I'm attempting to mimic the function used for creating CRC's in PNG files, I'm using the autodin II polynomial and the source code from:
http://www.opensource.apple.com/source/xnu/xnu-1456.1.26/bsd/libkern/crc32.c
My tests have all been for the IHDR chunk, so my parameters have been:
crc - 0xffffffff and 0 (both have been suggested)
buff - the address of the IHDR Chunk's type.
length - the IHDR Chunk's length + 4 (the length of the chunk's data + the length of the type)
I printed the calculated CRC in binary, which I compared to the actual CRC of the chunk. I can see no similarities (little-big endian, reversed bits, XOR'd, etc).
This is the data for the IHDR chunk (hexadecimal format):
length(big endian): d0 00 00 00 (13)
type: 49 48 44 52
data: 00 00 01 77 00 00 01 68 08 06 00 00 00
existing CRC: b0 bb 40 ac
If anyone can tell me why my calculations are off, or give me a CRC32 function that will work I would greatly appreciate it.
Thank-you!
The CRC-32 algorithm used in PNG images is described here: http://www.w3.org/TR/PNG-Structure.html#CRC-algorithm (there's also a link to C code for doing test calculations).
But as #Jigsore pointed out, you won't get sensible results from the data you posted here. You've given us a 4-byte type identifier and what looks like 7.5 bytes of data to follow it. There should be a total of 13 bytes according to the length header.
EDIT:
This works using the function from w3.org:
int main() {
char input[] = { 0x49,0x48,0x44,0x52,0x00,0x00,0x01,0x77,0x00,
0x00,0x01,0x68,0x08,0x06,0x00,0x00,0x00 };
printf("%08lx\n",crc(input,17));
return 0;
}
Output:
ac40bbb0
Related
I am trying to decode a 9-byte RS-485 packet that has a 3-byte CRC as shown below. Using a SW serial monitor I have decoded the parameter portion Using an online calculator.
I know can calculate Byte7 using 8-bit Checksum8 Modulo 256. My issue is
I cannot determine the last 2 Bytes of CRC.
I have attached my output from the reveng CRC decoder and it
finds a Model but not a name. I supplied 4 full 9 Byte packets.
reveng -w8 -sF 9BYTEPACKET1 9BYTEPACKET2 9BYTEPACKET3 9BYTEPACKET4
And as shown in the image link it returns:
width=8,poly=0x01,init=0x00,refin=false,refout=false,xorout=0x00,check=0x31,residue=0x00,name="(none)"
Sample Data:
00 32 2a ff ff cd 27 03 f1
00 20 03 ff ff cd ee 02 02
00 13 28 ff ff cd 06 03 f3
00 13 02 ff ff cd e0 02 3e
Byte1 = ID(Can Change)
Byte2 = Parameter(Can Change)
Byte3 = Parameter(Can Change)
Byte4 = 0xFF (Never changes)
Byte5 = 0xFF (Never changes)
Byte6 = 0xCD (Never changes)
Byte7 = Checksum8 Modulo 256 (I can calculate this from Byte1 to Byte6)
Byte8 = CRC (Changes with diff Parameter data)
Byte9 = CRC (Changes with diff Parameter data)
Any help would be appreciated.
Your Byte8 and Byte7 are simply the sum of the preceding bytes in little-endian order. Byte9 is then the exclusive-or of the preceding bytes.
There is no CRC.
Following hexdump shows some data made by device i have on my hands. It stores year, month, day, hour, minute, seconds, and lenght in weird way for me (4 bytes marks for single digit in reverse order).
de 07 00 00 01 00 00 00 16 00 00 00 10 00 00 00
24 00 00 00 1d 00 00 00 15 00 00 00 X X X X
For example:
Year is marked as "000007de" aka 0x07de (=2014). Now; problem i am having is how to properly handle this in c/c++. (first 4 bytes)
How do i read those 4 bytes with "reverse" order to make proper hexadecimal for handling afterwards with like ints/longs?
If you read the value as int on the same architecture it has been generated with then you don't need to do anything, as this is the natural format for your system.
You only need to do something about this if you want to read it on a different architecture, with a different binary format.
So you can read it simply with
int32_t n;
fread(&n, sizeof int32_t, 1, FILE);
Of course the file has to be opened in binary mode and you need a 32 bit int.
If you read it in the reverse order, you can then change the endianness with something like:
uint32_t before = 0xde070000;
uint32_t after = ((before<<24) & 0xff000000) |
((before<<8) & 0xff0000) |
((before>>8) & 0xff00) |
((before>>24) & 0xff);
Edit: as pointed out in comments, this is only defined for unsigned 32-bits conversions.
I am trying to read the file 'train-images-idx3-ubyte', which can be found here along with the corresponding file format description (at the bottom of the webpage). When I look at the bytes with od -t x1 train-images-idx3-ubyte | less (hexadecimal, bytewise), I get the following output:
adress bytes
0000000 00 00 08 03 00 00 ea 60 00 00 00 1c 00 00 00 1c
0000020 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
...
This is what I expected according to 1. But when I try to read the data with C++ I've got a problem. What I do is this:
std::fstream trainingData("minst/train-images-idx3-ubyte",
std::ios::in | std::ios::binary);
int8_t zero = 0, encoding = 0, dimension = 0;
int32_t samples = -1;
trainingData >> zero >> zero >> encoding >> dimension;
trainingData >> samples;
debugLogger << "training set image file, encoding = "
<< (int) encoding << ", dimension = "
<< (int) dimension << ", items = " << (int) samples << "\n";
But the output of these few lines of code is:
training set image file, encoding = 8, dimension = 3, items = 0
Everything but the number of instances (items, samples) is correct. I tried reading the next 4 bytes as int8_t and that gave me at least the same result as od. I cannot imagine how samples can be 0. What I actually wanted to read here was 10,000. Maybe you've got a clue?
As mentioned in other answers, you need to use unformatted input, i.e. istream::read(...) instead of operator>>. Translating your code above to use read yields:
trainingData.read(reinterpret_cast<char*>(&zero), sizeof(zero));
trainingData.read(reinterpret_cast<char*>(&zero), sizeof(zero));
trainingData.read(reinterpret_cast<char*>(&encoding), sizeof(encoding));
trainingData.read(reinterpret_cast<char*>(&dimension), sizeof(dimension));
trainingData.read(reinterpret_cast<char*>(&samples), sizeof(samples));
Which gets you most of the way there - but 00 00 ea 60 looks like it's in Big-endian format, so you'll have to pass it through ntohl to make sense of it if you're running on an intel-based machine:
samples = ntohl(samples);
which gives encoding = 8, dimension = 3, items = 60000.
The input is formatted, which will result in you reading wrong results from the file. Reading from an unformatted input will provide the correct results.
In the registry there is one ( or more ) key depending how many monitors you have HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Enum\DISPLAY\DEL404C\{Some Unique ID}\Device Parameters\EDID which is a REG_BINARY key. In my case this is :
00 FF FF FF FF FF FF 00 10 AC 4C 40 53 43 34 42 34 14 01 03 0A 2F 1E 78 EE EE 95 A3 54
4C 99 26 0F 50 54 A5 4B 00 71 4F 81 80 B3 00 01 01 01 01 01 01 01 01 01 01 21 39 90 30
62 1A 27 40 68 B0 36 00 DA 28 11 00 00 1C 00 00 00 FF 00 34 57 31 4D 44 30 43 53 42 34
43 53 0A 00 00 00 FC 00 44 45 4C 4C 20 50 32 32 31 30 0A 20 20 00 00 00 FD 00 38 4B 1E
53 10 00 0A 20 20 20 20 20 20 00 FA
This reg_binary value contains information (such as Serial Number and Type) about the connected monitor. I only need these two values. My question is how can i read these values using C or C++?
I have a VB script which can do this:
'you can tell If the location contains a serial number If it starts with &H00 00 00 ff
strSerFind=Chr(&H00) & Chr(&H00) & Chr(&H00) & Chr(&HfF)
'or a model description If it starts with &H00 00 00 fc
strMdlFind=Chr(&H00) & Chr(&H00) & Chr(&H00) & Chr(&Hfc)
This link also contains information about EDID: http://en.wikipedia.org/wiki/Extended_display_identification_data
Could someone help me, how can i do this in C? I can find only VB script examples, but unfortunately i don't understand them, and also it would be very important for me.
You mention wanting the "serial number" and "type". There is no "type" but there is a manufacturer ID and a product ID. For the most part these aren't stored as meaningful strings in the information you get back...they are just numeric values. And they're all in the first 16 bytes.
I'll decode the beginning according to the spec you cite.
Bytes 0,1,2,3,4,5,6,7 - Header information
This should be the literal string "00h FFh FFh FFh FFh FFh FFh 00h", which serves as a sanity check that we're looking at a valid EDID block. Your data starts off with exactly what we expect:
00 FF FF FF FF FF FF 00
Bytes 8 and 9 - Manufacturer ID.
These IDs are assigned by Microsoft, and are three-letter codes. Oh sure, they could have "wasted" three whole bytes in ASCII for this. But that would have been too sensible. So they frittered away eight bytes on an extremely "non-magic" number for the header, and invented an "ingenious" way to encode those three letters into the sixteen bits held by two bytes. How'd they pull it off?
+--------+--------+
| Byte 8 | Byte 9 |
--------+--------+--------+
Bit # 76543210 76543210
-----------------=---------
Meaning 0αααααββ βββγγγγγ
So the highest-order bit of Byte 8 is always zero, and the remaining 15 bits are divided into three groups of 5 bits (which I've called α, β, and γ). Each is interpreted as a letter, where "00001=A"; "00010=B"; ... "11010=Z".
You've got:
10 AC
And hexadecimal 10AC expressed as 16 binary bits is 0001000010101100. So let's bring that table back again:
+--------+--------+
| Byte 8 | Byte 9 |
--------+--------+--------+
Bit # 76543210 76543210
-----------------=---------
Meaning 0αααααββ βββγγγγγ
-----------------=---------
Yours 00010000 10101100
So α = 00100 (decimal 4), β = 00101 (decimal 5), γ = 01100 (decimal 12). Using those decimal numbers as indexes into the English alphabet we get D-E-L. By this arcane sorcery we have determined that your monitor is most likely made by Dell. :)
Bytes 10 and 11 - Product ID Code
This is a two-byte number, assigned by the manufacturer, stored as "LSB first". This is to say that the first byte is the least significant place value. You have:
4C 40
Which we need to interpret as the hexadecimal number 404C.
Bytes 12,13,14,15 - Serial Number.
This is a 32-bit value assigned by the manufacturer which has no requirement for the format. It is "usually stored as LSB first", but doesn't have to be.
53 43 34 42
You can interpret that as 0x53433442, or 0x42344353, or whatever...so long as you're consistent in comparing one value against another.
So now you see it's just three letters and some numbers. Once you get the bytes into a buffer there are a lot of ways to extract the information. #freerider provided some information on that, I'll just throw in a bit more.
The EDID standard says that what you get back as a description is 128 bytes. That is the case with the registry key here, and you can probably assume that if there are not exactly 128 bytes it is corrupt. So using the code provided by #freerider, there'd be no need to pass in anything larger than that...you could technically go down to just 16 if that's the only part of the EDID you're interested in:
#define EDID_BUFFER_SIZE 128
// in idiomatic C++ it's better to say:
// const size_t edidBufferSize = 128;
BYTE edidBuffer[EDID_BUFFER_SIZE];
DWORD nLength = GetLocalMachineProfileBuffer( Buffer, EDID_BUFFER_SIZE );
if (nLength != EDID_BUFFER_SIZE) {
// handle error case, not a valid EDID block
} else {
// valid EDID block, do extraction:
// * manufacturer ID
// * product ID
// * serial number
}
(Note: I prefer to avoid using the sizeof on arrays like #freerider's sizeof( Buffer ) above. While it will technically work in this case, it doesn't return the number of elements in the array...rather the number of bytes the array occupies in memory. In this case the elements happen to actually be bytes, so it will work...but you quickly run into problems, like when you pass an array to another function by pointer and suddenly it starts reporting its size as the size of a pointer...)
Beyond that, your question of how to extract structural data out of a buffer of bytes is a very general one, and is so foundational to C-style programming that if you don't know where to start on it then you should probably work through simpler programs. Getting the three five bit segments out of the manufacturer name involves things like bit shifting, bit masking, or bit fields. Going through the array deals with addresses and how to index arrays and things like that.
The closest parallel question I could find offhand right now is this:
extract IP from a buffer of bytes
Lots of ways to do it, but an interesting one is that you can define the layout of a structure in memory and then tell the program "hey, this block of memory I found is laid out just like the structure I defined. So let me extract information from it as simply as if I'd defined the object in my program"...
But then you have to be sensitive to issues like data structure alignment. That's because the way your compiler will naturally put objects into memory doesn't necessarily match what you think it would do:
http://en.wikipedia.org/wiki/Data_structure_alignment
With the information above you should at least be able to make a shot at reading some tutorials and seeing what works. If you can't figure out one part of the problem then break that little part out as its own question, and show what you tried and why it didn't work...
This previous question explains how to get EDID with C/C++/C#. It's not through the registry, but as long it works...
Win32 code to get EDID in Windows XP/7
If you want to still read the registry, use RegQueryValueEx and friends.
DWORD GetLocalMachineProfileBuffer(BYTE* pBuffer, DWORD nMaxLength )
{
CString szSubKey = "HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Enum\DISPLAY\DEL404C{Some Unique ID}\Device Parameters\EDID";
DWORD rc;
DWORD dwType;
HKEY hOpenedKey;
if( ERROR_SUCCESS == RegOpenKeyEx (
HKEY_LOCAL_MACHINE, // handle of open key
szSubKey, // address of name of subkey to open
0, // reserved
KEY_READ, // security access mask
&hOpenedKey // address of handle of open key
) )
{
rc = RegQueryValueEx(
hOpenedKey,
(const char*)szValueName,
0,
&dwType,
(LPBYTE)pBuffer,
&nMaxLength );
if( rc != ERROR_SUCCESS )
{
return (DWORD)-1;
}
else
{
ASSERT( dwType == REG_BINARY );
}
RegCloseKey( hOpenedKey );
return nMaxLength;
}
else
{
return (DWORD)-1;
}
}
call it like this:
BYTE Buffer[20000];
DWORD nLength = GetLocalMachineProfileBuffer( Buffer, sizeof( Buffer ) );
I'm trying to convert 8 bit char into hex view which looks like this:
00 03 80 45 E5 93 00 18 02 72 3B 90 88 64 11 00
45 FF 00 36 00 FF 45 00 00 34 7B FE 40 00 40 02
But some characters contain negative values which makes a larger hex value of more than 2 digits. how would i get each one as represented above?
I don't know what you are using for formatting, but make sure that you make your byte holding variable an unsigned char (assuming that char is 8-bits on your platform, which it is on all sane platforms), before formatting. If your platform has a sane BYTE typedef, use that. You can also use the boost::uint8_t type to store the byte and avoid these sorts of issues. For example:
char c=-25; // Oh no, this is one of those pesky "negative" characters
unsigned char byteVal=static_cast<unsigned char>(c); // FTFY
// Do the formatting with byteVal
"negative byte values" is an oxymoron, a byte is a number of bits without any sign typically an unsigned char which, when being 8 bits. can contain values 0-255 or in hex 00 to FF.