On Wed, 14 May 2025, Lukas Wunner wrote: > On Mon, May 12, 2025 at 03:08:57PM +0300, Ilpo Järvinen wrote: > > There are quite many reset and restore related functions in pci.c that > > barely depend on the other functions in pci.c. Create reset-restore.c > > for reset and restore related logic to keep those 1k lines in one place. > > > > Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@xxxxxxxxxxxxxxx> > > Hm, could I get a: > > Suggested-by: Lukas Wunner <lukas@xxxxxxxxx> > > ... per: > > https://lore.kernel.org/r/Z7hZZNT5NHYncZ3c@xxxxxxxxx/ Ah, I hadn't even noticed you suggested it (I recall reading the first paragraph of that but not the last one which made the suggestion). I made this patch first back in 2024 and have just sit on top of the change until there seems to be reasonably conflict free window. > > drivers/pci/Makefile | 4 +- > > drivers/pci/pci.c | 1015 +---------------------------------- > > drivers/pci/pci.h | 10 + > > drivers/pci/reset-restore.c | 1014 ++++++++++++++++++++++++++++++++++ > > I'd prefer reset.c for succinctness. I initially had reset.c but was worried the name is too narrow scoped, I can change back to reset.c. > That said, this patch conflicts with Mani's slot reset patches > which a lot of people seem to be interested in: > > https://lore.kernel.org/r/20250508-pcie-reset-slot-v4-0-7050093e2b50@xxxxxxxxxx/ > > Maybe it's better to give Mani's series the advantage and defer > this patch here to the next cycle. Fine for me but those conflicts looks quite simple. Next cycle will have its own share of conflicts, I'm sure :-). > > --- a/drivers/pci/pci.c > > +++ b/drivers/pci/pci.c > > @@ -69,15 +69,7 @@ struct pci_pme_device { > > */ > > #define PCI_RESET_WAIT 1000 /* msec */ > > I'd move PCI_RESET_WAIT, pci_dev_wait() and > pci_bridge_wait_for_secondary_bus() to reset.c as well. > Then pci_dev_d3_sleep() is the only function which is no longer static. Okay I'll move those as well but that static statement is not exactly true, I'll these need to do these as well: - move pci_bus_max_d3cold_delay() along with pci_bridge_wait_for_secondary_bus() to keep that static, or turn that into a non-static. - make pcie_wait_for_link_delay() non-static. -- i.
