Opened 3 years ago
Closed 3 years ago
#847 closed defect (fixed)
Drivers are not filtered by platform anymore
Reported by: | Jiri Svoboda | Owned by: | Jiri Svoboda |
---|---|---|---|
Priority: | major | Milestone: | 0.12.1 |
Component: | helenos/unspecified | Version: | mainline |
Keywords: | Cc: | ||
Blocker for: | Depends on: | ||
See also: |
Description
Drivers used to be filtered by platform, in other words, platform specific drivers would not be delivered on platforms they are not supposed to be used on. This was achieved either by not building them or not copying them to dist.
This functionality was lost with the switch to Meson. Currently we deliver incorrectly, for example, as part of amd64 image these drivers: adb-kbd, adb-mouse, amdm37x, amdm37x_dispc, arm64virt, cuda_adb, ddisk, gicv2, hikey960, icp, icp-ic, mac, malta, msim, msim-con, obio, pl011, pl050, ski, ski-con, sun4v, sun4v-con.
This delivering of many useless drivers bloats the OS images unnecessarily.
Change History (5)
comment:1 by , 3 years ago
comment:2 by , 3 years ago
Milestone: | → 0.12.1 |
---|
comment:3 by , 3 years ago
Actually it turns out the problem is a little different than I thought. If CONFIG_BAREBONE is on, we only get the drivers explicitly enabled in meson/arch/*/meson.build. If CONFIG_BAREBONE is off, we get all the drivers.
I think this is wrong. We don't want to ever include useless drivers in the ramdisk image. Of course that means Meson won't build them either, but I think we can live with that.
Now since we have platforms that have default CONFIG_BAREBONE=n and they can be built barebone with a restricted set of drivers (e.g. amd64), I will assume that we want to be able to get a different set of drivers depending on the barebone setting.
So, we need to introduce another variable, e.g. rd_drivers that will list the drivers to be included always and only install drivers from rd_drivers / rd_essential based on barebone configuration.
comment:4 by , 3 years ago
Owner: | set to |
---|---|
Status: | new → accepted |
comment:5 by , 3 years ago
Resolution: | → fixed |
---|---|
Status: | accepted → closed |
Fixed in changeset 581a54a616cde33e074a6a2e1641b98b682dbd7b
I'm not marking this as a stopper for 0.11.2 since it's not a new issue. I was somewhat aware of this issue (see also #843), but perhaps did not realize its full extent.
We should definitely fix this before the next release.