Hi,
On 4/4/22 21:56, Sathyanarayanan Kuppuswamy wrote:
> Hi Hans,
>
> On 4/4/22 3:07 AM, Hans de Goede wrote:
>>> +static int __init tdx_attest_init(void)
>>> +{
>>> + dma_addr_t handle;
>>> + long ret = 0;
>>> +
>>> + mutex_lock(&attestation_lock);
>>> +
>>> + ret = misc_register(&tdx_attest_device);
>>> + if (ret) {
>>> + pr_err("misc device registration failed\n");
>>> + mutex_unlock(&attestation_lock);
>>> + return ret;
>>> + }
>> Why not do this as the last thing of the probe?
>
> We need misc device reference in dma_alloc_coherent() and
> dma_set_coherent_mask() calls. This is the reason for keeping
> misc_register() at the beginining of the init function.
Erm, you are supposed to pass an actual device-device as
parameter to dma_alloc_coherent(), so that it can see
what bus/dma-domain the device is connected to and if
an ioMMU might be involved...
>> That will avoid the need to unregister this again in all
>> the error-exit paths and also fixes a possible deadlock.
>>
>
> Agree. But, unless we create another device locally, I don't
> think we can avoid this. Do you prefer this approach?
Yes, passing the "struct device" which is embedded inside
a miscdevice as device to dma_alloc_coherent() is just
wrong. Please make your module_init function register
a platform_device using platform_device_register_simple()
(on systems with TDX support) and then turn your code/driver
into a standard platform_driver using the platform_device
which the driver binds to as parameter to dma_alloc_coherent().
See: https://git.kernel.org/pub/scm/linux/kernel/git/pdx86/platform-drivers-x86.git/commit/?id=c4d2ff35d350428eca2ae1a1e2119e4c6297811d
for a simple (completely unrelated) driver which uses this
method to have a device to bind to for talking to a secondary
function of the embedded-controller on some platforms.
Regards,
Hans
