CHAPTER 6 MEMORY DISPLAY OPTIONS Memory Display Windows The debugger allows you to set up windows into your program memory space. Using these windows, you can view memory in a variety of formats. The windows will remain in effect until you cancel them; updating themselves automatically if the memory changes. There are six single-line memory windows always present, in the lower right portion of the screen. In addition, you can have the upper-right window display a 14-line page of memory values. Single-Line Memory Windows A single-line memory window line consists of a specification, typed in by you, followed by a display, supplied by the debugger. To type in a specification on any one of window lines 1 through 6, simply type the associated digit, 1 through 6, when the debugger is in its main command mode. The cursor will jump to the beginning of the memory-window line you are specifying. You then type in a display format specification, followed by the address of memory you want displayed. The simplest form of the display format specification is a single letter, signifying one of the display types available. The choices are: B for hexadecimal bytes W for hexadecimal words N for decimal bytes D for decimal words Q for octal bytes O for octal words T for text; each byte reduced to one ASCII display character A for ASCII text, each byte occupying 2 display characters (the exact display is spelled out later in this chapter) C for ASCII characters, occupying 2 bytes if needed, otherwise one A format specification of one of the above letters will cause the debugger to display the array of bytes starting at the address you specify, in the format indicated by the letter, as long as there is room on the line. All letters in a format specification (or in any other context in the debugger) can be typed in either upper or lower case. The format specification should be terminated by a comma. After the comma (and an optional space), you type the address of memory you want displayed. This consists of two values, the segment followed by the offset. The values should be separated by a comma. You can omit the segment value if you wish: in that case, the current value of DS is used. The debugger reminds you that you have specified this option by following what you type with 2 commas instead of one. 6-2 The value you type can take one of the following forms: a. a numeric constant, whose format is just as in the assembly language (leading zero means default hex, otherwise default decimal) b. a register name (IP is now accepted as a register name) c. a user symbol from the assembly language program being debugged. After you type the address specification, you hit the ENTER key, and the debugger fills out the rest of the line with the memory display. For example, if you want to display hex bytes starting at 01000 hex on display line 2, you type 2b,01000. The cursor jumps to the line immediately when you type the 2, and it displays the b,01000 on the line as you type it. The b says you want hex bytes, and the 01000 has a leading zero to signify a hexadecimal address, not decimal. When you press the ENTER key, the debugger displays two commas, followed by the hex bytes. If the memory is zeroed, you will see 00 00 00 00 00 etc. to the end of the display line. Erasing Memory Display Lines Any memory display window that you specify will remain in effect, always updated to show the latest memory contents, until you explicitly erase it. To erase a window, you type the number of the window, followed by a blank. The line will also be erased if you start typing a format specification, and you hit the ENTER key before you get to your address specification. In the coming sections, many of the examples assume (and they say so) that your display is blank before you type in the example. You can always get a blank display by typing in each number followed by a blank: "1 2 3 4 5 6 ". If you accidentally type a digit and DON'T want to erase the line your cursor has jumped to, press the ESC key to return to the main command mode. Continuation Lines You may continue a memory display window onto the next line, by placing a double quote mark " at the beginning of the next line. You may do this in one of two ways: you may type the number of the next line, followed by the double quote mark; or you may simply type the double quote mark at the command level. The first method allows you to specify which window line you want continued, if there is more than one. The second method is more convenient to use. It places the quote mark on the last blank line that immediately follows a non-blank line. 6-3 You may continue placing " marks on as many lines as you have, creating a multi-line display. The debugger follows the " mark with the address of memory being displayed, followed by the memory, according to the start of the type specification of the line being continued. The memory display is aligned with the display of above line. If you are creating a multi-line display, and your specification is a long one, you may wish to start the display at the beginning of the next line, rather than after the specification on the first line. This will often give you more room. You do this by terminating the format specification with a slash / instead of a comma. For example, to display many hex bytes at the array BYTE_ARRAY, type 1b/byte_array followed by five double quote marks, when the memory display is empty. Mixed Format Specification Instead of having all your bytes or words on a line displayed in the same format, you can mix your formats. You do this by providing more than one letter in your format specification. The debugger will display one memory unit for each letter-type you specify. The line will be filled out with the last type given. For example, if you type 3nwb,01000, you will get a display on memory line 3 of the decimal byte at 01000, the hex word at 01001, and an array of hex bytes starting at 01003. Numbers in a Format Specification You may precede any letter in a format specification with a number up to 255. The effect is the same as if you had repeated the letter the given number of times. For example, if you type 1 to go to memory line 1, followed by 4w10tb,02000, you will get 4 hex words at 02000, 10 text characters at 02008, and an array of hex bytes from 02012 filling out the rest of the line. You may also end your format specification with a number up to 255. This will cause the entire specification to be repeated the given number of times. If there is room on the line for the full number given, the display will stop there-- there will be no repeating of the last type byte. If there is not room on the line for the full number of global iterations, the debugger will stop at the end of the last iteration that would fit. For example, the specification b8 causes 8 hex bytes to be displayed, and the remainder of the display line to be blank. The specification b3w9 will cause the debugger to try to output 9 records, each consisting of a hex byte followed by 3 hex words. After putting out 2 such records, the debugger will see that there is not room for a third full record, so it will stop. This stopping at the record boundary allows you to continue the display, with correct alignment, on subsequent lines. 6-4 Spacing Between Memory Display Units In general, the debugger provides a space between each unit (byte or word) it displays. There is an exception, however: the debugger will not space between adjacent text characters (A,C, or T specifications). There are special specifiers G, J, and M, described in the section below, that allow you to override the debugger's spacing policies. Special-Action Format Specifiers In addition to the 9 letters already mentioned that specify data types, there are 10 other letters, and 2 other characters, that cause the debugger to perform special actions. Following is a complete description of all 21 non-digit characters that can occur within a format specification: = causes a display, using the format of the letter following =, of the current memory pointer value, instead of the contents of the memory location. If a letter does not follow the =, then W is used; i.e., the pointer is displayed as a 4-digit hex word. There are two uses for this feature that come to mind: * If your address specification is symbolic, you can display the equivalent numeric address with =w, telling you exactly where the symbol is in memory. Note that this display implicitly occurs at the beginning of continuation lines. If the format specification begins with =, then the implicit display is suppressed, because the same address is given explicitly by the L. * You can display the values of registers in a format other than hex. For example, in the 8086 debugger you can display AX as a decimal number by specifying =d1,ax on one of the memory display lines. @ causes the debugger to read the next byte it was going to display, and instead of displaying the byte, use it as a count, to repeat the next letter in the specification. The debugger uses only the bottom 7 bits of the memory byte for the count. For example, if the memory contains a length byte followed by that number of text characters, the text could be displayed by specifying @t (or @a or @c, depending on what you want the display to look like). If the memory contains 05 41 42 43 44 45, the @t would cause ABCDE to be displayed. A causes a display of a single ASCII byte, always using 2 display bytes. The following table shows what is displayed for unusual bytes: 6-5 range of N display of N Example 00--1F ^ followed by N+040 02 is ^B 22 "" 23 "# 24 "$ 5E "^ 7F ^r (r stands for rubout) 80--9F $ followed by N-080+040 081 is $A A0--FE # followed by N-080 0B1 is #1 FF $r All other bytes cause a display of a space following by the appropriate ASCII byte. The A specification is used when you need guaranteed display length for proper alignment of continuation lines; and you do not want the potential loss of information provided by the single-byte T specification. B causes a display of a single byte as a 2-digit hexadecimal number. Numbers less than hex 10 have a leading 0, so that the display is always 2 digits. C causes a display of a single ASCII character, just as the A specification, except that normal characters (not in the table) display as just one byte, without the preceding space. D causes the display of a 16-bit word as an unsigned positive decimal number. There will be no leading zeros in the display; so the length of the display depends on the size of the number. F causes the display of a floating point number, in one of the three formats recognized by the 8087. You must have a floating point chip (8087 or 287) installed in your computer for this to work. You specify which of the three formats you are reading by one of three letters immediately following the letter F: FD specifies a 4-byte Doubleword (single precision) number FQ specifies an 8-byte Quadword (double precision) number FT specifies a Ten-byte number-- 8087 extended precision. G causes a gap between the adjacent display formats, of one space more than there would have been without the G. For adjacent string bytes, this means a space where there would have been none. For other data types, this means two spaces where there would have been one. J (join) causes two adjacent data types, that would have had a space between them, to have no space. 6-6 L (line) causes the display of an entire text line, using the C-format for each character of the line. The debugger does not display the terminating carriage return; nor does it display the following linefeed if there is one. (If you want it to, specify LUC or LUUCC instead of L.) If a carriage return is not found and the display line fills, then the L-specifier is cut off in mid-string. Any continuation line will start up at the beginning of the format specification, at the mid-string place in memory. M (mark) causes a vertical-bars symbol to be displayed. The symbol will replace a separating space that would have been output in the position. If you want the space, you can provide G on either side (or both sides) of the M. N (number) causes the display of an 8-bit byte as an unsigned positive decimal number. There will be no leading zeroes in the display; so the length of the display depends on the size of the number. O causes a display of a 16-bit word as a 6-digit octal number. Numbers less than octal 100000 have one or more leading zeroes, so that the display is always 6 digits. Q causes a display of a single byte as a 3-digit octal number. Numbers less than octal 100 have one or more leading zeroes, so that the display is always 3 digits. S causes the display of an entire null-terminated string, using the C-format for each character of the string. The terminating null (hex 00) does not generate a display (if you want it to, specify SUC instead of S). If a null is not found and the display line fills, then the S-specifier is cut off in mid-string. Any continuation line would start up at the beginning of the format specification, at the mid-string place in memory. T causes the display of a single ASCII text byte, with a guaranteed display space of one character. The character displayed is the same as the second character of the A-format. This means that you will not be able to tell the difference between normal, displaying ASCII characters, and their control and non-ASCII counterparts. You gain a compact representation, but you also gain ambiguity. U (unskip) causes the memory display pointer to decrement by one byte. No display is generated by this command. This command is useful in several contexts: * displaying memory in more than one format. For example, the specification 8b8u2g8a gives a hex-and-ASCII side-by-side display, similar to that provided by many memory dump programs. 6-7 * displaying memory out of its sequence order. To test your understanding of the special-action letters in a format specification, you should convince yourself that the specification xb2ujbx99/ gives the same display on the following " continuation line as the specification w/ does. * displaying the count byte consumed by the @ character. For example, instead of @t, you could specify nu@t, which would display the string count as well as the string. If memory were 05 41 42 43 44 45, this would be 5 ABCDE. W causes a display of a 16-bit word as a 4-digit hex number. Numbers less than hex 1000 have one or more leading zeroes, so that the display is always 4 digits. X causes the debugger to skip over the memory byte currently pointed to, without displaying it. The memory pointer is thus incremented. Z is given immediately following another format letter. It causes the display to fill out the line with displays of the given preceding format; but instead of starting with the given address, the debugger starts with a lower address, and displays memory up to but not including the given address. The most common usage of Z is to display the memory just output by a moving output pointer. For example, in the 8086 debugger, you could specify bz,es,di to display the hex bytes most recently output by the STOSB instruction. Note that Z makes sense only in a limited number of contexts. You will almost certainly want to use Z only as the second letter of a two-letter specification, as in the example above. I further recommend that you use a format letter that generates a fixed length display; i.e., B,W,Q,O,T, or C. If you use a variable length display (N,D, or A), the debugger will be as pessimistic as possible about the number of display characters needed, so that the display will likely terminate before the end of the line. A continuation of a Z-line will produce the same output as the original line. If you want to continue beyond the address given, repeat the specification without the Z. I now discuss what happens if you use Z in other contexts. Unless your taste runs to the bizarre, you should skip this paragraph. Since Z fills out a line, there should be no specifiers after Z: they would be ignored. Also note that Z has an effect only on the single letter that precedes it. If you precede Z with more than one letter, you will get a confusing effect: the display would start out forward from the address, then it would retreat when it got to Z's preceding letter. The Z-array would run up to the address reached before, which is forward from the address you specified. 6-8 The Data Memory Window You may cause the switchable window in the upper right quadrant of the screen to display 14 lines of memory, continuing from the last of the lines you specified within the 6-line memory area. To do this, you: 1. Type a format-and-address specification on one of the numbered memory lines, as previously described in this chapter. For example, to display Bytes at location DS:0100, type 6b,ds,0100 followed by the ENTER key. 2. If you have already selected the memory window, you'll automatically have a continuation of the memory line you just specified, into the upper right quadrant of the screen. If not, you may select the window by pressing either the ctrl-N or ctrl-P keys. Once you have set up a 14-line window, you may page through memory with the ctrl-N (Next memory page) and ctrl-P (Previous memory page) keys, described in Chapter 4. If, after having pressed ctrl-N or ctrl-P several times, you wish to return to the first window continuing from the address you specified, you may do so by typing the digit (1 to 6) of the last specification line, followed immediately by the ESC key to preserve the specification settings. The memory window will be reset to its continuation value. For example, if your specification was on line 6, you type 6 followed immediately by the ESC key. If you want your 14-line memory window to start at a certain address instead of continuing from a 1-line display, you can separate the format and the address with a slash instead of a comma. For example, if you want Words at location ES:0, type 6 followed by w/es,0 followed by the ENTER key.