Looks good overall in preparing write functions for partial-write capability, e.g. taking a range and length rather than assuming 0 to total_size. I patched this in and it appears to be non-harmful on my NM10 setup, at least.
Note: There was one compilation issue ("it87spi.c:289: error: ‘it8716f_spi_page_program’ defined but not used") in the lines below due to the #ifdef.
+#ifdef TODO
+#warning This function needs to be converted to partial write
} else {
+ int i;
spi_disable_blockprotect();
- /* Erase first */
- msg_pinfo("Erasing flash before programming... ");
- if (erase_flash(flash)) {
- msg_perr("ERASE FAILED!\n");
- return -1;
- }
- msg_pinfo("done.\n");
- for (i = 0; i < total_size / 256; i++) {
+
+ /* FIXME: Handle chips which have max writechunk size >1 and <256. */
+ for (i = 0; i < flash->total_size * 1024 / 256; i++) {
it8716f_spi_page_program(flash, i, buf);
}
}
+#endif
I'd like to see this go in once the compilation issue is addressed so further progress can be made w.r.t. partial writes.
David Hendricks (dhendrix)
Systems Software Engineer, Google Inc.