[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <65bebe74-587b-eb5f-44f3-2e354fd2b82d@users.sourceforge.net>
Date: Mon, 23 Oct 2017 15:46:52 +0200
From: SF Markus Elfring <elfring@...rs.sourceforge.net>
To: Jarkko Sakkinen <jarkko.sakkinen@...ux.intel.com>,
linux-integrity@...r.kernel.org
Cc: Jason Gunthorpe <jgunthorpe@...idianresearch.com>,
Peter Hüwe <PeterHuewe@....de>,
LKML <linux-kernel@...r.kernel.org>,
kernel-janitors@...r.kernel.org
Subject: Re: [PATCH] tpm_i2c_nuvoton: Use common error handling code in
i2c_nuvoton_send()
>> @@ -457,12 +455,15 @@ static int i2c_nuvoton_send(struct tpm_chip *chip, u8 *buf, size_t len)
>> &priv->read_queue);
>> if (rc) {
>> dev_err(dev, "%s() timeout command duration\n", __func__);
>> - i2c_nuvoton_ready(chip);
>> - return rc;
>> + goto write_ready;
>> }
>>
>> dev_dbg(dev, "%s() -> %zd\n", __func__, len);
>> return len;
>> +
>> +write_ready:
>> + i2c_nuvoton_ready(chip);
>> + return rc;
>> }
>>
>> static bool i2c_nuvoton_req_canceled(struct tpm_chip *chip, u8 status)
>> --
>> 2.14.2
>>
>
> Setting the commandReady bit gives means to abort the command.
> Setting the bit is not the end goal.
Does your constructive feedback mean that there are any more implementation details
to consider besides the suggested code layout adjustment?
> Use something like err_cancel instead.
I am unsure about this suggestion. Will a simple replacement be sufficient
at the end?
Regards,
Markus
Powered by blists - more mailing lists