linux-ext4 - Re: [PATCH 1/2] kvm-xfstests: install-kconfig: Use $ARCH-config instead of $KERN

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 for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Thu, 12 Oct 2023 13:28:35 -0400
From:   "Theodore Ts'o" <tytso@....edu>
To:     "Ritesh Harjani (IBM)" <ritesh.list@...il.com>
Cc:     linux-ext4@...r.kernel.org, Ojaswin Mujoo <ojaswin@...ux.ibm.com>
Subject: Re: [PATCH 1/2] kvm-xfstests: install-kconfig: Use $ARCH-config
 instead of $KERN_ARCH-config

On Wed, Oct 11, 2023 at 12:46:30AM +0530, Ritesh Harjani (IBM) wrote:
> $KERN_ARCH is used for make arguments. For configs let's use
> $ARCH-config. This should not break anything since as of now we only
> have arm64-config for which $ARCH and $KERN_ARCH is same.
> 
> Signed-off-by: Ritesh Harjani (IBM) <ritesh.list@...il.com>

It would be much better to teach set_canoncalized_arch about the
powerpc64le architecture.  That way other scripts that use
set_canoicalized_arch can correctly depend on KERN_ARCH.  See
run-fstests/util/arch_funcs:

# There are multiple ways that CPU architectured can be named.
# KERN_ARCH is what is used when specifying ARCH=xxx when building the
# kernel.
#
# GCE_ARCH is what gets used when setting specifying the architecture
# when creating the GCE image using the --architecture flag
# ARCH (after being canonicalized by this function) is defined by
# the Debian architecture port names[1] and is used internally
# by gce-xfstests and kvm-xfstests.  So for example, when we add
# support for support for Power architectures, the ARCH name that should
# be used is ppc64 or ppc64el.

...

# This function takes as input a user-supplied architecture (which
# generally should be a Debian port name, but users might use
# a $(uname -m) instead.
#
function set_canonicalized_arch () {
    case "$1" in
	arm64|aarch64)
	    ARCH="arm64"
	    GCE_ARCH="ARM64"
	    KERN_ARCH="arm64"
	    ;;
    ...

So basically, after calling set_canoncalized_arch, ARCH should be
Debian architecture port name, GCE_ARCH (if not the empty string) is
the name of the GCE architecture name, and KERN_ARCH should be what
should be passed to the Kernel makefile as "make ARCH=$KERN_ARCH ..."

       	  	    	       		- Ted