linux-kernel - Re: [PATCH 6/9] nvme-apple: Add initial Apple SoC NVMe driver

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:   Sat, 02 Apr 2022 14:47:41 +0200
From:   "Sven Peter" <sven@...npeter.dev>
To:     "hch@....de" <hch@....de>
Cc:     "Keith Busch" <kbusch@...nel.org>, "axboe@...com" <axboe@...com>,
        "sagi@...mberg.me" <sagi@...mberg.me>,
        "Hector Martin" <marcan@...can.st>,
        "Alyssa Rosenzweig" <alyssa@...enzweig.io>,
        "Rob Herring" <robh+dt@...nel.org>,
        "Arnd Bergmann" <arnd@...db.de>, "Marc Zyngier" <maz@...nel.org>,
        devicetree@...r.kernel.org, linux-arm-kernel@...ts.infradead.org,
        linux-kernel@...r.kernel.org, linux-nvme@...ts.infradead.org
Subject: Re: [PATCH 6/9] nvme-apple: Add initial Apple SoC NVMe driver

Hi,

Thanks for the review!

On Thu, Mar 24, 2022, at 07:16, Christoph Hellwig wrote:
>> +
>> +//#define DEBUG
>
> This should not leak into the driverṡ

Agreed, I'll remove it.

>
>> +#include <linux/blk-integrity.h>
>
> As far as I can tell this driver does not support metadata or PI,
> so why is this include needed?

Yup, you're right, it's not required at all. It's a left-over from
when I hadn't realized that this controller doesn't support metadata
and still had it implemented.

>
>> +/* NVM Express NVM Command Set Specification, Revision 1.0a, Figure 18 */
>> +#define NVME_OPCODE_DATA_XFER_HOST_TO_CTRL BIT(0)
>> +#define NVME_OPCODE_DATA_XFER_CTRL_TO_HOST BIT(1)
>
> Please just use the nvme_is_write helper where you are using these.

I didn't realize that helper exists, will use it!


>
>> +static int apple_nvme_sart_dma_setup(void *cookie, struct apple_rtkit_shmem *bfr,
>
> Please avoid > 80 character lines.

Will fix it.

>
>> +static void apple_nvme_free_ctrl(struct nvme_ctrl *ctrl)
>> +{
>> +}
>
> So where are the apple specific resources free?  ->free_ctrl is
> the callback from the struct device release callback, so without
> one the resource release can't be tried to the device release.

The resources are all devres managed but I only just realized that
the nvme_ctrl creates its own struct device that has to be tied to
the platform device. I think I'll at least need a get_device in _probe
and a put_device here just like in pci.c to tie those two together.
I'll also double check to make sure that all resources are indeed
devres-managed.


Thanks,


Sven