[PATCH v4] bus: mhi: host: don't free bhie tables during suspend/hibernation

[Jeff Hugo]

On 5/6/2025 8:49 AM, Muhammad Usama Anjum wrote:
> Fix dma_direct_alloc() failure at resume time during bhie_table
> allocation because of memory pressure. There is a report where at
> resume time, the memory from the dma doesn't get allocated and MHI
> fails to re-initialize.
> 
> To fix it, don't free the memory at power down during suspend /
> hibernation. Instead, use the same allocated memory again after every
> resume / hibernation. This patch has been tested with resume and
> hibernation both.
> 
> The rddm is of constant size for a given hardware. While the fbc_image
> size depends on the firmware. If the firmware changes, we'll free and
> allocate new memory for it.

Why is it valid to load new firmware as a result of suspend?  I don't 
users would expect that.

> diff --git a/drivers/bus/mhi/host/boot.c b/drivers/bus/mhi/host/boot.c
> index efa3b6dddf4d2..bc8459798bbee 100644
> --- a/drivers/bus/mhi/host/boot.c
> +++ b/drivers/bus/mhi/host/boot.c
> @@ -584,10 +584,17 @@ void mhi_fw_load_handler(struct mhi_controller *mhi_cntrl)
>   	 * device transitioning into MHI READY state
>   	 */
>   	if (fw_load_type == MHI_FW_LOAD_FBC) {

Why is this FBC specific?

> -		ret = mhi_alloc_bhie_table(mhi_cntrl, &mhi_cntrl->fbc_image, fw_sz);
> -		if (ret) {
> -			release_firmware(firmware);
> -			goto error_fw_load;
> +		if (mhi_cntrl->fbc_image && fw_sz != mhi_cntrl->prev_fw_sz) {
> +			mhi_free_bhie_table(mhi_cntrl, mhi_cntrl->fbc_image);
> +			mhi_cntrl->fbc_image = NULL;
> +		}
> +		if (!mhi_cntrl->fbc_image) {
> +			ret = mhi_alloc_bhie_table(mhi_cntrl, &mhi_cntrl->fbc_image, fw_sz);
> +			if (ret) {
> +				release_firmware(firmware);
> +				goto error_fw_load;
> +			}
> +			mhi_cntrl->prev_fw_sz = fw_sz;
>   		}
>   
>   		/* Load the firmware into BHIE vec table */
> diff --git a/drivers/bus/mhi/host/pm.c b/drivers/bus/mhi/host/pm.c
> index e6c3ff62bab1d..107d71b4cc51a 100644
> --- a/drivers/bus/mhi/host/pm.c
> +++ b/drivers/bus/mhi/host/pm.c
> @@ -1259,10 +1259,19 @@ void mhi_power_down(struct mhi_controller *mhi_cntrl, bool graceful)
>   }
>   EXPORT_SYMBOL_GPL(mhi_power_down);
>   
> +static void __mhi_power_down_unprepare_keep_dev(struct mhi_controller *mhi_cntrl)
> +{
> +	mhi_cntrl->bhi = NULL;
> +	mhi_cntrl->bhie = NULL;

Why?

> +
> +	mhi_deinit_dev_ctxt(mhi_cntrl);
> +}