Jos - lab4 (Part A) Round-Robin Scheduling

Exercise 6.Implement round-robin scheduling in sched_yield()as described above. Don't forget to modifysyscall() to dispatchsys_yield().

这段code不算太多，不过一直碰到一个问题，每次sys_halt()以后就出现Gerneral Protection的exception：

HLT—Halt
Opcode Instruction Description
F4 HLT Halt
Description
Stops instruction execution and places the processor in a HALT state. An enabled interrupt
(including NMI and SMI), a debug exception, the BINIT# signal, the INIT# signal, or the
RESET# signal will resume execution. If an interrupt (including NMI) is used to resume execu-
tion after a HLT instruction, the saved instruction pointer (CS:EIP) points to the instruction
following the HLT instruction.
The HLT instruction is a privileged instruction. When the processor is running in protected or
virtual-8086 mode, the privilege level of a program or procedure must be 0 to execute the HLT
instruction.

看起来是发生的权限问题，而根据打出的栈信息来看，cs[bit1][bit0] = 00, 已经是最高级别了,所以看到权限访问上是没有问题的。

  trap 0x0000000d General Protection
  err  0x00000102
  eip  0xf0105499
  cs   0x----0008

flag 0x00000246

继续追踪，发现在原来的算法中如果当前指向的envs[0]是RUNNABLE状态的话，就会永远处于被忽略的状态，在原来的轮询中永远排不上，结果就导致了改进入monitor()的循环，却进了halt。至于最后为什么导致generation protection的问题，还没有完全搞清楚。粗线部分是后加的补丁。

 32         int i, pos = 0;
 33         struct Env *thisenv = thiscpu->cpu_env;
 34         if (thisenv == NULL)
 35                 thisenv = &envs[0];
 37         // find the position of thisenv in envs[]
 38         for (i = 0; i < NENV; i++) {
 39                 if (&envs[i] == thisenv) {
 40                         pos = i;
 41                         break;
 42                 }
 43         }
 45         for (i = (pos+1)%NENV; i != pos; i = (i+1)%NENV) {
 46                 if (envs[i].env_status == ENV_RUNNABLE) {
 47                         // Following are all done in env_run()
 48                         //c->cpu_env->env_status = ENV_RUNNABLE;
 49                         //c->cpu_env = &envs[i];
 50                         //unlock_kernel();
 52                         env_run(&envs[i]);
 53                 }
 54         }
 55         //if ((i == pos) && (thisenv->env_status == ENV_RUNNING)) {
 56         if ((thisenv->env_status == ENV_RUNNING) || (thisenv->env_status == ENV_RUNNABLE)) {
 59                         env_run(thisenv);
 60         }
 62         // sched_halt never returns
 63         sched_halt();

可是在cpu数目多的情况下还是会出现一定概率的page fault，从user在执行env_destroy的时候。

根据trap_init_percpu()的提示应该是设置某个cpu的TSS时出问题了。观察过几个log，发现enviroment[1000]在free后又被调用了，仔细review上面的代码逻辑，就会发现还是有问题的，如果thisenv为NULL的情况下默认把它设置成env[0],会造成一定概率下它被两个以上的cpu同时run起来（见56行），因此最后的版本如下：

 32         int i, pos = 0;
 33         struct Env *thisenv = thiscpu->cpu_env;
 63         if (thisenv != NULL) {
 65                 // find the position of thisenv in envs[]
 66                 for (i = 0; i < NENV; i++) {
 67                         if (&envs[i] == thisenv) {
 68                                 pos = i;
 69                                 break;
 70                         }
 71                 }
 73         }
 74         for (i = (pos+1)%NENV; i != pos; i = (i+1)%NENV) {
 75                 if (envs[i].env_status == ENV_RUNNABLE) {
 77                         env_run(&envs[i]);
 78                 }
 79         }
 80         if (envs[i].env_status == ENV_RUNNABLE) { // envs[pos] will be exclusive by above loop
 82                 env_run(&envs[i]);
 83         }
 84         if (thisenv != NULL) {
 85                 if (thisenv->env_status == ENV_RUNNING) {
 87                         env_run(thisenv);
 88                 }
 89         }
 92         // sched_halt never returns
 93         sched_halt();

最后再回头稍微提一下hlt造成的General Protection fault：

General Protection Fault

A General Protection Fault may occur for various reasons. The most common are:

• Segment error (privilege, type, limit, read/write rights).
• Executing a privileged instruction while CPL != 0.
• Writing a 1 in a reserved register field.
• Referencing or accessing a null-descriptor.

The saved instruction pointer points to the instruction which caused the exception.

Error code: The General Protection Fault sets an error code, which is thesegment selector index when the exception is segment related. Otherwise, 0.

Selector Error Code

 31         16   15         3   2   1   0+---+--  --+---+---+--  --+---+---+---+---+|   Reserved   |    Index     |  Tbl  | E |+---+--  --+---+---+--  --+---+---+---+---+

 LengthNameDescriptionE1 bitExternalWhen set, the exception originated externally to the processor.Tbl2 bitsIDT/GDT/LDT tableThis is one of the following values:ValueDescription0b00The Selector Index references a descriptor in the GDT.0b01The Selector Index references a descriptor in the IDT.0b10The Selector Index references a descriptor in the LDT.0b11The Selector Index references a descriptor in the IDT.Index13 bitsSelector IndexThe index in the GDT, IDT or LDT.

我们发现halt的error code 是102，根据上文的信息，这个异常还是和segment相关，并非halt的权限访问问题，并且tbl=0x01（IDT), Index = 0x102>>3 = 0x20, 可是idt[20]并不存在！！！到了这个份上，我只想怀疑是qemu模拟器的bug -_-|||

Question

3. In your implementation of env_run() you should havecalled lcr3(). Before and after the call tolcr3(), your code makes references (at least it should)to the variablee, the argument to env_run.Upon loading the %cr3 register, the addressing contextused by the MMU is instantly changed. But a virtualaddress (namelye) has meaning relative to a givenaddress context--the address context specifies the physical address towhich the virtual address maps. Why can the pointere bedereferenced both before and after the addressing switch?

A： For every userspace environment, they shared the same ENVS[] with each other(in fact the whole kernel part virtual memory), so the e is dereferenced to the same address.