linux-kernel - Re: [RFC PATCH v3 13/37] kvx: Add build infrastructure

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

Message-Id: <1a663b40-644b-48b3-880e-b24e3dd4827e@app.fastmail.com>
Date: Tue, 23 Jul 2024 09:46:01 +0000
From: "Arnd Bergmann" <arnd@...db.de>
To: "Yann Sionneau" <ysionneau@...rayinc.com>, linux-kernel@...r.kernel.org
Cc: "Jonathan Borne" <jborne@...rayinc.com>,
 "Julian Vetter" <jvetter@...rayinc.com>,
 "Clement Leger" <clement@...ment-leger.fr>,
 "Guillaume Thouvenin" <thouveng@...il.com>,
 "Jules Maselbas" <jmaselbas@...v.net>,
 Marc Poulhiès <dkm@...aplop.net>,
 "Marius Gligor" <mgligor@...rayinc.com>,
 "Samuel Jones" <sjones@...rayinc.com>,
 "Vincent Chardon" <vincent.chardon@...ys-design.com>, bpf@...r.kernel.org
Subject: Re: [RFC PATCH v3 13/37] kvx: Add build infrastructure

On Mon, Jul 22, 2024, at 09:41, ysionneau@...rayinc.com wrote:
> +
> +config GENERIC_CALIBRATE_DELAY
> +	def_bool y

You don't seem to define ARCH_HAS_READ_CURRENT_TIMER 
or calibrate_delay_is_known(), so it appears that you are
measuring the loops per jiffy. Since you are using a
cycle counter for your delay loop, you should be able
to just set the value directly instead.

> +	select GENERIC_IOMAP
> +	select GENERIC_IOREMAP

GENERIC_IOMAP is something you should normally not
set, it is a bit misnamed and not at all generic ;-)

> +config ARCH_SPARSEMEM_ENABLE
> +	def_bool y
> +
> +config ARCH_SPARSEMEM_DEFAULT
> +	def_bool ARCH_SPARSEMEM_ENABLE
> +
> +config ARCH_SELECT_MEMORY_MODEL
> +	def_bool ARCH_SPARSEMEM_ENABLE

There has been some discussion about removing traditional
sparsemem support in favor of just using SPARSEMEM_VMEMMAP,
which is simpler and more efficient on 64-bit architectures.

You should probably have a look at that.

> +config SMP
> +	bool "Symmetric multi-processing support"
> +	default n
> +	select GENERIC_SMP_IDLE_THREAD
> +	select GENERIC_IRQ_IPI
> +	select IRQ_DOMAIN_HIERARCHY

Do you have a use case for turning SMP off in practice?
Non-SMP kernels can be a little more efficient, but it's
getting rare these days to actually have systems with
just one CPU.

> +config KVX_PAGE_SHIFT
> +	int
> +	default 12

There is now a generic CONFIG_PAGE_SIZE and CONFIG_PAGE_SHIFT
setting, so you can remove the custom ones and just use
'select HAVE_PAGE_SIZE_4KB'

> --- /dev/null
> +++ b/arch/kvx/include/asm/Kbuild
> @@ -0,0 +1,20 @@
> +generic-y += asm-offsets.h
> +generic-y += clkdev.h
> +generic-y += auxvec.h
> +generic-y += bpf_perf_event.h
> +generic-y += cmpxchg-local.h
> +generic-y += errno.h
> +generic-y += extable.h
> +generic-y += export.h
...

You seem to have more entries here than necessary, e.g.
clkdev.h and auxvec.h are no longer in asm-generic, and
export.h is deprecated.

      Arnd