riotboot: add multislot support by kYc0o · Pull Request #10215 · RIOT-OS/RIOT

kYc0o · 2018-10-19T22:01:57Z

Contribution description

This PR adds the multislot capabilities for the riotboot bootloader. The bootloader now checks for the available slots headers and compares its version, the greatest version is kept and the corresponding slot is booted.

Testing procedure

To test the first slot:

BOARD=samr21-xpro APP_VER=$(date +%s) make -C tests/riotboot/ riotboot/flash-combined-slot0 test

For the second slot:

BOARD=samr21-xpro APP_VER=$(date +%s) make -C tests/riotboot/ riotboot/flash-slot1 test

Issues/PRs references

~~Depends on #10065~~

emmanuelsearch · 2018-12-18T14:07:49Z

@kYc0o needs a rebase on #10065

emmanuelsearch · 2018-12-18T14:08:45Z

(this PR seems to contain quite some commits that should not be here)

emmanuelsearch · 2018-12-18T18:28:46Z

cpu/cortexm_common/Makefile.include

+SLOT0_LEN ?= $(shell printf "0x%x" $$((($(ROM_LEN:%K=%*1024)-$(RIOTBOOT_LEN)) / $(NUM_SLOTS))))
+SLOT1_LEN ?= $(SLOT0_LEN)
 SLOT0_LEN := $(SLOT0_LEN)
+SLOT1_LEN := $(SLOT1_LEN)


@kYc0o @cladmi are lines 37-38 not redundant with what precedes?

emmanuelsearch · 2018-12-18T21:00:32Z

bootloaders/riotboot/main.c

+    int slot = -1;

-    if (slot_util_validate(slot) == 0) {
+    for (unsigned i = 0; i < slot_util_num_slots; i++) {


use riotboot_slot_numof instead?

emmanuelsearch · 2018-12-18T21:01:51Z

bootloaders/riotboot/main.c


-    if (slot_util_validate(slot) == 0) {
+    for (unsigned i = 0; i < slot_util_num_slots; i++) {
+        const riot_hdr_t *riot_hdr = slot_util_get_hdr(i);


riotboot_hdr_t and riotboot_slot_get_hdr instead

emmanuelsearch · 2018-12-18T21:02:34Z

bootloaders/riotboot/main.c

-    if (slot_util_validate(slot) == 0) {
+    for (unsigned i = 0; i < slot_util_num_slots; i++) {
+        const riot_hdr_t *riot_hdr = slot_util_get_hdr(i);
+        if (slot_util_validate(i)) {


riotboot_slot_validate instead

emmanuelsearch · 2018-12-18T21:03:15Z

bootloaders/riotboot/main.c

+            /* skip slot if metadata broken */
+            continue;
+        }
+        if (riot_hdr->start_addr != slot_util_get_image_startaddr(i)) {


riotboot_slot_get_image_startaddr instead

emmanuelsearch · 2018-12-18T21:03:42Z

bootloaders/riotboot/main.c

+    }
+
+    if (slot != -1) {
        slot_util_jump(slot);


riotboot_slot_jump instead

When a second slot is defined, the "partition table" gets a second field which represents the starting of the second slot, at a defined offset.

riotboot looks for valid, available slots and compares its version. The slot with the highest version is booted, otherwise if no valid slot is found it loops on `while(1);`

By erasing slot 1 header the slot gets invalidated. This is very useful while debugging, since we can force the bootloader to ignore anything on that slot.

A second slot is defined with a calculated size, from the remaining flash after the bootloader and the first slot. Both slots are defined as equal size, but it can be overriden.

kYc0o · 2019-01-02T17:15:11Z

Rebased to current master and solved issues. Should be ready for review.

emmanuelsearch · 2019-01-08T17:42:17Z

@kYc0o great!
For wannabe testers, how about slightly more documentation (in the README?)
For instance a concrete example of commands which

build and flash bootloader + an image in slot0
how to observe that slot0 is booted
build and flash a (distinguishable) image in slot1
how to observe that slot1 is booted

emmanuelsearch · 2019-01-08T20:20:33Z

@kYc0o testing on samr21: even if I flash a newer image in slot 1, it still boots slot 0. Can you reproduce the issue?

danpetry · 2019-01-09T11:31:33Z

It boots slot 1, for me...
Concretely, I did the following:

Plug in samr21
Run BOARD=samr21-xpro APP_VER=$(date +%s) make -C tests/riotboot/ riotboot/flash-combined-slot0 test from RIOT base dir
Observe "Current slot = 0" in output
Run BOARD=samr21-xpro APP_VER=$(date +%s) make -C tests/riotboot/ riotboot/flash-slot1 test from RIOT base dir
Observe "Current slot = 1" in output

However, it's not running pyexpect, it's just showing me the output. I'll have a closer look when I do a thorough review

emmanuelsearch · 2019-01-14T10:51:18Z

@kYc0o @danpetry tested again, can't reproduce my issue, so let's go forward as is for now, I'd say.

emmanuelsearch · 2019-01-14T10:57:28Z

tested to work on samr21 and iotlab-m3

kaspar030 · 2019-01-14T11:57:32Z

tests/riotboot/tests/01-run.py

    # Ask for current slot, should be 0 (riotboot slot)
    child.sendline("curslotnr")
-    child.expect_exact("Current slot=0")
+    child.expect("Current slot=[0-1]")


IMO if the regexp is changed, the comment above needs changing, too.

kaspar030 · 2019-01-14T11:58:41Z

Looks good to me. (No time to test, though).
@emmanuelsearch did you test the individual make tagets (flash0-combined, extended, ...)?

emmanuelsearch · 2019-01-14T12:17:10Z

@kaspar030 yes I tested the individual make targets, works like a charm.

emmanuelsearch · 2019-01-14T13:53:46Z

This PR was mislabeled as "major", which obviously it is not, hence removed the label.

Modify the test to accept slot values 0 and 1.

kYc0o · 2019-01-14T17:32:57Z

Updated and force-pushed.

emmanuelsearch · 2019-01-15T17:12:41Z

@kYc0o force-pushed some hopefully even better documentation.

kYc0o · 2019-01-15T17:32:55Z

Thanks @emmanuelsearch ! Yes, it looks better, however your commit introduced some whitespaces.

emmanuelsearch · 2019-01-15T17:45:22Z

@kYc0o I just fixed that hopefuly

bergzand · 2019-01-15T18:10:57Z

🎉 !

kYc0o · 2019-01-15T18:25:28Z

Yay! Thanks to everyone!

fedepell · 2019-01-16T03:21:35Z

Great stuff! Sorry I arrived a bit late ;)

Just tested a bit on my saml21-xpro everything works out of the box!

kYc0o requested review from cladmi, emmanuelsearch and kaspar030 October 19, 2018 22:01

emmanuelsearch requested changes Dec 18, 2018

View reviewed changes

emmanuelsearch reviewed Dec 18, 2018

View reviewed changes

Francisco Acosta and others added 4 commits January 2, 2019 17:56

riotboot_slot: add second slot compatibility

8f22175

When a second slot is defined, the "partition table" gets a second field which represents the starting of the second slot, at a defined offset.

riotboot: add support for multislot

ec24ea4

riotboot looks for valid, available slots and compares its version. The slot with the highest version is booted, otherwise if no valid slot is found it loops on `while(1);`

riotboot: allow to overwrite slot 1

8a0a0be

By erasing slot 1 header the slot gets invalidated. This is very useful while debugging, since we can force the bootloader to ignore anything on that slot.

cpu/cortex_common: add support for multislot

e8660b2

A second slot is defined with a calculated size, from the remaining flash after the bootloader and the first slot. Both slots are defined as equal size, but it can be overriden.

kYc0o force-pushed the pr/riotboot_multislot branch from 0b5f206 to e613cbf Compare January 2, 2019 17:13

kYc0o added CI: ready for build If set, CI server will compile all applications for all available boards for the labeled PR and removed CI: ready for build If set, CI server will compile all applications for all available boards for the labeled PR labels Jan 3, 2019

danpetry self-requested a review January 9, 2019 09:40

basilfx mentioned this pull request Jan 9, 2019

sys: stdio_null: add null driver #10741

Merged

kYc0o mentioned this pull request Jan 13, 2019

[WIP] [TF:OTA] OTA updates common work branch #9969

Closed

kaspar030 reviewed Jan 14, 2019

View reviewed changes

emmanuelsearch removed the Impact: major The PR changes a significant part of the code base. It should be reviewed carefully label Jan 14, 2019

emmanuelsearch added the Reviewed: 3-testing The PR was tested according to the maintainer guidelines label Jan 14, 2019

tests/riotboot: extend to accept slot 0 or 1

8e5ac62

Modify the test to accept slot values 0 and 1.

kYc0o force-pushed the pr/riotboot_multislot branch from e613cbf to e0c73a2 Compare January 14, 2019 17:32

emmanuelsearch force-pushed the pr/riotboot_multislot branch from e0c73a2 to bee0af3 Compare January 15, 2019 15:08

emmanuelsearch added the Reviewed: 5-documentation The documentation details of the PR were reviewed according to the maintainer guidelines label Jan 15, 2019

emmanuelsearch approved these changes Jan 15, 2019

View reviewed changes

bootloaders/riotboot: updated documentation on multislot

d8f64a3

emmanuelsearch force-pushed the pr/riotboot_multislot branch from bee0af3 to d8f64a3 Compare January 15, 2019 17:32

emmanuelsearch merged commit 209d90b into RIOT-OS:master Jan 15, 2019

kYc0o deleted the pr/riotboot_multislot branch January 21, 2019 10:56

aabadie added this to the Release 2019.01 milestone Jan 21, 2019

Conversation

kYc0o commented Oct 19, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Contribution description

Testing procedure

Issues/PRs references

Uh oh!

emmanuelsearch commented Dec 18, 2018

Uh oh!

emmanuelsearch commented Dec 18, 2018

Uh oh!

emmanuelsearch Dec 18, 2018

Choose a reason for hiding this comment

Uh oh!

emmanuelsearch Dec 18, 2018

Choose a reason for hiding this comment

Uh oh!

emmanuelsearch Dec 18, 2018

Choose a reason for hiding this comment

Uh oh!

emmanuelsearch Dec 18, 2018

Choose a reason for hiding this comment

Uh oh!

emmanuelsearch Dec 18, 2018

Choose a reason for hiding this comment

Uh oh!

emmanuelsearch Dec 18, 2018

Choose a reason for hiding this comment

Uh oh!

kYc0o commented Jan 2, 2019

Uh oh!

emmanuelsearch commented Jan 8, 2019

Uh oh!

emmanuelsearch commented Jan 8, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

danpetry commented Jan 9, 2019

Uh oh!

emmanuelsearch commented Jan 14, 2019

Uh oh!

emmanuelsearch commented Jan 14, 2019

Uh oh!

kaspar030 Jan 14, 2019

Choose a reason for hiding this comment

Uh oh!

kaspar030 commented Jan 14, 2019

Uh oh!

emmanuelsearch commented Jan 14, 2019

Uh oh!

emmanuelsearch commented Jan 14, 2019

Uh oh!

kYc0o commented Jan 14, 2019

Uh oh!

emmanuelsearch commented Jan 15, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

kYc0o commented Jan 15, 2019

Uh oh!

emmanuelsearch commented Jan 15, 2019

Uh oh!

bergzand commented Jan 15, 2019

Uh oh!

kYc0o commented Jan 15, 2019

Uh oh!

fedepell commented Jan 16, 2019

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

7 participants

kYc0o commented Oct 19, 2018 •

edited

Loading

emmanuelsearch commented Jan 8, 2019 •

edited

Loading

emmanuelsearch commented Jan 15, 2019 •

edited

Loading