[PATCH v4] tplink-safeloader: Patch to handle partitions with alternate names.

[Ole Kristian Lona]

Thanks!

I agree! In the end, this turned into a patch that ended up making some of the original code better, by moving all the hard-coded partition names into "defaults" that are all gathered in one function. Of course, further improvement could be to move the defaults into #define - statements in the beginning, but I guess tplink-safeloader will see a limited number of new devices going forward, as my newest devices, the X60, use UBI, with a different format for wrapping the image.


Best regards,

Ole Kristian

On 20/05/2022, 20:50, "Sander Vanheule" <sander at svanheule.net> wrote:

    Hi Ole Kristian,

    Thanks for the updates, this has turned into a pretty clean patch :)

    On Mon, 2022-05-16 at 15:50 +0200, Ole Kristian Lona wrote:
    > Some devices, specifically Deco M4R-v3 / M5 have partition names that
    > deviate from the scheme other devices use. They have an added "@0"
    > and "@1" for some partition names.
    > 
    > The devices have fallback partitions which will be used in case the
    > device determines that the primary partition set is unbootable.
    > 
    > This patch introduces an option to set these alternate partition names
    > in the device definition of tplink-safeloader.
    > 
    > Signed-off-by: Ole Kristian Lona <oklo at oklona.net>

    With one minor comment below:

    Reviewed-by: Sander Vanheule <sander at svanheule.net>

    > ---
    > 
    > @@ -2993,9 +3022,9 @@ static struct image_partition_entry make_partition_table(const struct
    > flash_part
    >         *(s++) = 0x00;
    >  
    >         size_t i;
    > -       for (i = 0; p[i].name; i++) {
    > +       for (i = 0; p->partitions[i].name; i++) {
    >                 size_t len = end-s;
    > -               size_t w = snprintf(s, len, "partition %s base 0x%05x size 0x%05x\n", p[i].name,
    > p[i].base, p[i].size);
    > +               size_t w = snprintf(s, len, "partition %s base 0x%05x size 0x%05x\n", p-
    > >partitions[i].name, p->partitions[i].base, p->partitions[i].size);

    If feel it would be better to split this line. There are long lines in this file, but 155 characters
    seems a bit excessive.

    As far as I'm concerned, you don't have to resubmit this patch if this is the only comment left.
    Reflowing a single line can be done when merging.


    Best,
    Sander