Commit Graph

67471 Commits

Author SHA1 Message Date
68eceded45
fixing things 2026-06-17 05:43:54 -04:00
e88d27b264
cal phy 12->13 2026-06-17 05:25:32 -04:00
9a2e403f67
fixing things 2026-06-17 04:32:23 -04:00
c9b4ef916a
fixing things 2026-06-17 04:14:26 -04:00
cf7f16b5cc
fuck me if leo is hallucinating :3 2026-06-17 03:47:37 -04:00
3ad9b58fbf
all ports exist, but crash on bringup lan3-8 2026-06-17 03:23:41 -04:00
bf63abe311
the fix? 2026-06-16 13:59:02 -04:00
b24d4f7e1e
the fix? 2026-06-16 12:34:44 -04:00
7e74b3ded9
the fix? 2026-06-16 12:25:22 -04:00
d7fba5c9d1
removed delete-node on our CPU PORT >.< 2026-06-16 10:52:19 -04:00
3f418cea29
whoops... status okay 2026-06-16 10:13:23 -04:00
53740be46a
reconfigured gmac & switches 2026-06-16 09:57:20 -04:00
19732afd10
switch work 2026-06-16 09:29:10 -04:00
cb230207f5
switch work 2026-06-16 08:06:16 -04:00
a9693fa8ff
switch work 2026-06-16 05:24:45 -04:00
f1f4940207
brain + 10 wrinkles added mdio pins assumed from hardware lineage 2026-06-16 01:10:09 -04:00
66831e86ca
compileable, unsure if working 2026-06-15 23:42:31 -04:00
0a66e440fe
commit first switch attempt, starting point 2026-06-15 22:20:07 -04:00
9b04d9eb3a
brain -5 wrinkles from ai correcting human 2026-06-15 22:04:54 -04:00
2492345313
brain +15 wrinkles from inline thinking breakthrough - re: reverse engineering / parallel reconstruction 2026-06-15 20:54:02 -04:00
db7494a98c
brain +15 wrinkles from inline thinking breakthrough - re: reverse engineering / parallel reconstruction 2026-06-15 20:50:19 -04:00
055763f312
brain +15 wrinkles from inline thinking breakthrough - re: reverse engineering / parallel reconstruction 2026-06-15 20:44:08 -04:00
fae78f3823
brain +15 wrinkles from inline thinking breakthrough - re: reverse engineering / parallel reconstruction 2026-06-15 20:31:30 -04:00
ea18b36d36
brain +5 wrinkles from inline thinking breakthroughs 2026-06-15 20:01:43 -04:00
bc28daaf0e
brain +5 wrinkles from inline thinking breakthroughs 2026-06-15 19:37:03 -04:00
6640928e30
brain +5 wrinkles from inline thinking breakthroughs 2026-06-15 19:32:55 -04:00
380fe229c9
brain +5 wrinkles from inline thinking breakthroughs 2026-06-15 19:29:24 -04:00
412928ebd6
brain +5 wrinkles from inline thinking breakthroughs 2026-06-15 19:20:39 -04:00
99d113d0e2
brain +5 wrinkles from inline thinking breakthroughs 2026-06-15 19:15:13 -04:00
fc0e6d3d1a
brain +2 wrinkles from inline thinking breakthroughs 2026-06-15 18:52:16 -04:00
aff963d26d
brain +1 wrinkle from inline thinking breakthroughs 2026-06-15 18:49:04 -04:00
55046c13b5
inline thought breakthrough aqcuired... 2026-06-15 18:43:14 -04:00
5fb1bb5969
update to inline thinking... 2026-06-15 18:21:02 -04:00
e6031426e1
update to inline thinking... 2026-06-15 18:12:16 -04:00
718b3af114
update to inline thinking... 2026-06-15 18:09:11 -04:00
e43c15a7c3
more inline thinking... 2026-06-15 17:56:03 -04:00
6bf54e27df
thinking... 2026-06-14 23:45:38 -04:00
2cd51e474e
thinking... 2026-06-14 22:58:07 -04:00
7750904cb1
fixed art->art_part 2026-06-14 17:51:14 -04:00
23ac153d57
dts - wifi updated 2026-06-14 17:32:50 -04:00
edaff9f631
update 2026-06-14 00:32:37 -04:00
c39f3c1a89
updated qcom-ipq4019-tew-829dru.dts 2026-06-11 10:33:54 -04:00
f303fe6dbb
updated .gitignore, generic.mk 2026-06-11 09:51:02 -04:00
f6ec836ca9
updated .gitignore 2026-06-11 09:47:29 -04:00
9b130b63a2
updated dts - spi flash works 2026-06-11 09:46:43 -04:00
17f091d59d
added wifi back, removed nand nodes - crashing 2026-06-11 02:02:30 -04:00
c8480bdb10
updated dts 2026-06-10 15:35:33 -04:00
cad39f2d5b
commit dts 2026-06-04 17:42:51 -04:00
311573b886
restoring workspace... 2026-06-04 17:30:32 -04:00
Guoxin Pu
78c88ce188 image: fix per device targz rootfs wrong suffix and redundant images
In commit d89cb72c23 , a new rootfs type "targz" was introduced, to correctly pack the rootfs AFTER `DEVICE_PACKAGES` installation (unlike the old simple `rootfs.tar.gz`) .

The expected release artifact shall be a single corresponding tar gz release accompanying or relacing the old simple `rootfs.tar.gz`.

However, if one take a look at the v25.12 series release download pages, e.g. [25.12.4/x86/64](https://downloads.openwrt.org/releases/25.12.4/targets/x86/64/), one could see that there're now four release artifacts related to rootfs targz:

- [generic-targz-combined-efi.img.gz](https://downloads.openwrt.org/releases/25.12.4/targets/x86/64/openwrt-25.12.4-x86-64-generic-targz-combined-efi.img.gz)
- [generic-targz-combined.img.gz](https://downloads.openwrt.org/releases/25.12.4/targets/x86/64/openwrt-25.12.4-x86-64-generic-targz-combined.img.gz)
- [generic-targz-rootfs.img.gz](https://downloads.openwrt.org/releases/25.12.4/targets/x86/64/openwrt-25.12.4-x86-64-generic-targz-rootfs.img.gz)
- [rootfs.tar.gz](https://downloads.openwrt.org/releases/25.12.4/targets/x86/64/openwrt-25.12.4-x86-64-rootfs.tar.gz)

It's obvious the new `targz` release actually reuses the same `TARGET_ROOTFS_{TYPE}` handler same as `squashfs`, `ext4` and alike. And the three `generic-targz` img.gz contains the same expected tar gz either as their second partition, or as the whole image. This could be verified by the following script:

```bash
URL_PARENT=https://downloads.openwrt.org/releases/25.12.4/targets/x86/64/openwrt-25.12.4-x86-64-
for TYPE in combined-efi combined rootfs; do
        curl -L "${URL_PARENT}generic-targz-${TYPE}.img.gz" | gzip -cd > "${TYPE}.img"
done
for TYPE in combined-efi combined; do
        INFO=$(sfdisk -d "${TYPE}.img" | sed -n 's/^'"${TYPE}.img"'2 : start= \+\([0-9]\+\), size= \+\([0-9]\+\),.\+.\+$/\1 \2/p')
        dd if="${TYPE}.img" of="${TYPE}.tar.gz" bs=512 skip="${INFO%% *}" count="${INFO##* }"
done
cp rootfs.img rootfs.tar.gz
sha256sum combined-efi.tar.gz combined.tar.gz rootfs.tar.gz
file rootfs.tar.gz
```

Output:
```log
  % Total    % Received % Xferd  Average Speed  Time    Time    Time   Current
                                 Dload  Upload  Total   Spent   Left   Speed
100 13.12M 100 13.12M   0      0 194.6M      0                              0
  % Total    % Received % Xferd  Average Speed  Time    Time    Time   Current
                                 Dload  Upload  Total   Spent   Left   Speed
100 12.93M 100 12.93M   0      0 190.9M      0                              0
  % Total    % Received % Xferd  Average Speed  Time    Time    Time   Current
                                 Dload  Upload  Total   Spent   Left   Speed
100  7.13M 100  7.13M   0      0 257.6M      0                              0
GPT PMBR size mismatch (246304 != 246334) will be corrected by write.
The backup GPT table is corrupt, but the primary appears OK, so that will be used.
The backup GPT table is not on the end of the device.
212992+0 records in
212992+0 records out
109051904 bytes (109 MB, 104 MiB) copied, 0.258064 s, 423 MB/s
212992+0 records in
212992+0 records out
109051904 bytes (109 MB, 104 MiB) copied, 0.261621 s, 417 MB/s
18b98d3562dc3067ae095ee44d4ff3158cf84e89063c3df66c0ef6638922ba71  combined-efi.tar.gz
18b98d3562dc3067ae095ee44d4ff3158cf84e89063c3df66c0ef6638922ba71  combined.tar.gz
18b98d3562dc3067ae095ee44d4ff3158cf84e89063c3df66c0ef6638922ba71  rootfs.tar.gz
rootfs.tar.gz: gzip compressed data, max compression, from Unix, original size modulo 2^32 0
```

The checksum of the extracted, actually expected .tar.gz are all same. And if we peek inside its content, it's indeed what we expected:

```
> tar -tzf rootfs.tar.gz
./
./bin/
./bin/ash
./bin/board_detect
./bin/busybox
./bin/cat
./bin/chgrp
./bin/chmod
./bin/chown
./bin/config_generate
./bin/cp
./bin/date
...
```

A disk image with a `tar.gz` as its second partiton doesn't boot anyway and I doubt if there's any mechanic to make it bootable. So `generic-targz-combined-efi.img.gz` and `generic-targz-combined.img.gz` are not needed at all, and `generic-targz-rootfs.img.gz` shall be renamed to have a `tar.gz` suffix instead.

Therefore work around this by skipping creating images for fs `targz` in the general loop, and only create the rootfs.tar.gz later for `targz` "fs"

This affects both the real builder and imagebuilder. I've tested this with multiple imagebuilders.

Tested with `openwrt-imagebuilder-25.12.4-mvebu-cortexa53.Linux-x86_64`:

```sh
rm -rf bin && make image PROFILE=generic
```

Without the fix:

```
> tree bin/
bin/
└── targets
    └── x86
        └── 64
            ├── openwrt-25.12.4-x86-64-generic.bom.cdx.json
            ├── openwrt-25.12.4-x86-64-generic-ext4-combined-efi.img.gz
            ├── openwrt-25.12.4-x86-64-generic-ext4-combined.img.gz
            ├── openwrt-25.12.4-x86-64-generic-ext4-rootfs.img.gz
            ├── openwrt-25.12.4-x86-64-generic-kernel.bin
            ├── openwrt-25.12.4-x86-64-generic.manifest
            ├── openwrt-25.12.4-x86-64-generic-rootfs.tar.gz
            ├── openwrt-25.12.4-x86-64-generic-squashfs-combined-efi.img.gz
            ├── openwrt-25.12.4-x86-64-generic-squashfs-combined.img.gz
            ├── openwrt-25.12.4-x86-64-generic-squashfs-rootfs.img.gz
            ├── openwrt-25.12.4-x86-64-generic-targz-combined-efi.img.gz
            ├── openwrt-25.12.4-x86-64-generic-targz-combined.img.gz
            ├── openwrt-25.12.4-x86-64-generic-targz-rootfs.img.gz
            ├── profiles.json
            └── sha256sums
```

With the fix:
```
> tree bin/
bin/
└── targets
    └── x86
        └── 64
            ├── openwrt-25.12.4-x86-64-generic.bom.cdx.json
            ├── openwrt-25.12.4-x86-64-generic-ext4-combined-efi.img.gz
            ├── openwrt-25.12.4-x86-64-generic-ext4-combined.img.gz
            ├── openwrt-25.12.4-x86-64-generic-ext4-rootfs.img.gz
            ├── openwrt-25.12.4-x86-64-generic-kernel.bin
            ├── openwrt-25.12.4-x86-64-generic.manifest
            ├── openwrt-25.12.4-x86-64-generic-rootfs.tar.gz
            ├── openwrt-25.12.4-x86-64-generic-squashfs-combined-efi.img.gz
            ├── openwrt-25.12.4-x86-64-generic-squashfs-combined.img.gz
            ├── openwrt-25.12.4-x86-64-generic-squashfs-rootfs.img.gz
            ├── openwrt-25.12.4-x86-64-generic-targz-rootfs.tar.gz
            ├── profiles.json
            └── sha256sums
```

And with `openwrt-imagebuilder-25.12.4-mvebu-cortexa53.Linux-x86_64`:

```sh
rm -rf bin && make image PROFILE=ripe_atlas-v5
```

Without the fix:
```
> tree bin/
bin/
└── targets
    └── mvebu
        └── cortexa53
            ├── openwrt-25.12.4-mvebu-cortexa53-ripe_atlas-v5.bom.cdx.json
            ├── openwrt-25.12.4-mvebu-cortexa53-ripe_atlas-v5-ext4-sdcard.img.gz
            ├── openwrt-25.12.4-mvebu-cortexa53-ripe_atlas-v5.manifest
            ├── openwrt-25.12.4-mvebu-cortexa53-ripe_atlas-v5-rootfs.tar.gz
            ├── openwrt-25.12.4-mvebu-cortexa53-ripe_atlas-v5-squashfs-sdcard.img.gz
            ├── openwrt-25.12.4-mvebu-cortexa53-ripe_atlas-v5-targz-sdcard.img.gz
            ├── profiles.json
            └── sha256sums
```

With the fix:
```
> tree bin/
bin/
└── targets
    └── mvebu
        └── cortexa53
            ├── openwrt-25.12.4-mvebu-cortexa53-ripe_atlas-v5.bom.cdx.json
            ├── openwrt-25.12.4-mvebu-cortexa53-ripe_atlas-v5-ext4-sdcard.img.gz
            ├── openwrt-25.12.4-mvebu-cortexa53-ripe_atlas-v5.manifest
            ├── openwrt-25.12.4-mvebu-cortexa53-ripe_atlas-v5-rootfs.tar.gz
            ├── openwrt-25.12.4-mvebu-cortexa53-ripe_atlas-v5-squashfs-sdcard.img.gz
            ├── openwrt-25.12.4-mvebu-cortexa53-ripe_atlas-v5-targz-rootfs.tar.gz
            ├── profiles.json
            └── sha256sums
```

Signed-off-by: Guoxin Pu <pugokushin@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/23570
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 246b216df5)
2026-05-29 13:11:15 +02:00