在 2024/9/7 0:00, Mathieu Poirier 写道:
You still have missed several instances. I am dropping this set.
Hi, Mathieu,
I have checked the subsystem again and does not find the missing
instances that can make such conversion.
Instance like this would require storing the resource size or the
res->start, so we can not conversion likt that:
------------------------code start---------------------------------
da8xx_remoteproc.c:
res = platform_get_resource_byname(pdev, IORESOURCE_MEM,
mem_names[i]);
drproc->mem[i].cpu_addr = devm_ioremap_resource(dev, res);
if (IS_ERR(drproc->mem[i].cpu_addr)) {
dev_err(dev, "failed to parse and map %s memory\n",
mem_names[i]);
return PTR_ERR(drproc->mem[i].cpu_addr);
}
drproc->mem[i].bus_addr = res->start;
drproc->mem[i].dev_addr =
res->start & DA8XX_RPROC_LOCAL_ADDRESS_MASK;
drproc->mem[i].size = resource_size(res);
------------------------------------------------------------------
I have thought about adding a
devm_platform_get_and_ioremap_resource_byname() to make conversion for
these instances, but the function name seems to be too long...
For other instance like below, we will have code logic broken, because
devm_platform_ioremap_resource_byname() will return error if res is NULL:
---------------------------code start------------------------------
mtk_scp.c:
/* l1tcm is an optional memory region */
res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "l1tcm");
if (res) {
scp_cluster->l1tcm_base = devm_ioremap_resource(dev, res);
if (IS_ERR(scp_cluster->l1tcm_base))
return dev_err_probe(dev,
PTR_ERR(scp_cluster->l1tcm_base),
"Failed to map l1tcm memory\n");
scp_cluster->l1tcm_size = resource_size(res);
scp_cluster->l1tcm_phys = res->start;
}
-------------------------------------------------------------------
Best Regards,
Zekun
