From: Hans Zhang <18255117159@163.com>
To: Lukas Wunner <lukas@wunner.de>
Cc: lpieralisi@kernel.org, kw@linux.com,
manivannan.sadhasivam@linaro.org, robh@kernel.org,
bhelgaas@google.com, jingoohan1@gmail.com,
thomas.richard@bootlin.com, linux-pci@vger.kernel.org,
linux-kernel@vger.kernel.org
Subject: Re: [v5 1/4] PCI: Introduce generic capability search functions
Date: Sat, 22 Mar 2025 23:47:18 +0800 [thread overview]
Message-ID: <e6e808b6-302b-4f3e-ad2d-5f9c4dce7394@163.com> (raw)
In-Reply-To: <Z92cgXEGwgYD2gau@wunner.de>
On 2025/3/22 01:06, Lukas Wunner wrote:
> On Sat, Mar 22, 2025 at 12:38:00AM +0800, Hans Zhang wrote:
>> Existing controller drivers (e.g., DWC, custom out-of-tree drivers)
>> duplicate logic for scanning PCI capability lists. This creates
>> maintenance burdens and risks inconsistencies.
>>
>> To resolve this:
>>
>> Add pci_host_bridge_find_*capability() in drivers/pci/pci.c, accepting
>> controller-specific read functions and device data as parameters.
> [...]
>> drivers/pci/pci.c | 86 +++++++++++++++++++++++++++++++++++++++++++++
>
> Please put this in a .c file which is only compiled and linked if
> one of the controller drivers using those new helpers is enabled
> in .config.
>
> If you put the helpers in drivers/pci/pci.c, they unnecessarily
> enlarge the kernel's .text section even if it's known already
> at compile time that they're never going to be used (e.g. on x86).
>
Hi Lukas,
Thanks your for reply. Increasing the size of the .text section was not
my intention. I see what you mean.
> You could put them in drivers/pci/controller/pci-host-common.c
> and then select PCI_HOST_COMMON for each driver using them.
> Or put them in a separate completely new file.
>
I add a drivers/pci/controller/pci-host-helpers.c file, how do you like
it? Below, I have rearranged the patch, please kindly review it, thank
you very much.
>
>> include/linux/pci.h | 16 ++++++++-
>
> Helpers that are only used internally in the PCI core should be
> declared in drivers/pci/pci.h. I'd assume this also applies to
> helpers used by controller drivers.
>
Will change.
> Thanks,
>
> Lukas
Next version patch:
drivers/pci/controller/Kconfig | 16 ++++
drivers/pci/controller/Makefile | 1 +
drivers/pci/controller/pci-host-helpers.c | 98 +++++++++++++++++++++++
drivers/pci/pci.h | 7 ++
4 files changed, 122 insertions(+)
create mode 100644 drivers/pci/controller/pci-host-helpers.c
diff --git a/drivers/pci/controller/Kconfig b/drivers/pci/controller/Kconfig
index 9800b7681054..662c775999a1 100644
--- a/drivers/pci/controller/Kconfig
+++ b/drivers/pci/controller/Kconfig
@@ -132,6 +132,22 @@ config PCI_HOST_GENERIC
Say Y here if you want to support a simple generic PCI host
controller, such as the one emulated by kvmtool.
+config PCI_HOST_HELPERS
+ bool "PCI Host Controller Helper Functions"
+ help
+ This provides common infrastructure for PCI host controller drivers to
+ handle PCI capability scanning and other shared operations. The helper
+ functions eliminate code duplication across controller drivers.
+
+ These functions are used by PCI controller drivers that need to scan
+ PCI capabilities using controller-specific access methods (e.g. when
+ the controller is behind a non-standard configuration space).
+
+ If you are using any PCI host controller drivers that require these
+ helpers (such as DesignWare, Cadence, etc), this will be
+ automatically selected. Say N unless you are developing a custom PCI
+ host controller driver.
+
config PCIE_HISI_ERR
depends on ACPI_APEI_GHES && (ARM64 || COMPILE_TEST)
bool "HiSilicon HIP PCIe controller error handling driver"
diff --git a/drivers/pci/controller/Makefile
b/drivers/pci/controller/Makefile
index 038ccbd9e3ba..e80091eb7597 100644
--- a/drivers/pci/controller/Makefile
+++ b/drivers/pci/controller/Makefile
@@ -12,6 +12,7 @@ obj-$(CONFIG_PCIE_RCAR_HOST) += pcie-rcar.o
pcie-rcar-host.o
obj-$(CONFIG_PCIE_RCAR_EP) += pcie-rcar.o pcie-rcar-ep.o
obj-$(CONFIG_PCI_HOST_COMMON) += pci-host-common.o
obj-$(CONFIG_PCI_HOST_GENERIC) += pci-host-generic.o
+obj-$(CONFIG_PCI_HOST_HELPERS) += pci-host-helpers.o
obj-$(CONFIG_PCI_HOST_THUNDER_ECAM) += pci-thunder-ecam.o
obj-$(CONFIG_PCI_HOST_THUNDER_PEM) += pci-thunder-pem.o
obj-$(CONFIG_PCIE_XILINX) += pcie-xilinx.o
diff --git a/drivers/pci/controller/pci-host-helpers.c
b/drivers/pci/controller/pci-host-helpers.c
new file mode 100644
index 000000000000..cd261a281c60
--- /dev/null
+++ b/drivers/pci/controller/pci-host-helpers.c
@@ -0,0 +1,98 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * PCI Host Controller Helper Functions
+ *
+ * Copyright (C) 2025 Hans Zhang
+ *
+ * Author: Hans Zhang <18255117159@163.com>
+ */
+
+#include <linux/pci.h>
+
+#include "../pci.h"
+
+/*
+ * These interfaces resemble the pci_find_*capability() interfaces, but
these
+ * are for configuring host controllers, which are bridges *to* PCI
devices but
+ * are not PCI devices themselves.
+ */
+static u8 __pci_host_bridge_find_next_cap(void *priv,
+ pci_host_bridge_read_cfg read_cfg,
+ u8 cap_ptr, u8 cap)
+{
+ u8 cap_id, next_cap_ptr;
+ u16 reg;
+
+ if (!cap_ptr)
+ return 0;
+
+ reg = read_cfg(priv, cap_ptr, 2);
+ cap_id = (reg & 0x00ff);
+
+ if (cap_id > PCI_CAP_ID_MAX)
+ return 0;
+
+ if (cap_id == cap)
+ return cap_ptr;
+
+ next_cap_ptr = (reg & 0xff00) >> 8;
+ return __pci_host_bridge_find_next_cap(priv, read_cfg, next_cap_ptr,
+ cap);
+}
+
+u8 pci_host_bridge_find_capability(void *priv,
+ pci_host_bridge_read_cfg read_cfg, u8 cap)
+{
+ u8 next_cap_ptr;
+ u16 reg;
+
+ reg = read_cfg(priv, PCI_CAPABILITY_LIST, 2);
+ next_cap_ptr = (reg & 0x00ff);
+
+ return __pci_host_bridge_find_next_cap(priv, read_cfg, next_cap_ptr,
+ cap);
+}
+EXPORT_SYMBOL_GPL(pci_host_bridge_find_capability);
+
+static u16 pci_host_bridge_find_next_ext_capability(
+ void *priv, pci_host_bridge_read_cfg read_cfg, u16 start, u8 cap)
+{
+ u32 header;
+ int ttl;
+ int pos = PCI_CFG_SPACE_SIZE;
+
+ /* minimum 8 bytes per capability */
+ ttl = (PCI_CFG_SPACE_EXP_SIZE - PCI_CFG_SPACE_SIZE) / 8;
+
+ if (start)
+ pos = start;
+
+ header = read_cfg(priv, pos, 4);
+ /*
+ * If we have no capabilities, this is indicated by cap ID,
+ * cap version and next pointer all being 0.
+ */
+ if (header == 0)
+ return 0;
+
+ while (ttl-- > 0) {
+ if (PCI_EXT_CAP_ID(header) == cap && pos != start)
+ return pos;
+
+ pos = PCI_EXT_CAP_NEXT(header);
+ if (pos < PCI_CFG_SPACE_SIZE)
+ break;
+
+ header = read_cfg(priv, pos, 4);
+ }
+
+ return 0;
+}
+
+u16 pci_host_bridge_find_ext_capability(void *priv,
+ pci_host_bridge_read_cfg read_cfg,
+ u8 cap)
+{
+ return pci_host_bridge_find_next_ext_capability(priv, read_cfg, 0, cap);
+}
+EXPORT_SYMBOL_GPL(pci_host_bridge_find_ext_capability);
diff --git a/drivers/pci/pci.h b/drivers/pci/pci.h
index 01e51db8d285..8d1c919cbfef 100644
--- a/drivers/pci/pci.h
+++ b/drivers/pci/pci.h
@@ -1034,4 +1034,11 @@ void pcim_release_region(struct pci_dev *pdev,
int bar);
(PCI_CONF1_ADDRESS(bus, dev, func, reg) | \
PCI_CONF1_EXT_REG(reg))
+typedef u32 (*pci_host_bridge_read_cfg)(void *priv, int where, int size);
+u8 pci_host_bridge_find_capability(void *priv,
+ pci_host_bridge_read_cfg read_cfg, u8 cap);
+u16 pci_host_bridge_find_ext_capability(void *priv,
+ pci_host_bridge_read_cfg read_cfg,
+ u8 cap);
+
#endif /* DRIVERS_PCI_H */
Best regards,
Hans
next prev parent reply other threads:[~2025-03-22 15:48 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-03-21 16:37 [v5 0/4] Introduce generic capability search functions Hans Zhang
2025-03-21 16:38 ` [v5 1/4] PCI: " Hans Zhang
2025-03-21 17:06 ` Lukas Wunner
2025-03-22 15:47 ` Hans Zhang [this message]
2025-03-22 16:11 ` Lukas Wunner
2025-03-23 15:36 ` Hans Zhang
2025-03-21 20:05 ` kernel test robot
2025-03-21 20:57 ` kernel test robot
2025-03-21 20:57 ` kernel test robot
2025-03-21 16:38 ` [v5 2/4] PCI: dwc: Use common PCI host bridge APIs for finding the capabilities Hans Zhang
2025-03-21 16:38 ` [v5 3/4] PCI: cadence: " Hans Zhang
2025-03-21 16:38 ` [v5 4/4] PCI: cadence: Use cdns_pcie_find_*capability to avoid hardcode Hans Zhang
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=e6e808b6-302b-4f3e-ad2d-5f9c4dce7394@163.com \
--to=18255117159@163.com \
--cc=bhelgaas@google.com \
--cc=jingoohan1@gmail.com \
--cc=kw@linux.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-pci@vger.kernel.org \
--cc=lpieralisi@kernel.org \
--cc=lukas@wunner.de \
--cc=manivannan.sadhasivam@linaro.org \
--cc=robh@kernel.org \
--cc=thomas.richard@bootlin.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox