Skip to contentSkip to conversation

S3 suspend and serial console

DISCUSSION Carl-Daniel Hailfinger 5 participants 13 messages Feb 3, 2005 – Feb 8, 2005 11 messages not in the archive ACPI

lore
Share & tools
lore

Carl-Daniel Hailfinger

author
#0 lore raw

Hi,

on resume from S3, my serial console prints garbage and stops working. I assume the BIOS resets the serial port on resume, but I have abolutely no idea how to find out what speed and parity the serial port has after resume. A quick try of 9600,57600,115200 didn't succeed and I now wonder if
- the kernel could reset the serial ports on resume
- the kernel could read the current parameters from the UART and print them out so I can adjust my serial console.

Oh well... after some investigation, it seems the serial console goes back to normal after issuing the command
echo foo >/dev/ttyS0

Could it be that the UART state is out-of-sync or something like that after resume and the serial console code doesn't notice?

Regards,
Carl-Daniel

Signature · 9 lines
http://www.hailfinger.org/



-------------------------------------------------------
This SF.Net email is sponsored by: IntelliVIEW -- Interactive Reporting
Tool for open source databases. Create drag-&-drop reports. Save time
by over 75%! Publish reports on the web. Export to DOC, XLS, RTF, etc.
Download a FREE copy at http://www.intelliview.com/go/osdn_nl

A message that is not in the archive (sent off-list or not yet received): [email protected]

author
#2 lore raw
Quoted 11 lines · Carl-Daniel Hailfinger

Hi,

on resume from S3, my serial console prints garbage and stops working. I assume the BIOS resets the serial port on resume, but I have abolutely no idea how to find out what speed and parity the serial port has after resume. A quick try of 9600,57600,115200 didn't succeed and I now wonder if
- the kernel could reset the serial ports on resume
- the kernel could read the current parameters from the UART and print them out so I can adjust my serial console.

Quoted 5 lines

Oh well... after some investigation, it seems the serial console goes back to normal after issuing the command
echo foo >/dev/ttyS0

/dev/ttyS0

fixes it, too. So you're right, Pavel.

Quoted 2 lines

Could it be that the UART state is out-of-sync or something like that after resume and the serial console code doesn't notice?

If you add an open()/close() to the S3 wakeup code, how can you find out which device you're supposed to open, that is which device the serial console is attached to? And how early can you open() the serial port? Before the "Back to C!" printk?

Once you have a patch, please mail it to me. I'll surely test it ;-)

Regards,
Carl-Daniel

Signature · 8 lines
http://www.hailfinger.org/


-------------------------------------------------------
This SF.Net email is sponsored by: IntelliVIEW -- Interactive Reporting
Tool for open source databases. Create drag-&-drop reports. Save time
by over 75%! Publish reports on the web. Export to DOC, XLS, RTF, etc.
Download a FREE copy at http://www.intelliview.com/go/osdn_nl

A message that is not in the archive (sent off-list or not yet received): [email protected]

#4 lore raw

Thus wrote Carl-Daniel Hailfinger:

Quoted 5 lines
Quoted 4 lines

goes back to normal after issuing the command
echo foo >/dev/ttyS0

/dev/ttyS0

fixes it, too. So you're right, Pavel.

echo > /dev/ttyS0 will echo a single '\n'. I guess echo -n would be more appropriate.
Best regards,

Signature · 9 lines
Karol 'sziwan' Kozimor
sziwan-DETuoxkZsSqrDJvtcaxF/[email protected]


-------------------------------------------------------
This SF.Net email is sponsored by: IntelliVIEW -- Interactive Reporting
Tool for open source databases. Create drag-&-drop reports. Save time
by over 75%! Publish reports on the web. Export to DOC, XLS, RTF, etc.
Download a FREE copy at http://www.intelliview.com/go/osdn_nl

A message that is not in the archive (sent off-list or not yet received): 20050205155845.GA20131-DETuoxkZsSqrDJvtcaxF/[email protected]

#6 lore raw
Quoted 1 line · Karol Kozimor

echo > /dev/ttyS0 will echo a single '\n'. I guess echo -n would be more

only "> /dev/ttyS0", no echo. This basically does...

open("/dev/tty", O_WRONLY|O_CREAT|O_TRUNC|O_LARGEFILE, 0666) = 3

fcntl64(1, F_GETFD)                     = 0
fcntl64(1, F_DUPFD, 10)                 = 10
fcntl64(1, F_GETFD)                     = 0
fcntl64(10, F_SETFD, FD_CLOEXEC)        = 0
dup2(3, 1)                              = 1
close(3)                                = 0

so it is pretty equivalent to "echo -n > /dev/ttyS0"

Signature · 9 lines
Stefan Seyfried



-------------------------------------------------------
This SF.Net email is sponsored by: IntelliVIEW -- Interactive Reporting
Tool for open source databases. Create drag-&-drop reports. Save time
by over 75%! Publish reports on the web. Export to DOC, XLS, RTF, etc.
Download a FREE copy at http://www.intelliview.com/go/osdn_nl

A message that is not in the archive (sent off-list or not yet received): 20050205220619.GA30739-l0tNAEGuAhhzZ8+rp42Dbp9+tswZ0GTaehPwdyo5hKaELgA04lAiVw@public.gmane.org

author
#8 lore raw
Quoted 4 lines · Stefan Seyfried

only "> /dev/ttyS0", no echo. This basically does...

open("/dev/tty", O_WRONLY|O_CREAT|O_TRUNC|O_LARGEFILE, 0666) = 3

Quoted 8 lines
fcntl64(1, F_GETFD)                     = 0
fcntl64(1, F_DUPFD, 10)                 = 10
fcntl64(1, F_GETFD)                     = 0
fcntl64(10, F_SETFD, FD_CLOEXEC)        = 0
dup2(3, 1)                              = 1
close(3)                                = 0

so it is pretty equivalent to "echo -n > /dev/ttyS0"

Erm... is it intentional that serial powermanagement code is never called on entering/leaving S3? When I added debugging statements to the serial code, this seemed very strange.

Regards,
Carl-Daniel

Signature · 8 lines
http://www.hailfinger.org/


-------------------------------------------------------
This SF.Net email is sponsored by: IntelliVIEW -- Interactive Reporting
Tool for open source databases. Create drag-&-drop reports. Save time
by over 75%! Publish reports on the web. Export to DOC, XLS, RTF, etc.
Download a FREE copy at http://www.intelliview.com/go/osdn_nl

A message that is not in the archive (sent off-list or not yet received): [email protected]

#10 lore raw

Hi!

Quoted 10 lines
Quoted 10 lines

only "> /dev/ttyS0", no echo. This basically does...

open("/dev/tty", O_WRONLY|O_CREAT|O_TRUNC|O_LARGEFILE, 0666) = 3

fcntl64(1, F_GETFD)                     = 0
fcntl64(1, F_DUPFD, 10)                 = 10
fcntl64(1, F_GETFD)                     = 0
fcntl64(10, F_SETFD, FD_CLOEXEC)        = 0
dup2(3, 1)                              = 1
close(3)                                = 0
Quoted 4 lines
Quoted 1 line

so it is pretty equivalent to "echo -n > /dev/ttyS0"

Erm... is it intentional that serial powermanagement code is never called on entering/leaving S3? When I added debugging statements

Probably noone was yet trying to use serial during suspend....

Pavel

Signature · 9 lines
People were complaining that M$ turns users into beta-testers...
...jr ghea gurz vagb qrirybcref, naq gurl frrz gb yvxr vg gung jnl!


-------------------------------------------------------
This SF.Net email is sponsored by: IntelliVIEW -- Interactive Reporting
Tool for open source databases. Create drag-&-drop reports. Save time
by over 75%! Publish reports on the web. Export to DOC, XLS, RTF, etc.
Download a FREE copy at http://www.intelliview.com/go/osdn_nl

A message that is not in the archive (sent off-list or not yet received): 20050207141655.GB8040-I/[email protected]

author
#12 lore raw

Hi!

Quoted 2 lines
Quoted 2 lines

Erm... is it intentional that serial powermanagement code is never called on entering/leaving S3? When I added debugging statements

Quoted 2 lines · Pavel Machek

Probably noone was yet trying to use serial during suspend....

You mean noone uses the serial console while testing suspend? What is the new way to debug suspend problems not requiring a serial console?

Regards,
Carl-Daniel

Signature · 8 lines
http://www.hailfinger.org/


-------------------------------------------------------
This SF.Net email is sponsored by: IntelliVIEW -- Interactive Reporting
Tool for open source databases. Create drag-&-drop reports. Save time
by over 75%! Publish reports on the web. Export to DOC, XLS, RTF, etc.
Download a FREE copy at http://www.intelliview.com/go/osdn_nl

A message that is not in the archive (sent off-list or not yet received): [email protected]

#14 lore raw

Hi!

Quoted 8 lines
Quoted 4 lines
Quoted 2 lines

Erm... is it intentional that serial powermanagement code is never called on entering/leaving S3? When I added debugging statements

Probably noone was yet trying to use serial during suspend....

You mean noone uses the serial console while testing suspend? What is the new way to debug suspend problems not requiring a serial console?

Well, when video works, you debug using video. I used hardware debugger, too...
Pavel

Signature · 9 lines
People were complaining that M$ turns users into beta-testers...
...jr ghea gurz vagb qrirybcref, naq gurl frrz gb yvxr vg gung jnl!


-------------------------------------------------------
This SF.Net email is sponsored by: IntelliVIEW -- Interactive Reporting
Tool for open source databases. Create drag-&-drop reports. Save time
by over 75%! Publish reports on the web. Export to DOC, XLS, RTF, etc.
Download a FREE copy at http://www.intelliview.com/go/osdn_nl

A message that is not in the archive (sent off-list or not yet received): 20050207143242.GE8040-I/[email protected]

author
#16 lore raw
Quoted 6 lines · Pavel Machek

Hi!

Quoted 3 lines
Quoted 2 lines
Quoted 2 lines

Erm... is it intentional that serial powermanagement code is never called on entering/leaving S3? When I added debugging statements

Quoted 8 lines
Quoted 5 lines
Quoted 1 line

Probably noone was yet trying to use serial during suspend....

You mean noone uses the serial console while testing suspend? What is the new way to debug suspend problems not requiring a serial console?

Well, when video works, you debug using video. I used hardware debugger, too...

Will there be any patch to actually use the powermanagement code in the serial core?

Regarding the original problem:
It seems the following call is sufficient to make the console work again. However, I haven't found a function which would give me the parameters needed to call ->set_termios.

serial/8250.c: serial8250_set_termios

Do you have any idea how to obtain the termios and tty structures for a given port, e.g. ttyS0?

Regards,
Carl-Daniel

Signature · 9 lines
http://www.hailfinger.org/



-------------------------------------------------------
This SF.Net email is sponsored by: IntelliVIEW -- Interactive Reporting
Tool for open source databases. Create drag-&-drop reports. Save time
by over 75%! Publish reports on the web. Export to DOC, XLS, RTF, etc.
Download a FREE copy at http://www.intelliview.com/go/osdn_nl

A message that is not in the archive (sent off-list or not yet received): [email protected]

#18 lore raw

Hi.

Quoted 25 lines · Carl-Daniel Hailfinger
Quoted 6 lines · Pavel Machek

Hi!

Quoted 3 lines
Quoted 2 lines
Quoted 2 lines

Erm... is it intentional that serial powermanagement code is never called on entering/leaving S3? When I added debugging statements

Quoted 8 lines
Quoted 5 lines
Quoted 1 line

Probably noone was yet trying to use serial during suspend....

You mean noone uses the serial console while testing suspend? What is the new way to debug suspend problems not requiring a serial console?

Well, when video works, you debug using video. I used hardware debugger, too...

Will there be any patch to actually use the powermanagement code in the serial core?

Regarding the original problem:
It seems the following call is sufficient to make the console work again. However, I haven't found a function which would give me the parameters needed to call ->set_termios.

Quoted 4 lines

serial/8250.c: serial8250_set_termios

Do you have any idea how to obtain the termios and tty structures for a given port, e.g. ttyS0?

We have code for 2.4 in Suspend2. I haven't seen any need for it in 2.6; it just works (except that you can't send keystrokes early in the boot process). Haven't tried S3, so I don't know the issues there.

Regards,

Nigel

Signature · 13 lines
Nigel Cunningham
Software Engineer, Canberra, Australia
http://www.cyclades.com

Ph: +61 (2) 6292 8028      Mob: +61 (417) 100 574



-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
#19 lore raw
In reply to a message not in the archive

Hi!

Quoted 9 lines

Regarding the original problem:
It seems the following call is sufficient to make the console work again. However, I haven't found a function which would give me the parameters needed to call ->set_termios.

serial/8250.c: serial8250_set_termios

Do you have any idea how to obtain the termios and tty structures for a given port, e.g. ttyS0?

Save some termios passed to it in static variable and than use that one as a quick hack...
Pavel

Signature · 9 lines
People were complaining that M$ turns users into beta-testers...
...jr ghea gurz vagb qrirybcref, naq gurl frrz gb yvxr vg gung jnl!


-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click

A message that is not in the archive (sent off-list or not yet received): 20050208125453.GA1622-I/[email protected]

author
#21 lore raw

Hi,

Quoted 6 lines · Pavel Machek
Quoted 6 lines

Regarding the original problem:
It seems the following call is sufficient to make the console work again. However, I haven't found a function which would give me the parameters needed to call ->set_termios.

Quoted 7 lines
Quoted 4 lines

serial/8250.c: serial8250_set_termios

Do you have any idea how to obtain the termios and tty structures for a given port, e.g. ttyS0?

Save some termios passed to it in static variable and than use that one as a quick hack...

Yes, that will hopefully work. The hack I'm using right now causes random oopses, but at least the serial console can display them :-|

drivers/acpi/sleep/main.c 2005-02-08 13:16:13 +01:00 renamed from drivers/acpi/sleep/main.c 2005-01-15 23:01:53 +01:00 +4 -0
@@ -18,6 +18,7 @@
 #include <acpi/acpi_bus.h>
 #include <acpi/acpi_drivers.h>
 #include "sleep.h"
+#include <linux/syscalls.h>
 
 u8 sleep_states[ACPI_S_STATE_COUNT];
 

@@ -82,6 +83,7 @@
        acpi_status status = AE_OK;
        unsigned long flags = 0;
        u32 acpi_state = acpi_suspend_states[pm_state];
+       struct file* filp;
 
        ACPI_FLUSH_CPU_CACHE();
 

@@ -126,6 +128,8 @@
        if (pm_state > PM_SUSPEND_STANDBY)
                acpi_restore_state_mem();
 
+       filp = filp_open("/dev/ttyS0", O_WRONLY, 0);
+       filp_close(filp, NULL);
 
        return ACPI_SUCCESS(status) ? 0 : -EFAULT;
 }

Any idea why that darn thing oopses on me at random locations?

First Oops on suspend:

divide error: 0000 [#1]

Modules linked in: nvram parport_pc lp parport edd cpufreq_userspace speedstep_s
CPU:    0
EIP:    0060:[<c011a8ed>]    Not tainted VLI
EFLAGS: 00000286   (2.6.11-rc3-latest-s3test14)

EIP is at release_console_sem+0x1d/0x90
eax: c0303fc0 ebx: 00000286 ecx: ffffffff edx: 00005536
esi: fffffe89 edi: 0000003e ebp: dfd8c200 esp: d97b5db4
ds: 0068 es: 007b ss: 0068
Process s3script (pid: 8845, threadinfo=d97b4000 task=dd45fa20)

Stack: c01b2744 00000001 00000246 c011a86d d97b5de8 c03abe60 dfb71900 d97b5e08
       c011a757 c02d77a4 d97b5de8 c021b6a5 c02d77a4 c013678f c03abe60 db35c9c0
       00000000 dd45fa20 c0117590 00000000 00000000 c03abe60 00000296 00000296
Call Trace:
 [<c01b2744>] vscnprintf+0x14/0x20
 [<c011a86d>] vprintk+0x10d/0x120
 [<c011a757>] printk+0x17/0x20
 [<c021b6a5>] uart_block_til_ready+0x1a5/0x1b0
 [<c013678f>] request_irq+0x7f/0xb0
 [<c0117590>] default_wake_function+0x0/0x10
 [<c0117590>] default_wake_function+0x0/0x10
 [<c0219c2c>] uart_startup+0x8c/0x120
 [<c021b8d8>] uart_open+0x118/0x140
 [<c0204cc6>] tty_open+0x166/0x2a0
 [<c015b609>] chrdev_open+0xa9/0x140
 [<c0154229>] get_empty_filp+0x59/0xe0
 [<c01529c6>] dentry_open+0x176/0x1e0
 [<c0152840>] filp_open+0x40/0x50
 [<c011a86d>] vprintk+0x10d/0x120
 [<c01f3795>] acpi_pm_enter+0x96/0xac
 [<c013153e>] suspend_enter+0x2e/0x50
 [<c01315ef>] enter_state+0x4f/0x90
 [<c01f380f>] acpi_suspend+0x25/0x32
 [<c01b369a>] copy_from_user+0x3a/0x80
 [<c01f3917>] acpi_system_write_sleep+0x69/0x7a
 [<c0153684>] vfs_write+0xd4/0x130
 [<c0153791>] sys_write+0x41/0x70
 [<c0102e0d>] sysenter_past_esp+0x52/0x75
Code: b5 37 c0 c3 8d 76 00 8d bc 27 00 00 00 00 56 31 f6 53 83 ec 04 eb 14 8d b

Reboot, next try to suspend:

Unable to handle kernel paging request at virtual address fffffe20
 printing eip:
c0152c03
*pde = 00002067
*pte = 00000000
Oops: 0000 [#1]
Modules linked in: nvram parport_pc lp parport cpufreq_userspace speedstep_cents
CPU:    0
EIP:    0060:[<c0152c03>]    Not tainted VLI
EFLAGS: 00010282   (2.6.11-rc3-latest-s3test14)

EIP is at filp_close+0x13/0x90
eax: fffffe00 ebx: fffffe00 ecx: dd1d00c0 edx: 00000000
esi: 00000002 edi: 00000000 ebp: 00000000 esp: d800def8
ds: 0068 es: 007b ss: 0068
Process s3script (pid: 8754, threadinfo=d800c000 task=df0d2520)

Stack: 00000002 00000003 00000002 00000046 c01f379c 00000000 00000282 00000003
       00000002 c013153e 00000000 00000003 d800df54 c01315ef c02bdaed 00000002
       c02b3088 c01f380f 00000000 00000001 00000000 00000003 00000004 40204000
Call Trace:
 [<c01f379c>] acpi_pm_enter+0x9d/0xac
 [<c013153e>] suspend_enter+0x2e/0x50
 [<c01315ef>] enter_state+0x4f/0x90
 [<c01f380f>] acpi_suspend+0x25/0x32
 [<c01b369a>] copy_from_user+0x3a/0x80
 [<c01f3917>] acpi_system_write_sleep+0x69/0x7a
 [<c0153684>] vfs_write+0xd4/0x130
 [<c0153791>] sys_write+0x41/0x70
 [<c0102e0d>] sysenter_past_esp+0x52/0x75
Code: 89 04 24 e8 40 ff ff ff 83 c4 0c c3 8d b6 00 00 00 00 8d bf 00 00 00 00 8

Is it generally unsafe to call anything during resume?

Regards,
Carl-Daniel

Signature · 8 lines
http://www.hailfinger.org/


-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click

A message that is not in the archive (sent off-list or not yet received): [email protected]

#23 lore raw

Hi!

Quoted 7 lines
Quoted 7 lines
Quoted 7 lines

Regarding the original problem:
It seems the following call is sufficient to make the console work again. However, I haven't found a function which would give me the parameters needed to call ->set_termios.

serial/8250.c: serial8250_set_termios

Quoted 8 lines
Quoted 5 lines
Quoted 2 lines

Do you have any idea how to obtain the termios and tty structures for a given port, e.g. ttyS0?

Save some termios passed to it in static variable and than use that one as a quick hack...

Yes, that will hopefully work. The hack I'm using right now causes random oopses, but at least the serial console can display them :-|

Are you using 8250.c? It seems to have some suspend/resume support. Perhaps you should talk to Russell King?

Quoted 26 lines
drivers/acpi/sleep/main.c 2005-02-08 13:16:13 +01:00 renamed from drivers/acpi/sleep/main.c 2005-01-15 23:01:53 +01:00 +4 -0
@@ -18,6 +18,7 @@
 #include <acpi/acpi_bus.h>
 #include <acpi/acpi_drivers.h>
 #include "sleep.h"
+#include <linux/syscalls.h>
 
 u8 sleep_states[ACPI_S_STATE_COUNT];
 

@@ -82,6 +83,7 @@
        acpi_status status = AE_OK;
        unsigned long flags = 0;
        u32 acpi_state = acpi_suspend_states[pm_state];
+       struct file* filp;
 
        ACPI_FLUSH_CPU_CACHE();
 

@@ -126,6 +128,8 @@
        if (pm_state > PM_SUSPEND_STANDBY)
                acpi_restore_state_mem();
 
+       filp = filp_open("/dev/ttyS0", O_WRONLY, 0);
+       filp_close(filp, NULL);
Quoted 1 line

Any idea why that darn thing oopses on me at random locations?

Yes, you can only do such stuff from process context etc... not from some random place like this.

Pavel

Signature · 9 lines
People were complaining that M$ turns users into beta-testers...
...jr ghea gurz vagb qrirybcref, naq gurl frrz gb yvxr vg gung jnl!


-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click