lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <2bbabd2c-24ef-493c-a199-594e5dada3da@broadcom.com>
Date: Fri, 22 Nov 2024 09:34:32 -0800
From: Florian Fainelli <florian.fainelli@...adcom.com>
To: Peng Fan <peng.fan@....com>, "Peng Fan (OSS)" <peng.fan@....nxp.com>,
 Thomas Renninger <trenn@...e.com>, Shuah Khan <shuah@...nel.org>,
 "John B. Wyatt IV" <jwyatt@...hat.com>, John Kacur <jkacur@...hat.com>
Cc: "linux-pm@...r.kernel.org" <linux-pm@...r.kernel.org>,
 "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v2 2/2] pm: cpupower: Makefile: Allow overriding
 cross-compiling env params

On 11/21/24 16:13, Peng Fan wrote:
>> Subject: Re: [PATCH v2 2/2] pm: cpupower: Makefile: Allow overriding
>> cross-compiling env params
>>
>> On 11/21/24 04:40, Peng Fan wrote:
>>>> Subject: Re: [PATCH v2 2/2] pm: cpupower: Makefile: Allow
>> overriding
>>>> cross-compiling env params
>>>>
>>>> Hi Peng,
>>>>
>>>> On 9/19/2024 5:08 AM, Peng Fan (OSS) wrote:
>>>>> From: Peng Fan <peng.fan@....com>
>>>>>
>>>>> Allow overriding the cross-comple env parameters to make it
>> easier
>>>> for
>>>>> Yocto users. Then cross-compiler toolchains to build cpupower
>> with
>>>>> only two steps:
>>>>> - source (toolchain path)/environment-setup-armv8a-poky-linux
>>>>> - make
>>>>
>>>> This patch breaks the way that buildroot builds cpupower:
>>>>
>>>>
>> https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2F
>>>> git.buildroot.net%2Fbuildroot%2Ftree%2Fpackage%2Flinux-
>>>> tools%2Flinux-tool-
>>>>
>> cpupower.mk.in&data=05%7C02%7Cpeng.fan%40nxp.com%7C246da9
>>>>
>> 2d8b6243d138c808dd09e6d644%7C686ea1d3bc2b4c6fa92cd99c5c3
>>>>
>> 01635%7C0%7C0%7C638677609234547728%7CUnknown%7CTWFpb
>>>>
>> GZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJX
>>>>
>> aW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ%3D%3D%7C0%7C%7C%7C&sdat
>>>>
>> a=nL1YUl%2F07Vd8F0GpW7uRqdpZT74avOku1ox9N3%2FFkUg%3D&r
>>>> eserved=0
>>>>
>>>> and after enabling verbose it becomes clear as to why, see below:
>>>>
>>>>    >>> linux-tools  Configuring
>>>>    >>> linux-tools  Building
>>>> GIT_DIR=.
>>>> PATH="/local/users/fainelli/buildroot-
>>>> upstream/output/arm/host/bin:/local/users/fainelli/buildroot-
>>>>
>> upstream/output/arm/host/sbin:/projects/firepath/tools/bin:/home/ff
>>>>
>> 944844/bin:/home/ff944844/.local/bin:/opt/stblinux/bin:/usr/local/sb
>>>> in:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local
>>>> /ga mes:/snap/bin:/opt/toolchains/metaware-vctools-0.4.1/bin/"
>>>> /usr/bin/make -j97
>>>> CROSS=/local/users/fainelli/buildroot-
>>>> upstream/output/arm/host/bin/arm-linux-
>>>> CPUFREQ_BENCH=false NLS=false LDFLAGS="-L.
>>>> -L/local/users/fainelli/buildroot-
>> upstream/output/arm/target/usr/lib"
>>>> DEBUG=false V=1 -C
>>>> /local/users/fainelli/buildroot-upstream/output/arm/build/linux-
>>>> custom/tools
>>>> cpupower
>>>> mkdir -p power/cpupower && /usr/bin/make
>> subdir=power/cpupower
>>>> --no-print-directory -C power/cpupower cc -DVERSION=\"6.12.0\" -
>>>> DPACKAGE=\"cpupower\"
>>>> -DPACKAGE_BUGREPORT=\"linux-pm@...r.kernel.org\" -
>> D_GNU_SOURCE -pipe
>>>> -Wall -Wchar-subscripts -Wpointer-arith -Wsign- compare
>>>> -Wno-pointer-sign -Wdeclaration-after-statement -Wshadow - Os
>>>> -fomit-frame-pointer -fPIC -o lib/cpufreq.o -c lib/cpufreq.c
>>>
>>> You are building on an ARM host? Or you are cross compiling with cc
>> is
>>> actually arm gcc?
>>
>> This is cross compiling targeting ARM, which is why CROSS is set to
>> ../arm-linux-
>>
>>>
>>>>
>>>> Here we use cc, aka host compiler, rather than $(CROSS)gcc as we
>>>> should, so we are no longer cross compiling at all.
>>>
>>> I not understand. CROSS was set, but using cc to compile for host?
>>
>> See below.
>>
>>>
>>>>
>>>> The issue is the use of the lazy set if absent for *all* of CC, LD,
>>>> AR, STRIP, RANLIB, rather than just for CROSS. The following fixes it
>> for me:
>>>>
>>>> diff --git a/tools/power/cpupower/Makefile
>>>> b/tools/power/cpupower/Makefile index
>>>> 175004ce44b2..96bb1e5f3970 100644
>>>> --- a/tools/power/cpupower/Makefile
>>>> +++ b/tools/power/cpupower/Makefile
>>>> @@ -87,11 +87,11 @@ INSTALL_SCRIPT = ${INSTALL} -m 644
>>>>     # to something more interesting, like "arm-linux-".  If you want
>>>>     # to compile vs uClibc, that can be done here as well.
>>>>     CROSS ?= #/usr/i386-linux-uclibc/usr/bin/i386-uclibc-
>>>> -CC ?= $(CROSS)gcc
>>>> -LD ?= $(CROSS)gcc
>>>> -AR ?= $(CROSS)ar
>>>> -STRIP ?= $(CROSS)strip
>>>> -RANLIB ?= $(CROSS)ranlib
>>>> +CC = $(CROSS)gcc
>>>> +LD = $(CROSS)gcc
>>>> +AR = $(CROSS)ar
>>>> +STRIP = $(CROSS)strip
>>>> +RANLIB = $(CROSS)ranlib
>>>
>>> The ? is just allow to override CC/LD/AR.., so in your env, CC is set,
>>> but should not be used for cpupower compling?
>>
>> Adding debug to show the origin of the CC variable shows the following:
>>
>> CROSS=/local/users/fainelli/buildroot-
>> upstream/output/arm/host/bin/arm-linux-
>> CC origin is (default) and value is (cc) LD origin is (default) and value is
>> (ld) CC=cc LD=ld AR=ar STRIP= RANLIB=
> 
> 
> How about
> CROSS ?= #/usr/i386-linux-uclibc/usr/bin/i386-uclibc-
> ifneq ($(CROSS), )
> CC = $(CROSS)gcc
> LD = $(CROSS)gcc
> AR = $(CROSS)ar
> STRIP = $(CROSS)strip
> RANLIB = $(CROSS)ranlib
> else
> CC ?= $(CROSS)gcc
> LD ?= $(CROSS)gcc
> AR ?= $(CROSS)ar
> STRIP ?= $(CROSS)strip
> RANLIB ?= $(CROSS)ranlib
> Endif

Yes, this works just as well, do you want to submit this if that works 
in your environment as well?
-- 
Florian

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ