Attention is currently required from: Edward O'Callaghan, Angel Pons, Arthur Heymans, Anastasia Klimchuk.

View Change

5 comments:

• Patchset:

  • Patch Set #1:

    I'm not convinced about unnecessary initialization. Want to have it
    discussed at least once...

• File jlink_spi.c:

  • Patch Set #1, Line 128: memset(buffer + writecnt, 0x00, readcnt);

    I wonder why this is done at all.

• File libflashrom.c:

  • Patch Set #1, Line 123: const char **supported_programmers = calloc(PROGRAMMER_INVALID + 1, sizeof(char *));

    This is probably what the original author intended. IIRC, I left a comment
    in Gerrit about this API (which is completely broken).

  • Patch Set #1, Line 144: calloc(flashchips_size, sizeof(*supported_flashchips));

    Hmmm, I'm not convinced to initialize things that are supposed to be completely
    overwritten later. I doubt that it's the case for calloc() but I know that some
    static analysis tools complain about explicit but unnecessary initialization.

    Failure to override all fields could be avoided, for instance, by using a
    struct initializer below, e.g.

        const struct flashrom_flashchip_info supported_flashchip = {
                .vendor = flashchips[i].vendor,
                /* ... */
        };
        supported_flashchips[i] = supported_flashchip;

• File mstarddc_spi.c:

  • Patch Set #1, Line 83: uint8_t *cmd = malloc((writecnt + 1) * sizeof(uint8_t));

    This is just `malloc(writecnt + 1);` AIUI.

To view, visit change 55269. To unsubscribe, or for help writing mail filters, visit settings.