Re: [RFC 2/5] drm/i915: Extract code mapping errno to vm fault code

From: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
To: Chris Wilson <chris@chris-wilson.co.uk>, Intel-gfx@lists.freedesktop.org
Subject: Re: [RFC 2/5] drm/i915: Extract code mapping errno to vm fault code
Date: Tue, 26 Jan 2016 16:24:46 +0000	[thread overview]
Message-ID: <56A79DCE.1000006@linux.intel.com> (raw)
In-Reply-To: <20160126151853.GO8281@nuc-i3427.alporthouse.com>

On 26/01/16 15:18, Chris Wilson wrote:
> On Tue, Jan 26, 2016 at 02:53:30PM +0000, Tvrtko Ursulin wrote:
>> From: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
>>
>> Will be used from multiple callers in a following patch.
>>
>> Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
>> ---
>>   drivers/gpu/drm/i915/i915_gem.c | 91 ++++++++++++++++++++++-------------------
>>   1 file changed, 49 insertions(+), 42 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
>> index af15c290c71d..dacf6a0013c5 100644
>> --- a/drivers/gpu/drm/i915/i915_gem.c
>> +++ b/drivers/gpu/drm/i915/i915_gem.c
>> @@ -1772,6 +1772,53 @@ i915_gem_mmap_ioctl(struct drm_device *dev, void *data,
>>   	return 0;
>>   }
>>
>> +static int
>> +i915_gem_ret_to_vm_ret(struct drm_i915_private *dev_priv, int ret)
>> +{
>> +	switch (ret) {
>> +	case -EIO:
>> +		/*
>> +		 * We eat errors when the gpu is terminally wedged to avoid
>> +		 * userspace unduly crashing (gl has no provisions for mmaps to
>> +		 * fail). But any other -EIO isn't ours (e.g. swap in failure)
>> +		 * and so needs to be reported.
>> +		 */
>> +		if (!i915_terminally_wedged(&dev_priv->gpu_error)) {
>> +			ret = VM_FAULT_SIGBUS;
>> +			break;
>> +		}
>> +	case -EAGAIN:
>> +		/*
>> +		 * EAGAIN means the gpu is hung and we'll wait for the error
>> +		 * handler to reset everything when re-faulting in
>> +		 * i915_mutex_lock_interruptible.
>> +		 */
>> +	case 0:
>> +	case -ERESTARTSYS:
>> +	case -EINTR:
>> +	case -EBUSY:
>> +		/*
>> +		 * EBUSY is ok: this just means that another thread
>> +		 * already did the job.
>> +		 */
>> +		ret = VM_FAULT_NOPAGE;
>> +		break;
>> +	case -ENOMEM:
>> +		ret = VM_FAULT_OOM;
>> +		break;
>> +	case -ENOSPC:
>> +	case -EFAULT:
>> +		ret = VM_FAULT_SIGBUS;
>> +		break;
>> +	default:
>> +		WARN_ONCE(ret, "unhandled error in page fault\n");
>> +		ret = VM_FAULT_SIGBUS;
>> +		break;
>> +	}
>
> So without having to pin (plus a few other similar changes), we only need
> to report
>
> -ENOMEM, -ENOSPC, -EIO (from shmemfs) and -EFAULT (get_pages).
>
> Given that I'd rather have the reasoning behind each explicit.

This is just existing code factored out so I am not sure what you are 
suggesting?

Regards,

Tvrtko

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx