On Mon, 2014-01-06 at 12:41 +0000, Mark Brown wrote:
On 6 January 2014 12:36, Jon Medhurst (Tixy) tixy@linaro.org wrote: On Mon, 2014-01-06 at 12:07 +0000, Mark Brown wrote: > Panda is included in our test coverage, though it's not got any out of tree > patches applied so depending on why the fix was needed To stop a #error during compilation. I'll check if it's still needed, what Panda config does LSK use, omap2plus_defconfig or one based on Linaro's config fragments?
Linaro fragments. Depending on where the issue is if it's something that could be triggered by options we might want to include it anyway in case users trigger it with their configs.
Actually, digging through two year old emails, the TI Landing team said of CONFIG_PERF_EVENTS "We turned this off in part to avoid errata in the core for which we have no workaround."
Now gator has an #error if this isn't enabled because it expects to use this feature, though it does compile and works - with much reduced functionality - if this config isn't available. So as a compromise, I downgraded the #error to a #warning and didn't make the KConfig for Gator depend on {HW_}PERF_EVENTS.
Looking at the config fragments I see that omap4.conf doesn't disable the {HW_}PERF_EVENTS configs set in linaro-base.conf so it is in fact now not hitting this code #error/#warning lines anyway.
So I would say that this hack should go now and I'll update the KConfig for gator to depend on the right configs to avoid the #error.
For LSK, I'll probably leave this change until the next Gator release, as it's not doing any harm at the moment.