From 71e4f3812982dba2efb471283d310224e8db363e Mon Sep 17 00:00:00 2001 From: "Kartik K. Agaram" Date: Wed, 3 Mar 2021 22:09:50 -0800 Subject: 7842 - new directory organization Baremetal is now the default build target and therefore has its sources at the top-level. Baremetal programs build using the phase-2 Mu toolchain that requires a Linux kernel. This phase-2 codebase which used to be at the top-level is now under the linux/ directory. Finally, the phase-2 toolchain, while self-hosting, has a way to bootstrap from a C implementation, which is now stored in linux/bootstrap. The bootstrap C implementation uses some literate programming tools that are now in linux/bootstrap/tools. So the whole thing has gotten inverted. Each directory should build one artifact and include the main sources (along with standard library). Tools used for building it are relegated to sub-directories, even though those tools are often useful in their own right, and have had lots of interesting programs written using them. A couple of things have gotten dropped in this process: - I had old ways to run on just a Linux kernel, or with a Soso kernel. No more. - I had some old tooling for running a single test at the cursor. I haven't used that lately. Maybe I'll bring it back one day. The reorg isn't done yet. Still to do: - redo documentation everywhere. All the README files, all other markdown, particularly vocabulary.md. - clean up how-to-run comments at the start of programs everywhere - rethink what to do with the html/ directory. Do we even want to keep supporting it? In spite of these shortcomings, all the scripts at the top-level, linux/ and linux/bootstrap are working. The names of the scripts also feel reasonable. This is a good milestone to take stock at. --- linux/bootstrap/029syscalls.cc | 126 +++++++++++++++++++++++++++++++++++++++++ 1 file changed, 126 insertions(+) create mode 100644 linux/bootstrap/029syscalls.cc (limited to 'linux/bootstrap/029syscalls.cc') diff --git a/linux/bootstrap/029syscalls.cc b/linux/bootstrap/029syscalls.cc new file mode 100644 index 00000000..84a6af34 --- /dev/null +++ b/linux/bootstrap/029syscalls.cc @@ -0,0 +1,126 @@ +:(before "End Initialize Op Names") +put_new(Name, "cd", "software interrupt (int)"); + +:(before "End Single-Byte Opcodes") +case 0xcd: { // int imm8 (software interrupt) + trace(Callstack_depth+1, "run") << "syscall" << end(); + uint8_t code = next(); + if (code != 0x80) { + raise << "Unimplemented interrupt code " << HEXBYTE << code << '\n' << end(); + raise << " Only `int 80h` supported for now.\n" << end(); + break; + } + process_int80(); + break; +} + +:(code) +void process_int80() { + switch (Reg[EAX].u) { + case 1: + exit(/*exit code*/Reg[EBX].u); + break; + case 3: + trace(Callstack_depth+1, "run") << "read: " << Reg[EBX].u << ' ' << Reg[ECX].u << ' ' << Reg[EDX].u << end(); + Reg[EAX].i = read(/*file descriptor*/Reg[EBX].u, /*memory buffer*/mem_addr_u8(Reg[ECX].u), /*size*/Reg[EDX].u); + trace(Callstack_depth+1, "run") << "result: " << Reg[EAX].i << end(); + if (Reg[EAX].i == -1) raise << "read: " << strerror(errno) << '\n' << end(); + break; + case 4: + trace(Callstack_depth+1, "run") << "write: " << Reg[EBX].u << ' ' << Reg[ECX].u << ' ' << Reg[EDX].u << end(); + trace(Callstack_depth+1, "run") << Reg[ECX].u << " => " << mem_addr_string(Reg[ECX].u, Reg[EDX].u) << end(); + Reg[EAX].i = write(/*file descriptor*/Reg[EBX].u, /*memory buffer*/mem_addr_u8(Reg[ECX].u), /*size*/Reg[EDX].u); + trace(Callstack_depth+1, "run") << "result: " << Reg[EAX].i << end(); + if (Reg[EAX].i == -1) raise << "write: " << strerror(errno) << '\n' << end(); + break; + case 5: { + check_flags(ECX); + check_mode(EDX); + trace(Callstack_depth+1, "run") << "open: " << Reg[EBX].u << ' ' << Reg[ECX].u << end(); + trace(Callstack_depth+1, "run") << Reg[EBX].u << " => " << mem_addr_kernel_string(Reg[EBX].u) << end(); + Reg[EAX].i = open(/*filename*/mem_addr_kernel_string(Reg[EBX].u), /*flags*/Reg[ECX].u, /*mode*/0640); + trace(Callstack_depth+1, "run") << "result: " << Reg[EAX].i << end(); + if (Reg[EAX].i == -1) raise << "open: " << strerror(errno) << '\n' << end(); + break; + } + case 6: + trace(Callstack_depth+1, "run") << "close: " << Reg[EBX].u << end(); + Reg[EAX].i = close(/*file descriptor*/Reg[EBX].u); + trace(Callstack_depth+1, "run") << "result: " << Reg[EAX].i << end(); + if (Reg[EAX].i == -1) raise << "close: " << strerror(errno) << '\n' << end(); + break; + case 8: + check_mode(ECX); + trace(Callstack_depth+1, "run") << "creat: " << Reg[EBX].u << end(); + trace(Callstack_depth+1, "run") << Reg[EBX].u << " => " << mem_addr_kernel_string(Reg[EBX].u) << end(); + Reg[EAX].i = creat(/*filename*/mem_addr_kernel_string(Reg[EBX].u), /*mode*/0640); + trace(Callstack_depth+1, "run") << "result: " << Reg[EAX].i << end(); + if (Reg[EAX].i == -1) raise << "creat: " << strerror(errno) << '\n' << end(); + break; + case 10: + trace(Callstack_depth+1, "run") << "unlink: " << Reg[EBX].u << end(); + trace(Callstack_depth+1, "run") << Reg[EBX].u << " => " << mem_addr_kernel_string(Reg[EBX].u) << end(); + Reg[EAX].i = unlink(/*filename*/mem_addr_kernel_string(Reg[EBX].u)); + trace(Callstack_depth+1, "run") << "result: " << Reg[EAX].i << end(); + if (Reg[EAX].i == -1) raise << "unlink: " << strerror(errno) << '\n' << end(); + break; + case 38: + trace(Callstack_depth+1, "run") << "rename: " << Reg[EBX].u << " -> " << Reg[ECX].u << end(); + trace(Callstack_depth+1, "run") << Reg[EBX].u << " => " << mem_addr_kernel_string(Reg[EBX].u) << end(); + trace(Callstack_depth+1, "run") << Reg[ECX].u << " => " << mem_addr_kernel_string(Reg[ECX].u) << end(); + Reg[EAX].i = rename(/*old filename*/mem_addr_kernel_string(Reg[EBX].u), /*new filename*/mem_addr_kernel_string(Reg[ECX].u)); + trace(Callstack_depth+1, "run") << "result: " << Reg[EAX].i << end(); + if (Reg[EAX].i == -1) raise << "rename: " << strerror(errno) << '\n' << end(); + break; + case 90: // mmap: allocate memory outside existing segment allocations + trace(Callstack_depth+1, "run") << "mmap: allocate new segment" << end(); + // Ignore most arguments for now: address hint, protection flags, sharing flags, fd, offset. + // We only support anonymous maps. + Reg[EAX].u = new_segment(/*length*/read_mem_u32(Reg[EBX].u+0x4)); + trace(Callstack_depth+1, "run") << "result: " << Reg[EAX].u << end(); + break; + case 0xa2: // nanosleep + cerr << "not sleeping\n"; + break; + default: + raise << HEXWORD << EIP << ": unimplemented syscall " << Reg[EAX].u << '\n' << end(); + } +} + +// SubX is oblivious to file permissions, directories, symbolic links, terminals, and much else besides. +// Also ignoring any concurrency considerations for now. +void check_flags(int reg) { + uint32_t flags = Reg[reg].u; + if (flags != ((flags & O_RDONLY) | (flags & O_WRONLY))) { + cerr << HEXWORD << EIP << ": most POSIX flags to the open() syscall are not supported. Just O_RDONLY and O_WRONLY for now. Zero concurrent access support.\n"; + exit(1); + } + if ((flags & O_RDONLY) && (flags & O_WRONLY)) { + cerr << HEXWORD << EIP << ": can't open a file for both reading and writing at once. See http://man7.org/linux/man-pages/man2/open.2.html.\n"; + exit(1); + } +} + +void check_mode(int reg) { + if (Reg[reg].u != 0600) { + cerr << HEXWORD << EIP << ": SubX is oblivious to file permissions; register " << reg << " must be 0x180.\n"; + exit(1); + } +} + +:(before "End Globals") +// Very primitive/fixed/insecure mmap segments for now. +uint32_t Segments_allocated_above = END_HEAP; +:(code) +// always allocate multiples of the segment size +uint32_t new_segment(uint32_t length) { + assert(length > 0); + uint32_t result = (Segments_allocated_above - length) & 0xff000000; // same number of zeroes as SEGMENT_ALIGNMENT + if (result <= START_HEAP) { + raise << "Allocated too many segments; the VM ran out of memory. " + << "Maybe SEGMENT_ALIGNMENT can be smaller?\n" << die(); + } + Mem.push_back(vma(result, result+length)); + Segments_allocated_above = result; + return result; +} -- cgit 1.4.1-2-gfad0