Closed nhahnetsrot closed 9 months ago
Thanks for looking into that! Still, it looks like there are still some todos remaining:
Keep in mind that, mid-term, we would like to get rid of the patch again by sharing our findings with TI and making them adopt them, directly or at least logically. Also for them, we need to explain why we did those changes based on the code, not just the observed stability improvement.
Thanks for looking into that! Still, it looks like there are still some todos remaining:
- resolve the open questions on the fixing patch itself - we really need to understand the issue when trying to fix it
- write the kernel patch like the patch adding it: subject, commit message, signed-off
Keep in mind that, mid-term, we would like to get rid of the patch again by sharing our findings with TI and making them adopt them, directly or at least logically. Also for them, we need to explain why we did those changes based on the code, not just the observed stability improvement.
TI already knows about this bug ( i emailed with Danish Anwar a0501179@ti.com, he ist responsible for the driver at TI side). He stated that a "cleaned up" version of the driver does not has this issue and will be going into the official kernel "soon". Then this patch is no longer required. (he did not define "soon" ...)
Will we get an update? Or will this bug have to wait for replacing the patches with their upstream versions?
Obsoleted by #491
Fixes the crash reported in https://github.com/siemens/meta-iot2050/issues/427 .
Allows now the removal / reload of the driver, network devices, links etc. are restored and working.