/[PAMELA software]/yoda/event/CRC.h
ViewVC logotype

Contents of /yoda/event/CRC.h

Parent Directory Parent Directory | Revision Log Revision Log


Revision 6.0 - (show annotations) (download)
Tue Feb 7 17:11:06 2006 UTC (18 years, 10 months ago) by kusanagi
Branch: MAIN
CVS Tags: YODA6_2/01, YODA6_2/00, YODA6_3/19, YODA6_3/18, YODA6_3/13, YODA6_3/12, YODA6_3/11, YODA6_3/10, YODA6_3/17, YODA6_3/16, YODA6_3/15, YODA6_3/14, YODA6_3/06, YODA6_1/00, YODA6_0/00, YODA6_3/04, YODA6_3/05, YODA6_3/20, YODA6_3/07, YODA6_3/00, YODA6_3/01, YODA6_3/02, YODA6_3/03, YODA6_3/08, YODA6_3/09, HEAD
Changes since 5.1: +4 -4 lines
File MIME type: text/plain
Error occurred while calculating annotation data.
Several new features in this revision:
a) all the packets are conform to the Mass Memory Format specifications (http://people.roma2.infn.it/~cpu/Mass_Memory_Format.html)
b) unpacking either using the old files structure OR the new one file unpacking.
c) parametrized root files compression factor
d) deleting of the following packet: TofTest, TrkTest, TrkEvent.
e) the Tracker routines now work without the use of temp files.

The point a) allow Yoda to unpack in the root file all the packets generated by the CPU. According to the MassMemoryFormat; that is three possible data are available:

1) almost explicit structure of the packet (like for Log, Tracker, Mcmd, etc....);
2) dummy data collection structure (InitHeader, InitTrailer, CalibHeader, CalibTrailer);
3) just the data of the packet (almost all Alarm and Init procedures). The class regarding this packets have only one parameters, a TArrayC class, which contain the data-block included in the packet (tat is the data below the packet Header).

The point b) has been implemented as a consequence of an agreement about a more compact structure of the unpacked data. Up to now the structure of each unpacked data consisted of a folder, named after the packet type, and three files: xxx.Header.root, xxx.NamePacket.root, xxx.Registry.root.
Starting from this release YODA, by default will unpack the data in a unique root file. The structure of this file will consist of:
- several TTree(s) named after the packet type;
- into each TTree are foreseen three TBranche(s):
    - 'Header'  (the old xxx.Header.root file)
    - 'NameOfThePacket' (the old xxx.Event.root file or the xxx.Event.DETECTOR.root)
    - 'Registry' (the old xxx.Registry.root file)

Anyway is still possible, but deprecated, to unpack using the old structure, passing to the "yoda" command the optional parameter "-multifile"

The point c) has been implemented because is well know that writing time in a TTree is as much fast as much lower is the compression factor for the root file; anyway for a PAMELA dat file, a compression equal to 0 will generate a root file which will be more than two times the original size. To modify the compression parameter just add the optional parameter "-c [0-9]" to the yoda command line.

1 /****************************************************************************
2 * F i l e D a t a
3 * $Id: CRC.h,v 5.1 2006/02/04 12:37:42 kusanagi Exp $
4 * $Revision: 5.1 $
5 * $Date: 2006/02/04 12:37:42 $
6 * $RCSfile: CRC.h,v $
7 *
8 ****************************************************************************
9 * S W D e v e l o p m e n t E n v i r o n m e n t
10 *
11 * $Author: kusanagi $
12 * :
13 *****************************************************************************/
14
15 #ifndef CRC_H
16 #define CRC_H
17
18 #define BYTE unsigned char
19 #define UINT32 unsigned int
20 #define UINT16 unsigned short
21
22
23 /** Example of CAST macro at work. public domain demo by Bob Stout.
24 *
25 * Example of CAST macro at work
26 *
27 * union {
28 * char ch[4];
29 * int i[2];
30 * } my_union;
31 *
32 * long longvar;
33 *
34 * longvar = (long)my_union; Illegal cast
35 * longvar = CAST(long, my_union); Legal cast
36 *
37 */
38 #define CM_CAST(new_type,old_object) (*((new_type *)&(old_object)))
39
40
41 #define CM_HI_UINT16(x) ( (BYTE)( ((x) & 0xff00)>>8 ) )
42 #define CM_LO_UINT16(x) ( (BYTE)( ((x) & 0x00ff) ) )
43
44 #define CM_HI_UINT32(x) ( (UINT16)( ((x) & 0xffff0000)>>16 ) )
45 #define CM_LO_UINT32(x) ( (UINT16)( ((x) & 0x0000ffff) ) )
46
47
48 #define CM_HIHI_UINT32(x) ( (BYTE)( ((x) & 0xff000000)>>24 ) )
49 #define CM_LOHI_UINT32(x) ( (BYTE)( ((x) & 0x00ff0000)>>16 ) )
50 #define CM_HILO_UINT32(x) ( (BYTE)( ((x) & 0x0000ff00)>> 8 ) )
51 #define CM_LOLO_UINT32(x) ( (BYTE)( ((x) & 0x000000ff) ) )
52
53
54
55 /**
56 Macro READ_NEXT_BITS_UINT(wordlen,p,offset,n,res)
57
58 this mascro scans bits for a 'wordlen'-bit long word. 'wordlen' can be 8,16,32...
59 It reads the next 'n' bits starting after the 'offset'-th bit of the 'wordlen'-bit word pointed by p.
60 store the result in 'res', and increment properly both 'offset' and 'p' such that they can be reused
61 in the text invocation of the macro.
62
63 \a p unsigned 'wordlen'-bit pointer (ie. wordlen==16, then unsinged short int * in i386 arch)
64 varibale name.
65 it points to the word to scan. it is incremented automatically by the macro.
66 \a offset unsigned 8 bit variable name (unsigned char?). Must be an l-value.
67 it holds the number of bit already read in the most significant part of the
68 current pointer (*p). Should be used always the same variable while a scanning
69 session. This value must be initialized to zero and should be always less then 'wordlen'.
70 \a n the number of bits to read. Can be a r-value. Must be 0 <= n <= 'wordlen'.
71 if n is zero (senseless), both 'p' and 'offset' are unchanged and 0 is returned in 'res'.
72 \a res unsigned 'wordlen'-bit variable name (i.e wordlen==16, then unsigned short int? on i386 arch)
73 in which to store the requested 'n' bits.
74 they are returned in the less significat part of 'res'. If n<'wordlen', then the most
75 significant bits of 'res' are padded to zero.
76
77 before using he macro the first time you have to: (let's assume 'wordlen' fixed to 16 and an i386 arch.)
78 1) choose an unsigned short pointer to pass as 'p', and initialize it to the sequence
79 of bits to scan from.
80 2) choose an unsigned char variable to pass as 'offset' and initialize it to zero
81 (or some other value X<16, if you want to skip the first X bits)
82 3) choose an unsigned short to return the result.
83
84 */
85
86 #define CM_READ_NEXT_BITS_UINT(wordlen,p,offset,n,res) \
87 do { \
88 res = (*p << offset); \
89 res >>= (wordlen-(n)); \
90 if(n<=wordlen-offset) { \
91 offset=(offset+(n))%wordlen; \
92 if(offset==0) \
93 p++; \
94 }else{ \
95 p++; \
96 res |= *p>>(wordlen*2-offset-(n)); \
97 offset+=(n)-wordlen; \
98 } \
99 }while(0)
100
101 /* Wrapper to 8,16,32,64 bit link wrapper to READ_NEXT_BITS_UINT : */
102 #define CM_READ_NEXT_BITS_UINT8(p,offset,n,res) CM_READ_NEXT_BITS_UINT(8 ,p,offset,n,res)
103 #define CM_READ_NEXT_BITS_UINT16(p,offset,n,res) CM_READ_NEXT_BITS_UINT(16,p,offset,n,res)
104 #define CM_READ_NEXT_BITS_UINT32(p,offset,n,res) CM_READ_NEXT_BITS_UINT(32,p,offset,n,res)
105 #define CM_READ_NEXT_BITS_UINT64(p,offset,n,res) CM_READ_NEXT_BITS_UINT(64,p,offset,n,res)
106
107 #define CM_GET_BIT(exp,n) (((exp) >> ((n)-1)) & 0x1)
108
109
110
111
112 /* Thees macros write a 16 or 32 bits in BigEndian fascion in a (unsigned char*) avoiding the
113 addressing alignement problem and also increment ptr
114 - ptr must be a (unsigned char*) pointer (l-value)
115 - byte must be a (unsigned char) (r-value)
116 - word must be a (unsigned short int) (r-value)
117 - dword must be a (unsigned int) (r-value)
118 */
119 #define CM_WRITE_BE_UINT8(ptr,byte) do { *ptr = (unsigned char)(byte); ptr++;} while(0)
120
121 #define CM_WRITE_BE_UINT16(ptr,word) do { \
122 CM_WRITE_BE_UINT8(ptr,((word)>>8)); \
123 CM_WRITE_BE_UINT8(ptr,(word)); \
124 } while(0)
125
126 #define CM_WRITE_BE_UINT32(ptr,dword) do { \
127 CM_WRITE_BE_UINT8(ptr,(dword)>>24); \
128 CM_WRITE_BE_UINT8(ptr,(dword)>>16); \
129 CM_WRITE_BE_UINT8(ptr,(dword)>>8); \
130 CM_WRITE_BE_UINT8(ptr,(dword)); \
131 } while(0)
132
133 /* Thees macros read a 16 or 32 bits in BigEndian fascion from a (unsigned char*) avoiding the
134 addressing alignement problem and also increment ptr
135 - ptr must be a (unsigned char*) pointer (l-value)
136 - byte and temp must be a (unsigned char) (l-value)
137 - word must be a (unsigned short int) (l-value)
138 - dword must be a (unsigned int) (l-value)
139 */
140 #define CM_READ_BE_UINT8(ptr,byte) do { byte = *(ptr); (ptr)++; } while(0)
141
142 #define CM_READ_BE_UINT16(ptr,word,temp) do { \
143 CM_READ_BE_UINT8(ptr,temp); \
144 word = ((unsigned short int)temp) << 8; \
145 CM_READ_BE_UINT8(ptr,temp); \
146 word |= temp; \
147 } while(0)
148
149 #define CM_READ_BE_UINT32(ptr,word,temp) do { \
150 CM_READ_BE_UINT8(ptr,temp); \
151 word = ((unsigned int)temp) << 24; \
152 CM_READ_BE_UINT8(ptr,temp); \
153 word |= ((unsigned int)temp) << 16; \
154 CM_READ_BE_UINT8(ptr,temp); \
155 word |= ((unsigned int)temp) << 8; \
156 CM_READ_BE_UINT8(ptr,temp); \
157 word |= ((unsigned int)temp); \
158 } while(0)
159
160
161 /** Univeral BIT converter to 2 registers set with different resolution.
162 Used in Microsecond to register converte for data time out and event time
163 out and other same-style register.
164
165 \a v is the value in some unit
166 \a unit_h is the resolution of the MSD-part. must be in same unit of v.
167 \a no_bit_h is the number of bits of the MSD-part.
168 \a unit_l is the resolution of the LSD-part. must be in same unit of v.
169 \a no_bit_l is the number of bits of the LSD-part.
170
171 \example (case of ETO)
172 Suppose to have 2 8-bit registers ETO1 (8 bit,LSB) and ETO2 (8 bit,HSB):
173 ETO2 have a resoluion of 65 microsec;
174 ETOHSB have a resolution of 16 millisec;
175
176 This will format a value of 1 millisecond (unit is in microsecond)
177 the usage is: CM_TIME2REG(1000,16*1000,8,65,8)
178
179 */
180
181 //#define CM_TIME2REG(v,unit_h,no_bit_h,unit_l,no_bit_l) ((((v)/(unit_h))<<(no_bit_l) ) | ((((v)%(unit_h))/(unit_l)) & ~((0xffffffff)<<((no_bit_l)+(no_bit_h)))))
182
183 //#define CM_INT_UNUSED 0
184
185
186 /**
187 Compute a 8 bit CRC based on a \a data, whith a \a old pre-computed crc.
188 */
189 /** old ---> the old pre-computed crc */
190 /** data ---> the data to compute crc for */
191 BYTE CM_crc8_8(BYTE old, BYTE data);
192
193 UINT32 CM_Compute_CRC8_8(UINT32 oldcrc
194 ,BYTE *buffer
195 ,UINT32 length
196 );
197
198
199 UINT16 CM_CRC16(BYTE* adrs, UINT16 Crc);
200 UINT16 CM_Compute_CRC16(UINT16 oldcrc,BYTE *buffer,UINT32 length);
201
202 BYTE* charToUnsignedChar(char buffer[], UINT32 length);
203
204 #endif /* CRC_H */
205
206
207

  ViewVC Help
Powered by ViewVC 1.1.23