Hi all! on Wed, 18 Dec 2024 13:35:07 -0300, Luis Claudio R. Goncalves wrote: >On Wed, Dec 18, 2024 at 08:57:53AM -0700, Keith Busch wrote: >> On Wed, Dec 18, 2024 at 04:48:38PM +0100, Sebastian Andrzej Siewior wrote: >> > On 2024-12-18 08:36:54 [-0700], Keith Busch wrote: >> > > On Wed, Dec 18, 2024 at 08:59:51PM +0900, Ryo Takakura wrote: >> > > > PCI config access is locked with pci_lock which serializes >> > > > pci_user/bus_write_config*() and pci_user/bus_read_config*(). >> > > > The subsequently invoked vmd_pci_write() and vmd_pci_read() are also >> > > > serialized as they are only invoked by them respectively. >> > > > >> > > > Remove cfg_lock which is taken by vmd_pci_write() and vmd_pci_read() >> > > > for their serialization as its already serialized by pci_lock. >> > > >> > > That's only true if CONFIG_PCI_LOCKLESS_CONFIG isn't set, so pci_lock >> > > won't help with concurrent kernel config access in such a setup. I think >> > > the previous change to raw lock proposal was the correct approach. >> > >> > I overlooked that. Wouldn't it make sense to let the vmd driver select >> > that option rather than adding/ having a lock for the same purpose? >> >> The arch/x86/Kconfig always selects PCI_LOCKESS_CONFIG, so I don't think >> the vmd driver can require it be turned off. Besides, no need to punish >> all PCI access if only this device requires it be serialized. > >Sorry, I also missed that and induced Ryo Takakura to rewrite the patch. My apologies I missed that as well. I'll send the first version of the patch once again! >Luis Sincerely, Ryo Takakura
