Skip to contentSkip to conversation

Sorry to be annoying, but "PATCH for shmdt"

DISCUSSION [email protected] 1 participant 1 message Dec 10, 2001

lore
Share & tools
lore
#0 lore raw

From: <[email protected]>

Unfortunately the patch I submitted for shmdt has been overlooked.
It's not a very exciting patch, but it does bring the code into agreement with the manpage. Here is an version of the patch updated for 2.4.16.

Yes, I agree. Referring to POSIX would have been more convincing, but it is true that shmdt is supposed to return EINVAL when the memory to be detached is not found.

Andries

    --- linux-2.4.16old/ipc/shm.c   Mon Dec  10 11:53:12 2001
    +++ linux-2.4.16/ipc/shm.c      Mon Dec  10 15:08:51 2001
    @@ -651,16 +651,19 @@
     {
            struct mm_struct *mm = current->mm;
            struct vm_area_struct *shmd, *shmdnext;
    +       int retcode = -EINVAL;

            down_write(&mm->mmap_sem);
            for (shmd = mm->mmap; shmd; shmd = shmdnext) {
                    shmdnext = shmd->vm_next;
                    if (shmd->vm_ops == &shm_vm_ops
    -                   && shmd->vm_start - (shmd->vm_pgoff << PAGE_SHIFT) == (ulong) shmaddr)
    +                   && shmd->vm_start - (shmd->vm_pgoff << PAGE_SHIFT) == (ulong) shmaddr) {
                            do_munmap(mm, shmd->vm_start, shmd->vm_end - shmd->vm_start);
    +                       retcode = 0;
    +               }
            }
            up_write(&mm->mmap_sem);
    -       return 0;
    +       return retcode;
     }

     #ifdef CONFIG_PROC_FS