diff mbox

[V1,00/18] mmc: Add Command Queue support

Message ID 84340004-5013-e934-c311-34eaa7760e61@rock-chips.com (mailing list archive)
State New, archived
Headers show

Commit Message

Shawn Lin March 9, 2017, 8:14 a.m. UTC 
On 2017/3/9 10:47, Shawn Lin wrote:
> On 2017/3/8 21:27, Jaehoon Chung wrote:
>> On 03/08/2017 05:08 PM, Adrian Hunter wrote:
>>> On 08/03/17 07:18, Jaehoon Chung wrote:
>>>> Hi Adrian,
>>>>
>>>> On 03/06/2017 06:10 PM, Adrian Hunter wrote:
>>>>> Hi
>>>>>
>>>>> Here are the hardware command queue patches without the software
>>>>> command
>>>>> queue patches or sdhci patches.
>>>>
>>>> I don't see your patches fully yet, but after applied your patches,
>>>> dwmmc host doesn't work.
>>>> So I'm finding what happen in dwmmc controller.
>>>> After finding problem, i will share it.
>>>> Well,.. maybe i missed something...
>>>
>>> Thanks for trying these patches.
>>>
>>> With no host controller yet using CQE, there isn't much changing in this
>>> patch set, but bisecting should reveal the first bad commit.
>>
>> https://patchwork.kernel.org/patch/9605315/
>>
>> It seems that this patch has a problem...Tomorrow i will check in more
>> detail.
>> Because I saw the different error logs..
>> Sometime occurred the NULL pointer dereference or I/O error..
>>
>> Didn't other guys see these problem?
>
> I was running linux-next with this pachset, but it can't boot up my
> Debain system, and mostly the log looks like
>
> [    4.137331] Hardware name: Rockchip (Device Tree)
> [    4.142585] task: ee753180 task.stack: ee7ca000
> [    4.142593] PC is at mmc_blk_rw_rq_prep+0x2c/0x3b0
> [    4.142595] LR is at mmc_blk_issue_rw_rq+0x64/0x340
> [    4.142599] pc : [<c0a85b74>]    lr : [<c0a86844>]    psr: 8001001
>
> so the first argument, namely mqrq is freed somehow.
>
> c0a85b48 <mmc_blk_rw_rq_prep>:
> c0a85b48:       e92d4ff0        push    {r4, r5, r6, r7, r8, r9, sl, fp,
> lr}
> c0a85b4c:       e1a0a003        mov     sl, r3
> c0a85b50:       e593301c        ldr     r3, [r3, #28]
> c0a85b54:       e24dd00c        sub     sp, sp, #12
> c0a85b58:       e2909004        adds    r9, r0, #4
> c0a85b5c:       e1a04000        mov     r4, r0
> c0a85b60:       e1a05001        mov     r5, r1
> c0a85b64:       e1a0b002        mov     fp, r2
> c0a85b68:       e5906000        ldr     r6, [r0]
> c0a85b6c:       e58d3004        str     r3, [sp, #4]
> c0a85b70:       0a0000d8        beq     c0a85ed8 <mmc_blk_rw_rq_prep+0x390>
> /* we trap here for fetching r6, and it was from [r0], the mqrq */
> c0a85b74:       e5968024        ldr     r8, [r6, #36]   ; 0x24
> c0a85b78:       e3a02802        mov     r2, #131072     ; 0x20000
>

Just update what I found when rebooting my board again, I am seeing
my userspace are trying to access RPMB and BOOT part endlessly...

I haven't enough time to look closely at what happened. But from the
commit msg of patch 1, I suspet is it correct for us to say the queue
was empty before switch partitions?(I need to check if there is a case 
that the queue was empty but the mrq was hold by host drivers, for
instance, dwmmc wouldqueue more mrqs when it is busy with doing
previous mrq)

Then I did a quick HACK intuitively by removing the BOOT AREA and it
works....

                         for (idx = 0; idx < MMC_NUM_BOOT_PARTITION; 
idx++) {
                                 part_size = ext_csd[EXT_CSD_BOOT_MULT] 
<< 17;
@@ -455,7 +456,7 @@ static int mmc_decode_ext_csd(struct mmc_card *card, 
u8 *ext_csd)
                                         "boot%d", idx, true,
                                         MMC_BLK_DATA_AREA_BOOT);
                         }
-               }
+               }*/


Thought?

>
>>
>> Best Regards,
>> Jaehoon Chung
>>
>>>
>>>>
>>>> Best Regards,
>>>> Jaehoon Chung
>>>>
>>>>>
>>>>>
>>>>> Changes since RFC:
>>>>>
>>>>>     Re-based on next.
>>>>>     Added comment about command queue priority.
>>>>>     Added some acks and reviews.
>>>>>
>>>>>
>>>>> Adrian Hunter (17):
>>>>>       mmc: queue: Share mmc request array between partitions
>>>>>       mmc: mmc: Add functions to enable / disable the Command Queue
>>>>>       mmc: mmc_test: Disable Command Queue while mmc_test is used
>>>>>       mmc: block: Disable Command Queue while RPMB is used
>>>>>       mmc: block: Change mmc_apply_rel_rw() to get block address
>>>>> from the request
>>>>>       mmc: block: Factor out data preparation
>>>>>       mmc: core: Factor out debug prints from mmc_start_request()
>>>>>       mmc: core: Factor out mrq preparation from mmc_start_request()
>>>>>       mmc: core: Add mmc_retune_hold_now()
>>>>>       mmc: core: Add members to mmc_request and mmc_data for CQE's
>>>>>       mmc: host: Add CQE interface
>>>>>       mmc: core: Turn off CQE before sending commands
>>>>>       mmc: core: Add support for handling CQE requests
>>>>>       mmc: mmc: Enable Command Queuing
>>>>>       mmc: mmc: Enable CQE's
>>>>>       mmc: block: Prepare CQE data
>>>>>       mmc: block: Add CQE support
>>>>>
>>>>> Venkat Gopalakrishnan (1):
>>>>>       mmc: cqhci: support for command queue enabled host
>>>>>
>>>>>  Documentation/mmc/mmc-dev-attrs.txt |    1 +
>>>>>  drivers/mmc/core/block.c            |  435 ++++++++++---
>>>>>  drivers/mmc/core/block.h            |    7 +
>>>>>  drivers/mmc/core/bus.c              |    7 +
>>>>>  drivers/mmc/core/core.c             |  203 ++++++-
>>>>>  drivers/mmc/core/host.c             |    6 +
>>>>>  drivers/mmc/core/host.h             |    1 +
>>>>>  drivers/mmc/core/mmc.c              |   39 +-
>>>>>  drivers/mmc/core/mmc_ops.c          |   28 +
>>>>>  drivers/mmc/core/mmc_ops.h          |    2 +
>>>>>  drivers/mmc/core/mmc_test.c         |   14 +
>>>>>  drivers/mmc/core/queue.c            |  535 +++++++++++++---
>>>>>  drivers/mmc/core/queue.h            |   45 +-
>>>>>  drivers/mmc/host/Kconfig            |   13 +
>>>>>  drivers/mmc/host/Makefile           |    1 +
>>>>>  drivers/mmc/host/cqhci.c            | 1148
>>>>> +++++++++++++++++++++++++++++++++++
>>>>>  drivers/mmc/host/cqhci.h            |  240 ++++++++
>>>>>  include/linux/mmc/card.h            |    8 +
>>>>>  include/linux/mmc/core.h            |   19 +-
>>>>>  include/linux/mmc/host.h            |   24 +
>>>>>  include/trace/events/mmc.h          |   17 +-
>>>>>  21 files changed, 2585 insertions(+), 208 deletions(-)
>>>>>  create mode 100644 drivers/mmc/host/cqhci.c
>>>>>  create mode 100644 drivers/mmc/host/cqhci.h
>>>>>
>>>>>
>>>>> Regards
>>>>> Adrian
>>>>>
>>>>>
>>>>>
>>>>
>>>>
>>>
>>> --
>>> To unsubscribe from this list: send the line "unsubscribe linux-mmc" in
>>> the body of a message to majordomo@vger.kernel.org
>>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>>>
>>>
>>
>> --
>> To unsubscribe from this list: send the line "unsubscribe linux-mmc" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>>
>
>
diff mbox

Patch

--- a/drivers/mmc/core/mmc.c
+++ b/drivers/mmc/core/mmc.c
@@ -447,6 +447,7 @@  static int mmc_decode_ext_csd(struct mmc_card *card, 
u8 *ext_csd)
                  * There are two boot regions of equal size, defined in
                  * multiples of 128K.
                  */
+               /*
                 if (ext_csd[EXT_CSD_BOOT_MULT] && 
mmc_boot_partition_access(card->host)) {