On Thu, 30 Apr 2026, Rosen Penev wrote: > Simplifies allocations slightly. > > Add __counted_by for extra runtime analysis. > > Signed-off-by: Rosen Penev <[email protected]> > --- > drivers/platform/x86/intel/plr_tpmi.c | 9 ++------- > 1 file changed, 2 insertions(+), 7 deletions(-) > > diff --git a/drivers/platform/x86/intel/plr_tpmi.c > b/drivers/platform/x86/intel/plr_tpmi.c > index 05727169f49c..2fd27e80bc1f 100644 > --- a/drivers/platform/x86/intel/plr_tpmi.c > +++ b/drivers/platform/x86/intel/plr_tpmi.c > @@ -57,9 +57,9 @@ struct tpmi_plr_die { > > struct tpmi_plr { > struct dentry *dbgfs_dir; > - struct tpmi_plr_die *die_info; > int num_dies; > struct auxiliary_device *auxdev; > + struct tpmi_plr_die die_info[] __counted_by(num_dies); > }; > > static const char * const plr_coarse_reasons[] = { > @@ -278,15 +278,10 @@ static int intel_plr_probe(struct auxiliary_device > *auxdev, const struct auxilia > if (!num_resources) > return -EINVAL; > > - plr = devm_kzalloc(&auxdev->dev, sizeof(*plr), GFP_KERNEL); > + plr = devm_kzalloc(&auxdev->dev, struct_size(plr, die_info, > num_resources), GFP_KERNEL);
Is there some particular reason why we don't have devm_kzalloc_flex() other than that nobody has yet bothered to added one. One of the most annoying thing with the various alloc APIs is that devm_* variants seem to never keep up with them. It would be nice if there would be some generic solution that when somebody decides a new alloc api func, the corresponding devm counterpart would be auto-generated for it. -- i.
