Skip to contentSkip to conversation

IDE interface on PPC has endian issue

U-Boot-Users Gregg 2 participants 3 messages Nov 15, 2005

lore
Share & tools
lore

Gregg

author
#0 lore raw

I have a board that has a 128 MB compact flash connected to a PPC 440GP peripheral bus using True IDE mode. Up until now, I have been successfully booting using NFS, and now I am trying to get the IDE interface to work.

When I issue an IDE reset command to u-boot, the information returned is byte-swapped:

Device 0: Model: aSDnsi KDSFC-J21 8 Firm: DH X.272 Ser#: <...>
Type: Removeable Hard Disk
Capacity: 24576.1 MB = 24.0 GB (50331860 x 512)

Note that "aSDnsi K" is "SanDisk ". If I modify cmd_ide.c to pretend I am little endian so that the input_swap_data function is disabled (#if defined (__LITTLE_ENDIAN) version is used), I get seemingly correct results:

Device 0: Model: SanDisk SDCFJ-128 Firm: HDX 2.27 Ser#: <...>
Type: Removeable Hard Disk
Capacity: 122.5 MB = 0.1 GB (250880 x 512)

I am confused why I need to do this, as it seems wrong. Furthermore, although I can get the IDE information, I cannot see any partitions on the device. I created a single linux partion (type 83), but when I issue "ide part 0", I am told "IDE device 0 not available".

The CF is connected to the PPC like this:

A0,A1,A2 on the CF ==> A30, A29, A28 on 440GP
D0..D15 on CF ==> D15..D0 on 440GP

I am using the following ATA configuration:

#define CFG_ATA_BASE_ADDR    0xa0300000
#define CFG_IDE_MAXDEVICE    1
#define CFG_IDE_MAXBUS       1
#define CFG_ATA_IDE0_OFFFSET 0
#define CFG_ATA_DATA_OFFSET  0
#define CFG_ATA_REG_OFFSET   1
#define CFG_ATA_ALT_OFFSET   0x00100001
#define CFG_ATA_STRIDE       2
#define ATA_RESET_TIME       5

Thanks.

Gregg

#1 lore raw

I had the same problem. It turns the idiot hardware engineer decided he would do byte swapping in hardware without telling me. To get around it, I just made the byte swapping code a noop (no byte swaping) and it worked.

Quoted 58 lines · --- Gregg

I have a board that has a 128 MB compact flash connected to a PPC 440GP
peripheral bus using True IDE mode. Up until now, I have been successfully
booting using NFS, and now I am trying to get the IDE interface to work.

When I issue an IDE reset command to u-boot, the information returned is
byte-swapped:

Device 0: Model: aSDnsi KDSFC-J21 8 Firm: DH X.272 Ser#: <...>
Type: Removeable Hard Disk
Capacity: 24576.1 MB = 24.0 GB (50331860 x 512)

Note that "aSDnsi K" is "SanDisk ". If I modify cmd_ide.c to pretend I am little
endian so that the input_swap_data function is disabled (#if defined
(__LITTLE_ENDIAN) version is used), I get seemingly correct results:

Device 0: Model: SanDisk SDCFJ-128 Firm: HDX 2.27 Ser#: <...>
Type: Removeable Hard Disk
Capacity: 122.5 MB = 0.1 GB (250880 x 512)

I am confused why I need to do this, as it seems wrong. Furthermore, although I
can get the IDE information, I cannot see any partitions on the device. I
created a single linux partion (type 83), but when I issue "ide part 0", I am
told "IDE device 0 not available".

The CF is connected to the PPC like this:

A0,A1,A2 on the CF ==> A30, A29, A28 on 440GP
D0..D15 on CF ==> D15..D0 on 440GP

I am using the following ATA configuration:

#define CFG_ATA_BASE_ADDR    0xa0300000
#define CFG_IDE_MAXDEVICE    1
#define CFG_IDE_MAXBUS       1
#define CFG_ATA_IDE0_OFFFSET 0
#define CFG_ATA_DATA_OFFSET  0
#define CFG_ATA_REG_OFFSET   1
#define CFG_ATA_ALT_OFFSET   0x00100001
#define CFG_ATA_STRIDE       2
#define ATA_RESET_TIME       5

Thanks.

Gregg

Quoted 12 lines

-------------------------------------------------------
This SF.Net email is sponsored by the JBoss Inc. Get Certified Today
Register for a JBoss Training Course. Free Certification Exam for All Training Attendees Through End of 2005. For more info visit:
http://ads.osdn.com/?ad_id=7628&alloc_id=16845&op=click
_______________________________________________
U-Boot-Users mailing list
U-Boot-Users at lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/u-boot-users

__________________________________
Start your day with Yahoo! - Make it your home page!
http://www.yahoo.com/r/hs

author
#2 lore raw
Quoted 5 lines · Frank <frannk_m1 <at> yahoo.com> writes:

I had the same problem. It turns the idiot hardware engineer decided he would do byte swapping in hardware without telling me. To get around it, I just made the byte swapping code a noop (no byte swaping) and it worked.

Doing an endian swap every place in cmd_ide.c where it is reading the data register seems to have fixed the problem. I can now see the partitions on the disk and load a file from it. I don't understand why the endianness is different from what the u-boot code expects though, given how the data bits are connected to the peripheral bus:

CF D0..D15 => PPC D15..D0

D15 is lsb on PPC 16-bit mode bus. It is as if the u-boot code expects

CF D0..D7 => PPC D7..D0
CF D8..D15 => PPC D15..D8

Thanks,

Gregg