Skip to contentSkip to conversation

common/console.c

U-Boot New Heiko Schocher 4 participants 8 messages Sep 1, 2008 – Sep 2, 2008

lore
Share & tools
lore
#0 lore raw

Hello Jean-Christophe,

I compilied actual u-boot for the muas3001 board and tried to start it, but it crashes in common/console.c console_init_r () ...

After reversing your Patch:

http://git.denx.de/?p=u-boot.git;a=commitdiff;h=c1de7a6daf9c657484e1c6d433f01fccd49a7f48#patch4

it works again fine.

It seems to me, that in console_init_r ()

struct list_head *list = device_get_list();

list points after device_get_list() not to a valid list ...

at least for MPC8260 systems!

bye
Heiko

Signature · 2 lines
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
#1 lore raw

PATCH 1/1 device: clone device on register

to prevent that the pointer will not be use for an other register

---
 common/devices.c  |   24 +++++++++++++++++++++++-
 include/devices.h |    1 +
 2 files changed, 24 insertions(+), 1 deletions(-)
common/devices.c +23 -1
index 2977436..8beebe2 100644
@@ -130,10 +130,32 @@ device_t* device_get_by_name(char* name)
 	return NULL;
 }
 
+device_t* device_clone(device_t *dev)
+{
+	device_t *_dev;
+
+	if(!dev)
+		return NULL;
+
+	_dev = calloc(1, sizeof(device_t));
+
+	if(!_dev)
+		return NULL;
+
+	memcpy(_dev, dev, sizeof(device_t));
+	strncpy(_dev->name, dev->name, 16);
+
+	return _dev;
+}
 
 int device_register (device_t * dev)
 {
-	list_add(&(dev->list), &(devs.list));
+	device_t *_dev;
+
+	_dev = device_clone(dev);
+	if(!_dev)
+		return -1;
+	list_add(&(_dev->list), &(devs.list));
 	return 0;
 }
 
include/devices.h +1 -0
index 490016b..6b78d58 100644
@@ -94,6 +94,7 @@ int	devices_init (void);
 int	device_deregister(char *devname);
 struct list_head* device_get_list(void);
 device_t* device_get_by_name(char* name);
+device_t* device_clone(device_t *dev);
 
 #ifdef CONFIG_LCD
 int	drv_lcd_init (void);
Signature · 1 line
1.5.6.3
#2 lore raw

PATCH 1/1 device: clone device on register

Quoted 3 lines · Jean-Christophe PLAGNIOL-VILLARD

to prevent that the pointer will not be use for an other register

Er. What? If a device is registered multiple times, that sounds like something which should be fixed, not papered over like this...

Btw, while I'm all for getting rid of common/lists.c, your cryptic one-line commit message is a shooting offense to anyone who tries to understand WTF is going on...

Haavard

#3 lore raw

PATCH 1/1 device: clone device on register

Quoted 7 lines · Haavard Skinnemoen
Quoted 3 lines · Jean-Christophe PLAGNIOL-VILLARD

to prevent that the pointer will not be use for an other register

Er. What? If a device is registered multiple times, that sounds like something which should be fixed, not papered over like this...

I take that back. I didn't realize how much of a f***ing trainwreck the device system is...apparently the callers expect device_register() to clone the device, but this fact was hidden well within the old list implementation. Your patches makes this much more apparent, which is obviously good.

Oh, and your patch fixes the breakage I was seeing on avr32. Thanks.

Haavard

#4 lore raw

PATCH 1/1 device: clone device on register

Quoted 9 lines · Haavard Skinnemoen
Quoted 7 lines · Haavard Skinnemoen
Quoted 3 lines · Jean-Christophe PLAGNIOL-VILLARD

to prevent that the pointer will not be use for an other register

Er. What? If a device is registered multiple times, that sounds like something which should be fixed, not papered over like this...

Quoted 7 lines

I take that back. I didn't realize how much of a f***ing trainwreck the device system is...apparently the callers expect device_register() to clone the device, but this fact was hidden well within the old list implementation. Your patches makes this much more apparent, which is obviously good.

Oh, and your patch fixes the breakage I was seeing on avr32. Thanks.

I've plan to cleanup this part during next release.

and redesign console management to simplify it and reduce its code.

I'm currently designing a generic bus/device/driver mecanism that will simplify the implementation without increase the size.

I'll send a announce about it

Best Regards,
J.

#5 lore raw

PATCH 1/1 device: clone device on register

Quoted 6 lines · Jean-Christophe PLAGNIOL-VILLARD

I've plan to cleanup this part during next release.

and redesign console management to simplify it and reduce its code.

I'm currently designing a generic bus/device/driver mecanism that will simplify the implementation without increase the size.

Cool. Something like that is definitely needed.

Haavard

#6 lore raw

PATCH 1/1 device: clone device on register

In reply to Haavard Skinnemoen

Dear Jean-Christophe,

In message <[email protected]>

Quoted 11 lines · Haavard Skinnemoen
Quoted 4 lines

to prevent that the pointer will not be use for an other register

Er. What? If a device is registered multiple times, that sounds like something which should be fixed, not papered over like this...

Btw, while I'm all for getting rid of common/lists.c, your cryptic one-line commit message is a shooting offense to anyone who tries to understand WTF is going on...

I agree with Haavard on both accounts.

Can you please look into this again? Thanks

Best regards,

Wolfgang Denk

Signature · 5 lines
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
The human race is a race of cowards; and I am not  only  marching  in
that procession but carrying a banner.                   - Mark Twain
#7 lore raw

PATCH 1/1 device: clone device on register

Quoted 7 lines · I wrote:
Quoted 7 lines
Quoted 3 lines

to prevent that the pointer will not be use for an other register

Er. What? If a device is registered multiple times, that sounds like something which should be fixed, not papered over like this...

Quoted 7 lines
Quoted 3 lines

Btw, while I'm all for getting rid of common/lists.c, your cryptic one-line commit message is a shooting offense to anyone who tries to understand WTF is going on...

I agree with Haavard on both accounts.

Can you please look into this again? Thanks

OK, I learned that item (1) was actually a non-issue.

But can we please use a commit message that can be understood? I cannot parse the current one.

Best regards,

Wolfgang Denk

Signature · 6 lines
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
The universe, they said, depended for its operation on the balance of
four forces which they identified as charm,  persuasion,  uncertainty
and bloody-mindedness.      -- Terry Pratchett, "The Light Fantastic"