that it works okay on x86_64. I'm currently working towards testing it on my old Omnibook. My P4 desktop won't resume from suspend to ram at all, and hasn't produced any beeps.
I needed to move the BEEP invocation to after the data segment is reloaded, so that the test could access the variable. That was pretty tricky to find - no oops or anything bad prior, it just didn't beep when expected.
A couple of notes:
- I'd like to put the BEEP macro somewhere that can be shared by x86 32 and 64. If that's a good idea, any suggestions on where? Nothing occurs to me straight off.
- I've just switched from Evo to Kmail. Please let me know if there's any mangling of the patch.
Regards,
Nigel
arch/i386/kernel/acpi/wakeup.S | 29 ++++++++++++++++++++++++++++- arch/x86_64/kernel/acpi/wakeup.S | 25 +++++++++++++++++++++++++ include/linux/acpi.h | 1 + kernel/power/main.c | 23 +++++++++++++++++++++++ 4 files changed, 77 insertions(+), 1 deletion(-)
arch/i386/kernel/acpi/wakeup.S +28 -1
@@ -11,7 +11,22 @@ # # If physical address of wakeup_code is 0x12345, BIOS should call us with # cs = 0x1234, eip = 0x05 -# +# + +#define BEEP \ + inb $97, %al; \ + outb %al, $0x80; \ + movb $3, %al; \ + outb %al, $97; \ + outb %al, $0x80; \ + movb $-74, %al; \ + outb %al, $67; \ + outb %al, $0x80; \ + movb $-119, %al; \ + outb %al, $66; \ + outb %al, $0x80; \ + movb $15, %al; \ + outb %al, $66; ALIGN .align 4096
@@ -31,6 +46,11 @@ wakeup_code: movw %cs, %ax movw %ax, %ds # Make ds:0 point to wakeup_start movw %ax, %ss + + testl $1, beep_flags - wakeup_code + jz 1f + BEEP +1: mov $(wakeup_stack - wakeup_code), %sp # Private stack is needed for ASUS board movw $0x0e00 + 'S', %fs:(0x12)
@@ -88,6 +108,10 @@ wakeup_code: cmpl $0x12345678, %eax jne bogus_real_magic + testl $2, beep_flags - wakeup_code + jz 1f + BEEP +1: ljmpl $__KERNEL_CS,$wakeup_pmode_return real_save_gdt: .word 0
@@ -98,6 +122,7 @@ real_save_cr4: .long 0 real_magic: .long 0 video_mode: .long 0 video_flags: .long 0 +beep_flags: .long 0 real_efer_save_restore: .long 0 real_save_efer_edx: .long 0 real_save_efer_eax: .long 0
@@ -261,6 +286,8 @@ ENTRY(acpi_copy_wakeup_routine) movl %edx, video_mode - wakeup_start (%eax) movl acpi_video_flags, %edx movl %edx, video_flags - wakeup_start (%eax) + movl s2ram_beep, %edx + movl %edx, beep_flags - wakeup_start (%eax) movl $0x12345678, real_magic - wakeup_start (%eax) movl $0x12345678, saved_magic ret
arch/x86_64/kernel/acpi/wakeup.S +25 -0
@@ -16,6 +16,21 @@ # cs = 0x1234, eip = 0x05 # +#define BEEP \ + inb $97, %al; \ + outb %al, $0x80; \ + movb $3, %al; \ + outb %al, $97; \ + outb %al, $0x80; \ + movb $-74, %al; \ + outb %al, $67; \ + outb %al, $0x80; \ + movb $-119, %al; \ + outb %al, $66; \ + outb %al, $0x80; \ + movb $15, %al; \ + outb %al, $66; + ALIGN .align 16
@@ -33,6 +48,13 @@ wakeup_code: movw %cs, %ax movw %ax, %ds # Make ds:0 point to wakeup_start movw %ax, %ss + + # Data segment must be set up before we can see whether to beep. + testl $1, beep_flags - wakeup_code + jz 1f + BEEP +1: + # Private stack is needed for ASUS board mov $(wakeup_stack - wakeup_code), %sp
@@ -229,6 +251,7 @@ gdt_48a: .long gdta - wakeup_code # gdt base (relocated in later) real_magic: .quad 0 +beep_flags: .quad 0 video_mode: .quad 0 video_flags: .quad 0
@@ -344,6 +367,8 @@ ENTRY(acpi_copy_wakeup_routine) pushq %rax pushq %rdx + movl s2ram_beep, %edx + movl %edx, beep_flags - wakeup_start (,%rdi) movl saved_video_mode, %edx movl %edx, video_mode - wakeup_start (,%rdi) movl acpi_video_flags, %edx
include/linux/acpi.h +1 -0
@@ -123,6 +123,7 @@ extern int pci_mmcfg_config_num;extern int sbf_port;extern unsigned long acpi_video_flags; +extern unsigned long s2ram_beep;#else /* !CONFIG_ACPI */
kernel/power/main.c +23 -0
@@ -308,6 +308,27 @@ static ssize_t state_store(struct kset *power_attr (state); +unsigned long s2ram_beep =0 ; + +static ssize_t s2ram_beep_show (struct kset *kset,char *buf) +{ +return sprintf (buf,"%d\n" , s2ram_beep); +} + +static ssize_t +s2ram_beep_store (struct kset *kset,const char *buf,size_t n) +{ +int val; + +if (sscanf (buf,"%d" , &val) >0 ) { + s2ram_beep = val; +return n; + } +return -EINVAL ; +} + +power_attr (s2ram_beep); +#ifdef CONFIG_PM_TRACE int pm_trace_enabled;
@@ -333,11 +354,13 @@ power_attr(pm_trace);static struct attribute * g[] = { &state_attr.attr, &pm_trace_attr.attr, + &s2ram_beep_attr.attr,NULL , };#else static struct attribute * g[] = { &state_attr.attr, + &s2ram_beep_attr.attr,NULL , };#endif /* CONFIG_PM_TRACE */
Signature · 4 lines
Nigel, Michelle and Alisdair Cunningham 5 Mitchell Street Cobden 3266 Victoria, Australia
Hi Nigel,
Quoted 13 lines · Nigel Cunningham
Hi all
Here's what I have after today's work.
I haven't yet been able to test on x86, but can confirm that it works okay on
x86_64. I'm currently working towards testing it on my old Omnibook. My P4 desktop won't resume from suspend to ram at all, and hasn't produced any beeps.I needed to move the BEEP invocation to after the data segment is reloaded, so that the test could access the variable. That was pretty tricky to find - no oops or anything bad prior, it just didn't beep when expected.
Quoted 5 lines
A couple of notes:
- I'd like to put the BEEP macro somewhere that can be shared by x86 32 and
64. If that's a good idea, any suggestions on where? Nothing occurs to me straight off.
I don't know either. I think Andi is the right person to ask (CC added).
Quoted 2 lines
- I've just switched from Evo to Kmail. Please let me know if there's any mangling of the patch.
The patch looks good to me. I'll try to run it on an i386 tomorrow.
Greetings,
Rafael
Signature · 1 line
"Premature optimization is the root of all evil." - Donald Knuth
Hi,
Quoted 14 lines · Rafael J. Wysocki
Quoted 13 lines · Nigel Cunningham
Hi all
Here's what I have after today's work.
I haven't yet been able to test on x86, but can confirm that it works okay on
x86_64. I'm currently working towards testing it on my old Omnibook. My P4 desktop won't resume from suspend to ram at all, and hasn't produced any beeps.I needed to move the BEEP invocation to after the data segment is reloaded, so that the test could access the variable. That was pretty tricky to find - no oops or anything bad prior, it just didn't beep when expected.
Quoted 10 lines
Quoted 5 lines
A couple of notes:
- I'd like to put the BEEP macro somewhere that can be shared by x86 32 and
64. If that's a good idea, any suggestions on where? Nothing occurs to me straight off.I don't know either. I think Andi is the right person to ask (CC added).
Quoted 2 lines
- I've just switched from Evo to Kmail. Please let me know if there's any mangling of the patch.
Unfortunately, the message is encoded as "quoted printable" which results in the whitespace being mangled.
Quoted 1 line
The patch looks good to me. I'll try to run it on an i386 tomorrow.
I've recoded it and rediffed against -rc5 with the hibernation and suspend patchset. The resut is at
Well, my i386 test box doesn't seem to be able to beep at all. I'll have to look for another one ...
Greetings,
Rafael
Signature · 1 line
"Premature optimization is the root of all evil." - Donald Knuth
Hi.
Quoted 33 lines · On Thursday 21 June 2007 08:09:26 Rafael J. Wysocki wrote:
Hi,
Quoted 14 lines · Rafael J. Wysocki
Quoted 13 lines · Nigel Cunningham
Hi all
Here's what I have after today's work.
I haven't yet been able to test on x86, but can confirm that it works okay on
x86_64. I'm currently working towards testing it on my old Omnibook. My P4 desktop won't resume from suspend to ram at all, and hasn't produced any beeps.I needed to move the BEEP invocation to after the data segment is reloaded, so that the test could access the variable. That was pretty tricky to find - no oops or anything bad prior, it just didn't beep when expected.
Quoted 10 lines
Quoted 5 lines
A couple of notes:
- I'd like to put the BEEP macro somewhere that can be shared by x86 32 and
64. If that's a good idea, any suggestions on where? Nothing occurs to me straight off.I don't know either. I think Andi is the right person to ask (CC added).
Quoted 2 lines
- I've just switched from Evo to Kmail. Please let me know if there's any mangling of the patch.
Unfortunately, the message is encoded as "quoted printable" which results in the whitespace being mangled.
Quoted 1 line
The patch looks good to me. I'll try to run it on an i386 tomorrow.
Quoted 7 lines
I've recoded it and rediffed against -rc5 with the hibernation and suspend patchset. The resut is at
Well, my i386 test box doesn't seem to be able to beep at all. I'll have to look for another one ...
Ok. No success with the Omnibook yet, though I did see it mentioned in Documentation/... as needing a vbepost.
Think I found the right option. Does this patch come through better?
Nigel
arch/i386/kernel/acpi/wakeup.S | 29 ++++++++++++++++++++++++++++- arch/x86_64/kernel/acpi/wakeup.S | 25 +++++++++++++++++++++++++ include/linux/acpi.h | 1 + kernel/power/main.c | 23 +++++++++++++++++++++++ 4 files changed, 77 insertions(+), 1 deletion(-)
arch/i386/kernel/acpi/wakeup.S +28 -1
@@ -11,7 +11,22 @@ # # If physical address of wakeup_code is 0x12345, BIOS should call us with # cs = 0x1234, eip = 0x05 -# +# + +#define BEEP \ + inb $97, %al; \ + outb %al, $0x80; \ + movb $3, %al; \ + outb %al, $97; \ + outb %al, $0x80; \ + movb $-74, %al; \ + outb %al, $67; \ + outb %al, $0x80; \ + movb $-119, %al; \ + outb %al, $66; \ + outb %al, $0x80; \ + movb $15, %al; \ + outb %al, $66; ALIGN .align 4096
@@ -31,6 +46,11 @@ wakeup_code: movw %cs, %ax movw %ax, %ds # Make ds:0 point to wakeup_start movw %ax, %ss + + testl $1, beep_flags - wakeup_code + jz 1f + BEEP +1: mov $(wakeup_stack - wakeup_code), %sp # Private stack is needed for ASUS board movw $0x0e00 + 'S', %fs:(0x12)
@@ -88,6 +108,10 @@ wakeup_code: cmpl $0x12345678, %eax jne bogus_real_magic + testl $2, beep_flags - wakeup_code + jz 1f + BEEP +1: ljmpl $__KERNEL_CS,$wakeup_pmode_return real_save_gdt: .word 0
@@ -98,6 +122,7 @@ real_save_cr4: .long 0 real_magic: .long 0 video_mode: .long 0 video_flags: .long 0 +beep_flags: .long 0 real_efer_save_restore: .long 0 real_save_efer_edx: .long 0 real_save_efer_eax: .long 0
@@ -261,6 +286,8 @@ ENTRY(acpi_copy_wakeup_routine) movl %edx, video_mode - wakeup_start (%eax) movl acpi_video_flags, %edx movl %edx, video_flags - wakeup_start (%eax) + movl s2ram_beep, %edx + movl %edx, beep_flags - wakeup_start (%eax) movl $0x12345678, real_magic - wakeup_start (%eax) movl $0x12345678, saved_magic ret
arch/x86_64/kernel/acpi/wakeup.S +25 -0
@@ -16,6 +16,21 @@ # cs = 0x1234, eip = 0x05 # +#define BEEP \ + inb $97, %al; \ + outb %al, $0x80; \ + movb $3, %al; \ + outb %al, $97; \ + outb %al, $0x80; \ + movb $-74, %al; \ + outb %al, $67; \ + outb %al, $0x80; \ + movb $-119, %al; \ + outb %al, $66; \ + outb %al, $0x80; \ + movb $15, %al; \ + outb %al, $66; + ALIGN .align 16
@@ -33,6 +48,13 @@ wakeup_code: movw %cs, %ax movw %ax, %ds # Make ds:0 point to wakeup_start movw %ax, %ss + + # Data segment must be set up before we can see whether to beep. + testl $1, beep_flags - wakeup_code + jz 1f + BEEP +1: + # Private stack is needed for ASUS board mov $(wakeup_stack - wakeup_code), %sp
@@ -229,6 +251,7 @@ gdt_48a: .long gdta - wakeup_code # gdt base (relocated in later) real_magic: .quad 0 +beep_flags: .quad 0 video_mode: .quad 0 video_flags: .quad 0
@@ -344,6 +367,8 @@ ENTRY(acpi_copy_wakeup_routine) pushq %rax pushq %rdx + movl s2ram_beep, %edx + movl %edx, beep_flags - wakeup_start (,%rdi) movl saved_video_mode, %edx movl %edx, video_mode - wakeup_start (,%rdi) movl acpi_video_flags, %edx
include/linux/acpi.h +1 -0
@@ -123,6 +123,7 @@ extern int pci_mmcfg_config_num;extern int sbf_port;extern unsigned long acpi_video_flags; +extern unsigned long s2ram_beep;#else /* !CONFIG_ACPI */
kernel/power/main.c +23 -0
@@ -308,6 +308,27 @@ static ssize_t state_store(struct kset *power_attr (state); +unsigned long s2ram_beep =0 ; + +static ssize_t s2ram_beep_show (struct kset *kset,char *buf) +{ +return sprintf (buf,"%d\n" , s2ram_beep); +} + +static ssize_t +s2ram_beep_store (struct kset *kset,const char *buf,size_t n) +{ +int val; + +if (sscanf (buf,"%d" , &val) >0 ) { + s2ram_beep = val; +return n; + } +return -EINVAL ; +} + +power_attr (s2ram_beep); +#ifdef CONFIG_PM_TRACE int pm_trace_enabled;
@@ -333,11 +354,13 @@ power_attr(pm_trace);static struct attribute * g[] = { &state_attr.attr, &pm_trace_attr.attr, + &s2ram_beep_attr.attr,NULL , };#else static struct attribute * g[] = { &state_attr.attr, + &s2ram_beep_attr.attr,NULL , };#endif /* CONFIG_PM_TRACE */
Signature · 4 lines
Nigel, Michelle and Alisdair Cunningham 5 Mitchell Street Cobden 3266 Victoria, Australia
Hi,
Quoted 35 lines · Nigel Cunningham
Quoted 34 lines · On Thursday 21 June 2007 08:09:26 Rafael J. Wysocki wrote:
Quoted 14 lines · Rafael J. Wysocki
Quoted 13 lines · Nigel Cunningham
Hi all
Here's what I have after today's work.
I haven't yet been able to test on x86, but can confirm that it works okay on
x86_64. I'm currently working towards testing it on my old Omnibook. My P4 desktop won't resume from suspend to ram at all, and hasn't produced any beeps.I needed to move the BEEP invocation to after the data segment is reloaded, so that the test could access the variable. That was pretty tricky to find - no oops or anything bad prior, it just didn't beep when expected.
Quoted 10 lines
Quoted 5 lines
A couple of notes:
- I'd like to put the BEEP macro somewhere that can be shared by x86 32 and
64. If that's a good idea, any suggestions on where? Nothing occurs to me straight off.I don't know either. I think Andi is the right person to ask (CC added).
Quoted 2 lines
- I've just switched from Evo to Kmail. Please let me know if there's any mangling of the patch.
Unfortunately, the message is encoded as "quoted printable" which results in the whitespace being mangled.
Quoted 1 line
The patch looks good to me. I'll try to run it on an i386 tomorrow.
I've recoded it and rediffed against -rc5 with the hibernation and suspend patchset. The resut is at
Quoted 8 lines
Quoted 4 lines
Well, my i386 test box doesn't seem to be able to beep at all. I'll have to look for another one ...
Ok. No success with the Omnibook yet, though I did see it mentioned in
Documentation/... as needing a vbepost.Think I found the right option. Does this patch come through better?
No, it's still "quoted printable", at least on my system (but I use Kmail too).
Greetings,
Rafael
Signature · 1 line
"Premature optimization is the root of all evil." - Donald Knuth
Hi!
Quoted 50 lines
Hi all
Here's what I have after today's work.
I haven't yet been able to test on x86, but can confirm that it works okay on
x86_64. I'm currently working towards testing it on my old Omnibook. My P4 desktop won't resume from suspend to ram at all, and hasn't produced any beeps.I needed to move the BEEP invocation to after the data segment is reloaded, so that the test could access the variable. That was pretty tricky to find - no oops or anything bad prior, it just didn't beep when expected.
A couple of notes:
- I'd like to put the BEEP macro somewhere that can be shared by x86 32 and 64. If that's a good idea, any suggestions on where? Nothing occurs to me straight off.
- I've just switched from Evo to Kmail. Please let me know if there's any mangling of the patch.Regards,
Nigel
arch/i386/kernel/acpi/wakeup.S | 29 ++++++++++++++++++++++++++++- arch/x86_64/kernel/acpi/wakeup.S | 25 +++++++++++++++++++++++++ include/linux/acpi.h | 1 + kernel/power/main.c | 23 +++++++++++++++++++++++ 4 files changed, 77 insertions(+), 1 deletion(-)diff -ruNp
970-str-beep.patch-old/arch/i386/kernel/acpi/wakeup.S970-str-beep.patch-new/arch/i386/kernel/acpi/wakeup.Sarch/i386/kernel/acpi/wakeup.S +16
-1@@ -11,7 +11,22 @@ # # If physical address of wakeup_code is 0x12345, BIOS should call us with # cs = 0x1234, eip = 0x05 -# +# + +#define BEEP \ + inb $97, %al; \ + outb %al, $0x80; \ + movb $3, %al; \ + outb %al, $97; \ + outb %al, $0x80; \ + movb $-74, %al; \ + outb %al, $67; \ + outb %al, $0x80; \ + movb $-119, %al; \ + outb %al, $66; \ + outb %al, $0x80; \ + movb $15, %al; \ + outb %al, $66; ALIGN .align 4096
arch/i386/kernel/acpi/wakeup.S wakeup_code: · line 53
Quoted 8 lines
arch/i386/kernel/acpi/wakeup.S +5
-0@@ -31,6 +46,11 @@ wakeup_code: movw %cs, %ax movw %ax, %ds # Make ds:0 point to wakeup_start movw %ax, %ss + + testl $1, beep_flags - wakeup_code + jz 1f + BEEP +1:
Can we rename/reuse existing flag variable?
Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
Hi.
Quoted 7 lines · On Friday 29 June 2007 00:25:32 Pavel Machek wrote:
Hi!
Quoted 5 lines
Hi all
Here's what I have after today's work.
I haven't yet been able to test on x86, but can confirm that it works okay
on x86_64. I'm currently working towards testing it on my old Omnibook. My P4 desktop won't resume from suspend to ram at all, and hasn't produced any beeps.
Quoted 2 lines
Quoted 2 lines
I needed to move the BEEP invocation to after the data segment is
reloaded, so that the test could access the variable. That was pretty tricky to find - no oops or anything bad prior, it just didn't beep when expected.
Quoted 4 lines
Quoted 4 lines
A couple of notes:
- I'd like to put the BEEP macro somewhere that can be shared by x86 32
and 64. If that's a good idea, any suggestions on where? Nothing occurs to me straight off.
Quoted 1 line
Quoted 1 line
- I've just switched from Evo to Kmail. Please let me know if there's any
mangling of the patch.
Quoted 11 lines
Quoted 11 lines
Regards,
Nigel
arch/i386/kernel/acpi/wakeup.S | 29 ++++++++++++++++++++++++++++- arch/x86_64/kernel/acpi/wakeup.S | 25 +++++++++++++++++++++++++ include/linux/acpi.h | 1 + kernel/power/main.c | 23 +++++++++++++++++++++++ 4 files changed, 77 insertions(+), 1 deletion(-)diff -ruNp
970-str-beep.patch-old/arch/i386/kernel/acpi/wakeup.S
970-str-beep.patch-new/arch/i386/kernel/acpi/wakeup.S
Quoted 1 line
Quoted 1 line
---
970-str-beep.patch-old/arch/i386/kernel/acpi/wakeup.S2007-06-19
12:15:25.000000000 +1000
Quoted 1 line
Quoted 1 line
+++
970-str-beep.patch-new/arch/i386/kernel/acpi/wakeup.S2007-06-19
21:14:49.000000000 +1000
Quoted 22 lines
Quoted 22 lines
arch/i386/kernel/acpi/wakeup.S +16
-1@@ -11,7 +11,22 @@ # # If physical address of wakeup_code is 0x12345, BIOS should call us with # cs = 0x1234, eip = 0x05 -# +# + +#define BEEP \ + inb $97, %al; \ + outb %al, $0x80; \ + movb $3, %al; \ + outb %al, $97; \ + outb %al, $0x80; \ + movb $-74, %al; \ + outb %al, $67; \ + outb %al, $0x80; \ + movb $-119, %al; \ + outb %al, $66; \ + outb %al, $0x80; \ + movb $15, %al; \ + outb %al, $66;
Quoted 13 lines
arch/i386/kernel/acpi/wakeup.S wakeup_code: · line 53
Quoted 11 lines
arch/i386/kernel/acpi/wakeup.S +5
-0ALIGN .align 4096@@ -31,6 +46,11 @@ wakeup_code: movw %cs, %ax movw %ax, %ds # Make ds:0 point to wakeup_start movw %ax, %ss + + testl $1, beep_flags - wakeup_code + jz 1f + BEEP +1:Can we rename/reuse existing flag variable?
Sorry, but I can't resist the opportunity to say "Send a patch!" :)
Seriously, though, I'd prefer not to. If we rename that acpi video flags variable (I assume this is what you're thinking of), we only create cause for confusion. A variable should for debugging or for controlling quirks, not for both at the same time.
Regards,
Nigel
Signature · 4 lines
Nigel, Michelle and Alisdair Cunningham 5 Mitchell Street Cobden 3266 Victoria, Australia
Quoted 8 lines · Nigel Cunningham
Quoted 1 line
Can we rename/reuse existing flag variable?
Sorry, but I can't resist the opportunity to say "Send a patch!" :)
Seriously, though, I'd prefer not to. If we rename that acpi video flags variable (I assume this is what you're thinking of), we only create cause for confusion. A variable should for debugging or for controlling quirks, not for both at the same time.
I agree. And video_flags is something totally different :-)
I just used that one in my ad-hoc hack (which actually was only to illustrate the idea) because a) it was enough to show the intent and b) i did not know how to do it better ;-)
Signature · 6 lines
Stefan Seyfried QA / R&D Team Mobile Devices | "Any ideas, John?" SUSE LINUX Products GmbH, Nürnberg | "Well, surrounding them's out." This footer brought to you by insane German lawmakers: SUSE Linux Products GmbH, GF: Markus Rex, HRB 16746 (AG Nürnberg)
Hi!
Quoted 14 lines
Quoted 5 lines
Quoted 5 lines
ALIGN
arch/i386/kernel/acpi/wakeup.S +0
-0.align 4096@@ -31,6 +46,11 @@ wakeup_code: movw %cs, %ax movw %ax, %ds # Make ds:0 point to wakeup_startQuoted 8 lines
arch/i386/kernel/acpi/wakeup.S wakeup_code: · line 53
Quoted 6 lines
arch/i386/kernel/acpi/wakeup.S +5
-0movw %ax, %ss + + testl $1, beep_flags - wakeup_code + jz 1f + BEEP +1:Can we rename/reuse existing flag variable?
Quoted 6 lines
Sorry, but I can't resist the opportunity to say "Send a patch!" :)
Seriously, though, I'd prefer not to. If we rename that acpi video flags variable (I assume this is what you're thinking of), we only create cause for confusion. A variable should for debugging or for controlling quirks, not for both at the same time.
Cause for confusion? We are currently using 2 bits of that variable, and we want to add one more bit. I seriously doubt that can confuse anyone.
Pavel
Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
Hi,
Quoted 18 lines · Pavel Machek
Hi!
Quoted 16 lines
Quoted 5 lines
Quoted 5 lines
ALIGN
arch/i386/kernel/acpi/wakeup.S +0
-0.align 4096@@ -31,6 +46,11 @@ wakeup_code: movw %cs, %ax movw %ax, %ds # Make ds:0 point to wakeup_startQuoted 8 lines
arch/i386/kernel/acpi/wakeup.S wakeup_code: · line 53
Quoted 6 lines
arch/i386/kernel/acpi/wakeup.S +5
-0movw %ax, %ss + + testl $1, beep_flags - wakeup_code + jz 1f + BEEP +1:Can we rename/reuse existing flag variable?
Sorry, but I can't resist the opportunity to say "Send a patch!" :)
Quoted 8 lines
Quoted 4 lines
Seriously, though, I'd prefer not to. If we rename that acpi video flags variable (I assume this is what you're thinking of), we only create cause for confusion. A variable should for debugging or for controlling quirks, not for both at the same time.
Cause for confusion? We are currently using 2 bits of that variable, and we want to add one more bit. I seriously doubt that can confuse anyone.
Well, indeed it would be more elegant to rename the existing flags variable and use another bit out of it, but I personally don't think it's _that_ important. At least, I don't think we should block the patch because of that.
BTW, has anyone confirmed that it works on i386?
BTW2, Nigel, please fix the formats in s2ram_beep_show()/_store(), they cause gcc to spit ugly warnings on some architectures.
Greetings,
Rafael
Signature · 1 line
"Premature optimization is the root of all evil." - Donald Knuth
Hi!
Quoted 7 lines
Quoted 7 lines
Quoted 6 lines
Sorry, but I can't resist the opportunity to say "Send a patch!" :)
Seriously, though, I'd prefer not to. If we rename that acpi video flags variable (I assume this is what you're thinking of), we only create cause for confusion. A variable should for debugging or for controlling quirks, not for both at the same time.
Quoted 8 lines
Quoted 3 lines
Cause for confusion? We are currently using 2 bits of that variable, and we want to add one more bit. I seriously doubt that can confuse anyone.
Well, indeed it would be more elegant to rename the existing flags variable and use another bit out of it, but I personally don't think it's _that_ important. At least, I don't think we should block the patch because of that.
It is not _that_ important.
Quoted 1 line
BTW, has anyone confirmed that it works on i386?
If you have patch somewhere nearby, I can test it on i386 and make it use just one flags variable.
Pavel
Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
Hi,
Quoted 18 lines · Pavel Machek
Hi!
Quoted 7 lines
Quoted 7 lines
Quoted 6 lines
Sorry, but I can't resist the opportunity to say "Send a patch!" :)
Seriously, though, I'd prefer not to. If we rename that acpi video flags variable (I assume this is what you're thinking of), we only create cause for confusion. A variable should for debugging or for controlling quirks, not for both at the same time.
Quoted 8 lines
Quoted 3 lines
Cause for confusion? We are currently using 2 bits of that variable, and we want to add one more bit. I seriously doubt that can confuse anyone.
Well, indeed it would be more elegant to rename the existing flags variable and use another bit out of it, but I personally don't think it's _that_ important. At least, I don't think we should block the patch because of that.
Quoted 6 lines
It is not _that_ important.
Quoted 1 line
BTW, has anyone confirmed that it works on i386?
If you have patch somewhere nearby, I can test it on i386 and make it use just one flags variable.
Greetings,
Rafael
Signature · 1 line
"Premature optimization is the root of all evil." - Donald Knuth
Hi!
Quoted 16 lines
Quoted 16 lines
Quoted 15 lines
Quoted 10 lines
Quoted 6 lines
Sorry, but I can't resist the opportunity to say "Send a patch!" :)
Seriously, though, I'd prefer not to. If we rename that acpi video flags variable (I assume this is what you're thinking of), we only create cause for confusion. A variable should for debugging or for controlling quirks, not for both at the same time.
Cause for confusion? We are currently using 2 bits of that variable, and we want to add one more bit. I seriously doubt that can confuse anyone.
Well, indeed it would be more elegant to rename the existing flags variable and use another bit out of it, but I personally don't think it's _that_ important. At least, I don't think we should block the patch because of that.
Quoted 9 lines
Quoted 6 lines
It is not _that_ important.
Quoted 1 line
BTW, has anyone confirmed that it works on i386?
If you have patch somewhere nearby, I can test it on i386 and make it use just one flags variable.
I can confirm that (Nigel's patch) works on i386; but it causes a warning:
kernel/power/main.c: In function 's2ram_beep_show':kernel/power/main.c:319: warning: format '%d' expects type 'int', but argument 3 has type 'long unsigned int'
Here's the version that uses just one variable, relative to Nigel's patch. Hmm, and it also closes nasty trap for the user in acpi_sleep_setup; order of parameters actually mattered there, acpi_sleep=s3_bios,s3_mode doing something different from acpi_sleep=s3_mode,s3_bios . It actually works here.
Will you fold it into patch28, or should I create a changelog? Testing welcome :-).
Pavel
arch/i386/kernel/acpi/sleep.c +8 -4
@@ -14,7 +14,7 @@ #include <asm/smp.h>/* address in low memory of the wakeup routine. */ unsigned long acpi_wakeup_address =0 ; -unsigned long acpi_video_flags; +unsigned long acpi_realmode_flags;extern char wakeup_start, wakeup_end;extern unsigned long FASTCALL (acpi_copy_wakeup_routine (unsigned long ));
@@ -68,9 +68,11 @@ static int __init acpi_sleep_setup(char {while ((str !=NULL ) && (*str !='\0' )) {if (strncmp (str,"s3_bios" ,7 ) ==0 ) - acpi_video_flags =1 ; + acpi_realmode_flags |=1 ;if (strncmp (str,"s3_mode" ,7 ) ==0 ) - acpi_video_flags |=2 ; + acpi_realmode_flags |=2 ; +if (strncmp (str,"s3_beep" ,7 ) ==0 ) + acpi_realmode_flags |=4 ; str =strchr (str,',' );if (str !=NULL ) str +=strspn (str,", \t" );
@@ -80,9 +82,11 @@ static int __init acpi_sleep_setup(char__setup ("acpi_sleep=" , acpi_sleep_setup); +/* Ouch, we want to delete this. We already have better version in userspace, in +s2ram from suspend.sf.net project */ static __init int reset_videomode_after_s3 (struct dmi_system_id *d) { - acpi_video_flags |=2 ; + acpi_realmode_flags |=2 ;return 0 ; }
arch/i386/kernel/acpi/wakeup.S +8 -10
@@ -50,7 +50,7 @@ # BEEP movw %ax, %ds # Make ds:0 point to wakeup_start movw %ax, %ss - testl $1, beep_flags - wakeup_code + testl $4, realmode_flags - wakeup_code jz 1f BEEP 1:
@@ -64,7 +64,7 @@ # BEEP cmpl $0x12345678, %eax jne bogus_real_magic - testl $1, video_flags - wakeup_code + testl $1, realmode_flags - wakeup_code jz 1f lcall $0xc000,$3 movw %cs, %ax
@@ -72,7 +72,7 @@ # BEEP movw %ax, %ss 1: - testl $2, video_flags - wakeup_code + testl $2, realmode_flags - wakeup_code jz 1f mov video_mode - wakeup_code, %ax call mode_set
@@ -111,11 +111,11 @@ # BEEP cmpl $0x12345678, %eax jne bogus_real_magic - testl $2, beep_flags - wakeup_code + testl $8, realmode_flags - wakeup_code jz 1f BEEP 1: - ljmpl $__KERNEL_CS,$wakeup_pmode_return + ljmpl $__KERNEL_CS, $wakeup_pmode_return real_save_gdt: .word 0 .long 0
@@ -124,7 +124,7 @@ real_save_cr3: .long 0 real_save_cr4: .long 0 real_magic: .long 0 video_mode: .long 0 -video_flags: .long 0 +realmode_flags: .long 0 beep_flags: .long 0 real_efer_save_restore: .long 0 real_save_efer_edx: .long 0
@@ -288,10 +288,8 @@ ENTRY(acpi_copy_wakeup_routine) movl saved_videomode, %edx movl %edx, video_mode - wakeup_start (%eax) - movl acpi_video_flags, %edx - movl %edx, video_flags - wakeup_start (%eax) - movl s2ram_beep, %edx - movl %edx, beep_flags - wakeup_start (%eax) + movl acpi_realmode_flags, %edx + movl %edx, realmode_flags - wakeup_start (%eax) movl $0x12345678, real_magic - wakeup_start (%eax) movl $0x12345678, saved_magic popl %ebx
arch/x86_64/kernel/acpi/sleep.c +5 -3
@@ -55,7 +55,7 @@ #ifdef CONFIG_ACPI_SLEEP/* address in low memory of the wakeup routine. */ unsigned long acpi_wakeup_address =0 ; -unsigned long acpi_video_flags; +unsigned long acpi_realmode_flags;extern char wakeup_start, wakeup_end;extern unsigned long acpi_copy_wakeup_routine (unsigned long );
@@ -103,9 +103,11 @@ static int __init acpi_sleep_setup(char {while ((str !=NULL ) && (*str !='\0' )) {if (strncmp (str,"s3_bios" ,7 ) ==0 ) - acpi_video_flags =1 ; + acpi_realmode_flags |=1 ;if (strncmp (str,"s3_mode" ,7 ) ==0 ) - acpi_video_flags |=2 ; + acpi_realmode_flags |=2 ; +if (strncmp (str,"s3_beep" ,7 ) ==0 ) + acpi_realmode_flags |=4 ; str =strchr (str,',' );if (str !=NULL ) str +=strspn (str,", \t" );
arch/x86_64/kernel/acpi/wakeup.S +6 -9
@@ -50,7 +50,7 @@ # Running in *copy* of this code, somewh movw %ax, %ss # Data segment must be set up before we can see whether to beep. - testl $1, beep_flags - wakeup_code + testl $4, realmode_flags - wakeup_code jz 1f BEEP 1:
@@ -70,7 +70,7 @@ # Running in *copy* of this code, somewh testl %eax, %eax jnz no_longmode - testl $1, video_flags - wakeup_code + testl $1, realmode_flags - wakeup_code jz 1f lcall $0xc000,$3 movw %cs, %ax
@@ -78,7 +78,7 @@ # Running in *copy* of this code, somewh movw %ax, %ss 1: - testl $2, video_flags - wakeup_code + testl $2, realmode_flags - wakeup_code jz 1f mov video_mode - wakeup_code, %ax call mode_seta
@@ -251,9 +251,8 @@ gdt_48a: .long gdta - wakeup_code # gdt base (relocated in later) real_magic: .quad 0 -beep_flags: .quad 0 video_mode: .quad 0 -video_flags: .quad 0 +realmode_flags: .quad 0 .code16 bogus_real_magic:
@@ -367,12 +366,10 @@ ENTRY(acpi_copy_wakeup_routine) pushq %rax pushq %rdx - movl s2ram_beep, %edx - movl %edx, beep_flags - wakeup_start (,%rdi) movl saved_video_mode, %edx movl %edx, video_mode - wakeup_start (,%rdi) - movl acpi_video_flags, %edx - movl %edx, video_flags - wakeup_start (,%rdi) + movl acpi_realmode_flags, %edx + movl %edx, realmode_flags - wakeup_start (,%rdi) movq $0x12345678, real_magic - wakeup_start (,%rdi) movq $0x123456789abcdef0, %rdx movq %rdx, saved_magic
include/linux/acpi.h +1 -2
@@ -122,8 +122,7 @@ extern struct acpi_mcfg_allocation *pci_extern int pci_mmcfg_config_num;extern int sbf_port; -extern unsigned long acpi_video_flags; -extern unsigned long s2ram_beep; +extern unsigned long acpi_realmode_flags;#else /* !CONFIG_ACPI */
kernel/power/main.c +0 -23
@@ -312,27 +312,6 @@ static ssize_t state_store(struct kset *power_attr (state); -unsigned long s2ram_beep =0 ; - -static ssize_t s2ram_beep_show (struct kset *kset,char *buf) -{ -return sprintf (buf,"%d\n" , s2ram_beep); -} - -static ssize_t -s2ram_beep_store (struct kset *kset,const char *buf,size_t n) -{ -int val; - -if (sscanf (buf,"%d" , &val) >0 ) { - s2ram_beep = val; -return n; - } -return -EINVAL ; -} - -power_attr (s2ram_beep); -#ifdef CONFIG_PM_TRACE int pm_trace_enabled;
@@ -358,13 +337,11 @@ power_attr(pm_trace);static struct attribute * g[] = { &state_attr.attr, &pm_trace_attr.attr, - &s2ram_beep_attr.attr,NULL , };#else static struct attribute * g[] = { &state_attr.attr, - &s2ram_beep_attr.attr,NULL , };#endif /* CONFIG_PM_TRACE */
kernel/sysctl.c +1 -1
@@ -569,7 +569,7 @@ #ifdef CONFIG_ACPI_SLEEP { .ctl_name = KERN_ACPI_VIDEO_FLAGS, .procname = "acpi_video_flags", - .data = &acpi_video_flags, + .data = &acpi_realmode_flags, .maxlen = sizeof (unsigned long), .mode = 0644, .proc_handler = &proc_doulongvec_minmax,
(s2ram patch for testing follows)
Index: s2ram.c
===================================================================
RCS file: /cvsroot/suspend/suspend/s2ram.c,v
retrieving revision 1.46
s2ram.c +11 -10
@@ -35,13 +35,14 @@/* flags for the whitelist */ #define S3_BIOS 0x01/* machine needs acpi_sleep=s3_bios */ #define S3_MODE 0x02/* machine needs acpi_sleep=s3_mode */ -#define VBE_SAVE 0x04/* machine needs "vbetool save / restore" */ -#define VBE_POST 0x08/* machine needs "vbetool post" */ -#define RADEON_OFF 0x10/* machine needs "radeontool light off" */ -#define UNSURE 0x20/* unverified entries from acpi-support 0.59 */ -#define NOFB 0x40/* must not use a frame buffer */ -#define VBE_MODE 0x80/* machine needs "vbetool vbemode get / set" */ -#define PCI_SAVE 0x100/* we need to save the VGA PCI registers */ +#define S3_BEEP 0x04/* debugging beep */ +#define VBE_SAVE 0x08/* machine needs "vbetool save / restore" */ +#define VBE_POST 0x10/* machine needs "vbetool post" */ +#define RADEON_OFF 0x20/* machine needs "radeontool light off" */ +#define UNSURE 0x40/* unverified entries from acpi-support 0.59 */ +#define NOFB 0x80/* must not use a frame buffer */ +#define VBE_MODE 0x100/* machine needs "vbetool vbemode get / set" */ +#define PCI_SAVE 0x200/* we need to save the VGA PCI registers */ #include "whitelist.c"
@@ -188,8 +189,8 @@int ret =0 ;/* 0 means: don't touch what was set on kernel commandline */ -if (flags & (S3_BIOS |S3_MODE )) - ret =set_acpi_video_mode (flags & (S3_BIOS |S3_MODE )); +if (flags & (S3_BIOS |S3_MODE |S3_BEEP )) + ret =set_acpi_video_mode (flags & (S3_BIOS |S3_MODE |S3_BEEP ));if (ret)return ret;
@@ -368,7 +369,7 @@ flags |=RADEON_OFF ;break ;case 'a' : - flags |= (atoi (optarg) & (S3_BIOS |S3_MODE )); + flags |= (atoi (optarg) & (S3_BIOS |S3_MODE |S3_BEEP ));break ;case 'v' : flags |=PCI_SAVE ;
Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
Hi,
Quoted 25 lines · Pavel Machek
Hi!
Quoted 23 lines
Quoted 22 lines
Quoted 7 lines
Quoted 7 lines
Quoted 6 lines
Sorry, but I can't resist the opportunity to say "Send a patch!" :)
Seriously, though, I'd prefer not to. If we rename that acpi video flags variable (I assume this is what you're thinking of), we only create cause for confusion. A variable should for debugging or for controlling quirks, not for both at the same time.
Quoted 8 lines
Quoted 3 lines
Cause for confusion? We are currently using 2 bits of that variable, and we want to add one more bit. I seriously doubt that can confuse anyone.
Well, indeed it would be more elegant to rename the existing flags variable and use another bit out of it, but I personally don't think it's _that_ important. At least, I don't think we should block the patch because of that.
It is not _that_ important.
Quoted 1 line
BTW, has anyone confirmed that it works on i386?
If you have patch somewhere nearby, I can test it on i386 and make it use just one flags variable.
Quoted 7 lines
Quoted 2 lines
I can confirm that (Nigel's patch) works on i386; but it causes a warning:
kernel/power/main.c: In function 's2ram_beep_show':kernel/power/main.c:319: warning: format '%d' expects type 'int', but argument 3 has type 'long unsigned int'
Yes, I've fixed the warning in the 2.6.22-rc7 series, but that's not a big deal.
Quoted 7 lines
Here's the version that uses just one variable, relative to Nigel's patch. Hmm, and it also closes nasty trap for the user in
acpi_sleep_setup; order of parameters actually mattered there, acpi_sleep=s3_bios,s3_mode doing something different from acpi_sleep=s3_mode,s3_bios . It actually works here.Will you fold it into patch28, or should I create a changelog?
Hmm, I think we should keep the record straight. Please add a changelog if that's not a problem.
Quoted 321 lines
Testing welcome :-).
Pavelarch/i386/kernel/acpi/sleep.c +8
-4@@ -14,7 +14,7 @@ #include <asm/smp.h>/* address in low memory of the wakeup routine. */ unsigned long acpi_wakeup_address =0 ; -unsigned long acpi_video_flags; +unsigned long acpi_realmode_flags;extern char wakeup_start, wakeup_end;extern unsigned long FASTCALL (acpi_copy_wakeup_routine (unsigned long ));@@ -68,9 +68,11 @@ static int __init acpi_sleep_setup(char {while ((str !=NULL ) && (*str !='\0' )) {if (strncmp (str,"s3_bios" ,7 ) ==0 ) - acpi_video_flags =1 ; + acpi_realmode_flags |=1 ;if (strncmp (str,"s3_mode" ,7 ) ==0 ) - acpi_video_flags |=2 ; + acpi_realmode_flags |=2 ; +if (strncmp (str,"s3_beep" ,7 ) ==0 ) + acpi_realmode_flags |=4 ; str =strchr (str,',' );if (str !=NULL ) str +=strspn (str,", \t" );@@ -80,9 +82,11 @@ static int __init acpi_sleep_setup(char__setup ("acpi_sleep=" , acpi_sleep_setup); +/* Ouch, we want to delete this. We already have better version in userspace, in +s2ram from suspend.sf.net project */ static __init int reset_videomode_after_s3 (struct dmi_system_id *d) { - acpi_video_flags |=2 ; + acpi_realmode_flags |=2 ;return 0 ; }arch/i386/kernel/acpi/wakeup.S +8
-10@@ -50,7 +50,7 @@ # BEEP movw %ax, %ds # Make ds:0 point to wakeup_start movw %ax, %ss - testl $1, beep_flags - wakeup_code + testl $4, realmode_flags - wakeup_code jz 1f BEEP 1:@@ -64,7 +64,7 @@ # BEEP cmpl $0x12345678, %eax jne bogus_real_magic - testl $1, video_flags - wakeup_code + testl $1, realmode_flags - wakeup_code jz 1f lcall $0xc000,$3 movw %cs, %ax@@ -72,7 +72,7 @@ # BEEP movw %ax, %ss 1: - testl $2, video_flags - wakeup_code + testl $2, realmode_flags - wakeup_code jz 1f mov video_mode - wakeup_code, %ax call mode_set@@ -111,11 +111,11 @@ # BEEP cmpl $0x12345678, %eax jne bogus_real_magic - testl $2, beep_flags - wakeup_code + testl $8, realmode_flags - wakeup_code jz 1f BEEP 1: - ljmpl $__KERNEL_CS,$wakeup_pmode_return + ljmpl $__KERNEL_CS, $wakeup_pmode_return real_save_gdt: .word 0 .long 0@@ -124,7 +124,7 @@ real_save_cr3: .long 0 real_save_cr4: .long 0 real_magic: .long 0 video_mode: .long 0 -video_flags: .long 0 +realmode_flags: .long 0 beep_flags: .long 0 real_efer_save_restore: .long 0 real_save_efer_edx: .long 0@@ -288,10 +288,8 @@ ENTRY(acpi_copy_wakeup_routine) movl saved_videomode, %edx movl %edx, video_mode - wakeup_start (%eax) - movl acpi_video_flags, %edx - movl %edx, video_flags - wakeup_start (%eax) - movl s2ram_beep, %edx - movl %edx, beep_flags - wakeup_start (%eax) + movl acpi_realmode_flags, %edx + movl %edx, realmode_flags - wakeup_start (%eax) movl $0x12345678, real_magic - wakeup_start (%eax) movl $0x12345678, saved_magic popl %ebxarch/x86_64/kernel/acpi/sleep.c +5
-3@@ -55,7 +55,7 @@ #ifdef CONFIG_ACPI_SLEEP/* address in low memory of the wakeup routine. */ unsigned long acpi_wakeup_address =0 ; -unsigned long acpi_video_flags; +unsigned long acpi_realmode_flags;extern char wakeup_start, wakeup_end;extern unsigned long acpi_copy_wakeup_routine (unsigned long );@@ -103,9 +103,11 @@ static int __init acpi_sleep_setup(char {while ((str !=NULL ) && (*str !='\0' )) {if (strncmp (str,"s3_bios" ,7 ) ==0 ) - acpi_video_flags =1 ; + acpi_realmode_flags |=1 ;if (strncmp (str,"s3_mode" ,7 ) ==0 ) - acpi_video_flags |=2 ; + acpi_realmode_flags |=2 ; +if (strncmp (str,"s3_beep" ,7 ) ==0 ) + acpi_realmode_flags |=4 ; str =strchr (str,',' );if (str !=NULL ) str +=strspn (str,", \t" );arch/x86_64/kernel/acpi/wakeup.S +6
-9@@ -50,7 +50,7 @@ # Running in *copy* of this code, somewh movw %ax, %ss # Data segment must be set up before we can see whether to beep. - testl $1, beep_flags - wakeup_code + testl $4, realmode_flags - wakeup_code jz 1f BEEP 1:@@ -70,7 +70,7 @@ # Running in *copy* of this code, somewh testl %eax, %eax jnz no_longmode - testl $1, video_flags - wakeup_code + testl $1, realmode_flags - wakeup_code jz 1f lcall $0xc000,$3 movw %cs, %ax@@ -78,7 +78,7 @@ # Running in *copy* of this code, somewh movw %ax, %ss 1: - testl $2, video_flags - wakeup_code + testl $2, realmode_flags - wakeup_code jz 1f mov video_mode - wakeup_code, %ax call mode_seta@@ -251,9 +251,8 @@ gdt_48a: .long gdta - wakeup_code # gdt base (relocated in later) real_magic: .quad 0 -beep_flags: .quad 0 video_mode: .quad 0 -video_flags: .quad 0 +realmode_flags: .quad 0 .code16 bogus_real_magic:@@ -367,12 +366,10 @@ ENTRY(acpi_copy_wakeup_routine) pushq %rax pushq %rdx - movl s2ram_beep, %edx - movl %edx, beep_flags - wakeup_start (,%rdi) movl saved_video_mode, %edx movl %edx, video_mode - wakeup_start (,%rdi) - movl acpi_video_flags, %edx - movl %edx, video_flags - wakeup_start (,%rdi) + movl acpi_realmode_flags, %edx + movl %edx, realmode_flags - wakeup_start (,%rdi) movq $0x12345678, real_magic - wakeup_start (,%rdi) movq $0x123456789abcdef0, %rdx movq %rdx, saved_magicinclude/linux/acpi.h +1
-2@@ -122,8 +122,7 @@ extern struct acpi_mcfg_allocation *pci_extern int pci_mmcfg_config_num;extern int sbf_port; -extern unsigned long acpi_video_flags; -extern unsigned long s2ram_beep; +extern unsigned long acpi_realmode_flags;#else /* !CONFIG_ACPI */ kernel/power/main.c +0
-23@@ -312,27 +312,6 @@ static ssize_t state_store(struct kset *power_attr (state); -unsigned long s2ram_beep =0 ; - -static ssize_t s2ram_beep_show (struct kset *kset,char *buf) -{ -return sprintf (buf,"%d\n" , s2ram_beep); -} - -static ssize_t -s2ram_beep_store (struct kset *kset,const char *buf,size_t n) -{ -int val; - -if (sscanf (buf,"%d" , &val) >0 ) { - s2ram_beep = val; -return n; - } -return -EINVAL ; -} - -power_attr (s2ram_beep); -#ifdef CONFIG_PM_TRACE int pm_trace_enabled;@@ -358,13 +337,11 @@ power_attr(pm_trace);static struct attribute * g[] = { &state_attr.attr, &pm_trace_attr.attr, - &s2ram_beep_attr.attr,NULL , };#else static struct attribute * g[] = { &state_attr.attr, - &s2ram_beep_attr.attr,NULL , };#endif /* CONFIG_PM_TRACE */ kernel/sysctl.c +1
-1@@ -569,7 +569,7 @@ #ifdef CONFIG_ACPI_SLEEP { .ctl_name = KERN_ACPI_VIDEO_FLAGS, .procname = "acpi_video_flags", - .data = &acpi_video_flags, + .data = &acpi_realmode_flags, .maxlen = sizeof (unsigned long), .mode = 0644, .proc_handler = &proc_doulongvec_minmax,(s2ram patch for testing follows)
Index:
s2ram.c
===================================================================
RCS file: /cvsroot/suspend/suspend/s2ram.c,v
retrieving revision 1.46
diff -u -u -r1.46s2ram.cs2ram.c +11
-10@@ -35,13 +35,14 @@/* flags for the whitelist */ #define S3_BIOS 0x01/* machine needs acpi_sleep=s3_bios */ #define S3_MODE 0x02/* machine needs acpi_sleep=s3_mode */ -#define VBE_SAVE 0x04/* machine needs "vbetool save / restore" */ -#define VBE_POST 0x08/* machine needs "vbetool post" */ -#define RADEON_OFF 0x10/* machine needs "radeontool light off" */ -#define UNSURE 0x20/* unverified entries from acpi-support 0.59 */ -#define NOFB 0x40/* must not use a frame buffer */ -#define VBE_MODE 0x80/* machine needs "vbetool vbemode get / set" */ -#define PCI_SAVE 0x100/* we need to save the VGA PCI registers */ +#define S3_BEEP 0x04/* debugging beep */ +#define VBE_SAVE 0x08/* machine needs "vbetool save / restore" */ +#define VBE_POST 0x10/* machine needs "vbetool post" */ +#define RADEON_OFF 0x20/* machine needs "radeontool light off" */ +#define UNSURE 0x40/* unverified entries from acpi-support 0.59 */ +#define NOFB 0x80/* must not use a frame buffer */ +#define VBE_MODE 0x100/* machine needs "vbetool vbemode get / set" */ +#define PCI_SAVE 0x200/* we need to save the VGA PCI registers */ #include "whitelist.c" @@ -188,8 +189,8 @@int ret =0 ;/* 0 means: don't touch what was set on kernel commandline */ -if (flags & (S3_BIOS |S3_MODE )) - ret =set_acpi_video_mode (flags & (S3_BIOS |S3_MODE )); +if (flags & (S3_BIOS |S3_MODE |S3_BEEP )) + ret =set_acpi_video_mode (flags & (S3_BIOS |S3_MODE |S3_BEEP ));if (ret)return ret;@@ -368,7 +369,7 @@ flags |=RADEON_OFF ;break ;case 'a' : - flags |= (atoi (optarg) & (S3_BIOS |S3_MODE )); + flags |= (atoi (optarg) & (S3_BIOS |S3_MODE |S3_BEEP ));break ;case 'v' : flags |=PCI_SAVE ;
Signature · 1 line
"Premature optimization is the root of all evil." - Donald Knuth
Hi!
Quoted 10 lines
Quoted 7 lines
Here's the version that uses just one variable, relative to Nigel's patch. Hmm, and it also closes nasty trap for the user in
acpi_sleep_setup; order of parameters actually mattered there, acpi_sleep=s3_bios,s3_mode doing something different from acpi_sleep=s3_mode,s3_bios . It actually works here.Will you fold it into patch28, or should I create a changelog?
Hmm, I think we should keep the record straight. Please add a changelog if that's not a problem.
(I also added documentation and removed the superfluous beep_flags).
---
Move "debug during resume from s2ram" into the variable we already use for real-mode flags to simplify code. It also closes nasty trap for the user in acpi_sleep_setup; order of parameters actually mattered there, acpi_sleep=s3_bios,s3_mode doing something different from acpi_sleep=s3_mode,s3_bios.
arch/i386/kernel/acpi/sleep.c +8 -4
@@ -14,7 +14,7 @@ #include <asm/smp.h>/* address in low memory of the wakeup routine. */ unsigned long acpi_wakeup_address =0 ; -unsigned long acpi_video_flags; +unsigned long acpi_realmode_flags;extern char wakeup_start, wakeup_end;extern unsigned long FASTCALL (acpi_copy_wakeup_routine (unsigned long ));
@@ -68,9 +68,11 @@ static int __init acpi_sleep_setup(char {while ((str !=NULL ) && (*str !='\0' )) {if (strncmp (str,"s3_bios" ,7 ) ==0 ) - acpi_video_flags =1 ; + acpi_realmode_flags |=1 ;if (strncmp (str,"s3_mode" ,7 ) ==0 ) - acpi_video_flags |=2 ; + acpi_realmode_flags |=2 ; +if (strncmp (str,"s3_beep" ,7 ) ==0 ) + acpi_realmode_flags |=4 ; str =strchr (str,',' );if (str !=NULL ) str +=strspn (str,", \t" );
@@ -80,9 +82,11 @@ static int __init acpi_sleep_setup(char__setup ("acpi_sleep=" , acpi_sleep_setup); +/* Ouch, we want to delete this. We already have better version in userspace, in +s2ram from suspend.sf.net project */ static __init int reset_videomode_after_s3 (struct dmi_system_id *d) { - acpi_video_flags |=2 ; + acpi_realmode_flags |=2 ;return 0 ; }
arch/i386/kernel/acpi/wakeup.S +8 -10
@@ -50,7 +50,7 @@ # BEEP movw %ax, %ds # Make ds:0 point to wakeup_start movw %ax, %ss - testl $1, beep_flags - wakeup_code + testl $4, realmode_flags - wakeup_code jz 1f BEEP 1:
@@ -64,7 +64,7 @@ # BEEP cmpl $0x12345678, %eax jne bogus_real_magic - testl $1, video_flags - wakeup_code + testl $1, realmode_flags - wakeup_code jz 1f lcall $0xc000,$3 movw %cs, %ax
@@ -72,7 +72,7 @@ # BEEP movw %ax, %ss 1: - testl $2, video_flags - wakeup_code + testl $2, realmode_flags - wakeup_code jz 1f mov video_mode - wakeup_code, %ax call mode_set
@@ -111,11 +111,11 @@ # BEEP cmpl $0x12345678, %eax jne bogus_real_magic - testl $2, beep_flags - wakeup_code + testl $8, realmode_flags - wakeup_code jz 1f BEEP 1: - ljmpl $__KERNEL_CS,$wakeup_pmode_return + ljmpl $__KERNEL_CS, $wakeup_pmode_return real_save_gdt: .word 0 .long 0
@@ -124,7 +124,7 @@ real_save_cr3: .long 0 real_save_cr4: .long 0 real_magic: .long 0 video_mode: .long 0 -video_flags: .long 0 +realmode_flags: .long 0 beep_flags: .long 0 real_efer_save_restore: .long 0 real_save_efer_edx: .long 0
@@ -288,10 +288,8 @@ ENTRY(acpi_copy_wakeup_routine) movl saved_videomode, %edx movl %edx, video_mode - wakeup_start (%eax) - movl acpi_video_flags, %edx - movl %edx, video_flags - wakeup_start (%eax) - movl s2ram_beep, %edx - movl %edx, beep_flags - wakeup_start (%eax) + movl acpi_realmode_flags, %edx + movl %edx, realmode_flags - wakeup_start (%eax) movl $0x12345678, real_magic - wakeup_start (%eax) movl $0x12345678, saved_magic popl %ebx
arch/x86_64/kernel/acpi/sleep.c +5 -3
@@ -55,7 +55,7 @@ #ifdef CONFIG_ACPI_SLEEP/* address in low memory of the wakeup routine. */ unsigned long acpi_wakeup_address =0 ; -unsigned long acpi_video_flags; +unsigned long acpi_realmode_flags;extern char wakeup_start, wakeup_end;extern unsigned long acpi_copy_wakeup_routine (unsigned long );
@@ -103,9 +103,11 @@ static int __init acpi_sleep_setup(char {while ((str !=NULL ) && (*str !='\0' )) {if (strncmp (str,"s3_bios" ,7 ) ==0 ) - acpi_video_flags =1 ; + acpi_realmode_flags |=1 ;if (strncmp (str,"s3_mode" ,7 ) ==0 ) - acpi_video_flags |=2 ; + acpi_realmode_flags |=2 ; +if (strncmp (str,"s3_beep" ,7 ) ==0 ) + acpi_realmode_flags |=4 ; str =strchr (str,',' );if (str !=NULL ) str +=strspn (str,", \t" );
arch/x86_64/kernel/acpi/wakeup.S +6 -9
@@ -50,7 +50,7 @@ # Running in *copy* of this code, somewh movw %ax, %ss # Data segment must be set up before we can see whether to beep. - testl $1, beep_flags - wakeup_code + testl $4, realmode_flags - wakeup_code jz 1f BEEP 1:
@@ -70,7 +70,7 @@ # Running in *copy* of this code, somewh testl %eax, %eax jnz no_longmode - testl $1, video_flags - wakeup_code + testl $1, realmode_flags - wakeup_code jz 1f lcall $0xc000,$3 movw %cs, %ax
@@ -78,7 +78,7 @@ # Running in *copy* of this code, somewh movw %ax, %ss 1: - testl $2, video_flags - wakeup_code + testl $2, realmode_flags - wakeup_code jz 1f mov video_mode - wakeup_code, %ax call mode_seta
@@ -251,9 +251,8 @@ gdt_48a: .long gdta - wakeup_code # gdt base (relocated in later) real_magic: .quad 0 -beep_flags: .quad 0 video_mode: .quad 0 -video_flags: .quad 0 +realmode_flags: .quad 0 .code16 bogus_real_magic:
@@ -367,12 +366,10 @@ ENTRY(acpi_copy_wakeup_routine) pushq %rax pushq %rdx - movl s2ram_beep, %edx - movl %edx, beep_flags - wakeup_start (,%rdi) movl saved_video_mode, %edx movl %edx, video_mode - wakeup_start (,%rdi) - movl acpi_video_flags, %edx - movl %edx, video_flags - wakeup_start (,%rdi) + movl acpi_realmode_flags, %edx + movl %edx, realmode_flags - wakeup_start (,%rdi) movq $0x12345678, real_magic - wakeup_start (,%rdi) movq $0x123456789abcdef0, %rdx movq %rdx, saved_magic
include/linux/acpi.h +1 -2
@@ -122,8 +122,7 @@ extern struct acpi_mcfg_allocation *pci_extern int pci_mmcfg_config_num;extern int sbf_port; -extern unsigned long acpi_video_flags; -extern unsigned long s2ram_beep; +extern unsigned long acpi_realmode_flags;#else /* !CONFIG_ACPI */
kernel/power/main.c +0 -23
@@ -312,27 +312,6 @@ static ssize_t state_store(struct kset *power_attr (state); -unsigned long s2ram_beep =0 ; - -static ssize_t s2ram_beep_show (struct kset *kset,char *buf) -{ -return sprintf (buf,"%d\n" , s2ram_beep); -} - -static ssize_t -s2ram_beep_store (struct kset *kset,const char *buf,size_t n) -{ -int val; - -if (sscanf (buf,"%d" , &val) >0 ) { - s2ram_beep = val; -return n; - } -return -EINVAL ; -} - -power_attr (s2ram_beep); -#ifdef CONFIG_PM_TRACE int pm_trace_enabled;
@@ -358,13 +337,11 @@ power_attr(pm_trace);static struct attribute * g[] = { &state_attr.attr, &pm_trace_attr.attr, - &s2ram_beep_attr.attr,NULL , };#else static struct attribute * g[] = { &state_attr.attr, - &s2ram_beep_attr.attr,NULL , };#endif /* CONFIG_PM_TRACE */
kernel/sysctl.c +1 -1
@@ -569,7 +569,7 @@ #ifdef CONFIG_ACPI_SLEEP { .ctl_name = KERN_ACPI_VIDEO_FLAGS, .procname = "acpi_video_flags", - .data = &acpi_video_flags, + .data = &acpi_realmode_flags, .maxlen = sizeof (unsigned long), .mode = 0644, .proc_handler = &proc_doulongvec_minmax,
Pavel
Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
Quoted 8 lines · Pavel Machek
Hi!
Quoted 6 lines
Quoted 6 lines
Here's the version that uses just one variable, relative to Nigel's patch. Hmm, and it also closes nasty trap for the user in
acpi_sleep_setup; order of parameters actually mattered there, acpi_sleep=s3_bios,s3_mode doing something different from acpi_sleep=s3_mode,s3_bios . It actually works here.
Quoted 6 lines
Quoted 4 lines
Quoted 1 line
Will you fold it into patch28, or should I create a changelog?
Hmm, I think we should keep the record straight. Please add a changelog if that's not a problem.
(I also added documentation and removed the superfluous
beep_flags).
OK, thanks.
Added to the patchset as
I think both patches are -mm-ready, aren't they?
Greetings,
Rafael
Signature · 1 line
"Premature optimization is the root of all evil." - Donald Knuth
Hi!
Quoted 13 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Here's the version that uses just one variable, relative to Nigel's patch. Hmm, and it also closes nasty trap for the user in
acpi_sleep_setup; order of parameters actually mattered there, acpi_sleep=s3_bios,s3_mode doing something different from acpi_sleep=s3_mode,s3_bios . It actually works here.Quoted 6 lines
Quoted 4 lines
Quoted 1 line
Will you fold it into patch28, or should I create a changelog?
Hmm, I think we should keep the record straight. Please add a changelog if that's not a problem.
(I also added documentation and removed the superfluous
beep_flags).
Quoted 7 lines
OK, thanks.
Added to the patchset as
I think both patches are -mm-ready, aren't they?
I'd say so.
Pavel
Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
Hi.
Quoted 5 lines · On Thursday 05 July 2007 07:29:07 Pavel Machek wrote:
Here's the version that uses just one variable, relative to Nigel's patch. Hmm, and it also closes nasty trap for the user in
acpi_sleep_setup; order of parameters actually mattered there, acpi_sleep=s3_bios,s3_mode doing something different from acpi_sleep=s3_mode,s3_bios . It actually works here.
Good catch.
Quoted 2 lines
Will you fold it into patch28, or should I create a changelog? Testing welcome :-).
I really dislike the overloading of acpi video flags in this way. You yourself have said in the past that new functionality should use /sys instead of /proc, and here you are overloading an existing variable for the sake of expediency. Please, stick with a new /sys/power entry.
Documentation is also an issue. Your patch should update the kernel_parameters file so users can know how to get the beeping to happen. It would be nice if it mentioned the proc entry too.
[...]
arch/i386/kernel/acpi/sleep.c static int __init acpi_sleep_setup(char · line 85
Quoted 5 lines
arch/i386/kernel/acpi/sleep.c +1
-0@@ -80,9 +82,11 @@ static int __init acpi_sleep_setup(char__setup ("acpi_sleep=" , acpi_sleep_setup); +/* Ouch, we want to delete this. We already have better version in
userspace, in
arch/i386/kernel/acpi/sleep.c static int __init acpi_sleep_setup(char · line 86
Quoted 1 line
arch/i386/kernel/acpi/sleep.c +1
-0+s2ram from suspend.sf.net project */
Do we? This version has advantages in not requiring any userspace app and in being able to work even if you can't yet get as far as having userspace working. [Reads more] Oh, do you mean a better way of setting this parameter (ie requiring the userspace app and then having it echo $VALUE
Quoted 1 line
/proc/sys/kernel/acpi_video_flags? Considering we're not going to suspend
to ram during init scripts, we have a way of setting/resetting it post-boot with or without s2m and the flag doesn't therefore need to be command line option, I can go with that. Does it need a patch to mark it as deprecated?
[...]
arch/i386/kernel/acpi/wakeup.S real_save_cr3: .long 0 · line 130
Quoted 9 lines
arch/i386/kernel/acpi/wakeup.S +1
-1@@ -124,7 +124,7 @@ real_save_cr3: .long 0 real_save_cr4: .long 0 real_magic: .long 0 video_mode: .long 0 -video_flags: .long 0 +realmode_flags: .long 0 beep_flags: .long 0 real_efer_save_restore: .long 0 real_save_efer_edx: .long 0
Beep_flags should be removed too if you're sticking with /proc.
Regards,
Nigel
Signature · 2 lines
See http://www.tuxonice.net for Howtos, FAQs, mailing lists, wiki and bugzilla info.
Hi!
Quoted 3 lines
Documentation is also an issue. Your patch should update the
kernel_parametersfile so users can know how to get the beeping to happen. It would be nice if it mentioned the proc entry too.
Fixed the docs.
Quoted 10 lines
arch/i386/kernel/acpi/sleep.c static int __init acpi_sleep_setup(char · line 85
Quoted 5 lines
arch/i386/kernel/acpi/sleep.c +1
-0@@ -80,9 +82,11 @@ static int __init acpi_sleep_setup(char__setup ("acpi_sleep=" , acpi_sleep_setup); +/* Ouch, we want to delete this. We already have better version in userspace, in
arch/i386/kernel/acpi/sleep.c static int __init acpi_sleep_setup(char · line 86
Quoted 1 line
arch/i386/kernel/acpi/sleep.c +1
-0+s2ram from suspend.sf.net project */ Do we? This version has advantages in not requiring any userspace app and in being able to work even if you can't yet get as far as having
Take a look at the file. It has whitelist with just one entry, too bad.
Quoted 11 lines
arch/i386/kernel/acpi/wakeup.S real_save_cr3: .long 0 · line 130
Quoted 9 lines
arch/i386/kernel/acpi/wakeup.S +1
-1@@ -124,7 +124,7 @@ real_save_cr3: .long 0 real_save_cr4: .long 0 real_magic: .long 0 video_mode: .long 0 -video_flags: .long 0 +realmode_flags: .long 0 beep_flags: .long 0 real_efer_save_restore: .long 0 real_save_efer_edx: .long 0
Beep_flagsshould be removed too if you're sticking with /proc.
Fixed.
Pavel
Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
Hi.
Quoted 3 lines · On Thursday 05 July 2007 08:48:59 Pavel Machek wrote:
Hi!
Quoted 1 line
Documentation is also an issue. Your patch should update the
kernel_parameters
Quoted 1 line
Quoted 1 line
file so users can know how to get the beeping to happen. It would be nice
if
Quoted 3 lines
Quoted 1 line
it mentioned the proc entry too.
Fixed the docs.
Ta.
Quoted 9 lines
Quoted 9 lines
arch/i386/kernel/acpi/sleep.c static int __init acpi_sleep_setup(char · line 85
Quoted 5 lines
arch/i386/kernel/acpi/sleep.c +1
-0@@ -80,9 +82,11 @@ static int __init acpi_sleep_setup(char__setup ("acpi_sleep=" , acpi_sleep_setup); +/* Ouch, we want to delete this. We already have better version in userspace, in
arch/i386/kernel/acpi/sleep.c static int __init acpi_sleep_setup(char · line 86
Quoted 1 line
arch/i386/kernel/acpi/sleep.c +1
-0+s2ram from suspend.sf.net project */ Do we? This version has advantages in not requiring any userspace app and
in
Quoted 4 lines
Quoted 1 line
being able to work even if you can't yet get as far as having
Take a look at the file. It has whitelist with just one entry, too bad.
The contents of the whitelist are irrelevant. My laptop needs this functionality, but I haven't bothered to send you a whitelist entry, in part because I don't use s2ram.
Regardless of that, if you had read the whole comment (you've deleted half of it), you would have noticed that I ended up changing my mind and instead saying "Why not just delete the __setup now, or at least put it in the deprecated file?"
Quoted 12 lines
Quoted 11 lines
arch/i386/kernel/acpi/wakeup.S real_save_cr3: .long 0 · line 130
Quoted 9 lines
arch/i386/kernel/acpi/wakeup.S +1
-1@@ -124,7 +124,7 @@ real_save_cr3: .long 0 real_save_cr4: .long 0 real_magic: .long 0 video_mode: .long 0 -video_flags: .long 0 +realmode_flags: .long 0 beep_flags: .long 0 real_efer_save_restore: .long 0 real_save_efer_edx: .long 0
Beep_flagsshould be removed too if you're sticking with /proc.
Quoted 1 line
Fixed.
Ta. But you didn't answer the question - why /proc and not sysfs?
Regards,
Nigel
Signature · 2 lines
See http://www.tuxonice.net for Howtos, FAQs, mailing lists, wiki and bugzilla info.
Hi!
Quoted 15 lines
Quoted 9 lines
Quoted 9 lines
arch/i386/kernel/acpi/sleep.c static int __init acpi_sleep_setup(char · line 85
Quoted 5 lines
arch/i386/kernel/acpi/sleep.c +1
-0@@ -80,9 +82,11 @@ static int __init acpi_sleep_setup(char__setup ("acpi_sleep=" , acpi_sleep_setup); +/* Ouch, we want to delete this. We already have better version in userspace, in
arch/i386/kernel/acpi/sleep.c static int __init acpi_sleep_setup(char · line 86
Quoted 1 line
arch/i386/kernel/acpi/sleep.c +1
-0+s2ram from suspend.sf.net project */ Do we? This version has advantages in not requiring any userspace app and
in
Quoted 4 lines
Quoted 1 line
being able to work even if you can't yet get as far as having
Take a look at the file. It has whitelist with just one entry, too bad.
Quoted 8 lines
The contents of the whitelist are irrelevant. My laptop needs this functionality, but I haven't bothered to send you a whitelist entry, in part because I don't use s2ram.
Regardless of that, if you had read the whole comment (you've deleted half of it), you would have noticed that I ended up changing my mind and instead saying "Why not just delete the
__setupnow, or at least put it in the deprecated file?"
That should be certainly done in separate patch, right? It is on my todolist somewhere now.
Quoted 12 lines
Quoted 12 lines
Quoted 11 lines
arch/i386/kernel/acpi/wakeup.S real_save_cr3: .long 0 · line 130
Quoted 9 lines
arch/i386/kernel/acpi/wakeup.S +1
-1@@ -124,7 +124,7 @@ real_save_cr3: .long 0 real_save_cr4: .long 0 real_magic: .long 0 video_mode: .long 0 -video_flags: .long 0 +realmode_flags: .long 0 beep_flags: .long 0 real_efer_save_restore: .long 0 real_save_efer_edx: .long 0
Beep_flagsshould be removed too if you're sticking with /proc.
Quoted 3 lines
Quoted 1 line
Fixed.
Ta. But you didn't answer the question - why /proc and not sysfs?
Do you seriously advocate setting two bits of one variable from /proc, and one more bit from /sys?
Pavel
Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
Hi.
Quoted 11 lines · On Thursday 05 July 2007 09:01:09 Pavel Machek wrote:
Hi!
Quoted 9 lines
Quoted 9 lines
Quoted 9 lines
arch/i386/kernel/acpi/sleep.c static int __init acpi_sleep_setup(char · line 85
Quoted 5 lines
arch/i386/kernel/acpi/sleep.c +1
-0@@ -80,9 +82,11 @@ static int __init acpi_sleep_setup(char__setup ("acpi_sleep=" , acpi_sleep_setup); +/* Ouch, we want to delete this. We already have better version in userspace, in
arch/i386/kernel/acpi/sleep.c static int __init acpi_sleep_setup(char · line 86
Quoted 1 line
arch/i386/kernel/acpi/sleep.c +1
-0+s2ram from suspend.sf.net project */ Do we? This version has advantages in not requiring any userspace app
and
Quoted 8 lines
Quoted 8 lines
in
Quoted 4 lines
Quoted 1 line
being able to work even if you can't yet get as far as having
Take a look at the file. It has whitelist with just one entry, too bad.
The contents of the whitelist are irrelevant. My laptop needs this functionality, but I haven't bothered to send you a whitelist entry, in
part
Quoted 3 lines
Quoted 3 lines
because I don't use s2ram.
Regardless of that, if you had read the whole comment (you've deleted half
of
Quoted 6 lines
Quoted 3 lines
it), you would have noticed that I ended up changing my mind and instead saying "Why not just delete the
__setupnow, or at least put it in the deprecated file?"That should be certainly done in separate patch, right? It is on my todolist somewhere now.
Yeah, agree.
Quoted 12 lines
Quoted 12 lines
Quoted 12 lines
Quoted 11 lines
arch/i386/kernel/acpi/wakeup.S real_save_cr3: .long 0 · line 130
Quoted 9 lines
arch/i386/kernel/acpi/wakeup.S +1
-1@@ -124,7 +124,7 @@ real_save_cr3: .long 0 real_save_cr4: .long 0 real_magic: .long 0 video_mode: .long 0 -video_flags: .long 0 +realmode_flags: .long 0 beep_flags: .long 0 real_efer_save_restore: .long 0 real_save_efer_edx: .long 0
Beep_flagsshould be removed too if you're sticking with /proc.
Quoted 6 lines
Quoted 3 lines
Quoted 1 line
Fixed.
Ta. But you didn't answer the question - why /proc and not sysfs?
Do you seriously advocate setting two bits of one variable from /proc, and one more bit from /sys?
That's partly why I had a separate variable - retaining proc only because it's existing functionality, using sysfs for the new code. Remember, too, that they're really distinct functionality. The only thing they share is that they're both used in real mode.
Regards,
Nigel
Signature · 4 lines
Nigel, Michelle and Alisdair Cunningham 5 Mitchell Street Cobden 3266 Victoria, Australia
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines
Quoted 4 lines
Quoted 2 lines
Beep_flagsshould be removed too if you're sticking with /proc.Fixed.
Quoted 7 lines
Quoted 4 lines
Quoted 1 line
Ta. But you didn't answer the question - why /proc and not sysfs?
Do you seriously advocate setting two bits of one variable from /proc, and one more bit from /sys?
That's partly why I had a separate variable - retaining proc only because it's existing functionality, using sysfs for the new code. Remember, too, that
/proc is not deprecated _that_ much, and notice that this is sysctl, not regular procfs code.
Yes, I see why you did it that way, but I also think you overdisgned it a bit.
Pavel
Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
Quoted 10 lines · Pavel Machek
Quoted 10 lines
Quoted 9 lines
Quoted 6 lines
Quoted 4 lines
Quoted 2 lines
Beep_flagsshould be removed too if you're sticking with /proc.Fixed.
Ta. But you didn't answer the question - why /proc and not sysfs?
Do you seriously advocate setting two bits of one variable from /proc, and one more bit from /sys?
Quoted 8 lines
Quoted 2 lines
That's partly why I had a separate variable - retaining proc only because it's existing functionality, using sysfs for the new code. Remember, too, that
/proc is not deprecated _that_ much, and notice that this is sysctl, not regular procfs code.
Yes, I see why you did it that way, but I also think you overdisgned it a bit.
Hmm, what about adding a second interface to acpi_realmode_flags in sysfs, in a separate patch, and scheduling the old one, in /proc, for removal?
Greetings,
Rafael
Signature · 1 line
"Premature optimization is the root of all evil." - Donald Knuth
Quoted 14 lines · Rafael J. Wysocki
Quoted 13 lines · Pavel Machek
Quoted 12 lines
Quoted 9 lines
Quoted 6 lines
Quoted 4 lines
Quoted 2 lines
Beep_flagsshould be removed too if you're sticking with /proc.Fixed.
Ta. But you didn't answer the question - why /proc and not sysfs?
Do you seriously advocate setting two bits of one variable from /proc, and one more bit from /sys?
That's partly why I had a separate variable - retaining proc only because it's existing functionality, using sysfs for the new code. Remember, too, that
Quoted 8 lines
Quoted 5 lines
/proc is not deprecated _that_ much, and notice that this is sysctl, not regular procfs code.
Yes, I see why you did it that way, but I also think you overdisgned it a bit.
Hmm, what about adding a second interface to
acpi_realmode_flagsin sysfs, in a separate patch, and scheduling the old one, in /proc, for removal?
Well.. it is sysctl... so we already have separate interface, sysctl(). Yes, sysctl will probably move to /sys one day, but that is probably bigger project than just suspend.
Pavel
Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html