Skip to content

Commit

Permalink
add more notes in computer architecture
Browse files Browse the repository at this point in the history
  • Loading branch information
@kekeandzeyu
kekeandzeyu committed Oct 20, 2024
1 parent 2fc9d33 commit 08ef8fe
Show file tree
Hide file tree
Showing 2 changed files with 200 additions and 1 deletion.
Binary file added Writerside/images_architecture/a6-4-1.png
View file
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
201 changes: 200 additions & 1 deletion Writerside/topics/Computer-Architecture.topic
View file
Original file line number Diff line number Diff line change
Expand Up @@ -2546,7 +2546,7 @@ Conversion"/>
</li>
</list>
<img src="../images_architecture/a6-2-2.png" alt="movq Operand Combinations"/>
<p><format color="BlueViolet">Simple Memory Addressing Mode</format></p>
<p><format color="BlueViolet">Memory Addressing Mode</format></p>
<list type="decimal">
<li>
<p>Normal (R) => pointer dereferencing in C</p>
Expand All @@ -2557,9 +2557,208 @@ Conversion"/>
</li>
<li>
<p>Displacement D(R) => accessing data at a fixed offset from a base address</p>
<p><code>movq 8(%rbp), %rdx</code> equals:</p>
<code-block lang="c">
rdx = *(((char*)rbp) + 8)
</code-block>
</li>
<li>
<p><format color="Fuchsia">Complete Mode:</format> D(Rb,Ri,S)</p>
<list type="bullet">
<li>
<p><format color="LawnGreen">D:</format> Constant "displacement" 1, 2, or 4 bytes</p>
</li>
<li>
<p><format color="LawnGreen">Rb:</format> Base register: Any of 16 integer registers</p>
</li>
<li>
<p><format color="LawnGreen">Ri:</format> Index register: Any, except for %rsp</p>
</li>
<li>
<p><format color="LawnGreen">S:</format> Scale: 1, 2, 4, or 8</p>
</li>
</list>
<p><format color="HotPink">Example:</format> <code>movl (&#37;rbx,&#37;rdi,4), %eax</code></p>
<list type="decimal">
<li>
<p>Calculate the address: %rbx + (4 * %rdi) (e.g., base address + 4 * 5 = base address
+ 20).</p>
</li>
<li>
<p>Access the 32-bit value at that calculated address.</p>
</li>
<li>
<p>Move the retrieved value into the %eax register.</p>
</li>
</list>
</li>
</list>
</chapter>
<chapter title="6.3.2 Address Computation Instructions" id="6-3-2-address-aomputation-instructions">
<p><code>leaq src, dst</code></p>
<code-block lang="plain text" ignore-vars="true">
leaq (%rdi,%rdi,2), %rax ; Calculate the address of the element at index rdx in an array of integers pointed to by rsi (assuming 4 bytes per integer)
</code-block>
<p>It can also be used for arithmetic expressions.</p>
<code-block lang="plain text" ignore-vars="true">
leaq (%rdi,%rdi,2), %rax ; Calculate rax = rdi + (rdi * 2) = 3 * rdi
</code-block>
</chapter>
<chapter title="6.3.3 Arithmetic Instructions" id="6-3-3-arithmetic-instructions">
<p><format color="BlueViolet">Two Operand Instructions:</format></p>
<table style="header-row">
<tr>
<td>Format</td>
<td>Computation</td>
</tr>
<tr>
<td><code>addq src, dest</code></td>
<td>dest = dest + src</td>
</tr>
<tr>
<td><code>subq src, dest</code></td>
<td>dest = dest - src</td>
</tr>
<tr>
<td><code>imulq src, dest</code></td>
<td>dest = dest * src</td>
</tr>
<tr>
<td><code>salq src, dest</code></td>
<td>dest = dest &lt;&lt; src</td>
</tr>
<tr>
<td><code>sarq src, dest</code> (Arithmetic Shift)</td>
<td>dest = dest &gt;&gt; src</td>
</tr>
<tr>
<td><code>shrq src, dest</code> (Logical Shift)</td>
<td>dest = dest &gt;&gt; src</td>
</tr>
<tr>
<td><code>xorq src, dest</code></td>
<td>dest = dest ^ src</td>
</tr>
<tr>
<td><code>andq src, dest</code></td>
<td>dest = dest &amp; src</td>
</tr>
<tr>
<td><code>orq src, dest</code></td>
<td>dest = dest | src</td>
</tr>
</table>
<note>
<p>Watch out for the argument order!</p>
<p>Intel docs use "op dest, src"!</p>
</note>
<p><format color="BlueViolet">One Operand Instructions:</format></p>
<table style="header-row">
<tr>
<td>Format</td>
<td>Computation</td>
</tr>
<tr>
<td><code>incq dest</code></td>
<td>dest = dest + 1</td>
</tr>
<tr>
<td><code>decq src, dest</code></td>
<td>dest = dest - 1</td>
</tr>
<tr>
<td><code>negq src, dest</code></td>
<td>dest = - dest</td>
</tr>
<tr>
<td><code>notq dest</code></td>
<td>dest = ~ dest</td>
</tr>
<tr>
<td><code>sarq src, dest</code> (Arithmetic Shift)</td>
<td>dest = dest &gt;&gt; src</td>
</tr>
<tr>
<td><code>shrq src, dest</code> (Logical Shift)</td>
<td>dest = dest &gt;&gt; src</td>
</tr>
<tr>
<td><code>xorq src, dest</code></td>
<td>dest = dest ^ src</td>
</tr>
<tr>
<td><code>andq src, dest</code></td>
<td>dest = dest &amp; src</td>
</tr>
<tr>
<td><code>orq src, dest</code></td>
<td>dest = dest | src</td>
</tr>
</table>
<p><format color="HotPink">Example</format></p>
<code-block lang="c" collapsible="true">
long arith(long x, long y, long z) {
long t1 = x + y;
long t2 = z + t1;
long t3 = x + 4;
long t4 = y * 48;
long t5 = t3 + t4;
long rval = t2 * t5;
return rval;
}
</code-block>
<code-block lang="plain text" collapsible="true" ignore-vars="true">
0000000000001149 &lt;arith&gt;:
1149: f3 0f 1e fa endbr64
114d: 55 push %rbp
114e: 48 89 e5 mov %rsp,%rbp
1151: 48 89 7d c8 mov %rdi,-0x38(%rbp) # Store x on the stack
1155: 48 89 75 c0 mov %rsi,-0x40(%rbp) # Store y on the stack
1159: 48 89 55 b8 mov %rdx,-0x48(%rbp) # Store z on the stack
115d: 48 8b 55 c8 mov -0x38(%rbp),%rdx # Load x into rdx
1161: 48 8b 45 c0 mov -0x40(%rbp),%rax # Load y into rax
1165: 48 01 d0 add %rdx,%rax # t1 = x + y
1168: 48 89 45 d0 mov %rax,-0x30(%rbp) # Store t1 on the stack
116c: 48 8b 55 b8 mov -0x48(%rbp),%rdx # Load z into rdx
1170: 48 8b 45 d0 mov -0x30(%rbp),%rax # Load t1 into rax
1174: 48 01 d0 add %rdx,%rax # t2 = z + t1
1177: 48 89 45 d8 mov %rax,-0x28(%rbp) # Store t2 on the stack
117b: 48 8b 45 c8 mov -0x38(%rbp),%rax # Load x into rax
117f: 48 83 c0 04 add $0x4,%rax # t3 = x + 4
1183: 48 89 45 e0 mov %rax,-0x20(%rbp) # Store t3 on the stack
1187: 48 8b 55 c0 mov -0x40(%rbp),%rdx # Load y into rdx
118b: 48 89 d0 mov %rdx,%rax # Copy y into rax (for multiplication)
118e: 48 01 c0 add %rax,%rax # Multiply y by 2 (left shift)
1191: 48 01 d0 add %rdx,%rax # Add y (original value) so now its 3y
1194: 48 c1 e0 04 shl $0x4,%rax # Multiply by 16 (left shift 4) 3y*16 = 48y
1198: 48 89 45 e8 mov %rax,-0x18(%rbp) # Store t4 = 48 * y on the stack
119c: 48 8b 55 e0 mov -0x20(%rbp),%rdx # Load t3 into rdx
11a0: 48 8b 45 e8 mov -0x18(%rbp),%rax # Load t4 into rax
11a4: 48 01 d0 add %rdx,%rax # t5 = t3 + t4
11a7: 48 89 45 f0 mov %rax,-0x10(%rbp) # Store t5 on the stack
11ab: 48 8b 45 d8 mov -0x28(%rbp),%rax # Load t2 into rax
11af: 48 0f af 45 f0 imul -0x10(%rbp),%rax # rval = t2 * t5
11b4: 48 89 45 f8 mov %rax,-0x8(%rbp) # Store rval on the stack
11b8: 48 8b 45 f8 mov -0x8(%rbp),%rax # Move rval into rax (return value)
11bc: 5d pop %rbp # Restore the caller's base pointer
11bd: c3 ret # Return from the function
</code-block>
</chapter>
</chapter>
<chapter title="6.4 C, Assembly &amp; Machine Code" id="6-4-c-assembly-machine-code" switcher-key="x86">
<p>Compile with the following code (debugging-friendly):</p>
<code-block lang="plain text">
gcc –Og p1.c -o p
</code-block>
<img src="../images_architecture/a6-4-1.png" alt="C, Assembly &amp; Machine Code"/>
<p>Use the following command to generate the assembly code:</p>
<code-block lang="plain text">
gcc –Og –S p1.c
</code-block>
<p>Use the following command to disassembly the machine code:</p>
<code-block lang="plain text">
objdump –d p1
</code-block>
</chapter>
</chapter>
<chapter id="4-risc-functions">
Expand Down

0 comments on commit 08ef8fe

Please sign in to comment.