Closed davidhuziji closed 3 months ago
Hi, I'm wondering if a flush-to-disk is missing before dumping the message ACS run is completed, in SystemReady IR init.sh: https://github.com/ARM-software/arm-systemready/blob/v23.09_SR_REL2.0.0_ES_REL1.3.0_IR_REL2.1.0/IR/Yocto/meta-woden/recipes-acs/install-files/files/init.sh#L149C6-L149C28.
ACS run is completed
It was observed that those test logs generated from ethtool-test.py and read_blk_devices.py could be lost, without a sync step. So far, we introduce a workaround to flush the log to disk in another thread: https://gitlab.arm.com/automotive-and-industrial/kronos-ref-stack/meta-arm/-/blob/v1.0/meta-arm-systemready/lib/oeqa/runtime/cases/arm_systemready_ir_acs.py#L34.
ethtool-test.py
read_blk_devices.py
sync
I notice that there is an additional sync command followed by a delay in https://github.com/ARM-software/arm-systemready/blob/v23.09_SR_REL2.0.0_ES_REL1.3.0_IR_REL2.1.0/common/ramdisk/init.sh#L177, before the script dumps ACS completion log. Could it be a more comprehensive implementation?
Thank you.
Hi, I'm wondering if a flush-to-disk is missing before dumping the message
ACS run is completed, in SystemReady IR init.sh: https://github.com/ARM-software/arm-systemready/blob/v23.09_SR_REL2.0.0_ES_REL1.3.0_IR_REL2.1.0/IR/Yocto/meta-woden/recipes-acs/install-files/files/init.sh#L149C6-L149C28.It was observed that those test logs generated from
ethtool-test.pyandread_blk_devices.pycould be lost, without asyncstep. So far, we introduce a workaround to flush the log to disk in another thread: https://gitlab.arm.com/automotive-and-industrial/kronos-ref-stack/meta-arm/-/blob/v1.0/meta-arm-systemready/lib/oeqa/runtime/cases/arm_systemready_ir_acs.py#L34.I notice that there is an additional
synccommand followed by a delay in https://github.com/ARM-software/arm-systemready/blob/v23.09_SR_REL2.0.0_ES_REL1.3.0_IR_REL2.1.0/common/ramdisk/init.sh#L177, before the script dumps ACS completion log. Could it be a more comprehensive implementation?Thank you.