Skip to contentSkip to conversation

Incorrect CONFIG_SYS_MONITOR_LEN on MPC85xx boards

U-Boot Felix Radensky 2 participants 7 messages Aug 20, 2009 – Aug 24, 2009

lore
Share & tools
lore
#0 lore raw

Hi,

All FSL MPC85xx boards define CONFIG_SYS_MONITOR_LEN as 256K although actual size of u-boot binary is 512K. XES Xpedite boards seem to do the right thing.

I was wandering whether CONFIG_SYS_MONITOR_LEN for 85xx boards can be defined in terms of CONFIG_SYS_MONITOR_BASE, similar to AMCC boards ?

#define CONFIG_SYS_MONITOR_LEN (0xFFFFFFFF - CONFIG_SYS_MONITOR_BASE + 1)

Felix.

#1 lore raw
Quoted 7 lines · Felix Radensky

Hi,

All FSL MPC85xx boards define CONFIG_SYS_MONITOR_LEN as 256K although actual size of u-boot binary is 512K. XES Xpedite boards seem to do
the right thing.

Quoted 6 lines

I was wandering whether CONFIG_SYS_MONITOR_LEN for 85xx boards can be defined in terms of CONFIG_SYS_MONITOR_BASE, similar to AMCC boards ?

#define CONFIG_SYS_MONITOR_LEN (0xFFFFFFFF -
CONFIG_SYS_MONITOR_BASE + 1)

I don't have any issue w/such a fix but would like to know what the implication is of having thing set the way we do on the FSL boards. Is there some bug we'd hit?

- k

#2 lore raw

Hi, Kumar

I don't see any immediate problem with current FSL definitions except they are confusing.

I think my proposal will not work on platforms like MPC8572, MPC8536, and P2020DS where TEXT_BASE is defined as 0xeff80000 instead of 0xfff80000. Can you please explain the reason why TEXT_BASE defined differently for these boards.

Thanks.

Felix.

Quoted 14 lines · Kumar Gala
Quoted 11 lines · Felix Radensky

Hi,

All FSL MPC85xx boards define CONFIG_SYS_MONITOR_LEN as 256K although actual size of u-boot binary is 512K. XES Xpedite boards seem to do
the right thing.

I was wandering whether CONFIG_SYS_MONITOR_LEN for 85xx boards can be defined in terms of CONFIG_SYS_MONITOR_BASE, similar to AMCC boards ?

Quoted 6 lines
Quoted 2 lines

#define CONFIG_SYS_MONITOR_LEN (0xFFFFFFFF -
CONFIG_SYS_MONITOR_BASE + 1)

I don't have any issue w/such a fix but would like to know what the implication is of having thing set the way we do on the FSL boards. Is there some bug we'd hit?

I don't think there's some bug

Quoted 2 lines

- k

#3 lore raw

Its defined differently on these boards (and all our future board ports) since after relocation and such we have the FLASH @ 0xe800_0000. And thus the u-boot image is at @ 0xeff8_0000.

- k

Quoted 38 lines · Felix Radensky

Hi, Kumar

I don't see any immediate problem with current FSL definitions except they are confusing.

I think my proposal will not work on platforms like MPC8572, MPC8536, and P2020DS where TEXT_BASE is defined as 0xeff80000 instead of 0xfff80000. Can you please explain the reason why TEXT_BASE defined differently for these boards.

Thanks.

Felix.

Quoted 14 lines · Kumar Gala
Quoted 11 lines · Felix Radensky

Hi,

All FSL MPC85xx boards define CONFIG_SYS_MONITOR_LEN as 256K although actual size of u-boot binary is 512K. XES Xpedite boards seem to do
the right thing.

I was wandering whether CONFIG_SYS_MONITOR_LEN for 85xx boards can be defined in terms of CONFIG_SYS_MONITOR_BASE, similar to AMCC boards ?

Quoted 6 lines
Quoted 2 lines

#define CONFIG_SYS_MONITOR_LEN (0xFFFFFFFF -
CONFIG_SYS_MONITOR_BASE + 1)

I don't have any issue w/such a fix but would like to know what the implication is of having thing set the way we do on the FSL boards. Is there some bug we'd hit?

I don't think there's some bug

Quoted 2 lines

- k

#4 lore raw

Hi, Kumar

Quoted 3 lines · Kumar Gala

Its defined differently on these boards (and all our future board ports) since after relocation and such we have the FLASH @ 0xe800_0000. And thus the u-boot image is at @ 0xeff8_0000.

Thanks for the explanation. Out of curiosity, I've tried to set FLASH physical address on MPC8536DS
to 0xf800_0000 and modify TEXT_BASE to 0xfff8_0000. But that resulted in non-bootable system
(no u-boot messages at all).

What did I miss ?

Thanks.

Felix.

Quoted 44 lines

- k

Quoted 38 lines · Felix Radensky

Hi, Kumar

I don't see any immediate problem with current FSL definitions except they are confusing.

I think my proposal will not work on platforms like MPC8572, MPC8536, and P2020DS where TEXT_BASE is defined as 0xeff80000 instead of 0xfff80000. Can you please explain the reason why TEXT_BASE defined differently for these boards.

Thanks.

Felix.

Quoted 14 lines · Kumar Gala
Quoted 11 lines · Felix Radensky

Hi,

All FSL MPC85xx boards define CONFIG_SYS_MONITOR_LEN as 256K although actual size of u-boot binary is 512K. XES Xpedite boards seem to do
the right thing.

I was wandering whether CONFIG_SYS_MONITOR_LEN for 85xx boards can be defined in terms of CONFIG_SYS_MONITOR_BASE, similar to AMCC boards ?

Quoted 6 lines
Quoted 2 lines

#define CONFIG_SYS_MONITOR_LEN (0xFFFFFFFF -
CONFIG_SYS_MONITOR_BASE + 1)

I don't have any issue w/such a fix but would like to know what the implication is of having thing set the way we do on the FSL boards. Is there some bug we'd hit?

I don't think there's some bug

Quoted 2 lines

- k

#5 lore raw
Quoted 11 lines · Felix Radensky

Thanks for the explanation. Out of curiosity, I've tried to set FLASH physical address on MPC8536DS
to 0xf800_0000 and modify TEXT_BASE to 0xfff8_0000. But that resulted in non-bootable system
(no u-boot messages at all).

What did I miss ?

Thanks.

Felix.

Did you get all the LAWs, TLB, etc associated with moving the FLASH address?

- k

#6 lore raw
Quoted 9 lines · Kumar Gala
Quoted 6 lines · Felix Radensky

Thanks for the explanation. Out of curiosity, I've tried to set FLASH physical address on MPC8536DS
to 0xf800_0000 and modify TEXT_BASE to 0xfff8_0000. But that resulted in non-bootable system
(no u-boot messages at all).

Quoted 8 lines
Quoted 5 lines

What did I miss ?

Thanks.

Felix.

Did you get all the LAWs, TLB, etc associated with moving the FLASH address?

I think so. TLB and LAW functions use CONFIG_SYS_FLASH_BASE_PHYS, which I set to 0xf0000000 instead of 0xe0000000. Maybe I've overridden some other mappings ?

Felix.