[Qemu-devel] Question about the source code.

[Qemu-devel] Question about the source code.

[Toni Nedialkov]

[-- Attachment #1: Type: text/plain, Size: 418 bytes --]

Hello, I have a question pertaining the source code, wondering if I've
discovered a bug or am just simply confused.

It's about this function in translate-all.c

static void page_flush_tb(void)
{
    int i;

    for (i = 0; i < V_L1_SIZE; i++) {
        page_flush_tb_1(V_L1_SHIFT / V_L2_BITS - 1, l1_map + i);
    }
}

Was wondering if l1_map + i is correct, or whether it should be l1_map +
i*sizeof(i).

Thank you.

[-- Attachment #2: Type: text/html, Size: 3263 bytes --]

Re: [Qemu-devel] Question about the source code.

[Peter Maydell]

On 6 November 2015 at 01:27, Toni Nedialkov <farmdve@gmail.com> wrote:
> Hello, I have a question pertaining the source code, wondering if I've
> discovered a bug or am just simply confused.
>
> It's about this function in translate-all.c
>
> static void page_flush_tb(void)
> {
>     int i;
>
>     for (i = 0; i < V_L1_SIZE; i++) {
>         page_flush_tb_1(V_L1_SHIFT / V_L2_BITS - 1, l1_map + i);
>     }
> }
>
> Was wondering if l1_map + i is correct, or whether it should be l1_map +
> i*sizeof(i).

If you look at the declaration of l1_map:
    static void *l1_map[V_L1_SIZE];
it is an array of pointers with V_L1_SIZE entries. This
loop is just indexing through that array, and "l1_map + i"
is the same as "&l1_map[i]".

i is just an index here, and the size of the int type
doesn't have any relation to the size of any of the
data structures involved in l1_map.

thanks
-- PMM