Re: [PATCH 3/4] iommu/exynos: Add iommu driver for Exynos4 Platforms

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi.

On Mon, Sep 26, 2011 at 4:59 PM, Ohad Ben-Cohen <ohad@xxxxxxxxxx> wrote:
> 2011/9/24 조경호 <pullip.cho@xxxxxxxxxxx>:
>> This is the System MMU driver and IOMMU API implementation for
>> Exynos4 SOC platforms. Exynos4 platforms has more than 10 System
>> MMUs dedicated for each multimedia accellerators.
>>
>> Signed-off-by: KyongHo Cho <pullip.cho@xxxxxxxxxxx>
> ...
>> +static int exynos_iommu_fault_handler(struct iommu_domain *domain,
>> +                       struct device *dev, unsigned long iova, int flags)
>> +{
>> +       struct exynos_iommu_domain *priv = domain->priv;
>> +
>> +       dev_err(priv->dev, "%s occured at %p(Page table base: %p)\n",
>> +                               sysmmu_fault_name[flags], (void *)iova,
>> +                               (void *)(__pa(priv->pgtable)));
>> +       dev_err(priv->dev, "\t\tGenerating Kernel OOPS...\n");
>> +       dev_err(priv->dev, "\t\tbecause it is unrecoverable.\n");
>> +       dev_err(priv->dev,
>> +               "\t\tSet Fault handler with iommu_set_fault_handler().\n");
>> +       dev_err(priv->dev, "\t\tto handle System MMU fault.\n");
>> +
>> +       BUG();
>> +
>> +       return 0;
>> +}
>> +
>> +static int exynos_iommu_domain_init(struct iommu_domain *domain)
>> +{
> ...
>> +       iommu_set_fault_handler(domain, &exynos_iommu_fault_handler);
>
> It doesn't make a lot of sense to set an iommu fault handler here; the
> intention of iommu_set_fault_handler() is to allow upper layers to do
> that.

As you've noticed, it is default behavior for iommu fault.
If an upper layer calls iommu_set_fault_handler(),
it is replaced with the new one.
>
> Moreover, exynos_sysmmu_irq() anyway knows whenever a fault occurs, so
> calling report_iommu_fault() just to have it call into the driver
> again seems redundant.
>
I think calling report_iommu_fault() in exynos_sysmmu_irq() is rather simple.
To do the default behavior in exynos_sysmmu_irq(),
it must check if domain->handler is NULL or not.
But the checking is also performed in report_iommu_fault() also.

I think just calling report_iommu_fault() reduces redundancy.

> If you want an exynos-specific behavior to occur whenever there's an
> iommu fault, you should do that in exynos_sysmmu_irq itself. If you
> just want a generic default logging behavior to occur whenever there's
> no other fault handler installed, you might even want to consider
> adding it to the IOMMU core.

It is neither a specific behavior nor a generic logging.

Just a default MMU fault report that can be replaceable with a fault
handler that a device driver provides.

Regards,
Cho KyongHo.
--
To unsubscribe from this list: send the line "unsubscribe linux-samsung-soc" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html


[Index of Archives]     [Linux SoC Development]     [Linux Rockchip Development]     [Linux USB Development]     [Video for Linux]     [Linux Audio Users]     [Linux SCSI]     [Yosemite News]

  Powered by Linux