This fixes regressed support for Tenda AC9 which got broken somewhere
between LEDE 17.01 and OpenWrt 18.06.
At some point bcm47xx_nvram_gpio_pin() started working on bcm53xx
devices and it resulted in swconfig b53 performing GPIO-based switch
reset. It exposed some bug as switch stopped working.
The problem is most likely to be some missing initialization step. Tenda
AC9 comes with BCM53125 rev 4 (pretty common chip) but apparently needs
some extra quirk.
As a workaround (it doesn't make much sense to debug swconfig driver
which has a working DSA replacement) simply avoid full switch reset.
So far every build of a single bcm53xx Target Profile (it means: when
NOT using CONFIG_TARGET_MULTI_PROFILE) resulted in all target devices
images being built. Now it only builds the one matching selected
profile.
Fixes: #13572 Suggested-by: Jonas Gorski <jonas.gorski@gmail.com> Signed-off-by: Rani Hod <rani.hod@gmail.com>
[rmilecki: update commit subject + body & move PROFILES line] Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
(cherry picked from commit 802a5f5cb4a7b42d25e82b787d7ab1323a20183f)
Arınç ÜNAL [Thu, 10 Aug 2023 12:23:09 +0000 (15:23 +0300)]
bcm53xx: add support for ASUS RT-AC3100
ASUS RT-AC3100 is ASUS RT-AC88U without the external switch.
OpenWrt forum users effortless and ktmakwana have confirmed that there are
revisions with either 4366b1 or 4366c0 wireless chips.
Therefore, include firmware for 4366b1 along with 4366c0. This way, all
hardware revisions of the router will be supported by having brcmfmac use
the firmware file for the wireless chip it detects.
ipq40xx: switch to performance governor by default
Doing a simple ping to my device shows this:
64 bytes from 10.0.253.101: icmp_seq=1 ttl=64 time=2.00 ms
64 bytes from 10.0.253.101: icmp_seq=2 ttl=64 time=2.02 ms
64 bytes from 10.0.253.101: icmp_seq=3 ttl=64 time=1.68 ms
64 bytes from 10.0.253.101: icmp_seq=4 ttl=64 time=1.91 ms
64 bytes from 10.0.253.101: icmp_seq=5 ttl=64 time=1.92 ms
64 bytes from 10.0.253.101: icmp_seq=6 ttl=64 time=2.04 ms
Some users even report higher values on older kernels:
64 bytes from 192.168.1.10: seq=0 ttl=64 time=0.612 ms
64 bytes from 192.168.1.10: seq=1 ttl=64 time=2.852 ms
64 bytes from 192.168.1.10: seq=2 ttl=64 time=2.719 ms
64 bytes from 192.168.1.10: seq=3 ttl=64 time=2.741 ms
64 bytes from 192.168.1.10: seq=4 ttl=64 time=2.808 ms
The problem is that the governor is set to Ondemand, which causes
the CPU to clock all the way down to 48MHz in some cases.
Switching to performance governor:
64 bytes from 10.0.253.101: icmp_seq=1 ttl=64 time=0.528 ms
64 bytes from 10.0.253.101: icmp_seq=2 ttl=64 time=0.561 ms
64 bytes from 10.0.253.101: icmp_seq=3 ttl=64 time=0.633 ms
64 bytes from 10.0.253.101: icmp_seq=4 ttl=64 time=0.526 ms
In theory, using the Performance governor should increase power draw,
but it looks like it really does not matter for this soc.
Using a calibrated precision DC power supply (cpu idle):
Hauke Mehrtens [Wed, 7 Jun 2023 20:37:47 +0000 (22:37 +0200)]
openssl: bump to 1.1.1u
Major changes between OpenSSL 1.1.1t and OpenSSL 1.1.1u [30 May 2023]
o Mitigate for very slow `OBJ_obj2txt()` performance with gigantic
OBJECT IDENTIFIER sub-identities. (CVE-2023-2650)
o Fixed documentation of X509_VERIFY_PARAM_add0_policy() (CVE-2023-0466)
o Fixed handling of invalid certificate policies in leaf certificates
(CVE-2023-0465)
o Limited the number of nodes created in a policy tree ([CVE-2023-0464])
Jonas Gorski [Sat, 9 Sep 2023 10:44:42 +0000 (12:44 +0200)]
x86: geode: fix hwrng register accesses
When the membase and pci_dev pointer were moved to a new struct in priv,
the actual membase users were left untouched, and they started reading
out arbitrary memory behind the struct instead of registers. This
unfortunately turned the RNG into a constant number generator, depending
on the content of what was at that offset.
To fix this, update geode_rng_data_{read,present}() to also get the
membase via amd_geode_priv, and properly read from the right addresses
again.
Closes #13417.
Reported-by: Timur I. Davletshin <timur.davletshin@gmail.com> Tested-by: Timur I. Davletshin <timur.davletshin@gmail.com> Suggested-by: Jo-Philipp Wich <jo@mein.io> Signed-off-by: Jonas Gorski <jonas.gorski@gmail.com>
(cherry picked from commit 09d13cd8d87cc50fde67bbe81c6cca4b799b2724)
What's interesting about jitterentropy-rngd v1.2.0 release is that it
bumps its copy of jitterentropy-library from v2.2.0 to the v3.0.0. That
bump includes a relevant commit 3130cd9 ("replace LSFR with SHA-3 256").
When initializing entropy jent calculates time delta. Time values are
obtained using clock_gettime() + CLOCK_REALTIME. There is no guarantee
from CLOCK_REALTIME of unique values and slow devices often return
duplicated ones.
A switch from jent_lfsr_time() to jent_hash_time() resulted in many less
cases of zero delta and avoids ECOARSETIME.
Long story short: on some system this fixes:
[ 6.722725] urngd: jent-rng init failed, err: 2
This is important change for BCM53573 which doesn't include hwrng and
seems to have arch_timer running at 36,8 Hz.
bcm53xx: add BCM53573 Ethernet fix sent upstream for v6.6
It seems that DSA-based b53 driver never worked with BCM53573 SoCs and
BCM53125.
In case of swconfig-based b53 this fixes a regression. Switching bgmac
from using mdiobus_register() to of_mdiobus_register() resulted in MDIO
device (BCM53125) having of_node set (see of_mdiobus_register_phy()).
That made downstream b53 driver read invalid data from DT and broke
Ethernet support.
Some device recipes remove default target packages. If user tries to add
them back they will be ignored, since packages list is processed in one
go. Process the device recipe packages first and do user ones later, so
additions won't get filtered out.
The DGND3700v2 renames the cferam bootloader from cferam to cfeXXX, where XXX
is the number of firmware upgrades performed by the bootloader. Other bcm63xx
devices rename cferam.000 to cferam.XXX, but this device is special because
the cferam name isn't changed on the first firmware flashing but it's changed
on the subsequent ones.
Therefore, we need to look for "cfe" instead of "cferam" to properly detect
the cferam partition and fix the bootlop.
Some devices rename cferam bootloader using specific patterns and don't follow
broadcom standards for renaming cferam files. This requires supporting
different cferam file names.
Apply two patches fixing low-severity vulnerabilities related to
certificate policies validation:
- Excessive Resource Usage Verifying X.509 Policy Constraints
(CVE-2023-0464)
Severity: Low
A security vulnerability has been identified in all supported versions
of OpenSSL related to the verification of X.509 certificate chains
that include policy constraints. Attackers may be able to exploit
this vulnerability by creating a malicious certificate chain that
triggers exponential use of computational resources, leading to a
denial-of-service (DoS) attack on affected systems.
Policy processing is disabled by default but can be enabled by passing
the `-policy' argument to the command line utilities or by calling the
`X509_VERIFY_PARAM_set1_policies()' function.
- Invalid certificate policies in leaf certificates are silently ignored
(CVE-2023-0465)
Severity: Low
Applications that use a non-default option when verifying certificates
may be vulnerable to an attack from a malicious CA to circumvent
certain checks.
Invalid certificate policies in leaf certificates are silently ignored
by OpenSSL and other certificate policy checks are skipped for that
certificate. A malicious CA could use this to deliberately assert
invalid certificate policies in order to circumvent policy checking on
the certificate altogether.
Policy processing is disabled by default but can be enabled by passing
the `-policy' argument to the command line utilities or by calling the
`X509_VERIFY_PARAM_set1_policies()' function.
Note: OpenSSL also released a fix for low-severity security advisory
CVE-2023-466. It is not included here because the fix only changes the
documentation, which is not built nor included in any OpenWrt package.
Due to the low-severity of these issues, there will be not be an
immediate new release of OpenSSL.
Signed-off-by: Eneas U de Queiroz <cotequeiroz@gmail.com>
Paul Spooren [Sun, 12 Mar 2023 15:56:41 +0000 (16:56 +0100)]
imagebuilder: allow to specific ROOTFS_PARTSIZE
Setting this options modifies the rootfs size of created images. When
installing a large number of packages it may become necessary to
increase the size to have enough storage.
This option is only useful for supported devices, i.e. with an attached
SD Card or installed on a hard drive.
The USB port on the MR8300 randomly fails to feed bus-powered devices.
This is caused by a misconfigured pinmux. The GPIO68 should be used to
enable the USB power (active low), but it's inside the NAND pinmux.
This GPIO pin was found in the original firmware at a startup script in
both MR8300 and EA8300. Therefore apply the fix for both boards.
Signed-off-by: Daniel González Cabanelas <dgcbueu@gmail.com> Reviewed-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit ed64c3323590e3c9fa8b423bf37689023a7a101f) Signed-off-by: Steffen Scheib <steffen@scheib.me>
Mathias Kresin [Sun, 27 Dec 2020 16:25:25 +0000 (17:25 +0100)]
lantiq: ltq-tapi: add kernel 5.10 compatiblity
Due to SCHED_FIFO being a broken scheduler model, all users of
sched_setscheduler() are converted to sched_set_fifo_low() upstream and
sched_setscheduler() is no longer exported.
The callback handling of the tasklet API was redesigned and the macros
using the old syntax renamed to _OLD.
Signed-off-by: Mathias Kresin <dev@kresin.me>
(cherry picked from commit 31f3f797004ad318a1de88ec9cfdece523ee46d9)
[Add DECLARE_TASKLET handling for kernel 5.4.235 too] Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
ca-certificates: fix python3-cryptography woes in certdata2pem.py
This patch is a revert of the upstream patch to Debian's ca-certificate
commit 033d52259172 ("mozilla/certdata2pem.py: print a warning for expired certificates.")
The reason is, that this change broke builds with the popular
Ubuntu 20.04 LTS (focal) releases which are shipping with an
older version of the python3-cryptography package that is not
compatible.
...or if the python3-cryptography was missing all together:
|Traceback (most recent call last):
| File "/certdata2pem.py", line 31, in <module>
| from cryptography import x509
|ModuleNotFoundError: No module named 'cryptography'
More concerns were raised by Jo-Philipp Wich:
"We don't want the build to depend on the local system time anyway.
Right now it seems to be just a warning but I could imagine that
eventually certs are simply omitted of found to be expired at
build time which would break reproducibility."
Link: <https://github.com/openwrt/openwrt/commit/7c99085bd697> Reported-by: Chen Minqiang <ptpt52@gmail.com> Reported-by: Shane Synan <digitalcircuit36939@gmail.com> Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
(cherry picked from commit 25bc66eb40ea2c062940778fba601032b2579734)
Rafał Miłecki [Wed, 1 Mar 2023 07:52:26 +0000 (08:52 +0100)]
kernel: support "linux,default-trigger" in leds-bcm63138
This driver is backported from the v6.0 which deals with
"linux,default-trigger" in leds core. For kernel 5.4 we need
leds-bcm63138 to read trigger on its own.
John Audia [Tue, 7 Feb 2023 19:56:52 +0000 (14:56 -0500)]
openssl: bump to 1.1.1t
Changes between 1.1.1s and 1.1.1t [7 Feb 2023]
*) Fixed X.400 address type confusion in X.509 GeneralName.
There is a type confusion vulnerability relating to X.400 address processing
inside an X.509 GeneralName. X.400 addresses were parsed as an ASN1_STRING
but subsequently interpreted by GENERAL_NAME_cmp as an ASN1_TYPE. This
vulnerability may allow an attacker who can provide a certificate chain and
CRL (neither of which need have a valid signature) to pass arbitrary
pointers to a memcmp call, creating a possible read primitive, subject to
some constraints. Refer to the advisory for more information. Thanks to
David Benjamin for discovering this issue. (CVE-2023-0286)
This issue has been fixed by changing the public header file definition of
GENERAL_NAME so that x400Address reflects the implementation. It was not
possible for any existing application to successfully use the existing
definition; however, if any application references the x400Address field
(e.g. in dead code), note that the type of this field has changed. There is
no ABI change.
[Hugo Landau]
*) Fixed Use-after-free following BIO_new_NDEF.
The public API function BIO_new_NDEF is a helper function used for
streaming ASN.1 data via a BIO. It is primarily used internally to OpenSSL
to support the SMIME, CMS and PKCS7 streaming capabilities, but may also
be called directly by end user applications.
The function receives a BIO from the caller, prepends a new BIO_f_asn1
filter BIO onto the front of it to form a BIO chain, and then returns
the new head of the BIO chain to the caller. Under certain conditions,
for example if a CMS recipient public key is invalid, the new filter BIO
is freed and the function returns a NULL result indicating a failure.
However, in this case, the BIO chain is not properly cleaned up and the
BIO passed by the caller still retains internal pointers to the previously
freed filter BIO. If the caller then goes on to call BIO_pop() on the BIO
then a use-after-free will occur. This will most likely result in a crash.
(CVE-2023-0215)
[Viktor Dukhovni, Matt Caswell]
*) Fixed Double free after calling PEM_read_bio_ex.
The function PEM_read_bio_ex() reads a PEM file from a BIO and parses and
decodes the "name" (e.g. "CERTIFICATE"), any header data and the payload
data. If the function succeeds then the "name_out", "header" and "data"
arguments are populated with pointers to buffers containing the relevant
decoded data. The caller is responsible for freeing those buffers. It is
possible to construct a PEM file that results in 0 bytes of payload data.
In this case PEM_read_bio_ex() will return a failure code but will populate
the header argument with a pointer to a buffer that has already been freed.
If the caller also frees this buffer then a double free will occur. This
will most likely lead to a crash.
The functions PEM_read_bio() and PEM_read() are simple wrappers around
PEM_read_bio_ex() and therefore these functions are also directly affected.
These functions are also called indirectly by a number of other OpenSSL
functions including PEM_X509_INFO_read_bio_ex() and
SSL_CTX_use_serverinfo_file() which are also vulnerable. Some OpenSSL
internal uses of these functions are not vulnerable because the caller does
not free the header argument if PEM_read_bio_ex() returns a failure code.
(CVE-2022-4450)
[Kurt Roeckx, Matt Caswell]
*) Fixed Timing Oracle in RSA Decryption.
A timing based side channel exists in the OpenSSL RSA Decryption
implementation which could be sufficient to recover a plaintext across
a network in a Bleichenbacher style attack. To achieve a successful
decryption an attacker would have to be able to send a very large number
of trial messages for decryption. The vulnerability affects all RSA padding
modes: PKCS#1 v1.5, RSA-OEAP and RSASVE.
(CVE-2022-4304)
[Dmitry Belyavsky, Hubert Kario]
Martin Kennedy [Tue, 30 Aug 2022 00:47:24 +0000 (20:47 -0400)]
mpc85xx: Drop pci aliases to avoid domain changes
As of upstream Linux commit 0fe1e96fef0a ("powerpc/pci: Prefer PCI
domain assignment via DT 'linux,pci-domain' and alias"), the PCIe
domain address is no longer numbered by the lowest 16 bits of the PCI
register address after a fallthrough. Instead of the fallthrough, the
enumeration process accepts the alias ID (as determined by
`of_alias_scan()`). This causes e.g.:
... which then causes the sysfs path of the netdev to change,
invalidating the `wifi_device.path`s enumerated in
`/etc/config/wireless`.
One other solution might be to migrate the uci configuration, as was
done for mvebu in commit 0bd5aa89fcf2 ("mvebu: Migrate uci config to
new PCIe path"). However, there are concerns that the sysfs path will
change once again once some upstream patches[^2][^3] are merged and
backported (and `CONFIG_PPC_PCI_BUS_NUM_DOMAIN_DEPENDENT` is enabled).
Instead, remove the aliases and allow the fallthrough to continue for
now. We will provide a migration in a later release.
Fixes: #10530 Tested-by: Martin Kennedy <hurricos@gmail.com>
[Tested on the Aerohive HiveAP 330 and Extreme Networks WS-AP3825i] Signed-off-by: Martin Kennedy <hurricos@gmail.com>
(cherry picked from commit 7f4b4c29f3489697dca7495216460d0ed5023e02) Signed-off-by: Fabian Bläse <fabian@blaese.de>
Hauke Mehrtens [Sat, 28 Jan 2023 18:25:32 +0000 (19:25 +0100)]
kernel: Reorder configuration
This was done by running these commands:
./scripts/kconfig.pl '+' target/linux/generic/config-5.4 /dev/null > target/linux/generic/config-5.4-new
mv target/linux/generic/config-5.4-new target/linux/generic/config-5.4
Rafał Miłecki [Mon, 23 Jan 2023 12:23:29 +0000 (13:23 +0100)]
kernel: expose (unhide) CONFIG_ASN1 as ksmbd requirement
OpenWrt provides kmod-asn1-decoder for CONFIG_ASN1 but selecting it
doesn't really work as expected. Kernel symbol is hidden and can be
actually selected only as a dependency. That works well for in-kernel
stuff but fails for external modules requiring ASN1 like ksmbd.
Modify kernel Kconfig to make CONFIG_ASN1 always selectable. It's
required to satisfy ksmbd dependencies cleanly (without hack like
selecting unrelated modules).
scripts/dl_github_archieve.py: fix generating unreproducible tar
Allign dl_github_archieve.py to 8252511dc0b5a71e9e64b96f233a27ad73e28b7f
change. On supported system the sigid bit is applied to files and tar
archieve that on tar creation. This cause unreproducible tar for these
system and these bit should be dropped to produce reproducible tar.
Add the missing option following the command options used in other
scripts.
Fixes: 75ab064d2b38 ("build: download code from github using archive API") Suggested-by: Eneas U de Queiroz <cotequeiroz@gmail.com> Tested-by: Robert Marko <robimarko@gmail.com> Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
(cherry picked from commit 5f1758ef14575df4e86896526b1c2035c231899e)
Josef Schlehofer [Mon, 12 Dec 2022 22:08:05 +0000 (23:08 +0100)]
opkg: add patch to avoid remove package repeatly with force
This patch was taken from the OpenWrt-devel mailing list:
https://www.mail-archive.com/openwrt-devel@lists.openwrt.org/msg59794.html
It is included already in OpenWrt master branch and OpenWrt 22.03
release as it was included in opkg-lede repository:
https://git.openwrt.org/?p=project/opkg-lede.git;a=commit;h=9c44557a776da993c2ab80cfac4dbd8d59807d01
However, it is not included in OpenWrt 21.02, where the same issue is
happening.
Fixes: CI for https://github.com/openwrt/packages/pull/20074 Signed-off-by: Josef Schlehofer <pepe.schlehofer@gmail.com>
Targets which allow booting from NVMe (x86, maybe some mvebu boards come
to mind) should have it built-in, so rootfs can be mounted from there.
For targets without NVMe support in bootloader or BIOS/firmware it's
sufficient to provide the kernel module package.
On targets having the NVMe driver built-in the resulting kmod package
is an empty dummy. In any case, depending on or installing kmod-nvme
results in driver support being available (either because it was already
built-in or because the relevant kernel modules are added and loaded).
Hauke Mehrtens [Sun, 7 Aug 2022 11:32:31 +0000 (13:32 +0200)]
kernel: kmod-w1-slave-ds2760: Remove package
The w1_ds2760.ko driver was merged into the ds2760_battery.ko driver.
The driver was removed and this package was never build any more.
This happened with kernel 4.19.
CI: build: fix external toolchain use with release tag tests
When a new tag for a release is created, the just checkout repo from
github actions will already have such tag locally created.
This will result in git fetch --tags failing with error rejecting the
remote tag with (would clobber existing tag).
Add -f option to overwrite any local tags and always fetch them from
remote.
Fixes: e24a1e6f6d7f ("CI: build: add support for external toolchains from stable branch") Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
(cherry picked from commit f655923b362e9f2d70672eee9c1fa82550a145a6)
Hauke Mehrtens [Tue, 1 Nov 2022 14:17:03 +0000 (15:17 +0100)]
dnsmasq: Backport DHCPv6 server fix (CVE-2022-0934)
This backports a commit from upstream dnsmasq to fix CVE-2022-0934.
CVE-2022-0934 description:
A single-byte, non-arbitrary write/use-after-free flaw was found in
dnsmasq. This flaw allows an attacker who sends a crafted packet
processed by dnsmasq, potentially causing a denial of service.
CI: build: skip sdk adapt to external toolchain on cache hit
On cache hit, skip sdk adapt to external toolchain. This is needed because we
cache the already extracted sdk and that is already adapted to be used
as external toolchain.
Rerunning the adap step will result in the test to fail for missing file
as the file are already got wrapped to the external toolchain format.
Fixes: 42f0ab028e2e ("CI: build: fix use of sdk as toolchain") Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
(cherry picked from commit 99eaedfe3966b1ca812e8a962197cf91286247f7)
The toolchain included in a sdk have a different format than an external
toolchain tar.
Since sdk is a more integrated setup doesn't use and include wrapper bin
that use the external toolchain config and use an alternative and more
standard way to include all the toolchain headers.
External toolchain use wrapper.sh to append the configured include
header when each tool is called.
Fix the sdk toolchain by reverting their own sdk wrapper scripts and to
simulate an external toolchain build copying what is done in the
toolchain target makefile.
This handle compilation error and warning caused by not using fortify
header on building packages.
Fixes: 006e52545d14 ("CI: build: add support to fallback to sdk for external toolchain") Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
(cherry picked from commit 42f0ab028e2eae0d4e7acf9db7fd68b256f23503)
toolchain: Select USE_SSTRIP with external musl toolchain
When we use the internal toolchain USE_SSTRIP will be selected by
default for musl libc and USE_STRIP when glibc is used. Do the same when
an external toolchain is used. USE_GLIBC will also be set for external
toolchain builds based on the EXTERNAL_TOOLCHAIN_LIBC_USE_GLIBC setting.
CI: build: fix matching for openwrt release branch for toolchain parsing
The current match logic doesn't handle test for push events related to
stable release (example openwrt-22.03) but only fork with the related
prefix (example openwrt-22.03-fixup)
Fix wrong matching and while at it also add extra checks to other
matching (check if the branch name actually start with the requested
prefix)
Fixes: e24a1e6f6d7f ("CI: build: add support for external toolchains from stable branch") Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
(cherry picked from commit e3cf2b84e5f8708ca17d931ef60746516c8a2fe4)
CI: fix matching for openwrt release branch for container selection
The current match logic doesn't handle test for push events related to
stable release (example openwrt-22.03) but only fork with the related
prefix (example openwrt-22.03-fixup)
Fix wrong matching and while at it also add extra checks to other
matching (check if the branch name actually start with the requested
prefix)
Fixes: abe8a4824210 ("CI: build: add support for per branch tools container") Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
(cherry picked from commit 65c3d19c4b28ccac0d08d916de0ffa4c0e7b3dc2)
CI: labeler: fix wrong label for pr targeting stable branch
The label used for stable branch is in the form of
release/[0-9][0-9].[0-9][0-9]
Currently we apply the name of the target branch as the label, fix this
and correctly use the current label.
Add support to tag pr targeting stable branch matching the simple regex
of openwrt-[0-9][0-9].[0-9][0-9]. The tag that will be added will match
the pr target branch.
Ansuel Smith [Mon, 10 Jan 2022 16:02:30 +0000 (17:02 +0100)]
kernel: split kernel version to dedicated files
Move the kernel versions and hash to dedicated files.
This makes kernel bump quicker and fix some annoying
problem with rebasing when multiple kernel bump are proposed.
Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
[Rebased on top of current master] Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
(cherry picked from commit 0765466a42f46f7357e260866a4284ed567bb7ad) Signed-off-by: Josef Schlehofer <pepe.schlehofer@gmail.com>
[Rebased on top of current openwrt-21.02]
build: handle directory with whitespace in AUTOREMOVE clean
Package with whitespace in their build directory are not correctly
removed when CONFIG_AUTOREMOVE is enabled. This is caused by xargs that
use whitespace as delimiters. To handle this use \0 as the delimiter and
set find to use \0 as the delimiter.
projects / thirdparty / openwrt.git / log
