gem5/src/sim/process.hh
Brandon Potter a5802c823f syscall_emul: [patch 13/22] add system call retry capability
This changeset adds functionality that allows system calls to retry without
affecting thread context state such as the program counter or register values
for the associated thread context (when system calls return with a retry
fault).

This functionality is needed to solve problems with blocking system calls
in multi-process or multi-threaded simulations where information is passed
between processes/threads. Blocking system calls can cause deadlock because
the simulator itself is single threaded. There is only a single thread
servicing the event queue which can cause deadlock if the thread hits a
blocking system call instruction.

To illustrate the problem, consider two processes using the producer/consumer
sharing model. The processes can use file descriptors and the read and write
calls to pass information to one another. If the consumer calls the blocking
read system call before the producer has produced anything, the call will
block the event queue (while executing the system call instruction) and
deadlock the simulation.

The solution implemented in this changeset is to recognize that the system
calls will block and then generate a special retry fault. The fault will
be sent back up through the function call chain until it is exposed to the
cpu model's pipeline where the fault becomes visible. The fault will trigger
the cpu model to replay the instruction at a future tick where the call has
a chance to succeed without actually going into a blocking state.

In subsequent patches, we recognize that a syscall will block by calling a
non-blocking poll (from inside the system call implementation) and checking
for events. When events show up during the poll, it signifies that the call
would not have blocked and the syscall is allowed to proceed (calling an
underlying host system call if necessary). If no events are returned from the
poll, we generate the fault and try the instruction for the thread context
at a distant tick. Note that retrying every tick is not efficient.

As an aside, the simulator has some multi-threading support for the event
queue, but it is not used by default and needs work. Even if the event queue
was completely multi-threaded, meaning that there is a hardware thread on
the host servicing a single simulator thread contexts with a 1:1 mapping
between them, it's still possible to run into deadlock due to the event queue
barriers on quantum boundaries. The solution of replaying at a later tick
is the simplest solution and solves the problem generally.
2015-07-20 09:15:21 -05:00

210 lines
7.2 KiB
C++

/*
* Copyright (c) 2014-2016 Advanced Micro Devices, Inc.
* Copyright (c) 2001-2005 The Regents of The University of Michigan
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions are
* met: redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer;
* redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution;
* neither the name of the copyright holders nor the names of its
* contributors may be used to endorse or promote products derived from
* this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
* "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
* LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
* A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
* OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
* SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
* LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
* DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
* THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*
* Authors: Nathan Binkert
* Steve Reinhardt
* Brandon Potter
*/
#ifndef __PROCESS_HH__
#define __PROCESS_HH__
#include <inttypes.h>
#include <map>
#include <string>
#include <vector>
#include "arch/registers.hh"
#include "base/statistics.hh"
#include "base/types.hh"
#include "config/the_isa.hh"
#include "mem/se_translating_port_proxy.hh"
#include "sim/fd_array.hh"
#include "sim/fd_entry.hh"
#include "sim/sim_object.hh"
struct ProcessParams;
class EmulatedDriver;
class ObjectFile;
class PageTableBase;
class SyscallDesc;
class SyscallReturn;
class System;
class ThreadContext;
class Process : public SimObject
{
public:
struct WaitRec
{
Addr waitChan;
ThreadContext *waitingContext;
WaitRec(Addr chan, ThreadContext *ctx)
: waitChan(chan), waitingContext(ctx)
{ }
};
Process(ProcessParams *params, ObjectFile *obj_file);
void serialize(CheckpointOut &cp) const override;
void unserialize(CheckpointIn &cp) override;
void initState() override;
DrainState drain() override;
void syscall(int64_t callnum, ThreadContext *tc, Fault *fault);
virtual TheISA::IntReg getSyscallArg(ThreadContext *tc, int &i) = 0;
virtual TheISA::IntReg getSyscallArg(ThreadContext *tc, int &i, int width);
virtual void setSyscallArg(ThreadContext *tc, int i,
TheISA::IntReg val) = 0;
virtual void setSyscallReturn(ThreadContext *tc,
SyscallReturn return_value) = 0;
virtual SyscallDesc *getDesc(int callnum) = 0;
inline uint64_t uid() { return _uid; }
inline uint64_t euid() { return _euid; }
inline uint64_t gid() { return _gid; }
inline uint64_t egid() { return _egid; }
inline uint64_t pid() { return _pid; }
inline uint64_t ppid() { return _ppid; }
const char *progName() const { return executable.c_str(); }
std::string fullPath(const std::string &filename);
std::string getcwd() const { return cwd; }
/**
* Find an emulated device driver.
*
* @param filename Name of the device (under /dev)
* @return Pointer to driver object if found, else NULL
*/
EmulatedDriver *findDriver(std::string filename);
// This function acts as a callback to update the bias value in
// the object file because the parameters needed to calculate the
// bias are not available when the object file is created.
void updateBias();
Addr getBias();
Addr getStartPC();
ObjectFile *getInterpreter();
// override of virtual SimObject method: register statistics
void regStats() override;
void allocateMem(Addr vaddr, int64_t size, bool clobber = false);
/// Attempt to fix up a fault at vaddr by allocating a page on the stack.
/// @return Whether the fault has been fixed.
bool fixupStackFault(Addr vaddr);
// After getting registered with system object, tell process which
// system-wide context id it is assigned.
void
assignThreadContext(ContextID context_id)
{
contextIds.push_back(context_id);
}
// Find a free context to use
ThreadContext *findFreeContext();
/**
* Does mmap region grow upward or downward from mmap_end? Most
* platforms grow downward, but a few (such as Alpha) grow upward
* instead, so they can override this method to return false.
*/
virtual bool mmapGrowsDown() const { return true; }
/**
* Maps a contiguous range of virtual addresses in this process's
* address space to a contiguous range of physical addresses.
* This function exists primarily to expose the map operation to
* python, so that configuration scripts can set up mappings in SE mode.
*
* @param vaddr The starting virtual address of the range.
* @param paddr The starting physical address of the range.
* @param size The length of the range in bytes.
* @param cacheable Specifies whether accesses are cacheable.
* @return True if the map operation was successful. (At this
* point in time, the map operation always succeeds.)
*/
bool map(Addr vaddr, Addr paddr, int size, bool cacheable = true);
// list of all blocked contexts
std::list<WaitRec> waitList;
// thread contexts associated with this process
std::vector<ContextID> contextIds;
// system object which owns this process
System *system;
Addr brk_point; // top of the data segment
Addr stack_base; // stack segment base
unsigned stack_size; // initial stack size
Addr stack_min; // furthest address accessed from stack base
Addr max_stack_size; // the maximum size allowed for the stack
Addr next_thread_stack_base; // addr for next region w/ multithreaded apps
Addr mmap_end; // base of automatic mmap region allocs
Stats::Scalar num_syscalls; // track how many system calls are executed
bool useArchPT; // flag for using architecture specific page table
bool kvmInSE; // running KVM requires special initialization
PageTableBase* pTable;
SETranslatingPortProxy initVirtMem; // memory proxy for initial image load
ObjectFile *objFile;
std::vector<std::string> argv;
std::vector<std::string> envp;
std::string cwd;
std::string executable;
// Id of the owner of the process
uint64_t _uid;
uint64_t _euid;
uint64_t _gid;
uint64_t _egid;
// pid of the process and it's parent
uint64_t _pid;
uint64_t _ppid;
// Emulated drivers available to this process
std::vector<EmulatedDriver *> drivers;
std::shared_ptr<FDArray> fds;
};
#endif // __PROCESS_HH__