Re: [PATCH v3] firmware: qcom_scm: Clear download bit during reboot

From: Mukesh Ojha
Date: Tue May 23 2023 - 05:42:46 EST




On 5/22/2023 3:34 PM, Robert Marko wrote:
On Mon, 22 May 2023 at 11:26, Robert Marko <robimarko@xxxxxxxxx> wrote:

On Mon, 22 May 2023 at 11:11, Mukesh Ojha <quic_mojha@xxxxxxxxxxx> wrote:



On 5/22/2023 2:29 PM, Robert Marko wrote:
On Mon, 22 May 2023 at 08:11, Mukesh Ojha <quic_mojha@xxxxxxxxxxx> wrote:



On 5/18/2023 3:45 PM, Robert Marko wrote:

On 16. 03. 2023. 16:14, Mukesh Ojha wrote:
During normal restart of a system download bit should
be cleared irrespective of whether download mode is
set or not.

Fixes: 8c1b7dc9ba22 ("firmware: qcom: scm: Expose download-mode control")
Signed-off-by: Mukesh Ojha <quic_mojha@xxxxxxxxxxx>

Hi, this has been backported to 5.15.111, however it seems to be
breaking reboot
on IPQ4019 by causing the board to then hang in SBL with:
root@OpenWrt:/# reboot
root@OpenWrt:/# [ 76.473541] device lan1 left promiscuous mode
[ 76.474204] br-lan: port 1(lan1) entered disabled state
[ 76.527975] device lan2 left promiscuous mode
[ 76.530301] br-lan: port 2(lan2) entered disabled state
[ 76.579376] device lan3 left promiscuous mode
[ 76.581698] br-lan: port 3(lan3) entered disabled state
[ 76.638434] device lan4 left promiscuous mode
[ 76.638777] br-lan: port 4(lan4) entered disabled state
[ 76.978489] qca8k-ipq4019 c000000.switch wan: Link is Down
[ 76.978883] device eth0 left promiscuous mode
[ 76.987077] ipqess-edma c080000.ethernet eth0: Link is Down
[
Format: Log Type - Time(microsec) - Message - Optional Info
Log Type: B - Since Boot(Power On Reset), D - Delta, S - Statistic
S - QC_IMAGE_VERSION_STRING=BOOT.BF.3.1.1-00123
S - IMAGE_VARIANT_STRING=DAABANAZA
S - OEM_IMAGE_VERSION_STRING=CRM
S - Boot Config, 0x00000021
S - Reset status Config, 0x00000010
S - Core 0 Frequency, 0 MHz
B - 261 - PBL, Start
B - 1339 - bootable_media_detect_entry, Start
B - 1679 - bootable_media_detect_success, Start
B - 1693 - elf_loader_entry, Start
B - 5076 - auth_hash_seg_entry, Start
B - 7223 - auth_hash_seg_exit, Start
B - 578349 - elf_segs_hash_verify_entry, Start
B - 696356 - PBL, End
B - 696380 - SBL1, Start
B - 787236 - pm_device_init, Start
D - 7 - pm_device_init, Delta
B - 788701 - boot_flash_init, Start
D - 52782 - boot_flash_init, Delta
B - 845625 - boot_config_data_table_init, Start
D - 3836 - boot_config_data_table_init, Delta - (419 Bytes)
B - 852841 - clock_init, Start
D - 7566 - clock_init, Delta
B - 864883 - CDT version:2,Platform ID:9,Major ID:0,Minor
ID:0,Subtype:64
B - 868413 - sbl1_ddr_set_params, Start
B - 873402 - cpr_init, Start
D - 2 - cpr_init, Delta
B - 877842 - Pre_DDR_clock_init, Start
D - 4 - Pre_DDR_clock_init, Delta
D - 13234 - sbl1_ddr_set_params, Delta
B - 891155 - pm_driver_init, Start
D - 2 - pm_driver_init, Delta
B - 909105 - Image Load, Start
B - 1030210 - Boot error ocuured!. Error code: 303d

Reverting the commit fixes rebooting.

Hi Robert,

Can you check if disable SDI [1] works with this issue

https://lore.kernel.org/linux-arm-msm/20230518140224.2248782-1-robimarko@xxxxxxxxx/

[1]


diff --git a/drivers/firmware/qcom_scm.c b/drivers/firmware/qcom_scm.c
index fde33acd46b7..01496ceb7136 100644
--- a/drivers/firmware/qcom_scm.c
+++ b/drivers/firmware/qcom_scm.c
@@ -1508,6 +1508,7 @@ static int qcom_scm_probe(struct platform_device
*pdev)
static void qcom_scm_shutdown(struct platform_device *pdev)
{
/* Clean shutdown, disable download mode to allow normal restart */
+ qcom_scm_disable_sdi();
qcom_scm_set_download_mode(false);
}

Hi,
I can confirm reboot works this way as well.

That's great, So, i don't need to revert the patch and you can
add this in your patch without target specific check ?

Oh, you mean IPQ4019 not rebooting?
I haven't tested that, give me couple of minutes to try that out.
Cause, the link was just back to the SDI patchset.

And, I can confirm that IPQ4019 does not reboot even with SDI disabled if dload
mode was set so it still needs a revert.

Ok, So, before we go for revert of the change.

- How do you generally collect the ram dump on your device on crash ?
did you check if you get any error when qcom_scm_set_download_mode()
get called.


-- Mukesh


Regards,
Robert

Regards,
Robert

-- Mukesh


Regards,
Robert


-- Mukesh


Regards,
Robert

---
Changes in v3:
- Added Fixes tag.
- Removed it from below patch series, as it makes sense to go this
independently.

https://lore.kernel.org/lkml/1677664555-30191-1-git-send-email-quic_mojha@xxxxxxxxxxx/

Changes in v2:
- No change.

drivers/firmware/qcom_scm.c | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/drivers/firmware/qcom_scm.c b/drivers/firmware/qcom_scm.c
index 468d4d5..3e020d1 100644
--- a/drivers/firmware/qcom_scm.c
+++ b/drivers/firmware/qcom_scm.c
@@ -1506,8 +1506,7 @@ static int qcom_scm_probe(struct platform_device
*pdev)
static void qcom_scm_shutdown(struct platform_device *pdev)
{
/* Clean shutdown, disable download mode to allow normal restart */
- if (download_mode)
- qcom_scm_set_download_mode(false);
+ qcom_scm_set_download_mode(false);
}
static const struct of_device_id qcom_scm_dt_match[] = {