Re: [PATCH -next 2/4] spi: mockup: Add writeable tracepoint for spi transfer

From: "weiyongjun (A)" <weiyongjun1@huawei.com>
To: Mark Brown <broonie@kernel.org>
Cc: Steven Rostedt <rostedt@goodmis.org>,
	Ingo Molnar <mingo@redhat.com>, <linux-kernel@vger.kernel.org>,
	<linux-spi@vger.kernel.org>
Subject: Re: [PATCH -next 2/4] spi: mockup: Add writeable tracepoint for spi transfer
Date: Wed, 31 Aug 2022 12:43:50 +0800	[thread overview]
Message-ID: <bc4dd16b-e2f8-9734-8db8-78500f361f5a@huawei.com> (raw)
In-Reply-To: <Yw5Ta8sO3fMsPLDY@sirena.org.uk>

On 2022/8/31 2:14, Mark Brown wrote:
> On Fri, Aug 26, 2022 at 02:43:39PM +0000, Wei Yongjun wrote:
> 
>> +#define SPI_BUFSIZ_MAX		0x1000
>> +
>> +struct spi_msg_ctx {
>> +	int ret;
>> +	__u8 data[SPI_BUFSIZ_MAX];
>> +};
> 
> This has a hard limit on the size of messages of only 4k on the
> size of messages, that seems a bit on the low side especially
> considering that the example is for a flash device.  There's also
> things like firmware downloads which can easily exceed 4k.  The
> driver should at the very least report this limit via the SPI
> API, and probably also complain loudly if the limit is exceeded
> since it's unlikely that most drivers will actually be checking
> for such a low limit given how rare they are.
> 

We need a limit because of BPF writeble tracepoint only support const
size buffer write now. This limit can be fixed after make BPF support
dynptr write.

Regards,
Wei Yongjun