Hello,

I translated the following struct

struct iphdr {
#if defined(__LITTLE_ENDIAN_BITFIELD)
	__u8	ihl:4,
		version:4;
#elif defined (__BIG_ENDIAN_BITFIELD)
	__u8	version:4,
  		ihl:4;
#else
#error	"Please fix <asm/byteorder.h>"
#endif
	__u8	tos;
	__be16	tot_len;
	__be16	id;
	__be16	frag_off;
	__u8	ttl;
	__u8	protocol;
	__sum16	check;
	__be32	saddr;
	__be32	daddr;
	/*The options start here. */
};

to the following Pascal record:

iphdr    = record
{$IFDEF ENDIAN_LITTLE}
              jhl     : Cardinal; // __u8	ihl:4,
              version : Cardinal; // version:4
{$ELSE}
              version : Cardinal; // __u8 version:4,
              jhl     : Cardinal; // version:4
{$ENDIF}
              tos      : cuint8;
              tot_len  : cint16;
              id       : cint16;
              frag_off : cint16;
              ttl      : cint8;
              protocol : cint8;
              check    : cuint16;
              saddr    : cint32;
              daddr    : cint32;
  end;

The problem is, that I get weird data in the fields, so I think I did not translate it properly, but can't figure out what have I missed.

Thanks,
Ido

Answering myself:

                           cuchar; // instead of Cardinal

The ":4" in C means "4 bits", not 4 bytes. The correct translation in the general case for bitfields is hard, because FPC does not support C-compatible bitpacking (it depends on the type of the bitfield, e.g. "int a:4;" and "char a:4" have different alignment rules). In this case, this should match the C layout:

{$packrecords c}
iphdr    = record
             bitfields = bitpacked record
{$IFDEF ENDIAN_LITTLE}
               jhl: 0..(1 shl 4)-1;
               version: 0..(1 shl 4)-1;
{$ELSE}
               jhl: 0..(1 shl 4)-1;
               version: 0..(1 shl 4)-1;
{$ENDIF}
             end;
             tos      : cuint8;
             tot_len  : cint16;
             id       : cint16;
             frag_off : cint16;
             ttl      : cint8;
             protocol : cint8;
             check    : cuint16;
             saddr    : cint32;
             daddr    : cint32;
 end;

You will then of course have to use iphdrvar.bitfields.jhl instead of iphdrvar.jhl.

Jonas

I'ved replaced it to

iphdr    = bitpacked record
{$IFDEF ENDIAN_LITTLE}
              jhl     : 0..(1 shl 4)-1; // __u8	ihl:4,
              version : 0..(1 shl 4)-1; // version:4
{$ELSE}
              version : 0..(1 shl 4)-1; // __u8 version:4,
              jhl     : 0..(1 shl 4)-1; // version:4
{$ENDIF}
              tos      : cuint8;
              tot_len  : cint16;
              id       : cint16;
              frag_off : cint16;
              ttl      : cint8;
              protocol : cint8;
              check    : cuint16;
              saddr    : cint32;
              daddr    : cint32;
  end;

And now all of the problems that I continue to find gone away.

Thank you

ik wrote

I'ved replaced it to iphdr = bitpacked record

This is wrong. The record is a C record, and hence follows {$packrecord c} rules. "bitpacked record" overrides any current {$packrecords x} setting. That is the reason why I used the nested bitpacked record.

And now all of the problems that I continue to find gone away.

You may get lucky, but it's still wrong.

Jonas

I understand, so I'll change it, to your way