* controller/mediatek:
PCI: mediatek-gen3: Enable async probe by default
PCI: mediatek-gen3: Avoid PCIe resetting via PERST# for Airoha EN7581 SoC
PCI: mediatek-gen3: Rely on msleep() in mtk_pcie_en7581_power_up()
PCI: mediatek-gen3: Move reset delay in mtk_pcie_en7581_power_up()
PCI: mediatek-gen3: Add comment about initialization order in mtk_pcie_en7581_power_up()
PCI: mediatek-gen3: Move reset/assert callbacks in .power_up()
PCI: mediatek-gen3: Rely on clk_bulk_prepare_enable() in mtk_pcie_en7581_power_up()
* controller/dwc:
PCI: dwc: Clean up some unnecessary codes in dw_pcie_suspend_noirq()
PCI: dwc: Always stop link in the dw_pcie_suspend_noirq
PCI: dwc: Fix potential truncation in dw_pcie_edma_irq_verify()
PCI: dw-rockchip: Remove redundant calls to dev_err()
PCI: dw-rockchip: Replace magic values with defines
PCI: dw-rockchip: Don't wait for link since we can detect Link Up
PCI: dw-rockchip: Enumerate endpoints based on dll_link_up IRQ
PCI: qcom: Update ICC and OPP values after Link Up event
PCI: qcom: Don't wait for link if we can detect Link Up
PCI: dwc: Skip waiting for link up if driver can detect Link Up event
* endpoint:
PCI: endpoint: pci-epf-test: Set dma_chan_rx pointer to NULL on error
misc: pci_endpoint_test: Add support for capabilities
PCI: endpoint: pci-epf-test: Add support for capabilities
PCI: endpoint: Verify that requested BAR size is a power of two
PCI: endpoint: Add size check for fixed size BARs in pci_epc_set_bar()
PCI: artpec6: Implement dw_pcie_ep operation get_features
PCI: dwc: ep: Add 'address' alignment to 'size' check in dw_pcie_prog_ep_inbound_atu()
PCI: dwc: ep: Prevent changing BAR size/flags in pci_epc_set_bar()
PCI: dwc: ep: Write BAR_MASK before iATU registers in pci_epc_set_bar()
PCI: endpoint: Finish virtual EP removal in pci_epf_remove_vepf()
PCI: endpoint: Simplify pci_epc_get()
PCI: endpoint: Destroy the EPC device in devm_pci_epc_destroy()
PCI: endpoint: Replace magic number '6' by PCI_STD_NUM_BARS
* devres:
HID: amd_sfh: Use always-managed version of pcim_intx()
wifi: qtnfmac: use always-managed version of pcim_intx()
ata: Use always-managed version of pci_intx()
PCI/MSI: Use never-managed version of pci_intx()
vfio/pci: Use never-managed version of pci_intx()
misc: Use never-managed version of pci_intx()
ntb: Use never-managed version of pci_intx()
drivers/xen: Use never-managed version of pci_intx()
PCI: Export pci_intx_unmanaged() and pcim_intx()
Add MODULE_DEVICE_TABLE(), so modules could be properly autoloaded based
on the alias from of_device_id table.
Link: https://lore.kernel.org/r/20240903132311.961646-1-liaochen4@huawei.com
Signed-off-by: Liao Chen <liaochen4@huawei.com>
Signed-off-by: Krzysztof Wilczyński <kwilczynski@kernel.org>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Before sending PME_TURN_OFF, don't test the LTSSM state. Since it's
safe to send PME_TURN_OFF message regardless of whether the link is
up or down.
Thus, there is no need to test the LTSSM state before sending the
PME_TURN_OFF message, and only print the message when ltssm_stat is
not in DETECT and POLL. There shouldn't be an error message shown
when no endpoint is connected at all.
Link: https://lore.kernel.org/r/20241210081557.163555-3-hongxing.zhu@nxp.com
Signed-off-by: Richard Zhu <hongxing.zhu@nxp.com>
[kwilczynski: commit log]
Signed-off-by: Krzysztof Wilczyński <kwilczynski@kernel.org>
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
On the i.MX8QM, PCIe link can't be re-established again in
dw_pcie_resume_noirq(), if the LTSSM_EN bit is not cleared
properly in dw_pcie_suspend_noirq().
So, add dw_pcie_stop_link() to dw_pcie_suspend_noirq() to fix
this issue and to align the suspend/resume functions since there
is dw_pcie_start_link() in dw_pcie_resume_noirq() already.
Fixes: 4774faf854f5 ("PCI: dwc: Implement generic suspend/resume functionality")
Link: https://lore.kernel.org/r/20241210081557.163555-2-hongxing.zhu@nxp.com
Signed-off-by: Richard Zhu <hongxing.zhu@nxp.com>
[kwilczynski: commit log]
Signed-off-by: Krzysztof Wilczyński <kwilczynski@kernel.org>
Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Increase the size of the string buffer to avoid potential truncation in
dw_pcie_edma_irq_verify().
This fixes the following build warning when compiling with W=1:
drivers/pci/controller/dwc/pcie-designware.c: In function ‘dw_pcie_edma_detect’:
drivers/pci/controller/dwc/pcie-designware.c:989:50: warning: ‘%d’ directive output may be truncated writing between 1 and 11 bytes into a region of size 3 [-Wformat-truncation=]
989 | snprintf(name, sizeof(name), "dma%d", pci->edma.nr_irqs);
| ^~
Link: https://lore.kernel.org/r/20250104002119.2681246-2-cassel@kernel.org
Signed-off-by: Niklas Cassel <cassel@kernel.org>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Krzysztof Wilczyński <kwilczynski@kernel.org>
Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Usage of the dev_err() is redundant because platform_get_irq_byname()
already prints an error. So, remove the redundant call to dev_err().
This will address the following Coccinelle warning:
drivers/pci/controller/dwc/pcie-dw-rockchip.c:454:2-9:line 454 is redundant because platform_get_irq() already prints an error
Link: https://lore.kernel.org/r/327718207d3cd72847c079ff9d56eb246744c182.1736126067.git.xiaopei01@kylinos.cn
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202412312343.najrW1Db-lkp@intel.com
Signed-off-by: Pei Xiao <xiaopei01@kylinos.cn>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
[kwilczynski: commit log, drop the Fixes: tag]
Signed-off-by: Krzysztof Wilczyński <kwilczynski@kernel.org>
Reviewed-by: Niklas Cassel <cassel@kernel.org>
The Root Complex specific device tree binding for pcie-dw-rockchip has the
'sys' interrupt marked as required.
The driver requests the 'sys' IRQ unconditionally, and errors out if not
provided.
Thus, we can unconditionally set 'use_linkup_irq', so dw_pcie_host_init()
doesn't wait for the link to come up.
This will skip the wait for link up (since the bus will be enumerated once
the link up IRQ is triggered), which reduces the bootup time.
Link: https://lore.kernel.org/r/20250113-rockchip-no-wait-v1-1-25417f37b92f@kernel.org
Signed-off-by: Niklas Cassel <cassel@kernel.org>
[bhelgaas: commit log]
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Krzysztof Wilczyński <kwilczynski@kernel.org>
Most boards using the pcie-dw-rockchip PCIe controller lack standard
hotplug support.
Thus, when an endpoint is attached to the SoC, users have to rescan the bus
manually to enumerate the device. This can be avoided by using the
'dll_link_up' interrupt in the combined system interrupt 'sys'.
Once the 'dll_link_up' IRQ is received, the bus underneath the host bridge
is scanned to enumerate PCIe endpoint devices.
This implements the same functionality that was implemented in the DWC
based pcie-qcom driver in 4581403f6792 ("PCI: qcom: Enumerate endpoints
based on Link up event in 'global_irq' interrupt").
The Root Complex specific device tree binding for pcie-dw-rockchip already
has the 'sys' interrupt marked as required, so there is no need to update
the device tree binding. This also means that we can request the 'sys' IRQ
unconditionally.
Link: https://lore.kernel.org/r/20241127145041.3531400-2-cassel@kernel.org
Signed-off-by: Niklas Cassel <cassel@kernel.org>
[bhelgaas: commit log]
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Krzysztof Wilczyński <kwilczynski@kernel.org>
Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
4581403f6792 ("PCI: qcom: Enumerate endpoints based on Link up event in
'global_irq' interrupt") added the Link Up-based enumeration support, but
did not update the ICC/OPP vote once link is up. Before that, the update
happened during probe and the endpoints may or may not be enumerated at
that time, so the ICC/OPP vote was not guaranteed to be accurate.
With Link Up-based enumeration support, the driver can request the accurate
vote based on the PCIe link.
Call qcom_pcie_icc_opp_update() in qcom_pcie_global_irq_thread() after
enumerating the endpoints.
Fixes: 4581403f6792 ("PCI: qcom: Enumerate endpoints based on Link up event in 'global_irq' interrupt")
Link: https://lore.kernel.org/r/20241123-remove_wait2-v5-3-b5f9e6b794c2@quicinc.com
Signed-off-by: Krishna chaitanya chundru <quic_krichai@quicinc.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
[kwilczynski: commit log]
Signed-off-by: Krzysztof Wilczyński <kwilczynski@kernel.org>
Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Reviewed-by: Niklas Cassel <cassel@kernel.org>
If we have a 'global' IRQ for Link Up events, we need not wait for the
link to be up during PCI initialization, which reduces startup time.
Check for 'global' IRQ, and if present, set 'use_linkup_irq',
so dw_pcie_host_init() doesn't wait for the link to come up.
Link: https://lore.kernel.org/r/20241123-remove_wait2-v5-2-b5f9e6b794c2@quicinc.com
Signed-off-by: Krishna chaitanya chundru <quic_krichai@quicinc.com>
[bhelgaas: commit log]
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Krzysztof Wilczyński <kwilczynski@kernel.org>
Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Reviewed-by: Niklas Cassel <cassel@kernel.org>
If the driver can detect the Link Up event and enumerate downstream devices
at that time, we need not wait here.
Skip waiting for link to come up if the driver supports 'use_linkup_irq'.
Link: https://lore.kernel.org/r/20241123-remove_wait2-v5-1-b5f9e6b794c2@quicinc.com
Signed-off-by: Krishna chaitanya chundru <quic_krichai@quicinc.com>
[bhelgaas: wrap comment, update commit log]
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Krzysztof Wilczyński <kwilczynski@kernel.org>
Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Reviewed-by: Niklas Cassel <cassel@kernel.org>
Toggling memory enable is free on bare metal, but potentially expensive in
virtualized environments as the device MMIO spaces are added and removed
from the VM address space, including DMA mapping of those spaces through
the IOMMU where peer-to-peer is supported. Currently memory decode is
disabled around sizing each individual BAR, even for SR-IOV BARs while VF
Enable is cleared.
This can be better optimized for virtual environments by sizing a set of
BARs at once, stashing the resulting mask into an array, while only
toggling memory enable once. This also naturally improves the SR-IOV path
as the caller becomes responsible for any necessary decode disables while
sizing BARs, therefore SR-IOV BARs are sized relying only on the VF Enable
rather than toggling the PF memory enable in the command register.
Link: https://lore.kernel.org/r/20250111210652.402845-1-alex.williamson@redhat.com
Reported-by: Mitchell Augustin <mitchell.augustin@canonical.com>
Link: https://lore.kernel.org/all/CAHTA-uYp07FgM6T1OZQKqAdSA5JrZo0ReNEyZgQZub4mDRrV5w@mail.gmail.com
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Tested-by: Mitchell Augustin <mitchell.augustin@canonical.com>
Reviewed-by: Mitchell Augustin <mitchell.augustin@canonical.com>
Refactor rockchip_pcie_disable_clocks() to accept a struct rockchip_pcie
pointer instead of a void pointer thus improving type safety and code
readability by explicitly specifying the expected data type.
Link: https://lore.kernel.org/r/20241202151150.7393-4-linux.amoon@gmail.com
Signed-off-by: Anand Moon <linux.amoon@gmail.com>
Signed-off-by: Lorenzo Pieralisi <lpieralisi@kernel.org>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Currently, the driver acquires and asserts/deasserts the resets
individually thereby making the driver complex to read.
This can be simplified by using the reset_control_bulk() APIs.
Use devm_reset_control_bulk_get_exclusive() API to acquire all the resets
and use reset_control_bulk_{assert/deassert}() APIs to assert/deassert them
in bulk.
Following the recommendations in 'Rockchip RK3399 TRM v1.3 Part2':
1. Split the reset controls into two groups as per section '17.5.8.1.1
PCIe as Root Complex'.
2. Deassert the 'Pipe, MGMT Sticky, MGMT, Core' resets in groups as per
section '17.5.8.1.1 PCIe as Root Complex'. This is accomplished using
the reset_control_bulk APIs.
Link: https://lore.kernel.org/r/20241202151150.7393-3-linux.amoon@gmail.com
Signed-off-by: Anand Moon <linux.amoon@gmail.com>
Signed-off-by: Lorenzo Pieralisi <lpieralisi@kernel.org>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Currently, the driver acquires clocks and prepare/enable/disable/unprepare
the clocks individually thereby making the driver complex to read.
The driver can be simplified by using the clk_bulk*() APIs.
Use:
- devm_clk_bulk_get_all() API to acquire all the clocks
- clk_bulk_prepare_enable() to prepare/enable clocks
- clk_bulk_disable_unprepare() APIs to disable/unprepare them in bulk
Link: https://lore.kernel.org/r/20241202151150.7393-2-linux.amoon@gmail.com
Signed-off-by: Anand Moon <linux.amoon@gmail.com>
Signed-off-by: Lorenzo Pieralisi <lpieralisi@kernel.org>
[bhelgaas: squash error handling fix from https://lore.kernel.org/r/20250106153041.55267-1-linux.amoon@gmail.com]
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
When compiling the Rockchip endpoint driver with -W=1, the following
warnings can be seen in the output:
drivers/pci/controller/pcie-rockchip-ep.c:59: warning: Function parameter or struct member 'perst_irq' not described in 'rockchip_pcie_ep'
drivers/pci/controller/pcie-rockchip-ep.c:59: warning: Function parameter or struct member 'perst_asserted' not described in 'rockchip_pcie_ep'
drivers/pci/controller/pcie-rockchip-ep.c:59: warning: Function parameter or struct member 'link_up' not described in 'rockchip_pcie_ep'
drivers/pci/controller/pcie-rockchip-ep.c:59: warning: Function parameter or struct member 'link_training' not described in 'rockchip_pcie_ep'
Fix these warnings by adding the missing field descriptions in
struct rockchip_pcie_ep kernel-doc comment.
Fixes: a7137cbf6bd5 ("PCI: rockchip-ep: Handle PERST# signal in EP mode")
Fixes: bd6e61df4b2e ("PCI: rockchip-ep: Improve link training")
Link: https://lore.kernel.org/r/20241216133404.540736-1-dlemoal@kernel.org
Reported-by: Bjorn Helgaas <helgaas@kernel.org>
Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
Signed-off-by: Krzysztof Wilczyński <kwilczynski@kernel.org>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
If dma_chan_tx allocation fails, set dma_chan_rx to NULL after it is
freed.
Fixes: 8353813c88ef ("PCI: endpoint: Enable DMA tests for endpoints with DMA capabilities")
Link: https://lore.kernel.org/r/20241227160841.92382-1-khalfella@gmail.com
Signed-off-by: Mohamed Khalfella <khalfella@gmail.com>
[kwilczynski: commit log]
Signed-off-by: Krzysztof Wilczyński <kwilczynski@kernel.org>
Reviewed-by: Niklas Cassel <cassel@kernel.org>
Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
The rcar_pcie_parse_outbound_ranges() uses the devm_request_mem_region()
macro to request a needed resource. A string variable that lives on the
stack is then used to store a dynamically computed resource name, which
is then passed on as one of the macro arguments. This can lead to
undefined behavior.
Depending on the current contents of the memory, the manifestations of
errors may vary. One possible output may be as follows:
$ cat /proc/iomem
30000000-37ffffff :
38000000-3fffffff :
Sometimes, garbage may appear after the colon.
In very rare cases, if no NULL-terminator is found in memory, the system
might crash because the string iterator will overrun which can lead to
access of unmapped memory above the stack.
Thus, fix this by replacing outbound_name with the name of the previously
requested resource. With the changes applied, the output will be as
follows:
$ cat /proc/iomem
30000000-37ffffff : memory2
38000000-3fffffff : memory3
Fixes: 2a6d0d63d999 ("PCI: rcar: Add endpoint mode support")
Link: https://lore.kernel.org/r/tencent_DBDCC19D60F361119E76919ADAB25EC13C06@qq.com
Tested-by: Lad Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
Signed-off-by: King Dix <kingdix10@qq.com>
[kwilczynski: commit log]
Signed-off-by: Krzysztof Wilczyński <kwilczynski@kernel.org>
Reviewed-by: Lad Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
The mediatek-gen3 driver can run its probe routine fairly slow on some
hardware, which adds to the total time it takes for the system start up.
Thus, turn on async mode for the probe to avoid blocking the rest of the
system.
Link: https://lore.kernel.org/r/20241220145205.1.Ibf2563896c3b1fc133bb46d3fc96ad0041763922@changeid
Signed-off-by: Douglas Anderson <dianders@chromium.org>
[kwilczynski: commit log]
Signed-off-by: Krzysztof Wilczyński <kwilczynski@kernel.org>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Airoha EN7581 has a hw bug asserting/releasing PERST# signal causing
occasional PCIe link down issues. In order to overcome the problem,
PERST# signal is not asserted/released during device probe or
suspend/resume phase and the PCIe block is reset using
en7523_reset_assert() and en7581_pci_enable().
Introduce flags field in the mtk_gen3_pcie_pdata struct in order to
specify per-SoC capabilities.
Link: https://lore.kernel.org/r/20250109-pcie-en7581-rst-fix-v4-1-4a45c89fb143@kernel.org
Tested-by: Hui Ma <hui.ma@airoha.com>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Krzysztof Wilczyński <kwilczynski@kernel.org>
Since mtk_pcie_en7581_power_up() runs in non-atomic context, rely on
msleep() routine instead of mdelay().
Link: https://lore.kernel.org/r/20250108-pcie-en7581-fixes-v6-5-21ac939a3b9b@kernel.org
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Krzysztof Wilczyński <kwilczynski@kernel.org>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Airoha EN7581 has a hw bug asserting/releasing PCIE_PE_RSTB signal
causing occasional PCIe link down issues. In order to overcome the
problem, PCIe block is reset using REG_PCI_CONTROL (0x88) and
REG_RESET_CONTROL (0x834) registers available in the clock module
running clk_bulk_prepare_enable() in mtk_pcie_en7581_power_up().
In order to make the code more readable, move the wait for the time
needed to complete the PCIe reset from en7581_pci_enable() to
mtk_pcie_en7581_power_up().
Reduce reset timeout from 250ms to the standard PCIE_T_PVPERL_MS value
(100ms) since it has no impact on the driver behavior.
Link: https://lore.kernel.org/r/20250108-pcie-en7581-fixes-v6-4-21ac939a3b9b@kernel.org
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Krzysztof Wilczyński <kwilczynski@kernel.org>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Acked-by: Stephen Boyd <sboyd@kernel.org>
Add a comment in mtk_pcie_en7581_power_up() to clarify, unlike the other
MediaTek Gen3 controllers, the Airoha EN7581 requires PHY initialization
and power-on before PHY reset deassert.
Link: https://lore.kernel.org/r/20250108-pcie-en7581-fixes-v6-3-21ac939a3b9b@kernel.org
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Krzysztof Wilczyński <kwilczynski@kernel.org>
Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
In order to make the code more readable, the reset_control_bulk_assert()
function for PHY reset lines is moved to make it pair with
reset_control_bulk_deassert() in mtk_pcie_power_up() and
mtk_pcie_en7581_power_up(). The same change is done for
reset_control_assert() used to assert MAC reset line.
Introduce PCIE_MTK_RESET_TIME_US macro for the time needed to
complete PCIe reset on MediaTek controller.
Link: https://lore.kernel.org/r/20250108-pcie-en7581-fixes-v6-2-21ac939a3b9b@kernel.org
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Krzysztof Wilczyński <kwilczynski@kernel.org>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Replace clk_bulk_prepare() and clk_bulk_enable() with
clk_bulk_prepare_enable() in mtk_pcie_en7581_power_up() routine.
Link: https://lore.kernel.org/r/20250108-pcie-en7581-fixes-v6-1-21ac939a3b9b@kernel.org
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Krzysztof Wilczyński <kwilczynski@kernel.org>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
The interrupt handler for bandwidth notifications, pcie_bwnotif_irq(),
dereferences a "data" pointer.
On unbind, that pointer is set to NULL by pcie_bwnotif_remove(). However
the interrupt handler may still be invoked afterwards and will dereference
that NULL pointer.
That's because the interrupt is requested using a devm_*() helper and the
driver core releases devm_*() resources *after* calling ->remove().
pcie_bwnotif_remove() does clear the Link Bandwidth Management Interrupt
Enable and Link Autonomous Bandwidth Interrupt Enable bits in the Link
Control Register, but that won't prevent execution of pcie_bwnotif_irq():
The interrupt for bandwidth notifications may be shared with AER, DPC,
PME, and hotplug. So pcie_bwnotif_irq() may be executed as long as the
interrupt is requested.
There's a similar race on bind: pcie_bwnotif_probe() requests the
interrupt when the "data" pointer still points to NULL. A NULL pointer
deref may thus likewise occur if AER, DPC, PME or hotplug raise an
interrupt in-between the bandwidth controller's call to devm_request_irq()
and assignment of the "data" pointer.
Drop the devm_*() usage and reorder requesting of the interrupt to fix the
issue.
While at it, drop a stray but harmless no_free_ptr() invocation when
assigning the "data" pointer in pcie_bwnotif_probe().
Ilpo points out that the locking on unbind and bind needs to be symmetric,
so move the call to pcie_bwnotif_disable() inside the critical section
protected by pcie_bwctrl_setspeed_rwsem and pcie_bwctrl_lbms_rwsem.
Evert reports a hang on shutdown of an ASUS ROG Strix SCAR 17 G733PYV.
The issue is no longer reproducible with the present commit.
Evert found that attaching a USB-C monitor prevented the hang. The
machine contains an ASMedia USB 3.2 controller below a hotplug-capable
Root Port. So one possible explanation is that the controller gets
hot-removed on shutdown unless something is connected. And the ensuing
hotplug interrupt occurs exactly when the bandwidth controller is
unregistering. The precise cause could not be determined because the
screen had already turned black when the hang occurred.
Link: https://lore.kernel.org/r/ae2b02c9cfbefff475b6e132b0aa962aaccbd7b2.1736162539.git.lukas@wunner.de
Fixes: 665745f27487 ("PCI/bwctrl: Re-add BW notification portdrv as PCIe BW controller")
Reported-by: Evert Vorster <evorster@gmail.com>
Closes: https://bugzilla.kernel.org/show_bug.cgi?id=219629
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Tested-by: Evert Vorster <evorster@gmail.com>
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
The helper devm_clk_bulk_get_all_enable() missed to return the number of
clocks stored in the clk_bulk_data table referenced by the clks
argument and, therefore, will be dropped.
Use the newly introduced devm_clk_bulk_get_all_enabled() variant
instead, which is consistent with devm_clk_bulk_get_all() in terms of
the returned value:
> 0 if one or more clocks have been stored
= 0 if there are no clocks
< 0 if an error occurred
Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Signed-off-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
Link: https://lore.kernel.org/r/20241217-clk_bulk_ena_fix-v5-2-aafbbb245155@collabora.com
Acked-by: Krzysztof Wilczyński <kwilczynski@kernel.org>
Signed-off-by: Stephen Boyd <sboyd@kernel.org>
struct bus_type has a new callback for retrieving the IRQ affinity for a
device. Hook this callback up for PCI based devices.
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Reviewed-by: John Garry <john.g.garry@oracle.com>
Signed-off-by: Daniel Wagner <wagi@kernel.org>
Link: https://lore.kernel.org/r/20241202-refactor-blk-affinity-helpers-v6-2-27211e9c2cd5@kernel.org
Signed-off-by: Jens Axboe <axboe@kernel.dk>
-----BEGIN PGP SIGNATURE-----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=zLos
-----END PGP SIGNATURE-----
Merge tag 'pci-v6.13-fixes-2' of git://git.kernel.org/pub/scm/linux/kernel/git/pci/pci
Pull PCI fixes from Krzysztof Wilczyński:
"Two small patches that are important for fixing boot time hang on
Intel JHL7540 'Titan Ridge' platforms equipped with a Thunderbolt
controller.
The boot time issue manifests itself when a PCI Express bandwidth
control is unnecessarily enabled on the Thunderbolt controller
downstream ports, which only supports a link speed of 2.5 GT/s in
accordance with USB4 v2 specification (p. 671, sec. 11.2.1, "PCIe
Physical Layer Logical Sub-block").
As such, there is no need to enable bandwidth control on such
downstream port links, which also works around the issue.
Both patches were tested by the original reporter on the hardware on
which the failure origin golly manifested itself. Both fixes were
proven to resolve the reported boot hang issue, and both patches have
been in linux-next this week with no reported problems"
* tag 'pci-v6.13-fixes-2' of git://git.kernel.org/pub/scm/linux/kernel/git/pci/pci:
PCI/bwctrl: Enable only if more than one speed is supported
PCI: Honor Max Link Speed when determining supported speeds
The header clearly states that it does not want to be included directly,
only via 'device.h'. The 'platform_device.h' works equally well.
Thus, remove the direct inclusion.
Link: https://lore.kernel.org/r/20241118072917.3853-12-wsa+renesas@sang-engineering.com
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Signed-off-by: Krzysztof Wilczyński <kwilczynski@kernel.org>
If a PCIe port only supports a single speed, enabling bandwidth control
is pointless: There's no need to monitor autonomous speed changes, nor
can the speed be changed.
Not enabling it saves a small amount of memory and compute resources,
but also fixes a boot hang reported by Niklas: It occurs when enabling
bandwidth control on Downstream Ports of Intel JHL7540 "Titan Ridge 2018"
Thunderbolt controllers. The ports only support 2.5 GT/s in accordance
with USB4 v2 sec 11.2.1, so the present commit works around the issue.
PCIe r6.2 sec 8.2.1 prescribes that:
"A device must support 2.5 GT/s and is not permitted to skip support
for any data rates between 2.5 GT/s and the highest supported rate."
Consequently, bandwidth control is currently only disabled if a port
doesn't support higher speeds than 2.5 GT/s. However the Implementation
Note in PCIe r6.2 sec 7.5.3.18 cautions:
"It is strongly encouraged that software primarily utilize the
Supported Link Speeds Vector instead of the Max Link Speed field,
so that software can determine the exact set of supported speeds on
current and future hardware. This can avoid software being confused
if a future specification defines Links that do not require support
for all slower speeds."
In other words, future revisions of the PCIe Base Spec may allow gaps
in the Supported Link Speeds Vector. To be future-proof, don't just
check whether speeds above 2.5 GT/s are supported, but rather check
whether *more than one* speed is supported.
Fixes: 665745f27487 ("PCI/bwctrl: Re-add BW notification portdrv as PCIe BW controller")
Closes: https://lore.kernel.org/r/db8e457fcd155436449b035e8791a8241b0df400.camel@kernel.org
Link: https://lore.kernel.org/r/3564908a9c99fc0d2a292473af7a94ebfc8f5820.1734428762.git.lukas@wunner.de
Reported-by: Niklas Schnelle <niks@kernel.org>
Tested-by: Niklas Schnelle <niks@kernel.org>
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Krzysztof Wilczyński <kwilczynski@kernel.org>
Reviewed-by: Jonathan Cameron <Jonthan.Cameron@huawei.com>
Reviewed-by: Mario Limonciello <mario.limonciello@amd.com>
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
The Supported Link Speeds Vector in the Link Capabilities 2 Register
indicates the *supported* link speeds. The Max Link Speed field in the
Link Capabilities Register indicates the *maximum* of those speeds.
pcie_get_supported_speeds() neglects to honor the Max Link Speed field and
will thus incorrectly deem higher speeds as supported. Fix it.
One user-visible issue addressed here is an incorrect value in the sysfs
attribute "max_link_speed".
But the main motivation is a boot hang reported by Niklas: Intel JHL7540
"Titan Ridge 2018" Thunderbolt controllers supports 2.5-8 GT/s speeds,
but indicate 2.5 GT/s as maximum. Ilpo recalls seeing this on more
devices. It can be explained by the controller's Downstream Ports
supporting 8 GT/s if an Endpoint is attached, but limiting to 2.5 GT/s
if the port interfaces to a PCIe Adapter, in accordance with USB4 v2
sec 11.2.1:
"This section defines the functionality of an Internal PCIe Port that
interfaces to a PCIe Adapter. [...]
The Logical sub-block shall update the PCIe configuration registers
with the following characteristics: [...]
Max Link Speed field in the Link Capabilities Register set to 0001b
(data rate of 2.5 GT/s only).
Note: These settings do not represent actual throughput. Throughput
is implementation specific and based on the USB4 Fabric performance."
The present commit is not sufficient on its own to fix Niklas' boot hang,
but it is a prerequisite: A subsequent commit will fix the boot hang by
enabling bandwidth control only if more than one speed is supported.
The GENMASK() macro used herein specifies 0 as lowest bit, even though
the Supported Link Speeds Vector ends at bit 1. This is done on purpose
to avoid a GENMASK(0, 1) macro if Max Link Speed is zero. That macro
would be invalid as the lowest bit is greater than the highest bit.
Ilpo has witnessed a zero Max Link Speed on Root Complex Integrated
Endpoints in particular, so it does occur in practice. (The Link
Capabilities Register is optional on RCiEPs per PCIe r6.2 sec 7.5.3.)
Fixes: d2bd39c0456b ("PCI: Store all PCIe Supported Link Speeds")
Closes: https://lore.kernel.org/r/70829798889c6d779ca0f6cd3260a765780d1369.camel@kernel.org
Link: https://lore.kernel.org/r/fe03941e3e1cc42fb9bf4395e302bff53ee2198b.1734428762.git.lukas@wunner.de
Reported-by: Niklas Schnelle <niks@kernel.org>
Tested-by: Niklas Schnelle <niks@kernel.org>
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Krzysztof Wilczyński <kwilczynski@kernel.org>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
