Hi, > Currently, dw_pcie_msi_init() allocates and maps page for msi, then > program the PCIE_MSI_ADDR_LO and PCIE_MSI_ADDR_HI. The Root Complex > may lose power during suspend-to-RAM, so when we resume, we want to > redo the latter but not the former. If designware based driver (for > example, pcie-tegra194.c) calls dw_pcie_msi_init() in resume path, the > previous msi page will be leaked. From another side, except > pci-dra7xx.c we can move the dw_pcie_msi_init() from each users to > designware host, I.E move the msi page allocation and mapping to > dw_pcie_host_init() and move the PCIE_MSI_ADDR_* programming to > dw_pcie_setup_rc(). After this moving, we solve the msi page leakage > as well. [...] A small nitpick. All the "designware" should be "DesignWare" both in the commit message and the subject. Similarly, "msi" would be "MSI", and "I.E" would become "i.e.,". If you ever sent another version of the patch, that is. See the following for reference: https://lore.kernel.org/linux-pci/20171026223701.GA25649@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx/ Krzysztof
