linux-kernel - Re: [PATCH v1] arm64: dts: actions: Fix smp Bringing up secondary CPUs

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]

Date:   Mon, 6 Jul 2020 00:09:39 +0200
From:   Andreas Färber <afaerber@...e.de>
To:     Matheus Castello <matheus@...tello.eng.br>
Cc:     manivannan.sadhasivam@...aro.org, robh+dt@...nel.org,
        linux-arm-kernel@...ts.infradead.org, devicetree@...r.kernel.org,
        linux-kernel@...r.kernel.org, linux-actions@...ts.infradead.org
Subject: Re: [PATCH v1] arm64: dts: actions: Fix smp Bringing up secondary
 CPUs

Hi Matheus,

Am 05.07.20 um 21:19 schrieb Matheus Castello:
> Change the enable-method to fix the failed to boot errors:
> 
> [    0.040330] smp: Bringing up secondary CPUs ...
> [    0.040683] psci: failed to boot CPU1 (-22)
> [    0.040691] CPU1: failed to boot: -22
> [    0.041062] psci: failed to boot CPU2 (-22)
> [    0.041071] CPU2: failed to boot: -22
> [    0.041408] psci: failed to boot CPU3 (-22)
> [    0.041417] CPU3: failed to boot: -22
> [    0.041443] smp: Brought up 1 node, 1 CPU
> [    0.041451] SMP: Total of 1 processors activated.
> 
> Tested on Caninos Labrador v3 based on Actions Semi S700.
> 
> Signed-off-by: Matheus Castello <matheus@...tello.eng.br>
> ---
>   arch/arm64/boot/dts/actions/s700.dtsi | 33 +++++++++++++++++++--------
>   1 file changed, 23 insertions(+), 10 deletions(-)

NACK.

For starters, if this were an actual fix, it should have a Fixes header.

Don't do random changes in a single patch and call it a "fix". I don't 
see what changing the cell size has to do with smp, nor adding L2 cache. 
The latter could be a patch of its own, after fixes are applied (to 
avoid conflicts when backporting that fix to older branches). A cell 
size of two used to be perfectly valid, please checking the DT binding.

Finally, you're changing generic S700 code here - you can't just break 
Cubieboard7 just because your Labrador has too old BL31 firmware. Can't 
you just update its TF-A firmware and use the standard PSCI interface 
for Linux? If not, then you should add your workaround to your 
module's/board's .dts(i) instead of the SoC's .dtsi.

Thanks,
Andreas

-- 
SUSE Software Solutions Germany GmbH
Maxfeldstr. 5, 90409 Nürnberg, Germany
GF: Felix Imendörffer
HRB 36809 (AG Nürnberg)