Hi
I am attempting to download a 23M RAM debug image to my PXA255 based
board with 32M RAM onboard over ethernet. After about 16.1M, it fails
repeatedly at same location each time. I use routines used in the
XScale BSP EbootInitEtherTransport, and EbootEtherReadData, but my
bootloader code is otherwise modified from XSC1BD.

My download fails after 16.1M and with error message from the PB the
Download failed, probably due to my device stopping to repond. I am
connected with a HW debugger and when hitting break after download
failed, I see that it is somewhere in the ne2000 routines, but the
stack pointer and cpsr looks OK, so the system looks like it has
entered an infinite loop or something.

When trying to download other images, they pass if they are smaller
release images, but the big debug images all fail, each at its own
dedicated location.

I know since I am using a "personal" bootloader, it may be a longshot,
but have anybody seen similar issues?

My boot.bib and config.bib MEMORY sections looks as follows when
downloading nk to RAM , maybe the error lies in here. Note that the
bootloader is using physical space, while the image is using virtual
adressing mode.


boot.bib
MEMORY
EBOOT 00001000 0007E000 RAMIMAGE
RAM 04204000 00110000 RAM

TRANSTABL 04000000 00004000 RESERVED
STACK 04004000 001FC000 RESERVED
IRQSTACK 04200000 00004000 RESERVED


config.bib
; Name|Virtual/Cached |Size| Region Type
MEMORY
NK 84314000 0173C000 RAMIMAGE
RAM 85A50000 005B0000 RAM
EBOOT 80000000 00080000 RESERVED

Regards,
Inger

Re: Ethernet download of large nk.bin fails at same location by K

K
Wed Jun 23 10:22:02 CDT 2004

Would you please also post the config section of boot.bib and config.bib
also the OEMAddressTable
so that it migh help to clear your problem.

"Inger" <inger72@online.no>
???????:41a44da6.0406230710.762bbff1@posting.google.com...
> Hi
> I am attempting to download a 23M RAM debug image to my PXA255 based
> board with 32M RAM onboard over ethernet. After about 16.1M, it fails
> repeatedly at same location each time. I use routines used in the
> XScale BSP EbootInitEtherTransport, and EbootEtherReadData, but my
> bootloader code is otherwise modified from XSC1BD.
>
> My download fails after 16.1M and with error message from the PB the
> Download failed, probably due to my device stopping to repond. I am
> connected with a HW debugger and when hitting break after download
> failed, I see that it is somewhere in the ne2000 routines, but the
> stack pointer and cpsr looks OK, so the system looks like it has
> entered an infinite loop or something.
>
> When trying to download other images, they pass if they are smaller
> release images, but the big debug images all fail, each at its own
> dedicated location.
>
> I know since I am using a "personal" bootloader, it may be a longshot,
> but have anybody seen similar issues?
>
> My boot.bib and config.bib MEMORY sections looks as follows when
> downloading nk to RAM , maybe the error lies in here. Note that the
> bootloader is using physical space, while the image is using virtual
> adressing mode.
>
>
> boot.bib
> MEMORY
> EBOOT 00001000 0007E000 RAMIMAGE
> RAM 04204000 00110000 RAM
>
> TRANSTABL 04000000 00004000 RESERVED
> STACK 04004000 001FC000 RESERVED
> IRQSTACK 04200000 00004000 RESERVED
>
>
> config.bib
> ; Name|Virtual/Cached |Size| Region Type
> MEMORY
> NK 84314000 0173C000 RAMIMAGE
> RAM 85A50000 005B0000 RAM
> EBOOT 80000000 00080000 RESERVED
>
> Regards,
> Inger



Re: Ethernet download of large nk.bin fails at same location by inger72

inger72
Thu Jun 24 01:50:05 CDT 2004

Will do:
memmap.inc
OEMAddressTable
DCD FLASH_BASE_C_VIRTUAL, FLASH_BASE_PHYSICAL, 32
DCD PSRAM_BASE_C_VIRTUAL, PSRAM_BASE_PHYSICAL, 32
DCD CPLD_BASE_C_VIRTUAL, CPLD_BASE_PHYSICAL, 1
DCD COMPFLASH_IO_BASE_C_VIRTUAL, COMPFLASH_IO_BASE_PHYSICAL,
256
DCD PERIF_BASE_C_VIRTUAL, PERIF_BASE_PHYSICAL, 64
DCD LCD_BASE_C_VIRTUAL, LCD_BASE_PHYSICAL, 64
DCD MEMC_BASE_C_VIRTUAL, MEMC_BASE_PHYSICAL, 64
DCD 0x00000000, 0x00000000, 0

defined as:
FLASH_BASE_PHYSICAL EQU 0x00000000
FLASH_BASE_C_VIRTUAL EQU 0x80000000
FLASH_BASE_U_VIRTUAL EQU (FLASH_BASE_C_VIRTUAL +
CACHED_TO_UNCACHED_OFFSET)

PSRAM_BASE_PHYSICAL EQU 0x04000000
PSRAM_BASE_C_VIRTUAL EQU 0x84000000
PSRAM_BASE_U_VIRTUAL EQU (PSRAM_BASE_C_VIRTUAL +
CACHED_TO_UNCACHED_OFFSET)

CPLD_BASE_PHYSICAL EQU 0x08000000
CPLD_BASE_C_VIRTUAL EQU 0x88000000
CPLD_BASE_U_VIRTUAL EQU (CPLD_BASE_C_VIRTUAL +
CACHED_TO_UNCACHED_OFFSET)

CACHED_TO_UNCACHED_OFFSET EQU 0x30000000

PERIF_BASE_PHYSICAL EQU 0x40000000
PERIF_BASE_C_VIRTUAL EQU 0x98500000
PERIF_BASE_U_VIRTUAL EQU (PERIF_BASE_C_VIRTUAL +
CACHED_TO_UNCACHED_OFFSET)

LCD_BASE_PHYSICAL EQU 0x44000000
LCD_BASE_C_VIRTUAL EQU 0x9C500000
LCD_BASE_U_VIRTUAL EQU (LCD_BASE_C_VIRTUAL +
CACHED_TO_UNCACHED_OFFSET)

MEMC_BASE_PHYSICAL EQU 0x48000000
MEMC_BASE_C_VIRTUAL EQU 0xA0500000
MEMC_BASE_U_VIRTUAL EQU (MEMC_BASE_C_VIRTUAL +
CACHED_TO_UNCACHED_OFFSET)




config.bib CONFIG section:
CONFIG
COMPRESSION=ON
PROFILE=ON
KERNELFIXUPS=ON
ROMFLAGS=0
AUTOSIZE=ON
ROMSTART=80000000
ROMWIDTH=32
ROMSIZE=02000000
ROMOFFSET=80000000

boot.bib CONFIG section:
COMPRESSION= ON
PROFILE= OFF
KERNELFIXUPS= ON
SRE= ON
ROMSTART= 00000000
ROMWIDTH= 32
ROMSIZE= 00080000


the ROMsize in boot.bib was reduced to decrease the size of eboot.bin
that I program using an UniSTAC II ICE.

Regards,
Inger


"K. S. Huang" <ks_huang@alphanetworks.com.remove.this> wrote in message news:<eQh6IWTWEHA.644@tk2msftngp13.phx.gbl>...
> Would you please also post the config section of boot.bib and config.bib
> also the OEMAddressTable
> so that it migh help to clear your problem.
>
> "Inger" <inger72@online.no>
> ???????:41a44da6.0406230710.762bbff1@posting.google.com...
> > Hi
> > I am attempting to download a 23M RAM debug image to my PXA255 based
> > board with 32M RAM onboard over ethernet. After about 16.1M, it fails
> > repeatedly at same location each time. I use routines used in the
> > XScale BSP EbootInitEtherTransport, and EbootEtherReadData, but my
> > bootloader code is otherwise modified from XSC1BD.
> >
> > My download fails after 16.1M and with error message from the PB the
> > Download failed, probably due to my device stopping to repond. I am
> > connected with a HW debugger and when hitting break after download
> > failed, I see that it is somewhere in the ne2000 routines, but the
> > stack pointer and cpsr looks OK, so the system looks like it has
> > entered an infinite loop or something.
> >
> > When trying to download other images, they pass if they are smaller
> > release images, but the big debug images all fail, each at its own
> > dedicated location.
> >
> > I know since I am using a "personal" bootloader, it may be a longshot,
> > but have anybody seen similar issues?
> >
> > My boot.bib and config.bib MEMORY sections looks as follows when
> > downloading nk to RAM , maybe the error lies in here. Note that the
> > bootloader is using physical space, while the image is using virtual
> > adressing mode.
> >
> >
> > boot.bib
> > MEMORY
> > EBOOT 00001000 0007E000 RAMIMAGE
> > RAM 04204000 00110000 RAM
> >
> > TRANSTABL 04000000 00004000 RESERVED
> > STACK 04004000 001FC000 RESERVED
> > IRQSTACK 04200000 00004000 RESERVED
> >
> >
> > config.bib
> > ; Name|Virtual/Cached |Size| Region Type
> > MEMORY
> > NK 84314000 0173C000 RAMIMAGE
> > RAM 85A50000 005B0000 RAM
> > EBOOT 80000000 00080000 RESERVED
> >
> > Regards,
> > Inger

Re: Ethernet download of large nk.bin fails at same location by K

K
Thu Jun 24 02:45:51 CDT 2004

Maybe you can write some simple SDRAM testing code (Write something and Read
it back) in your EBoot to make sure the RAM between 0x04314_000 and
0x05FF_FFFF is all accessable.

if failed you may beed to check out the SDRAM is configred properly!

"Inger" <inger72@online.no>
???????:41a44da6.0406232250.6da83d18@posting.google.com...
> Will do:
> memmap.inc
> OEMAddressTable
> DCD FLASH_BASE_C_VIRTUAL, FLASH_BASE_PHYSICAL, 32
> DCD PSRAM_BASE_C_VIRTUAL, PSRAM_BASE_PHYSICAL, 32
> DCD CPLD_BASE_C_VIRTUAL, CPLD_BASE_PHYSICAL, 1
> DCD COMPFLASH_IO_BASE_C_VIRTUAL, COMPFLASH_IO_BASE_PHYSICAL,
> 256
> DCD PERIF_BASE_C_VIRTUAL, PERIF_BASE_PHYSICAL, 64
> DCD LCD_BASE_C_VIRTUAL, LCD_BASE_PHYSICAL, 64
> DCD MEMC_BASE_C_VIRTUAL, MEMC_BASE_PHYSICAL, 64
> DCD 0x00000000, 0x00000000, 0
>
> defined as:
> FLASH_BASE_PHYSICAL EQU 0x00000000
> FLASH_BASE_C_VIRTUAL EQU 0x80000000
> FLASH_BASE_U_VIRTUAL EQU (FLASH_BASE_C_VIRTUAL +
> CACHED_TO_UNCACHED_OFFSET)
>
> PSRAM_BASE_PHYSICAL EQU 0x04000000
> PSRAM_BASE_C_VIRTUAL EQU 0x84000000
> PSRAM_BASE_U_VIRTUAL EQU (PSRAM_BASE_C_VIRTUAL +
> CACHED_TO_UNCACHED_OFFSET)
>
> CPLD_BASE_PHYSICAL EQU 0x08000000
> CPLD_BASE_C_VIRTUAL EQU 0x88000000
> CPLD_BASE_U_VIRTUAL EQU (CPLD_BASE_C_VIRTUAL +
> CACHED_TO_UNCACHED_OFFSET)
>
> CACHED_TO_UNCACHED_OFFSET EQU 0x30000000
>
> PERIF_BASE_PHYSICAL EQU 0x40000000
> PERIF_BASE_C_VIRTUAL EQU 0x98500000
> PERIF_BASE_U_VIRTUAL EQU (PERIF_BASE_C_VIRTUAL +
> CACHED_TO_UNCACHED_OFFSET)
>
> LCD_BASE_PHYSICAL EQU 0x44000000
> LCD_BASE_C_VIRTUAL EQU 0x9C500000
> LCD_BASE_U_VIRTUAL EQU (LCD_BASE_C_VIRTUAL +
> CACHED_TO_UNCACHED_OFFSET)
>
> MEMC_BASE_PHYSICAL EQU 0x48000000
> MEMC_BASE_C_VIRTUAL EQU 0xA0500000
> MEMC_BASE_U_VIRTUAL EQU (MEMC_BASE_C_VIRTUAL +
> CACHED_TO_UNCACHED_OFFSET)
>
>
>
>
> config.bib CONFIG section:
> CONFIG
> COMPRESSION=ON
> PROFILE=ON
> KERNELFIXUPS=ON
> ROMFLAGS=0
> AUTOSIZE=ON
> ROMSTART=80000000
> ROMWIDTH=32
> ROMSIZE=02000000
> ROMOFFSET=80000000
>
> boot.bib CONFIG section:
> COMPRESSION= ON
> PROFILE= OFF
> KERNELFIXUPS= ON
> SRE= ON
> ROMSTART= 00000000
> ROMWIDTH= 32
> ROMSIZE= 00080000
>
>
> the ROMsize in boot.bib was reduced to decrease the size of eboot.bin
> that I program using an UniSTAC II ICE.
>
> Regards,
> Inger
>
>
> "K. S. Huang" <ks_huang@alphanetworks.com.remove.this> wrote in message
news:<eQh6IWTWEHA.644@tk2msftngp13.phx.gbl>...
> > Would you please also post the config section of boot.bib and config.bib
> > also the OEMAddressTable
> > so that it migh help to clear your problem.
> >
> > "Inger" <inger72@online.no>
> > ???????:41a44da6.0406230710.762bbff1@posting.google.com...
> > > Hi
> > > I am attempting to download a 23M RAM debug image to my PXA255 based
> > > board with 32M RAM onboard over ethernet. After about 16.1M, it fails
> > > repeatedly at same location each time. I use routines used in the
> > > XScale BSP EbootInitEtherTransport, and EbootEtherReadData, but my
> > > bootloader code is otherwise modified from XSC1BD.
> > >
> > > My download fails after 16.1M and with error message from the PB the
> > > Download failed, probably due to my device stopping to repond. I am
> > > connected with a HW debugger and when hitting break after download
> > > failed, I see that it is somewhere in the ne2000 routines, but the
> > > stack pointer and cpsr looks OK, so the system looks like it has
> > > entered an infinite loop or something.
> > >
> > > When trying to download other images, they pass if they are smaller
> > > release images, but the big debug images all fail, each at its own
> > > dedicated location.
> > >
> > > I know since I am using a "personal" bootloader, it may be a longshot,
> > > but have anybody seen similar issues?
> > >
> > > My boot.bib and config.bib MEMORY sections looks as follows when
> > > downloading nk to RAM , maybe the error lies in here. Note that the
> > > bootloader is using physical space, while the image is using virtual
> > > adressing mode.
> > >
> > >
> > > boot.bib
> > > MEMORY
> > > EBOOT 00001000 0007E000 RAMIMAGE
> > > RAM 04204000 00110000 RAM
> > >
> > > TRANSTABL 04000000 00004000 RESERVED
> > > STACK 04004000 001FC000 RESERVED
> > > IRQSTACK 04200000 00004000 RESERVED
> > >
> > >
> > > config.bib
> > > ; Name|Virtual/Cached |Size| Region Type
> > > MEMORY
> > > NK 84314000 0173C000 RAMIMAGE
> > > RAM 85A50000 005B0000 RAM
> > > EBOOT 80000000 00080000 RESERVED
> > >
> > > Regards,
> > > Inger



Re: Ethernet download of large nk.bin fails at same location by inger72

inger72
Fri Jun 25 07:47:21 CDT 2004

... now this should be renamed to "Ethernet download of large nk.bin
fails after 10 minutes, regardless."

I was doing some experiments and found that my downloads always seem
to halt after 10 minutes, when PB4.2 gives the error message "Download
failed. Please check your download configuration..." etc. By inserting
larger and larger delays I could see that the number of bytes
downloaded did not matter, but rather the download time was the common
factor.

1) Is there a timeout in PB4.2 that I need to set higher for my images

2) I observe only about 30Kb/s download speed in the Download window
from PB, which seems low. What is the common speed for ethernet
download from PB given in the "Download OS Image" window?

3) I also see the messages I have seen that others are seeing: "TFTP:
Desktop losing ACK, block number = 3014, Ack again." Does this matter
for the downloaded data, will the downloaded image still be correct?
Is this what is slowing me down, or the effect of a slow response
either from my device or from the network?


Regards,
Inger



"K. S. Huang" <ks_huang@AlphaNetworks.com_remove.this> wrote in message news:<#pxe#8bWEHA.1152@TK2MSFTNGP09.phx.gbl>...
> Maybe you can write some simple SDRAM testing code (Write something and Read
> it back) in your EBoot to make sure the RAM between 0x04314_000 and
> 0x05FF_FFFF is all accessable.
>
> if failed you may beed to check out the SDRAM is configred properly!
>
> "Inger" <inger72@online.no>
> ???????:41a44da6.0406232250.6da83d18@posting.google.com...
> > Will do:
> > memmap.inc
> > OEMAddressTable
> > DCD FLASH_BASE_C_VIRTUAL, FLASH_BASE_PHYSICAL, 32
> > DCD PSRAM_BASE_C_VIRTUAL, PSRAM_BASE_PHYSICAL, 32
> > DCD CPLD_BASE_C_VIRTUAL, CPLD_BASE_PHYSICAL, 1
> > DCD COMPFLASH_IO_BASE_C_VIRTUAL, COMPFLASH_IO_BASE_PHYSICAL,
> > 256
> > DCD PERIF_BASE_C_VIRTUAL, PERIF_BASE_PHYSICAL, 64
> > DCD LCD_BASE_C_VIRTUAL, LCD_BASE_PHYSICAL, 64
> > DCD MEMC_BASE_C_VIRTUAL, MEMC_BASE_PHYSICAL, 64
> > DCD 0x00000000, 0x00000000, 0
> >
> > defined as:
> > FLASH_BASE_PHYSICAL EQU 0x00000000
> > FLASH_BASE_C_VIRTUAL EQU 0x80000000
> > FLASH_BASE_U_VIRTUAL EQU (FLASH_BASE_C_VIRTUAL +
> > CACHED_TO_UNCACHED_OFFSET)
> >
> > PSRAM_BASE_PHYSICAL EQU 0x04000000
> > PSRAM_BASE_C_VIRTUAL EQU 0x84000000
> > PSRAM_BASE_U_VIRTUAL EQU (PSRAM_BASE_C_VIRTUAL +
> > CACHED_TO_UNCACHED_OFFSET)
> >
> > CPLD_BASE_PHYSICAL EQU 0x08000000
> > CPLD_BASE_C_VIRTUAL EQU 0x88000000
> > CPLD_BASE_U_VIRTUAL EQU (CPLD_BASE_C_VIRTUAL +
> > CACHED_TO_UNCACHED_OFFSET)
> >
> > CACHED_TO_UNCACHED_OFFSET EQU 0x30000000
> >
> > PERIF_BASE_PHYSICAL EQU 0x40000000
> > PERIF_BASE_C_VIRTUAL EQU 0x98500000
> > PERIF_BASE_U_VIRTUAL EQU (PERIF_BASE_C_VIRTUAL +
> > CACHED_TO_UNCACHED_OFFSET)
> >
> > LCD_BASE_PHYSICAL EQU 0x44000000
> > LCD_BASE_C_VIRTUAL EQU 0x9C500000
> > LCD_BASE_U_VIRTUAL EQU (LCD_BASE_C_VIRTUAL +
> > CACHED_TO_UNCACHED_OFFSET)
> >
> > MEMC_BASE_PHYSICAL EQU 0x48000000
> > MEMC_BASE_C_VIRTUAL EQU 0xA0500000
> > MEMC_BASE_U_VIRTUAL EQU (MEMC_BASE_C_VIRTUAL +
> > CACHED_TO_UNCACHED_OFFSET)
> >
> >
> >
> >
> > config.bib CONFIG section:
> > CONFIG
> > COMPRESSION=ON
> > PROFILE=ON
> > KERNELFIXUPS=ON
> > ROMFLAGS=0
> > AUTOSIZE=ON
> > ROMSTART=80000000
> > ROMWIDTH=32
> > ROMSIZE=02000000
> > ROMOFFSET=80000000
> >
> > boot.bib CONFIG section:
> > COMPRESSION= ON
> > PROFILE= OFF
> > KERNELFIXUPS= ON
> > SRE= ON
> > ROMSTART= 00000000
> > ROMWIDTH= 32
> > ROMSIZE= 00080000
> >
> >
> > the ROMsize in boot.bib was reduced to decrease the size of eboot.bin
> > that I program using an UniSTAC II ICE.
> >
> > Regards,
> > Inger
> >
> >
> > "K. S. Huang" <ks_huang@alphanetworks.com.remove.this> wrote in message
> news:<eQh6IWTWEHA.644@tk2msftngp13.phx.gbl>...
> > > Would you please also post the config section of boot.bib and config.bib
> > > also the OEMAddressTable
> > > so that it migh help to clear your problem.
> > >
> > > "Inger" <inger72@online.no>
> > > ???????:41a44da6.0406230710.762bbff1@posting.google.com...
> > > > Hi
> > > > I am attempting to download a 23M RAM debug image to my PXA255 based
> > > > board with 32M RAM onboard over ethernet. After about 16.1M, it fails
> > > > repeatedly at same location each time. I use routines used in the
> > > > XScale BSP EbootInitEtherTransport, and EbootEtherReadData, but my
> > > > bootloader code is otherwise modified from XSC1BD.
> > > >
> > > > My download fails after 16.1M and with error message from the PB the
> > > > Download failed, probably due to my device stopping to repond. I am
> > > > connected with a HW debugger and when hitting break after download
> > > > failed, I see that it is somewhere in the ne2000 routines, but the
> > > > stack pointer and cpsr looks OK, so the system looks like it has
> > > > entered an infinite loop or something.
> > > >
> > > > When trying to download other images, they pass if they are smaller
> > > > release images, but the big debug images all fail, each at its own
> > > > dedicated location.
> > > >
> > > > I know since I am using a "personal" bootloader, it may be a longshot,
> > > > but have anybody seen similar issues?
> > > >
> > > > My boot.bib and config.bib MEMORY sections looks as follows when
> > > > downloading nk to RAM , maybe the error lies in here. Note that the
> > > > bootloader is using physical space, while the image is using virtual
> > > > adressing mode.
> > > >
> > > >
> > > > boot.bib
> > > > MEMORY
> > > > EBOOT 00001000 0007E000 RAMIMAGE
> > > > RAM 04204000 00110000 RAM
> > > >
> > > > TRANSTABL 04000000 00004000 RESERVED
> > > > STACK 04004000 001FC000 RESERVED
> > > > IRQSTACK 04200000 00004000 RESERVED
> > > >
> > > >
> > > > config.bib
> > > > ; Name|Virtual/Cached |Size| Region Type
> > > > MEMORY
> > > > NK 84314000 0173C000 RAMIMAGE
> > > > RAM 85A50000 005B0000 RAM
> > > > EBOOT 80000000 00080000 RESERVED
> > > >
> > > > Regards,
> > > > Inger

Re: Ethernet download of large nk.bin fails at same location by K

K
Fri Jun 25 09:42:03 CDT 2004

It seems that your download speed is rarely slow!
In general scenario, the download rate of a 10Mb NIC will be about
300-500KB, and for 100Mb NIC it will be about 1 or 2MB.
I think there might be some packet lost in your ethdbg driver...

"Inger" <inger72@online.no>
???????:41a44da6.0406250447.745d3cc6@posting.google.com...
> ... now this should be renamed to "Ethernet download of large nk.bin
> fails after 10 minutes, regardless."
>
> I was doing some experiments and found that my downloads always seem
> to halt after 10 minutes, when PB4.2 gives the error message "Download
> failed. Please check your download configuration..." etc. By inserting
> larger and larger delays I could see that the number of bytes
> downloaded did not matter, but rather the download time was the common
> factor.
>
> 1) Is there a timeout in PB4.2 that I need to set higher for my images
>
> 2) I observe only about 30Kb/s download speed in the Download window
> from PB, which seems low. What is the common speed for ethernet
> download from PB given in the "Download OS Image" window?
>
> 3) I also see the messages I have seen that others are seeing: "TFTP:
> Desktop losing ACK, block number = 3014, Ack again." Does this matter
> for the downloaded data, will the downloaded image still be correct?
> Is this what is slowing me down, or the effect of a slow response
> either from my device or from the network?
>
>
> Regards,
> Inger
>
>
>