Commit Graph

48562 Commits (master)
 

Author SHA1 Message Date
Adrian Schmutzler 3635f7efc1 ipq806x: really fix aliases node for Qualcomm IPQ8064/DB149
Having looked at this again, it appears that only gsbi2_serial
is actually enabled for this device, so the entry in the broken
aliases node was correct.

Therefore, this needs to set its own serial0 instead of inheriting
"serial0 = &gsbi4_serial;" from DTSI. Do this with the correctly
named aliases node now.

Fixes: c83f7b6d21 ("ipq806x: fix aliases node name for Qualcomm
IPQ8064/DB149")

Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
4 years ago
Adrian Schmutzler e363470d1a ipq806x: use consistent DT labels for gsbiX_serial
Kernel uses the label gsbiX_serial, so let's adjust our labels to
this naming scheme.

This is cosmetic, and actually only already existing gsbi4_serial
has been used at all.

Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
4 years ago
Adrian Schmutzler c83f7b6d21 ipq806x: fix aliases node name for Qualcomm IPQ8064/DB149
This device uses a node incorrectly named "alias" instead of
"aliases" since it was introduced.

Remove it without replacement, as the definitions in it don't
seem to be required anyway:

The serial0 definition has never been effective anyway and this
would be the only device deviating from the common setting
"serial0 = &gsbi4_serial;" for ipq8064. (So, maybe the wrong
node prevented us from finding out about the wrong serial
definition?)

The mdio-gpio0 alias was supposed to be removed in d2a2eb7e48
anyway, the redundant definition in the alias node was just
overlooked back then.

Fixes: 0fd202f3e5 ("ipq806x: add db149 dts files")
Fixes: d2a2eb7e48 ("ipq806x: replace caf nss-gmac driver by upstream stmmac")

Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
4 years ago
Daniel Bailey 5792f6a104 procd: allow optional watchdog instance parameter
Optional instance watchdog timeout and watchdog mode can be set by
adding: procd_set_param $mode $timeout

$mode is an integer [0-1] representing instance watchdog mode of
operation:
0 = disabled
1 = passive mode, client must periodically poke watchdog via ubus

$timeout is an integer representing how often, in seconds, the watchdog must be poked.

Signed-off-by: Daniel Bailey <danielb@meshplusplus.com>
4 years ago
Daniel Golle 732b70c5bd procd: update to git HEAD
639df57 uxc: fix build with uClibc-ng
 b2230e4 procd: add service instance watchdog

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
4 years ago
Roger Pueyo Centelles 26fda610f3 ath79: use rgmii-id instead of rgmii in ubnt,lap-120
Since commit 6f2e1b7485 (ath79: disable delays on AT803X config init)
the incoming incoming traffic on the ubnt,lap-120 devices Ethernet
port was not making it through. Using rgmii-id instead of rgmii (same
configuration as ubnt,litebeam-ac-gen2) fixes it.

Fixes FS#2893.

Signed-off-by: Roger Pueyo Centelles <roger.pueyo@guifi.net>
4 years ago
Paul Spooren 5b8b42c78b build,json: fix compatibility with Python 3.5
The f-string feature was introduced in Python 3.6. As Buildbots may run
on Debian 9, which comes per default with Python 3.5, this would cause
an issue. Instead of f-strings use the *legacy* `.format()` function.

Signed-off-by: Paul Spooren <mail@aparcar.org>
4 years ago
Mathieu Martin-Borret d7233b0a6a ramips: add support for D-Link DIR-878 A1
Specifications:

    SoC: MT7621AT
    RAM: 128MB
    Flash: 16MB NOR SPI flash
    WiFi: MT7615N (2.4GHz) and MT7615N (5Ghz)
    LAN: 5x1000M
    Firmware layout is Uboot with extra 96 bytes in header
    Base PCB is AP-MTKH7-0002
    LEDs Power Green,Power Orange,Internet Green,Internet Orange
    LEDs "2.4G" Green & "5G" Green connected directly to wifi module
    Buttons Reset,WPS,WIFI

Flashing instructions:

Upload image via emergency recovery mode
Push and hold reset button (on the back of the device) until power led
starts flashing (about 10 secs or so) while powering the device on.
Give it ~30 seconds, to boot the recovery mode GUI
Connect your client computer to LAN1 of the device
Set your client IP address manually to 192.168.0.2 / 255.255.255.0.
Call the recovery page for the device at http://192.168.0.1
Use the provided emergency web GUI to upload and flash a new firmware to
the device. Some browsers/OS combinations are known not to work, so if
you don't see the percentage complete displayed and moving within a few
seconds, restart the procedure from scratch and try anoher one,
or try the command line way.

Alternative method using command line on Linux:
curl -v -i -F "firmware=@openwrt-xxxx-squashfs-factory.bin" 192.168.0.1

Signed-off-by: Mathieu Martin-Borret <mathieu.mb@protonmail.com>
[use of generic uimage-padhdr in image generation code]
Signed-off-by: Petr Štetiar <ynezz@true.cz>
4 years ago
Mathieu Martin-Borret 96761ecc2f kernel: mtdsplit_uimage: add SGE parser
Adding SGE parser for SGE (T&W) Shenzhen Gongjin Electronics
to add 96 bytes padding the the firmware image

Signed-off-by: Mathieu Martin-Borret <mathieu.mb@protonmail.com>
4 years ago
Hans Dedecker a44eff6ad3 glibc: fix avx2 strncmp offset compare condition check [BZ #25933]
4e8a33a959 NEWS: Mention BZ 25933 fix
fd15ba932d Fix avx2 strncmp offset compare condition check [BZ #25933

Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
4 years ago
David Bauer 0485fb3925 rockchip: don't disable timer LED trigger
The timer LED trigger is enabled in all targets (except for lantiq
xway-legacy). It's necessary for the OpenWrt preinit LED pattern to
work.

Signed-off-by: David Bauer <mail@david-bauer.net>
4 years ago
David Bauer 875a04478c rockchip: use downstream GPIO button implementation
Use the OpenWrt netlink GPIO button implementation to forward button
presses to procd. This is necessary to make failsafe-mode access
using a button possible.

Signed-off-by: David Bauer <mail@david-bauer.net>
4 years ago
Daniel Golle 79fd87ef9d procd: update to git HEAD
aed7fb3 procd: fix compilation with uClibc-ng
 9d0f831 jail: fix segfault with len(uidmap/gidmap) > 1
 42a6217 jail: consider PATH for argv in OCI container
 83f4b72 jail: actually chdir into OCI defined CWD
 fc9f614 jail: parse and run OCI hooks
 02eec92 jail: memory allocation fixes
 71e75f4 jail: refactor mount support to cover OCI spec
 b586e7d jail: don't make mount source read-only
 dacab12 uxc: fix 'stop' command

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
4 years ago
Adrian Schmutzler eb15b0e0c9 bcm47xx: disable image for Linksys WRT160n v3
The device can only hold 3.6 MB, but newer images (since 18.06)
are bigger, so flashing a new version fails.

This disables default build for this device based on the bug report
referenced below.

Fixes: FS#1963

Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
4 years ago
Adrian Schmutzler 11478524a2 ramips: consolidate Netgear devices for mt7628
This creates a common DTSI and shared image definition for the
relatively similar Netgear devices for mt7628 platform.

As a side effect, this raises SPI flash frequency for the R6120,
as it's expected to work there as well if it works for R6080 and
R6020.

Based on the data from the other devices, it also seems probable
the 5g MAC address for R6120 could be extracted from the caldata,
and the mtd-mac-address there could be dropped.

Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
4 years ago
Jason A. Donenfeld ad81e2da08 wireguard: bump to 1.0.20200712
This release brings parity with the commits Linus released a few hours
ago into 5.8-rc5.

* receive: account for napi_gro_receive never returning GRO_DROP

The napi_gro_receive function no longer returns GRO_DROP ever, making
handling GRO_DROP dead code. This commit removes that dead code.
Further, it's not even clear that device drivers have any business in
taking action after passing off received packets; that's arguably out of
their hands.

* device: implement header_ops->parse_protocol for AF_PACKET

WireGuard uses skb->protocol to determine packet type, and bails out if
it's not set or set to something it's not expecting. For AF_PACKET
injection, we need to support its call chain of:

    packet_sendmsg -> packet_snd -> packet_parse_headers ->
      dev_parse_header_protocol -> parse_protocol

Without a valid parse_protocol, this returns zero, and wireguard then
rejects the skb. So, this wires up the ip_tunnel handler for layer 3
packets for that case.

* queueing: make use of ip_tunnel_parse_protocol

Now that wg_examine_packet_protocol has been added for general
consumption as ip_tunnel_parse_protocol, it's possible to remove
wg_examine_packet_protocol and simply use the new
ip_tunnel_parse_protocol function directly.

* compat: backport ip_tunnel_parse_protocol and ip_tunnel_header_ops

These are required for moving wg_examine_packet_protocol out of
wireguard and into upstream.

Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
4 years ago
Kevin Darbyshire-Bryant 8242c6de29 scripts/env: Fix 56f813674a scripts/env: use command -v instead of which
We don't need to see how git will be executed and it produces non silent
output on 'scripts/env diff' commands when there are no differences
unlike before.

Re-introduce original silent behaviour.

Signed-off-by: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
4 years ago
Chuanhong Guo 28893f5878 ath79: fix dsa binding for TP-Link TL-WR941ND v2
upstream changed dt-bindings for marvell 88e6060 to use mdio-device
and dropped support for legacy bindings.
fix it in our local dts.

Fixes: FS#2524
Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
4 years ago
Chuanhong Guo 0997342f24 ath79: qihoo,c301: use phy-supply for usb vbus
Linux phy subsystem provides support for a phy regulator defined via
phy-supply property. Use it to turn on usb power only when usb is
probed.

Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
4 years ago
Tomasz Maciej Nowak 2dc5ce622a mvebu: add support for MACCHIATObin Single Shot
Add support for Marvell MACCHIATObin Single Shot, cortex-a72 based
Marvell ARMADA 8040 Community board. Single Shot was broken as the
device tree is different on the Double Shot Board.

Specifications:
- Quad core Cortex-A72 (up to 2GHz)
- DDR4 DIMM slot with optional ECC and single/dual chip select support
- Dual 10GbE (1/2.5/10GbE) SFP+
  2.5GbE (1/2.5GbE) via SFP
  1GbE via copper
- SPI Flash
- 3 X SATA 3.0 connectors
- MicroSD connector
- eMMC
- PCI x4 3.0 slot
- USB 2.0 Headers (Internal)
- USB 3.0 connector
- Console port (UART) over microUSB connector
- 20-pin Connector for CPU JTAG debugger
- 2 X UART Headers
- 12V input via DC Jack
- ATX type power connector
- Form Factor: Mini-ITX (170 mm x 170 mm)

More details at http://macchiatobin.net

Installation:

Write the Image to your Micro SD Card and insert it in the
MACCHIATObin Single Shot SD Card Slot.

In the U-Boot Environment:
   1. reset U-Boot environment:
      env default -a
      saveenv

   2. prepare U-Boot with boot script:
      setenv bootcmd "load mmc 1:1 0x4d00000 boot.scr; source 0x4d00000"
      saveenv

   or manually (hanging lines indicate wrapped one-line command):
      setenv fdt_name armada-8040-mcbin-singleshot.dtb
      setenv image_name Image
      setenv bootcmd 'mmc dev 1; ext4load mmc 1:1 $kernel_addr
         $image_name;ext4load mmc 1:1 $fdt_addr $fdt_name;setenv
         bootargs $console root=/dev/mmcblk1p2 rw rootwait; booti
         $kernel_addr - $fdt_addr'
      saveenv

   On newer Bootloaders (18.12) the Variables have been changed, use:
      setenv fdt_name armada-8040-mcbin-singleshot.dtb
      setenv image_name Image
      setenv bootcmd 'mmc dev 1; ext4load mmc 1:1 $kernel_addr_r
         $image_name;ext4load mmc 1:1 $fdt_addr_r $fdt_name;setenv
         bootargs $console root=/dev/mmcblk1p2 rw rootwait; booti
         $kernel_addr_r - $fdt_addr_r'

Reported-by: Alexandra Alth <alexandra@alth.de>
Signed-off-by: Tomasz Maciej Nowak <tomek_n@o2.pl>
Tested-by: Alexandra Alth <alexandra@alth.de>
[add specs and installation as provided by Alexandra Alth]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
4 years ago
Adrian Schmutzler 52cdd6185e scripts/mkits.sh: fix use of printf
Due to a line break, printf was accidentally called with three
arguments instead of two, causing a different output than before.

Fix it by splitting the printf command into two lines.

Fixes: 907053193a ("scripts/mkits.sh: replace echo -e with printf")

Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
4 years ago
Adrian Schmutzler 479adb0929 ramips: consolidate recipes with uimage_padhdr
There are already two very similar recipes using uimage_padhdr
in ramips target, and a third one is about to be added.

Make the recipe more generic, so redefinitions are not necessary
anymore.

Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Tested-by: Bjørn Mork <bjorn@mork.no> [Zyxel WAP6805]
4 years ago
Adrian Schmutzler 9faf9f8f23 mvebu: fix support for Marvell 8040 MACCHIATOBin
Between kernels 4.20 and 5.0, a new variant of this board has been
introduced ("Single Shot"), and the existing one has been renamed
with the appendix "Double Shot". [1]
This also adjusted the first compatible in the list:

marvell,armada8040-mcbin -> marvell,armada8040-mcbin-doubleshot

This patch updates the OpenWrt implementation of this device by
adjusting the relevant references to that compatible (i.e., our
board name).

To still provide support for 4.19 with our setup, this adds a
small patch to change the compatible there as well.

[1] https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=b1f0bbe2700051886b954192b6c1751233fe0f52

Cc: Tomasz Maciej Nowak <tomek_n@o2.pl>
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Reviewed-by: Tomasz Maciej Nowak <tomek_n@o2.pl>
4 years ago
Daniel González Cabanelas cef9e5a49f bcm63xx: R5010UNv2: fix flash partitions for 16MB flash
The router Nucom R5010UN v2 has the partitions defined for a 8MB flash,
but the flash chip is 16MB size. We are wasting half of the flash.

Fix it and use generic names for partitions.

Fixes: 474cde6123 ("brcm63xx: probe SPI flash through DT")

Signed-off-by: Daniel González Cabanelas <dgcbueu@gmail.com>
4 years ago
Roman Yeryomin 2ca084ccaa build: improve ccache support
Set CCACHE_DIR to $(TOPDIR)/.ccache and CCACHE_BASEDIR to $(TOPDIR).
This allows to do clean and dirclean. Cache hit rate for test build
after dirclean is ~65%.
If CCACHE is enabled stats are printed out at the end of building process.
CCACHE_DIR config variable allows to override default, which could be useful
when sharing cache with many builds.
cacheclean make target allows to clean the cache.

Changes from v1:
- remove ccache directory using CCACHE_DIR variable
- remove ccache leftovers from sdk and toolchain make files
- introduce CONFIG_CCACHE_DIR variable
- introduce cacheclean make target

Signed-off-by: Roman Yeryomin <roman@advem.lv>
4 years ago
DENG Qingfang 23de93a521 ramips: mt7621: refactor set affinity script
The current one only looks for mt76x2e and mt7603e, and
does not work for 2 or more same Wi-Fi chips.
Refactor the script to cover those cases.

Signed-off-by: DENG Qingfang <dengqf6@mail2.sysu.edu.cn>
4 years ago
Sebastian Kemper ff8b8fee89 ltq-vmmc: update permission handling
The firmware is currently just copied. It can end up with o= on the
device (this is the case for voice_ar9_firmware.bin for instance).
Instead of copying it the Makefile is changed to use the macro
"$(INSTALL_DATA)" in order for the file to be world-readable.

While at it refactor the device node creation in the init script with
loop.

Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
[removed 2nd part with custom group handling for device nodes]
Signed-off-by: Petr Štetiar <ynezz@true.cz>
4 years ago
Yangbo Lu 2a5460b11b layerscape: define IMAGE_SIZE and LS_SYSUPGRADE_IMAGE_SIZE
Define 64m IMAGE_SIZE for flash firmware.bin since the flash size
is 64MB. Define 48m LS_SYSUPGRADE_IMAGE_SIZE for flash sysupgrade.bin
which contains maximum 16MB kernel and 32MB rootfs according to
memory map.

Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
4 years ago
Yangbo Lu 0841b68c91 layerscape: support sysupgrade for SD card ext4 rootfs
Support sysupgrade for SD card ext4 rootfs.

Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
4 years ago
Yangbo Lu c4d0e57e16 layerscape: support sysupgrade for squashfs rootfs
Support sysupgrade for all Layerscape boards with squashfs
rootfs.

Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
4 years ago
Yangbo Lu 3a99b624d8 layerscape: fix board name for SD card boot
Many Layerscape boards support both flash boot and SD card boot.
And different firmware and sysupgrade.bin are built for the two
boot methods. To identify them, a fix could be done on board name
by adding a postfix "-sdboot" for SD card boot.

Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
4 years ago
Yangbo Lu 3a3b9ded50 layerscape: add compatible for board ls1043ardb
Backport a dts patch which added compatible for ls1043ardb.

Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
4 years ago
Yangbo Lu 5ca8903631 layerscape: convert to squashfs rootfs for QSPI NOR boot
There had been an issue in Layerscape QSPI driver for very long
time, which made squashfs,jffs2 rootfs not work on QSPI NOR.
And the ubifs had been used as a workaround.
Now the issue has been fixed. So convert to use squashfs,jffs2
rootfs on QSPI NOR for Layerscape boards (LS1012ARDB/LS1046ARDB/
LS1088ARDB), and update u-boot bootargs for booting.

Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
4 years ago
Rui Salvaterra eb24a57b7f busybox: store applet usage messages uncompressed
The rootfs squashfs is already highly (XZ) compressed. Storing the applet
messages in compressed form will increase the entropy and reduce the overall
image compression ratio.

Size diffs (compressed vs uncompressed):

busybox (the executable): 364596 vs 384804 bytes.

OpenWrt target images (the kernel image is unchanged, obviously):

 omnia-medkit-openwrt-mvebu-cortexa9-cznic_turris-omnia-initramfs.tar.gz:

  9163597 vs 9162531 bytes (1066 bytes difference).

openwrt-mvebu-cortexa9-cznic_turris-omnia-initramfs-kernel.bin:

  9161688 vs 9160600 bytes (1088 bytes difference).

openwrt-mvebu-cortexa9-cznic_turris-omnia-sysupgrade.img.gz:

 9729550 vs 9729230 bytes (320 bytes difference).

All in all, we save just a little bit over 1 kiB. As an added bonus, we
also don't have to decompress the messages twice, (first from squashfs,
then from the bzip2 message storage).

Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
[added additional size comparision diff detaisl]
Signed-off-by: Petr Štetiar <ynezz@true.cz>
4 years ago
Rosen Penev 8e6f2c029f base-files/functions.sh: use command -v instead of which
which must be executed. command -v is a shell builtin.

https://github.com/koalaman/shellcheck/wiki/SC2230

Signed-off-by: Rosen Penev <rosenp@gmail.com>
4 years ago
Petr Štetiar dec9ae6d9e scripts/mkits.sh: fix remaining shellcheck warning
Fixes following shellcheck warning:

 In scripts/mkits.sh line 19:
 		 "-k kernel [-D name -d dtb] -o its_file" "$(basename $0)"
                                                                       ^-- SC2086: Double quote to prevent globbing and word splitting.

Signed-off-by: Petr Štetiar <ynezz@true.cz>
4 years ago
Rosen Penev 3d418dd51e scripts/mkits.sh: switch from bash to sh
This no longer needs bash.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
4 years ago
Rosen Penev 7aefc85cab scripts/mkits.sh: fix improper string and array concatenation
Found with shellcheck.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
4 years ago
Rosen Penev 765858f7ad scripts/mkits.sh: add missing quotes
Found with shellcheck.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
4 years ago
Rosen Penev 907053193a scripts/mkits.sh: replace echo -e with printf
echo flags are not POSIX. printf does the same with added \n.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
4 years ago
Petr Štetiar 92d7cedcad scripts/gen_image_generic.sh: fix more shellcheck warnings
Fixes following shellcheck warnings:

 In scripts/gen_image_generic.sh line 20:
 cyl=$(( (KERNELSIZE + ROOTFSSIZE) * 1024 * 1024 / (head * sect * 512)))
 ^-^ SC2034: cyl appears unused. Verify use (or export if used externally).
 --
 In scripts/gen_image_generic.sh line 34:
     [ -n "$PADDING" ] && dd if=/dev/zero of="$OUTPUT" bs=512 seek="$(($ROOTFSOFFSET + $ROOTFSSIZE))" conv=notrunc count="$sect"
                                                                       ^-----------^ SC2004: $/${} is unnecessary on arithmetic variables.
 --
 In scripts/gen_image_generic.sh line 35:
     mkfs.fat -n kernel -C "$OUTPUT.kernel" -S 512 "$(($KERNELSIZE / 1024))"
                                                       ^---------^ SC2004: $/${} is unnecessary on arithmetic variables.

Signed-off-by: Petr Štetiar <ynezz@true.cz>
4 years ago
Rosen Penev aee169b06d scripts/gen_image_generic.sh: use /bin/sh
This has nothing bash specific.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
4 years ago
Rosen Penev 3819337cdf scripts/gen_image_generic.sh: replace -o with if/&&
-o is not well defined.

Found with shellcheck.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
4 years ago
Petr Štetiar 227e47b3ca scripts/env: fix remaining shellcheck warning
Fixes following shellcheck warning:

 In scripts/env line 25:
 	exit ${1:-1}
             ^-----^ SC2086: Double quote to prevent globbing and word splitting.

Signed-off-by: Petr Štetiar <ynezz@true.cz>
4 years ago
Rosen Penev 5b7cace9e3 scripts/env: use read -r instead of read
read mangles backslashes.

Found with shellcheck.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
4 years ago
Rosen Penev f80a540dbf scripts/env: exit in case of failure to cd
Found with shellcheck.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
4 years ago
Rosen Penev e68810ba53 scripts/env: replace \! with !
The latter is more standard. The former throws an error under
shellcheck.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
4 years ago
Rosen Penev 56f813674a scripts/env: use command -v instead of which
Simpler and built in to the shell.

Found with shellcheck.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
4 years ago
Eneas U de Queiroz 11192484fa build: reduce number of files passed to ipk-remove
Instead of using xargs to pass a huge number of files to
script/ipkg-remove, which will usually pick only one, use a more
restrictive wildcard so that, currently, at the most 325 files are
examined, instead of up to over 2,300.  The 325-file package is python,
which is picking up python3* ipks.  It is about to be removed.
Runner-up is ddns-scripts with 7 files.

This makes a second run of make package/luci/compile go from
real 16.40s; user 17.42s; sys 2.73s
to
real 10.71s; user 9.51s; sys 1.27s

There is a caveat though: if one were to remove the ABI_VERSION of a
package that ends in a digit [0-9], then the old package ipk will not be
removed from the bin directory by make package/abc2/clean.

Signed-off-by: Eneas U de Queiroz <cotequeiroz@gmail.com>
4 years ago
Eneas U de Queiroz 78d1f3ac61 build: call ipkg-remove using xargs if #args>=512
The wildcard call to clean up luci package (luci*) can pick up over
2,300 files when the full tree is built. Running make package/luci/clean
or a second run of make package/luci/compile would fail with an
'Argument list too long' error.

To avoid that, a maybe_use_xargs function was created that runs the
command straight as usual if the number of arguments is < 512, or saves
the list in a temporary file and feeds it to xargs otherwise.

Signed-off-by: Eneas U de Queiroz <cotequeiroz@gmail.com>
4 years ago