Re: [PATCH v4 0/2] PCI: brcmstb: Add panic/die handler to driver

[Bjorn Helgaas <helgaas@kernel.org>]

On Wed, Oct 29, 2025 at 03:36:13PM -0400, Jim Quinlan wrote:
> v4 Changes:
>   -- Commit "Add panic/die handler to driver"
>     o s/__MASK/_MASK/g -- (Alok)
>     o Remove underscore of _brcm_pcie_dump_err -- (Bjorn)
>     o Remove _MASK suffix from OUTB constants -- (Bjorn)
>     o Use standard notation %04x:%02x:%02x.%d" -- (Bjorn)
>     o Use BIT(i) instead of hex number (Ilpo)
>     o Remove unused constant (Ilpo)
>     o Use str_read_write() (Ilpo)
> 
> v3 Changes:
>   -- Commit "Add a way to indicate if PCIe bridge is active"
>     o Implement Bjorn's V1 suggestion properly (Bjorn, Mani)
>     o Remove unrelated change in commit (Mani)
>     o Remove an "inline" directive (Mani)
>     o s/bridge_on/bridge_in_reset/ (Mani)
>   -- Commit "Add panic/die handler to driver"
>     o dev_err(...) message changed from "handling" error (Mani)
> 
> v2 Changes:
>   -- Commit "Add a way to indicate if PCIe bridge is active"
>     o Set "bridge_on" correctly when bridge is reset (Bjorn)
>     o Return 0 instead "return ret" and skip ret init (Bjorn)
>     o Use u32p_replace_bits(...) instead of shifts and AND/OR (Bjorn)
>     o Reword error statement regarding bridge reset (Bjorn)
> 
> The first commit sets up a field variable and spinlock to indicate whether
> the PCIe bridge is active.  The second commit builds upon the first and
> adds a "die" handler to the driver, which, when invoked, prints out a
> summary of any pending PCIe errors.  The "die" handler is careful not to
> access any registers unless the bridge is active.
> 
> Jim Quinlan (2):
>   PCI: brcmstb: Add a way to indicate if PCIe bridge is active
>   PCI: brcmstb: Add panic/die handler to driver
> 
>  drivers/pci/controller/pcie-brcmstb.c | 198 +++++++++++++++++++++++++-
>  1 file changed, 192 insertions(+), 6 deletions(-)

Applied to pci/controller/brcmstb for v6.19, thanks!