Re: [PATCH linux-next] thermal/drivers/qcom-spmi-mbg-tm: Add module namespace import for IIO_CONSUMER
Jonathan Cameron <[email protected]>
| Newsgroups | org.kernel.vger.linux-pm,org.kernel.vger.linux-arm-msm,org.kernel.vger.linux-iio,org.kernel.vger.linux-kernel |
|---|---|
| Message-ID | <20260813023543.6936b188@jic23-huawei> |
On Wed, 12 Aug 2026 18:17:22 -0700 Nathan Chancellor <[email protected]> wrote: > Commit ebf1d03dab96 ("iio: inkern: Use namespaced exports") in the iio > tree restricts certain exported core functions that a driver added in > commit c3dce117333c ("thermal/drivers/qcom: Add support for Qualcomm MBG > thermal monitoring") from the thermal tree uses, causing modpost to > warn (or error without CONFIG_MODULE_ALLOW_MISSING_NAMESPACE_IMPORTS): > > ERROR: modpost: drivers/thermal/qcom/qcom-spmi-mbg-tm.ko: module uses symbol 'devm_iio_channel_get' from namespace 'IIO_CONSUMER', but does not import it. > ERROR: modpost: drivers/thermal/qcom/qcom-spmi-mbg-tm.ko: module uses symbol 'iio_read_channel_processed' from namespace 'IIO_CONSUMER', but does not import it. > > Add the IIO_CONSUMER namespace import to clear up the error. > > Signed-off-by: Nathan Chancellor <[email protected]> Ah. I'd forgotten we had the namespace addition earlier this cycle. Too much going on at the moment :( Sorry about that - I should have given you a heads up. Another win for linux-next testing things! > --- > I am not sure how this should be handled. As far as I can tell, there is > no error if this patch is applied to the thermal tree directly but it > would leave an "orphaned" module namespace import until the iio and > thermal trees are combined, which may not be preferable. Otherwise, it > would need to be applied to whatever merge happens second (hence > including Mark and Jonathan in this conversation). I did plan for this being a possible issue: Ideal is merge https://git.kernel.org/pub/scm/linux/kernel/git/jic23/iio.git #iio-inkern-namespace-ib to the thermal tree. If I'd called that out originally (or a thermal rebase is a possibility) it would be cleaner to do this before the qcom patch adds another user then squash what you have here in the appropriate patch. Alternative is do it as a merge conflict resolution whilst merging in that tree. If the second, I'd also mention that in the thermal pull request to Linus given it's a little unusual. Jonathan > --- > drivers/thermal/qcom/qcom-spmi-mbg-tm.c | 1 + > 1 file changed, 1 insertion(+) > > diff --git a/drivers/thermal/qcom/qcom-spmi-mbg-tm.c b/drivers/thermal/qcom/qcom-spmi-mbg-tm.c > index 0492d5eeca01..a05bb444627a 100644 > --- a/drivers/thermal/qcom/qcom-spmi-mbg-tm.c > +++ b/drivers/thermal/qcom/qcom-spmi-mbg-tm.c > @@ -254,3 +254,4 @@ module_platform_driver(mbg_tm_driver); > > MODULE_DESCRIPTION("PMIC MBG Temperature monitor driver"); > MODULE_LICENSE("GPL"); > +MODULE_IMPORT_NS("IIO_CONSUMER"); > > --- > base-commit: 28d012efb4327f9c75d5e042a7c91e9a542efa98 > change-id: 20260812-qcom-spmi-mbg-tm-ns-modpost-error-fad8be2f985c > > Best regards, > -- > Cheers, > Nathan >