Linux ARM-MSM sub-architecture
 help / color / mirror / Atom feed
From: Krishna Chaitanya Chundru <quic_krichai@quicinc.com>
To: Muhammad Usama Anjum <usama.anjum@collabora.com>,
	Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>,
	Johannes Berg <johannes@sipsolutions.net>,
	Jeff Johnson <jjohnson@kernel.org>,
	Jeffrey Hugo <quic_jhugo@quicinc.com>,
	Yan Zhen <yanzhen@vivo.com>,
	Youssef Samir <quic_yabdulra@quicinc.com>,
	Qiang Yu <quic_qianyu@quicinc.com>, Alex Elder <elder@kernel.org>,
	Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	"Kunwu Chan" <chentao@kylinos.cn>
Cc: <kernel@collabora.com>, <mhi@lists.linux.dev>,
	<linux-arm-msm@vger.kernel.org>, <linux-kernel@vger.kernel.org>,
	<linux-wireless@vger.kernel.org>, <ath11k@lists.infradead.org>
Subject: Re: [PATCH v2] bus: mhi: host: don't free bhie tables during suspend/hibernation
Date: Fri, 11 Apr 2025 14:09:11 +0530	[thread overview]
Message-ID: <e0159cb8-fe21-7f71-1ebe-744ed26bd698@quicinc.com> (raw)
In-Reply-To: <fc9ca0da-9f6a-42b5-aa79-abcd43c97043@collabora.com>



On 4/11/2025 12:32 PM, Muhammad Usama Anjum wrote:
> On 4/11/25 8:37 AM, Krishna Chaitanya Chundru wrote:
>>
>>
>> On 4/10/2025 8:26 PM, Muhammad Usama Anjum wrote:
>>> Fix dma_direct_alloc() failure at resume time during bhie_table
>>> allocation. There is a crash report where at resume time, the memory
>>> from the dma doesn't get allocated and MHI fails to re-initialize.
>>> There may be fragmentation of some kind which fails the allocation
>>> call.
>>>
>>> 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
>> If firmware image will change between suspend and resume ?
> Yes, correct.
> 
why the firmware image size will change between suspend & resume?
who will update the firmware image after bootup?
It is not expected behaviour.

- Krishna chaitanya.
>>> allocate new memory for it.
>>>
>>> Here are the crash logs:
>>>
>>> [ 3029.338587] mhi mhi0: Requested to power ON
>>> [ 3029.338621] mhi mhi0: Power on setup success
>>> [ 3029.668654] kworker/u33:8: page allocation failure: order:7,
>>> mode:0xc04(GFP_NOIO|GFP_DMA32), nodemask=(null),cpuset=/,mems_allowed=0
>>> [ 3029.668682] CPU: 4 UID: 0 PID: 2744 Comm: kworker/u33:8 Not tainted
>>> 6.11.11-valve10-1-neptune-611-gb69e902b4338
>>> #1ed779c892334112fb968aaa3facf9686b5ff0bd7
>>> [ 3029.668690] Hardware name: Valve Galileo/Galileo, BIOS F7G0112
>>> 08/01/2024
>>> [ 3029.668694] Workqueue: mhi_hiprio_wq mhi_pm_st_worker [mhi]
>>> [ 3029.668717] Call Trace:
>>> [ 3029.668722]  <TASK>
>>> [ 3029.668728]  dump_stack_lvl+0x4e/0x70
>>> [ 3029.668738]  warn_alloc+0x164/0x190
>>> [ 3029.668747]  ? srso_return_thunk+0x5/0x5f
>>> [ 3029.668754]  ? __alloc_pages_direct_compact+0xaf/0x360
>>> [ 3029.668761]  __alloc_pages_slowpath.constprop.0+0xc75/0xd70
>>> [ 3029.668774]  __alloc_pages_noprof+0x321/0x350
>>> [ 3029.668782]  __dma_direct_alloc_pages.isra.0+0x14a/0x290
>>> [ 3029.668790]  dma_direct_alloc+0x70/0x270
>>> [ 3029.668796]  mhi_alloc_bhie_table+0xe8/0x190 [mhi
>>> faa917c5aa23a5f5b12d6a2c597067e16d2fedc0]
>>> [ 3029.668814]  mhi_fw_load_handler+0x1bc/0x310 [mhi
>>> faa917c5aa23a5f5b12d6a2c597067e16d2fedc0]
>>> [ 3029.668830]  mhi_pm_st_worker+0x5c8/0xaa0 [mhi
>>> faa917c5aa23a5f5b12d6a2c597067e16d2fedc0]
>>> [ 3029.668844]  ? srso_return_thunk+0x5/0x5f
>>> [ 3029.668853]  process_one_work+0x17e/0x330
>>> [ 3029.668861]  worker_thread+0x2ce/0x3f0
>>> [ 3029.668868]  ? __pfx_worker_thread+0x10/0x10
>>> [ 3029.668873]  kthread+0xd2/0x100
>>> [ 3029.668879]  ? __pfx_kthread+0x10/0x10
>>> [ 3029.668885]  ret_from_fork+0x34/0x50
>>> [ 3029.668892]  ? __pfx_kthread+0x10/0x10
>>> [ 3029.668898]  ret_from_fork_asm+0x1a/0x30
>>> [ 3029.668910]  </TASK>
>>>
>>> Tested-on: QCNFA765 WLAN.HSP.1.1-03926.13-
>>> QCAHSPSWPL_V2_SILICONZ_CE-2.52297.6
>>>
>>> Signed-off-by: Muhammad Usama Anjum <usama.anjum@collabora.com>
>>> ---
>>> Changes sice v1:
>>> - Don't free bhie tables during suspend/hibernation only
>>> - Handle fbc_image changed size correctly
>>> - Remove fbc_image getting set to NULL in *free_bhie_table()
>>> ---
>>>    drivers/bus/mhi/host/boot.c           | 15 +++++++++++----
>>>    drivers/bus/mhi/host/init.c           | 13 ++++++++++---
>>>    drivers/net/wireless/ath/ath11k/mhi.c |  9 +++++----
>>>    include/linux/mhi.h                   |  7 +++++++
>>>    4 files changed, 33 insertions(+), 11 deletions(-)
>>>
>>> diff --git a/drivers/bus/mhi/host/boot.c b/drivers/bus/mhi/host/boot.c
>>> index 9dcc7184817d5..0df26100c8f9c 100644
>>> --- a/drivers/bus/mhi/host/boot.c
>>> +++ b/drivers/bus/mhi/host/boot.c
>>> @@ -487,10 +487,17 @@ void mhi_fw_load_handler(struct mhi_controller
>>> *mhi_cntrl)
>>>         * device transitioning into MHI READY state
>>>         */
>>>        if (mhi_cntrl->fbc_download) {
>>> -        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/init.c b/drivers/bus/mhi/host/init.c
>>> index a9b1f8beee7bc..09b946b86ac46 100644
>>> --- a/drivers/bus/mhi/host/init.c
>>> +++ b/drivers/bus/mhi/host/init.c
>>> @@ -1173,8 +1173,9 @@ int mhi_prepare_for_power_up(struct
>>> mhi_controller *mhi_cntrl)
>>>            /*
>>>             * Allocate RDDM table for debugging purpose if specified
>>>             */
>>> -        mhi_alloc_bhie_table(mhi_cntrl, &mhi_cntrl->rddm_image,
>>> -                     mhi_cntrl->rddm_size);
>>> +        if (!mhi_cntrl->rddm_image)
>>> +            mhi_alloc_bhie_table(mhi_cntrl, &mhi_cntrl->rddm_image,
>>> +                         mhi_cntrl->rddm_size);
>>>            if (mhi_cntrl->rddm_image) {
>>>                ret = mhi_rddm_prepare(mhi_cntrl,
>>>                               mhi_cntrl->rddm_image);
>>> @@ -1212,12 +1213,18 @@ void mhi_unprepare_after_power_down(struct
>>> mhi_controller *mhi_cntrl)
>>>            mhi_cntrl->rddm_image = NULL;
>>>        }
>>>    +    mhi_partial_unprepare_after_power_down(mhi_cntrl);
>>> +}
>>> +EXPORT_SYMBOL_GPL(mhi_unprepare_after_power_down);
>>> +
>>> +void mhi_partial_unprepare_after_power_down(struct mhi_controller
>>> *mhi_cntrl)
>>> +{
>>>        mhi_cntrl->bhi = NULL;
>>>        mhi_cntrl->bhie = NULL;
>>>          mhi_deinit_dev_ctxt(mhi_cntrl);
>>>    }
>>> -EXPORT_SYMBOL_GPL(mhi_unprepare_after_power_down);
>>> +EXPORT_SYMBOL_GPL(mhi_partial_unprepare_after_power_down);
>>>    
>> Instead of adding new API you can free memory from the unregister
>> controller also.
>>
>> - Krishna Chaitanya.
>>>    static void mhi_release_device(struct device *dev)
>>>    {
>>> diff --git a/drivers/net/wireless/ath/ath11k/mhi.c b/drivers/net/
>>> wireless/ath/ath11k/mhi.c
>>> index acd76e9392d31..f77cec79b5b80 100644
>>> --- a/drivers/net/wireless/ath/ath11k/mhi.c
>>> +++ b/drivers/net/wireless/ath/ath11k/mhi.c
>>> @@ -460,12 +460,13 @@ void ath11k_mhi_stop(struct ath11k_pci *ab_pci,
>>> bool is_suspend)
>>>         * workaround, otherwise ath11k_core_resume() will timeout
>>>         * during resume.
>>>         */
>>> -    if (is_suspend)
>>> +    if (is_suspend) {
>>>            mhi_power_down_keep_dev(ab_pci->mhi_ctrl, true);
>>> -    else
>>> +        mhi_partial_unprepare_after_power_down(ab_pci->mhi_ctrl);
>>> +    } else {
>>>            mhi_power_down(ab_pci->mhi_ctrl, true);
>>> -
>>> -    mhi_unprepare_after_power_down(ab_pci->mhi_ctrl);
>>> +        mhi_unprepare_after_power_down(ab_pci->mhi_ctrl);
>>> +    }
>>>    }
>>>      int ath11k_mhi_suspend(struct ath11k_pci *ab_pci)
>>> diff --git a/include/linux/mhi.h b/include/linux/mhi.h
>>> index 059dc94d20bb6..65a47c712b3a0 100644
>>> --- a/include/linux/mhi.h
>>> +++ b/include/linux/mhi.h
>>> @@ -382,6 +382,7 @@ struct mhi_controller {
>>>        const char *fw_image;
>>>        const u8 *fw_data;
>>>        size_t fw_sz;
>>> +    size_t prev_fw_sz;
>>>        const char *edl_image;
>>>        size_t rddm_size;
>>>        size_t sbl_size;
>>> @@ -662,6 +663,12 @@ void mhi_power_down_keep_dev(struct
>>> mhi_controller *mhi_cntrl, bool graceful);
>>>     */
>>>    void mhi_unprepare_after_power_down(struct mhi_controller *mhi_cntrl);
>>>    +/**
>>> + * mhi_partial_unprepare_after_power_down - Free any allocated memory
>>> after power down partially
>>> + * @mhi_cntrl: MHI controller
>>> + */
>>> +void mhi_partial_unprepare_after_power_down(struct mhi_controller
>>> *mhi_cntrl);
>>> +
>>>    /**
>>>     * mhi_pm_suspend - Move MHI into a suspended state
>>>     * @mhi_cntrl: MHI controller
> 
> 

  reply	other threads:[~2025-04-11  8:39 UTC|newest]

Thread overview: 24+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-04-10 14:56 [PATCH v2] bus: mhi: host: don't free bhie tables during suspend/hibernation Muhammad Usama Anjum
2025-04-10 17:00 ` Greg Kroah-Hartman
2025-04-11 18:47   ` Muhammad Usama Anjum
2025-04-11  3:37 ` Krishna Chaitanya Chundru
2025-04-11  7:02   ` Muhammad Usama Anjum
2025-04-11  8:39     ` Krishna Chaitanya Chundru [this message]
2025-04-11 18:32       ` Muhammad Usama Anjum
2025-04-12  1:22         ` Krishna Chaitanya Chundru
2025-04-14  7:32           ` Muhammad Usama Anjum
2025-04-14 14:14             ` Jeff Hugo
2025-04-18  8:10               ` Muhammad Usama Anjum
2025-04-18 14:08                 ` Jeff Hugo
2025-04-22  7:23                   ` Muhammad Usama Anjum
2025-04-22 14:22                     ` Jeff Hugo
2025-04-23  6:41                       ` Muhammad Usama Anjum
2025-04-11 16:10 ` Jeff Hugo
2025-04-11 19:10   ` Muhammad Usama Anjum
2025-04-25  7:04 ` Manivannan Sadhasivam
2025-04-25  7:14   ` Muhammad Usama Anjum
2025-04-25  7:32     ` Manivannan Sadhasivam
2025-04-25  7:42       ` Muhammad Usama Anjum
2025-04-25  8:59         ` Manivannan Sadhasivam
2025-04-25 11:41           ` Muhammad Usama Anjum
2025-04-25 14:47             ` Manivannan Sadhasivam

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=e0159cb8-fe21-7f71-1ebe-744ed26bd698@quicinc.com \
    --to=quic_krichai@quicinc.com \
    --cc=ath11k@lists.infradead.org \
    --cc=chentao@kylinos.cn \
    --cc=elder@kernel.org \
    --cc=gregkh@linuxfoundation.org \
    --cc=jjohnson@kernel.org \
    --cc=johannes@sipsolutions.net \
    --cc=kernel@collabora.com \
    --cc=linux-arm-msm@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-wireless@vger.kernel.org \
    --cc=manivannan.sadhasivam@linaro.org \
    --cc=mhi@lists.linux.dev \
    --cc=quic_jhugo@quicinc.com \
    --cc=quic_qianyu@quicinc.com \
    --cc=quic_yabdulra@quicinc.com \
    --cc=usama.anjum@collabora.com \
    --cc=yanzhen@vivo.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox