Re: [PATCH v27 2/6] fs/proc/task_mmu: Implement IOCTL to get and optionally clear info about PTEs

[Muhammad Usama Anjum <usama.anjum@collabora.com>]

On 8/9/23 12:55 AM, Andrei Vagin wrote:
> On Tue, Aug 8, 2023 at 12:35 PM Muhammad Usama Anjum
> <usama.anjum@collabora.com> wrote:
>>
>> On 8/9/23 12:21 AM, Andrei Vagin wrote:
>>> On Tue, Aug 8, 2023 at 3:43 AM Muhammad Usama Anjum
>>> <usama.anjum@collabora.com> wrote:
>>>
>>> ....
>>>
>>>> +static int pagemap_scan_output(unsigned long categories,
>>>> +                              struct pagemap_scan_private *p,
>>>> +                              unsigned long addr, unsigned long *end)
>>>> +{
>>>> +       unsigned long n_pages, total_pages;
>>>> +       int ret = 0;
>>>> +
>>>> +       if (!p->vec_buf)
>>>> +               return 0;
>>>> +
>>>> +       categories &= p->arg.return_mask;
>>>> +
>>>> +       n_pages = (*end - addr) / PAGE_SIZE;
>>>> +       if (check_add_overflow(p->found_pages, n_pages, &total_pages) || //TODO
>>>
>>> Need to fix this TODO.
>> Sorry, I forgot to remove the "//TODO". As far as I've understood, the last
>> discussion ended in keeping the check_add_overflow(). [1] I'll just remove
>> the TODO.
>>
>> https://lore.kernel.org/all/CABb0KFEfmRz+Z_-7GygTL12E5Y254dvoUfWe4uSv9-wOx+Cs8w@mail.gmail.com
>>
>>
>>>
>>>> +           total_pages > p->arg.max_pages) {
>>>> +               size_t n_too_much = total_pages - p->arg.max_pages;
>>>> +               *end -= n_too_much * PAGE_SIZE;
>>>> +               n_pages -= n_too_much;
>>>> +               ret = -ENOSPC;
>>>> +       }
>>>> +
>>>> +       if (!pagemap_scan_push_range(categories, p, addr, *end)) {
>>>> +               *end = addr;
>>>> +               n_pages = 0;
>>>> +               ret = -ENOSPC;
>>>> +       }
>>>> +
>>>> +       p->found_pages += n_pages;
>>>> +       if (ret)
>>>> +               p->walk_end_addr = *end;
>>>> +
>>>> +       return ret;
>>>> +}
>>>> +
>>>
>>> ...
>>>
>>>> +static long do_pagemap_scan(struct mm_struct *mm, unsigned long uarg)
>>>> +{
>>>> +       struct mmu_notifier_range range;
>>>> +       struct pagemap_scan_private p;
>>>> +       unsigned long walk_start;
>>>> +       size_t n_ranges_out = 0;
>>>> +       int ret;
>>>> +
>>>> +       memset(&p, 0, sizeof(p));
>>>> +       ret = pagemap_scan_get_args(&p.arg, uarg);
>>>> +       if (ret)
>>>> +               return ret;
>>>> +
>>>> +       p.masks_of_interest = MASKS_OF_INTEREST(p.arg);
>>>> +       ret = pagemap_scan_init_bounce_buffer(&p);
>>>> +       if (ret)
>>>> +               return ret;
>>>> +
>>>> +       /* Protection change for the range is going to happen. */
>>>> +       if (p.arg.flags & PM_SCAN_WP_MATCHING) {
>>>> +               mmu_notifier_range_init(&range, MMU_NOTIFY_PROTECTION_VMA, 0,
>>>> +                                       mm, p.arg.start, p.arg.end);
>>>> +               mmu_notifier_invalidate_range_start(&range);
>>>> +       }
>>>> +
>>>> +       walk_start = p.arg.start;
>>>> +       for (; walk_start < p.arg.end; walk_start = p.arg.walk_end) {
>>>> +               int n_out;
>>>> +
>>>> +               if (fatal_signal_pending(current)) {
>>>> +                       ret = -EINTR;
>>>> +                       break;
>>>> +               }
>>>> +
>>>> +               ret = mmap_read_lock_killable(mm);
>>>> +               if (ret)
>>>> +                       break;
>>>> +               ret = walk_page_range(mm, walk_start, p.arg.end,
>>>> +                                     &pagemap_scan_ops, &p);
>>>> +               mmap_read_unlock(mm);
>>>> +
>>>> +               n_out = pagemap_scan_flush_buffer(&p);
>>>> +               if (n_out < 0)
>>>> +                       ret = n_out;
>>>> +               else
>>>> +                       n_ranges_out += n_out;
>>>> +
>>>> +               if (ret != -ENOSPC || p.arg.vec_len - 1 == 0 ||
>>>> +                   p.found_pages == p.arg.max_pages) {
>>>> +                       p.walk_end_addr = p.arg.end;
>>>
>>> You should not change p.walk_end_addr If ret is ENOSPC. Pls add a test
>>> case to check this.
>> Yeah, I'm not setting walk_end_addr if ret is ENOSPC.
>>
>> I'm setting walk_end_addr only when ret = 0. I'd added this as a result of
>> a test case in my local test application. I can look at adding some tests
>> in pagemap_ioctl.c kselftest as well.
> 
> I am not sure that I understand what you mean here. ENOSPC can be returned
> when the vec array is full and in this case, walk_end_addr should be
> the address when it stops scanning.
I'll copy a test case in kselftest to prove or dis-prove the correctness of
walk_end address.

> 
>>
>>>
>>>> +                       break;
>>>> +               }
>>>> +       }
>>>> +
>>>> +       if (p.cur_buf.start != p.cur_buf.end) {
>>>> +               if (copy_to_user(p.vec_out, &p.cur_buf, sizeof(p.cur_buf)))
>>>> +                       ret = -EFAULT;
>>>> +               else
>>>> +                       ++n_ranges_out;
>>>> +       }
>>>> +
>>>> +       /* ENOSPC signifies early stop (buffer full) from the walk. */
>>>> +       if (!ret || ret == -ENOSPC)
>>>> +               ret = n_ranges_out;
>>>> +
>>>> +       p.arg.walk_end = p.walk_end_addr ? p.walk_end_addr : walk_start;
>>>> +       if (pagemap_scan_writeback_args(&p.arg, uarg))
>>>> +               ret = -EFAULT;
>>>> +
>>>> +       if (p.arg.flags & PM_SCAN_WP_MATCHING)
>>>> +               mmu_notifier_invalidate_range_end(&range);
>>>> +
>>>> +       kfree(p.vec_buf);
>>>> +       return ret;
>>>> +}
>>>
>>> Thanks,
>>> Andrei
>>
>> --
>> BR,
>> Muhammad Usama Anjum

-- 
BR,
Muhammad Usama Anjum