On 2024-07-28 21:45, Andrew Lunn wrote:
Basically, the way I see it, weakdeps are the right solution for the
problem at hand, i.e. for the generation of the initial ramdisk with
all the possible PHY driver modules.
Sorry, but i still don't see how this works.
Say you get this one patch merged. What then? You throw all other PHY
drivers which don't have a weakdep out of the initramfs? That
potentially breaks around 200 MAC drivers who need those PHYs you have
discarded.
Actually, no PHY drivers can be thrown out of an initial ramdisk by this
patch, [1] simply because no PHY drivers not needed for a specific
system
shouldn't be included there before, unless they were added specifically
by the utilities that created the initial ramdisk, but that's beyond the
dependencies that the kernel provides in /lib/modules, and beyond what
can be obtained by automatic detection using /sys/devices on a
particular
system. That's all a result of this specific module dependency being
unknown before, at the kernel level.
In other words, this patch doesn't subtract anything. Instead, it just
adds a weakdep link between the lan78xx and micrel modules, so the
kernel
itself can report that dependency, which may actually result in one more
PHY driver added to a generated initial ramdisk.
[1]
https://lore.kernel.org/netdev/20240724102349.430078-1-jtornosm@xxxxxxxxxx/
