又遇到vmware不能启动了（vmware12/Deepin15.4(4.9.0kernel))

最新推荐文章于 2024-08-16 02:26:49 发布

sjg20010414

最新推荐文章于 2024-08-16 02:26:49 发布

阅读量1k

点赞数

有可能是内核从4.4升级到了4.9的缘故，原来可以工作的vmware12不能启动了，还是老样子，模块vmmon和vmnet

参考 https://communities.vmware.com/message/2604831

https://communities.vmware.com/thread/536977?start=0&tstart=0

使用脚本 https://bugfixes.promisedev.com/

但出现问题：

Starting VMware services:
Virtual machine monitor failed
Virtual machine communication interface done
VM communication interface socket family done
Blocking file system done
Virtual ethernet failed
VMware Authentication Daemon done

仔细看编译过程，还是 get_user_pages 有问题，猜想有可能是因为4.9 新内核，使用 /usr/src/linux-headers-4.9.0-deepin4-common/include/linux/*.h等头文件和/usr/lib/vmware/modules/source/vmnet-only or vmmon-only/*.c 文件存在版本匹配问题

上面的参考并没有解决问题，但 http://www.jianshu.com/p/df30c0c3889b 一文提出了问题所在，就是内核更新后，同一个函数名，参数个数和形式有变化

在文件 /usr/src/linux-headers-4.9.0-deepin4-common/include/linux/mm.h 中（即新内核中）

long get_user_pages_remote(struct task_struct *tsk, struct mm_struct *mm,
unsigned long start, unsigned long nr_pages,
unsigned int gup_flags, struct page **pages,
struct vm_area_struct **vmas);
long get_user_pages(unsigned long start, unsigned long nr_pages,
unsigned int gup_flags, struct page **pages,
struct vm_area_struct **vmas);

在 vmnet-only/userif.c 中（即 vmware 12 的模块中）

struct page *page = NULL;
int retval;

down_read(&current->mm->mmap_sem);
retval = get_user_pages(current, current->mm, addr,
1, 1, 0, &page, NULL);

可以发现，即使把 get_user_pages 修改为 get_user_pages_remote，仍然存在参数个数不匹配的问题（get_user_pages_remote只有7个参数）

current 是宏，就是 get_current()，将返回 struct task_struct * 指针类型

仔细比对一下 4.4.0 内核中 mm.h 文件中的 get_user_pages，其形式如下：

long __get_user_pages(struct task_struct *tsk, struct mm_struct *mm,
unsigned long start, unsigned long nr_pages,
unsigned int foll_flags, struct page **pages,
struct vm_area_struct **vmas, int *nonblocking);
long get_user_pages(struct task_struct *tsk, struct mm_struct *mm,
unsigned long start, unsigned long nr_pages,
int write, int force, struct page **pages,
struct vm_area_struct **vmas);

就是原来多一个 __get_user_pages，但没有 get_user_pages_remote，而 get_user_pages 有8个参数。怎么把8个参数中的 int write 和 int force 与7个参数中的 unsigned int gup_flags对应起来是个问题

花了不少时间，终于发现这些函数的实现在4.9内核的gup.c文件中（参考linux内核源码 https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable.git/tree/mm/gup.c?h=linux-4.9.y ）

long get_user_pages_remote(struct task_struct *tsk, struct mm_struct *mm,
unsigned long start, unsigned long nr_pages,
unsigned int gup_flags, struct page **pages,
struct vm_area_struct **vmas)
{
return __get_user_pages_locked(tsk, mm, start, nr_pages, pages, vmas,
NULL, false,
gup_flags | FOLL_TOUCH | FOLL_REMOTE);
}

my god！其实是调用了 __get_user_pages_locked函数：

static __always_inline long __get_user_pages_locked(struct task_struct *tsk,
struct mm_struct *mm,
unsigned long start,
unsigned long nr_pages,
struct page **pages,
struct vm_area_struct **vmas,
int *locked, bool notify_drop,
unsigned int flags)
{

.................

if (pages)
flags |= FOLL_GET; -------------->注意这里喽！

....................

ret = __get_user_pages(tsk, mm, start, nr_pages, flags, pages,
vmas, locked);

....................

ret = __get_user_pages(tsk, mm, start, 1, flags | FOLL_TRIED,
pages, NULL, NULL);

...............

}

继续看 __get_user_pages函数的第5个参数：

static long __get_user_pages(struct task_struct *tsk, struct mm_struct *mm,
unsigned long start, unsigned long nr_pages,
unsigned int gup_flags, struct page **pages,
struct vm_area_struct **vmas, int *nonblocking)
{

.............................

VM_BUG_ON(!!pages != !!(gup_flags & FOLL_GET));

if (!(gup_flags & FOLL_FORCE))
gup_flags |= FOLL_NUMA;

do {
struct page *page;
unsigned int foll_flags = gup_flags;

.............................

从这里，似乎可以看到 gup_flags 的相应位，和 FOLL_XXX 掩码定义是有关系的，特别关注 FOLL_FORCE 和 FOLL_WRITE，4.9内核的mm.h中

#define FOLL_WRITE 0x01/* check pte is writable */
#define FOLL_TOUCH 0x02/* mark page accessed */
#define FOLL_GET 0x04/* do get_page on page */
#define FOLL_DUMP 0x08/* give error on hole if it would be zero */
#define FOLL_FORCE 0x10/* get_user_pages read/write w/o permission */
#define FOLL_NOWAIT 0x20/* if a disk transfer is needed, start the IO
* and return without waiting upon it */

再去看老的内核中对应的函数

long get_user_pages(struct task_struct *tsk, struct mm_struct *mm,
unsigned long start, unsigned long nr_pages, int write,
int force, struct page **pages, struct vm_area_struct **vmas)
{
return __get_user_pages_locked(tsk, mm, start, nr_pages, write, force,
pages, vmas, NULL, false, FOLL_TOUCH);
}

跟踪 __get_user_pages_locked 并关注第5\6参数 write \force

static __always_inline long __get_user_pages_locked(struct task_struct *tsk,
struct mm_struct *mm,
unsigned long start,
unsigned long nr_pages,
int write, int force,
struct page **pages,
struct vm_area_struct **vmas,
int *locked, bool notify_drop,
unsigned int flags)
{