@@ 653,15 653,17 @@ ise_erase(ulong addr)
* the flash spec claimes writing goes faster if we use
* the write buffer. We fill the write buffer and then
* issue the write request. After the write request,
- * subsequent reads will yield the status register or,
- * since error bits are sticky, another write buffer can
- * be filled and written.
+ * subsequent reads will yield the status register.
*
- * On timeout, we issue a read status register request so
- * that the status register can be read no matter how we
- * exit.
+ * returns the status, even on timeouts.
*
- * returns the status.
+ * NOTE: I tried starting back to back buffered writes
+ * without reading the status in between, as the
+ * flowchart in the intel data sheet suggests.
+ * However, it always responded with an illegal
+ * command sequence, so I must be missing something.
+ * If someone learns better, please email me, though
+ * I doubt it will be much faster. - presotto@bell-labs.com
*/
static ulong
ise_wbwrite(ulong *p, int n, ulong off, ulong baddr, ulong *status)