In function kwboot_xm_sendblock() the loop that implements retrying to
send a boot block might be quit if kwboot_tty_send or kwboot_tty_recv
return a failure. In this case the value of the variable c that is
expected to hold the response byte is uninitialized and most likely
still holds an ACK from the previous call to kwboot_xm_sendblock().
So the right thing to do is not to clobber rc if it's already != 0.
The result of this patch in my current scenario is that kwboot dies with
xmodem: Connection timed out
when the SoC stops replying instead of continuing to try sending the
remaining blocks which results in select() blocking for one second for
each block.
Fixes: 0535713bbfa0 ("scripts: add kwboot tool")
Signed-off-by: Uwe Kleine-König <uwe@xxxxxxxxxxxxxxxxx>
--
Cc: Thomas Petazzoni <thomas.petazzoni@xxxxxxxxxxxxxxxxxx>
Cc: Gregory CLEMENT <gregory.clement@xxxxxxxxxxxxxxxxxx>
Cc: Ezequiel Garcia <ezequiel.garcia@xxxxxxxxxxxxxxxxxx>
Cc: Sebastian Hesselbarth <sebastian.hesselbarth@xxxxxxxxx>
I also tried to increase the timeout, but without success :-(. Will test
a different USB-to-RS232 adapter ...
---
scripts/kwboot.c | 30 ++++++++++++++++--------------
1 file changed, 16 insertions(+), 14 deletions(-)
diff --git a/scripts/kwboot.c b/scripts/kwboot.c
index e06866052ba9..46328d8ed006 100644
--- a/scripts/kwboot.c
+++ b/scripts/kwboot.c
@@ -376,21 +376,23 @@ kwboot_xm_sendblock(int fd, struct kwboot_block *block)
} while (c == NAK && retries-- > 0);
- rc = -1;
+ if (!rc) {
+ rc = -1;
- switch (c) {
- case ACK:
- rc = 0;
- break;
- case NAK:
- errno = EBADMSG;
- break;
- case CAN:
- errno = ECANCELED;
- break;
- default:
- errno = EPROTO;
- break;
+ switch (c) {
+ case ACK:
+ rc = 0;
+ break;
+ case NAK:
+ errno = EBADMSG;
+ break;
+ case CAN:
+ errno = ECANCELED;
+ break;
+ default:
+ errno = EPROTO;
+ break;
+ }
}
return rc;
--
2.1.3
_______________________________________________
barebox mailing list
barebox@xxxxxxxxxxxxxxxxxxx
http://lists.infradead.org/mailman/listinfo/barebox
