Bjorn Helgaas <helgaas@xxxxxxxxxx> writes: > On Wed, Jun 09, 2021 at 04:07:20PM -0700, Vinicius Costa Gomes wrote: >> Hi Paul, >> >> >> >> >>> Regarding my comment, I did not mean returning an error but the log >> >>> *level* of the message. So, `dmesg --level err` would show that message. >> >>> But if there are PCI controllers not supporting that, it’s not an error, >> >>> but a warning at most. So, I’d use: >> >>> >> >>> dev_warn(&pdev->dev, "PTM not supported by PCI bus/controller >> >>> (pci_enable_ptm() failed)\n"); >> >> >> >> I will use you suggestion for the message, but I think that warn is a >> >> bit too much, info or notice seem to be better. >> > >> > I do not know, if modern PCI(e)(?) controllers normally support PTM or >> > not. If recent controllers should support it, then a warning would be >> > warranted, otherwise a notice. >> >> From the Intel side, it seems that it's been supported for a few years. >> So, fair enough, let's go with a warn. > > I'm not sure about this. I think "warning" messages interrupt distro > graphical boot scenarios and cause user complaints. In this case, > there is nothing broken and the user can do nothing about it; it's > merely a piece of missing optional functionality. So I think "info" > is a more appropriate level. Good point. "info" it is, then. Cheers, -- Vinicius
