uboot: Import hush fixes from kkit/u-boot-2023.07.y

patches/uboot/2023.07.02/0001-net-mv88e6xxx-fix-missing-SMI-address-initialization.patch

@@ -1,7 +1,7 @@
 From e4213fc62d3d6a6733e3026b63dd6314a6f09ee2 Mon Sep 17 00:00:00 2001
 From: =?UTF-8?q?Marek=20Moj=C3=ADk?= <marek.mojik@nic.cz>
 Date: Wed, 6 Dec 2023 15:35:56 +0100
-Subject: [PATCH 1/6] net: mv88e6xxx: fix missing SMI address initialization
+Subject: [PATCH 1/8] net: mv88e6xxx: fix missing SMI address initialization
 MIME-Version: 1.0
 Content-Type: text/plain; charset=UTF-8
 Content-Transfer-Encoding: 8bit

patches/uboot/2023.07.02/0002-net-mv88e6xxx-Add-reset-gpios-support.patch

@@ -1,7 +1,7 @@
 From 80ddeb093f24cf1d8a99e92564e07a7180791f8c Mon Sep 17 00:00:00 2001
 From: Tobias Waldekranz <tobias@waldekranz.com>
 Date: Tue, 17 Oct 2023 13:15:58 +0200
-Subject: [PATCH 2/6] net: mv88e6xxx: Add reset-gpios support
+Subject: [PATCH 2/8] net: mv88e6xxx: Add reset-gpios support
 Organization: Addiva Elektronik
 
 If the switch's RESETn signal is under the CPUs control, release it

patches/uboot/2023.07.02/0003-net-mv88e6xxx-Support-clause-45-addressing-on-intern.patch

@@ -1,7 +1,7 @@
 From c31794d78337da82cf41ffbcb9e696e42759aa12 Mon Sep 17 00:00:00 2001
 From: Tobias Waldekranz <tobias@waldekranz.com>
 Date: Tue, 17 Oct 2023 13:20:19 +0200
-Subject: [PATCH 3/6] net: mv88e6xxx: Support clause 45 addressing on internal
+Subject: [PATCH 3/8] net: mv88e6xxx: Support clause 45 addressing on internal
  MDIO bus
 Organization: Addiva Elektronik
 

patches/uboot/2023.07.02/0004-net-mv88e6xxx-Add-support-for-6393X.patch

@@ -1,7 +1,7 @@
 From c94723526672f1e5555b7545c13256edb7bd7557 Mon Sep 17 00:00:00 2001
 From: Tobias Waldekranz <tobias@waldekranz.com>
 Date: Tue, 17 Oct 2023 13:22:00 +0200
-Subject: [PATCH 4/6] net: mv88e6xxx: Add support for 6393X
+Subject: [PATCH 4/8] net: mv88e6xxx: Add support for 6393X
 Organization: Addiva Elektronik
 
 Only built-in copper PHYs are supported as access ports. Support

patches/uboot/2023.07.02/0005-i2c-pcf8575-Properly-address-chip.patch

@@ -1,7 +1,7 @@
 From c39245a2d92d6f851dfcb8f573af873a05d4a5b8 Mon Sep 17 00:00:00 2001
 From: Tobias Waldekranz <tobias@waldekranz.com>
 Date: Fri, 20 Oct 2023 13:19:41 +0200
-Subject: [PATCH 5/6] i2c: pcf8575: Properly address chip
+Subject: [PATCH 5/8] i2c: pcf8575: Properly address chip
 Organization: Addiva Elektronik
 
 These devices hold a simple shift register that is accessed without

patches/uboot/2023.07.02/0006-arm64-mvebu-a8k-Add-eFuse-support.patch

@@ -1,7 +1,7 @@
 From 967f7f88c88e1a72d7cf389b5b3b6f184788006f Mon Sep 17 00:00:00 2001
 From: Tobias Waldekranz <tobias@waldekranz.com>
 Date: Tue, 24 Oct 2023 22:23:44 +0200
-Subject: [PATCH 6/6] arm64: mvebu: a8k: Add eFuse support
+Subject: [PATCH 6/8] arm64: mvebu: a8k: Add eFuse support
 Organization: Addiva Elektronik
 
 Add support for reading and burning LD and HD fuses. Use a specialized

patches/uboot/2023.07.02/0007-cmd-sleep-Consume-Ctrl-C-when-exiting-early.patch

@@ -0,0 +1,55 @@
+From ed733077b87b1f3da6ac956840c927406c268a3d Mon Sep 17 00:00:00 2001
+From: Tobias Waldekranz <tobias@waldekranz.com>
+Date: Mon, 10 Jun 2024 12:59:21 +0200
+Subject: [PATCH 7/8] cmd/sleep: Consume Ctrl-C when exiting early
+Organization: Addiva Elektronik
+
+When sleep aborts early due to Ctrl-C input from the user, that state
+is transferred to the exit status of the program and should therefore
+be cleared from the global state.
+
+Before this change, the lingering global state would lead to very
+unintuitive execution flows in scripts due to the U-Boot specific
+modification to the hush interpreter which checks for Ctrl-C in all
+loop constructs.
+
+This script, for example...
+
+   if sleep 3; then
+       echo "Had a good night's sleep"
+   else
+       echo "That was a rude awakening! I refuse to go on..."
+       while true; do
+           true
+       done
+
+       echo "...or maybe I will"
+   fi
+
+...would just skip over the infinite loop (indeed, all following
+loops) when prematurely interrupting the sleep.
+
+Signed-off-by: Tobias Waldekranz <tobias@waldekranz.com>
+---
+ cmd/sleep.c | 4 +++-
+ 1 file changed, 3 insertions(+), 1 deletion(-)
+
+diff --git a/cmd/sleep.c b/cmd/sleep.c
+index c741b4aa02..b3511e13b8 100644
+--- a/cmd/sleep.c
++++ b/cmd/sleep.c
+@@ -39,8 +39,10 @@ static int do_sleep(struct cmd_tbl *cmdtp, int flag, int argc,
+ 	delay += mdelay;
+ 
+ 	while (get_timer(start) < delay) {
+-		if (ctrlc())
++		if (ctrlc()) {
++			clear_ctrlc();
+ 			return CMD_RET_FAILURE;
++		}
+ 
+ 		udelay(100);
+ 	}
+-- 
+2.34.1
+

patches/uboot/2023.07.02/0008-hush-Remove-Ctrl-C-detection-in-loops.patch

@@ -0,0 +1,41 @@
+From 989ef8436df2ee48d308981098c44b1b8257c23b Mon Sep 17 00:00:00 2001
+From: Tobias Waldekranz <tobias@waldekranz.com>
+Date: Mon, 10 Jun 2024 13:25:31 +0200
+Subject: [PATCH 8/8] hush: Remove Ctrl-C detection in loops
+Organization: Addiva Elektronik
+
+Assume that the original intent was to emulate SIGINT to a shell. This
+only works as expected if the loop in question is the ouermost, and
+last, statement. In all other cases, it completely breaks the expected
+execution flow. It more or less resurrects Visual Basic's "On Error
+Resume Next".
+
+Disable this behavior and delegate the problem of loop termination to
+the writer of the script instead.
+
+Signed-off-by: Tobias Waldekranz <tobias@waldekranz.com>
+---
+ common/cli_hush.c | 7 -------
+ 1 file changed, 7 deletions(-)
+
+diff --git a/common/cli_hush.c b/common/cli_hush.c
+index 171069f5f4..d6d487893f 100644
+--- a/common/cli_hush.c
++++ b/common/cli_hush.c
+@@ -1796,13 +1796,6 @@ static int run_list_real(struct pipe *pi)
+ 	for (; pi; pi = (flag_restore != 0) ? rpipe : pi->next) {
+ 		if (pi->r_mode == RES_WHILE || pi->r_mode == RES_UNTIL ||
+ 			pi->r_mode == RES_FOR) {
+-#ifdef __U_BOOT__
+-				/* check Ctrl-C */
+-				ctrlc();
+-				if ((had_ctrlc())) {
+-					return 1;
+-				}
+-#endif
+ 				flag_restore = 0;
+ 				if (!rpipe) {
+ 					flag_rep = 0;
+-- 
+2.34.1
+