ichspi: fix missing set_addr on erases and possible crossings of 256 B boundaries
Apparently the erase function did never set any address before issuing the
erase commands. How could this ever work?
Also, according to PCH documentation crossing 256 byte boundaries is invalid
and may cause wraparound due to the flash chip's pages. Check for this on
reads as well as writes.
Thanks to Vladimir 'φ-coder/phcoder' Serbinenko for noticing these issues
and providing the initial patch.
Corresponding to flashrom svn r1837.
Signed-off-by: Stefan Tauner <stefan.tauner@alumni.tuwien.ac.at>
Acked-by: Stefan Tauner <stefan.tauner@alumni.tuwien.ac.at>
diff --git a/ichspi.c b/ichspi.c
index 90bb0de..5d37d06 100644
--- a/ichspi.c
+++ b/ichspi.c
@@ -1272,6 +1272,7 @@
}
msg_pdbg("Erasing %d bytes starting at 0x%06x.\n", len, addr);
+ ich_hwseq_set_addr(addr);
/* make sure FDONE, FCERR, AEL are cleared by writing 1 to them */
REGWRITE16(ICH9_REG_HSFS, REGREAD16(ICH9_REG_HSFS));
@@ -1307,7 +1308,11 @@
REGWRITE16(ICH9_REG_HSFS, REGREAD16(ICH9_REG_HSFS));
while (len > 0) {
+ /* Obey programmer limit... */
block_len = min(len, flash->mst->opaque.max_data_read);
+ /* as well as flash chip page borders as demanded in the Intel datasheets. */
+ block_len = min(block_len, 256 - (addr & 0xFF));
+
ich_hwseq_set_addr(addr);
hsfc = REGREAD16(ICH9_REG_HSFC);
hsfc &= ~HSFC_FCYCLE; /* set read operation */
@@ -1345,7 +1350,10 @@
while (len > 0) {
ich_hwseq_set_addr(addr);
+ /* Obey programmer limit... */
block_len = min(len, flash->mst->opaque.max_data_write);
+ /* as well as flash chip page borders as demanded in the Intel datasheets. */
+ block_len = min(block_len, 256 - (addr & 0xFF));
ich_fill_data(buf, block_len, ICH9_REG_FDATA0);
hsfc = REGREAD16(ICH9_REG_HSFC);
hsfc &= ~HSFC_FCYCLE; /* clear operation */