* [PATCH 0/2] PCI: rcar: Fix of_find_matching_node() reference leak @ 2022-06-22 2:57 Bjorn Helgaas 2022-06-22 2:57 ` [PATCH 1/2] PCI: rcar: Add dev struct for of_device_get_match_data() Bjorn Helgaas 2022-06-22 2:57 ` [PATCH 2/2] PCI: rcar: Resolve of_find_matching_node() reference leak Bjorn Helgaas 0 siblings, 2 replies; 5+ messages in thread From: Bjorn Helgaas @ 2022-06-22 2:57 UTC (permalink / raw) To: Marek Vasut, Yoshihiro Shimoda Cc: Liang He, Geert Uytterhoeven, Lorenzo Pieralisi, Rob Herring, Krzysztof Wilczyński, linux-pci, linux-renesas-soc, linux-kernel, Bjorn Helgaas From: Bjorn Helgaas <bhelgaas@google.com> Alternate proposal for Liang's patch [1]. This is a bulkier fix, but removes the redundant rcar_pcie_abort_handler_of_match[] table. It also simplifies the unusual device_initcall()/builtin_platform_driver() construct, which seems unnecessary as far as I can tell. Compile-tested only. [1] https://lore.kernel.org/r/20220621070145.4080147-1-windhl@126.com Bjorn Helgaas (2): PCI: rcar: Add dev struct for of_device_get_match_data() PCI: rcar: Resolve of_find_matching_node() reference leak drivers/pci/controller/pcie-rcar-host.c | 113 +++++++++++++----------- 1 file changed, 60 insertions(+), 53 deletions(-) -- 2.25.1 ^ permalink raw reply [flat|nested] 5+ messages in thread
* [PATCH 1/2] PCI: rcar: Add dev struct for of_device_get_match_data() 2022-06-22 2:57 [PATCH 0/2] PCI: rcar: Fix of_find_matching_node() reference leak Bjorn Helgaas @ 2022-06-22 2:57 ` Bjorn Helgaas 2022-06-22 2:57 ` [PATCH 2/2] PCI: rcar: Resolve of_find_matching_node() reference leak Bjorn Helgaas 1 sibling, 0 replies; 5+ messages in thread From: Bjorn Helgaas @ 2022-06-22 2:57 UTC (permalink / raw) To: Marek Vasut, Yoshihiro Shimoda Cc: Liang He, Geert Uytterhoeven, Lorenzo Pieralisi, Rob Herring, Krzysztof Wilczyński, linux-pci, linux-renesas-soc, linux-kernel, Bjorn Helgaas, Geert Uytterhoeven From: Bjorn Helgaas <bhelgaas@google.com> Add a struct rcar_variant to hold details about differences between the different R-Car variants we support. No functional change intended. The benefit is that we can find all this information with the existing single of_device_get_match_data() call. A subsequent commit will remove an of_find_matching_node() call by extending this structure. Signed-off-by: Bjorn Helgaas <bhelgaas@google.com> Cc: Liang He <windhl@126.com> Cc: Geert Uytterhoeven <geert@linux-m68k.org> --- drivers/pci/controller/pcie-rcar-host.c | 53 +++++++++++++++---------- 1 file changed, 33 insertions(+), 20 deletions(-) diff --git a/drivers/pci/controller/pcie-rcar-host.c b/drivers/pci/controller/pcie-rcar-host.c index 997c4df6a1e7..ccf13aafa1e5 100644 --- a/drivers/pci/controller/pcie-rcar-host.c +++ b/drivers/pci/controller/pcie-rcar-host.c @@ -56,13 +56,19 @@ static void __iomem *pcie_base; static struct device *pcie_dev; #endif +struct rcar_pcie_host; + +struct rcar_variant { + int (*phy_init_fn)(struct rcar_pcie_host *host); +}; + /* Structure representing the PCIe interface */ struct rcar_pcie_host { - struct rcar_pcie pcie; - struct phy *phy; - struct clk *bus_clk; - struct rcar_msi msi; - int (*phy_init_fn)(struct rcar_pcie_host *host); + struct rcar_pcie pcie; + struct phy *phy; + struct clk *bus_clk; + struct rcar_msi msi; + const struct rcar_variant *variant; }; static DEFINE_SPINLOCK(pmsr_lock); @@ -958,19 +964,26 @@ static int rcar_pcie_parse_map_dma_ranges(struct rcar_pcie_host *host) return err; } +static const struct rcar_variant rcar_h1_data = { + .phy_init_fn = rcar_pcie_phy_init_h1, +}; + +static const struct rcar_variant rcar_gen2_data = { + .phy_init_fn = rcar_pcie_phy_init_gen2, +}; + +static const struct rcar_variant rcar_gen3_data = { + .phy_init_fn = rcar_pcie_phy_init_gen3, +}; + static const struct of_device_id rcar_pcie_of_match[] = { - { .compatible = "renesas,pcie-r8a7779", - .data = rcar_pcie_phy_init_h1 }, - { .compatible = "renesas,pcie-r8a7790", - .data = rcar_pcie_phy_init_gen2 }, - { .compatible = "renesas,pcie-r8a7791", - .data = rcar_pcie_phy_init_gen2 }, - { .compatible = "renesas,pcie-rcar-gen2", - .data = rcar_pcie_phy_init_gen2 }, - { .compatible = "renesas,pcie-r8a7795", - .data = rcar_pcie_phy_init_gen3 }, - { .compatible = "renesas,pcie-rcar-gen3", - .data = rcar_pcie_phy_init_gen3 }, + { .compatible = "renesas,pcie-r8a7779", .data = &rcar_h1_data }, + { .compatible = "renesas,pcie-r8a7790", .data = &rcar_gen2_data }, + { .compatible = "renesas,pcie-r8a7791", .data = &rcar_gen2_data }, + { .compatible = "renesas,pcie-rcar-gen2", .data = &rcar_gen2_data }, + + { .compatible = "renesas,pcie-r8a7795", .data = &rcar_gen3_data }, + { .compatible = "renesas,pcie-rcar-gen3", .data = &rcar_gen3_data }, {}, }; @@ -1015,8 +1028,8 @@ static int rcar_pcie_probe(struct platform_device *pdev) if (err) goto err_clk_disable; - host->phy_init_fn = of_device_get_match_data(dev); - err = host->phy_init_fn(host); + host->variant = of_device_get_match_data(dev); + err = host->variant->phy_init_fn(host); if (err) { dev_err(dev, "failed to init PCIe PHY\n"); goto err_clk_disable; @@ -1084,7 +1097,7 @@ static int __maybe_unused rcar_pcie_resume(struct device *dev) return 0; /* Failure to get a link might just be that no cards are inserted */ - err = host->phy_init_fn(host); + err = host->variant->phy_init_fn(host); if (err) { dev_info(dev, "PCIe link down\n"); return 0; -- 2.25.1 ^ permalink raw reply related [flat|nested] 5+ messages in thread
* [PATCH 2/2] PCI: rcar: Resolve of_find_matching_node() reference leak 2022-06-22 2:57 [PATCH 0/2] PCI: rcar: Fix of_find_matching_node() reference leak Bjorn Helgaas 2022-06-22 2:57 ` [PATCH 1/2] PCI: rcar: Add dev struct for of_device_get_match_data() Bjorn Helgaas @ 2022-06-22 2:57 ` Bjorn Helgaas 2022-06-22 7:45 ` Geert Uytterhoeven 1 sibling, 1 reply; 5+ messages in thread From: Bjorn Helgaas @ 2022-06-22 2:57 UTC (permalink / raw) To: Marek Vasut, Yoshihiro Shimoda Cc: Liang He, Geert Uytterhoeven, Lorenzo Pieralisi, Rob Herring, Krzysztof Wilczyński, linux-pci, linux-renesas-soc, linux-kernel, Bjorn Helgaas, Geert Uytterhoeven From: Bjorn Helgaas <bhelgaas@google.com> Previously, rcar_pcie_init() used of_find_matching_node() to search the entire device tree for compatible strings for which we need to install an abort handler. If we found one, we got a device_node with refcount incremented, but we discarded the pointer and never released that reference. Extend the struct rcar_variant to indicate whether each variant requires an abort handler. Install the handler in rcar_pcie_probe() when needed. Signed-off-by: Bjorn Helgaas <bhelgaas@google.com> Cc: Liang He <windhl@126.com> Cc: Geert Uytterhoeven <geert@linux-m68k.org> --- drivers/pci/controller/pcie-rcar-host.c | 60 +++++++++++-------------- 1 file changed, 27 insertions(+), 33 deletions(-) diff --git a/drivers/pci/controller/pcie-rcar-host.c b/drivers/pci/controller/pcie-rcar-host.c index ccf13aafa1e5..139a0274b9e0 100644 --- a/drivers/pci/controller/pcie-rcar-host.c +++ b/drivers/pci/controller/pcie-rcar-host.c @@ -60,6 +60,7 @@ struct rcar_pcie_host; struct rcar_variant { int (*phy_init_fn)(struct rcar_pcie_host *host); + bool hook_aborts; }; /* Structure representing the PCIe interface */ @@ -964,12 +965,35 @@ static int rcar_pcie_parse_map_dma_ranges(struct rcar_pcie_host *host) return err; } +#ifdef CONFIG_ARM +static int rcar_pcie_aarch32_abort_handler(unsigned long addr, + unsigned int fsr, struct pt_regs *regs) +{ + return !fixup_exception(regs); +} +#endif + +static void rcar_pcie_hook_aborts(void) +{ +#ifdef CONFIG_ARM +#ifdef CONFIG_ARM_LPAE + hook_fault_code(17, rcar_pcie_aarch32_abort_handler, SIGBUS, 0, + "asynchronous external abort"); +#else + hook_fault_code(22, rcar_pcie_aarch32_abort_handler, SIGBUS, 0, + "imprecise external abort"); +#endif +#endif +} + static const struct rcar_variant rcar_h1_data = { .phy_init_fn = rcar_pcie_phy_init_h1, + .hook_aborts = true, }; static const struct rcar_variant rcar_gen2_data = { .phy_init_fn = rcar_pcie_phy_init_gen2, + .hook_aborts = true, }; static const struct rcar_variant rcar_gen3_data = { @@ -1035,6 +1059,9 @@ static int rcar_pcie_probe(struct platform_device *pdev) goto err_clk_disable; } + if (host->variant->hook_aborts) + rcar_pcie_hook_aborts(); + /* Failure to get a link might just be that no cards are inserted */ if (rcar_pcie_hw_init(pcie)) { dev_info(dev, "PCIe link down\n"); @@ -1153,37 +1180,4 @@ static struct platform_driver rcar_pcie_driver = { }, .probe = rcar_pcie_probe, }; - -#ifdef CONFIG_ARM -static int rcar_pcie_aarch32_abort_handler(unsigned long addr, - unsigned int fsr, struct pt_regs *regs) -{ - return !fixup_exception(regs); -} - -static const struct of_device_id rcar_pcie_abort_handler_of_match[] __initconst = { - { .compatible = "renesas,pcie-r8a7779" }, - { .compatible = "renesas,pcie-r8a7790" }, - { .compatible = "renesas,pcie-r8a7791" }, - { .compatible = "renesas,pcie-rcar-gen2" }, - {}, -}; - -static int __init rcar_pcie_init(void) -{ - if (of_find_matching_node(NULL, rcar_pcie_abort_handler_of_match)) { -#ifdef CONFIG_ARM_LPAE - hook_fault_code(17, rcar_pcie_aarch32_abort_handler, SIGBUS, 0, - "asynchronous external abort"); -#else - hook_fault_code(22, rcar_pcie_aarch32_abort_handler, SIGBUS, 0, - "imprecise external abort"); -#endif - } - - return platform_driver_register(&rcar_pcie_driver); -} -device_initcall(rcar_pcie_init); -#else builtin_platform_driver(rcar_pcie_driver); -#endif -- 2.25.1 ^ permalink raw reply related [flat|nested] 5+ messages in thread
* Re: [PATCH 2/2] PCI: rcar: Resolve of_find_matching_node() reference leak 2022-06-22 2:57 ` [PATCH 2/2] PCI: rcar: Resolve of_find_matching_node() reference leak Bjorn Helgaas @ 2022-06-22 7:45 ` Geert Uytterhoeven 2022-06-22 11:05 ` Bjorn Helgaas 0 siblings, 1 reply; 5+ messages in thread From: Geert Uytterhoeven @ 2022-06-22 7:45 UTC (permalink / raw) To: Bjorn Helgaas Cc: Marek Vasut, Yoshihiro Shimoda, Liang He, Lorenzo Pieralisi, Rob Herring, Krzysztof Wilczyński, linux-pci, Linux-Renesas, Linux Kernel Mailing List, Bjorn Helgaas Hi Bjorn, On Wed, Jun 22, 2022 at 4:57 AM Bjorn Helgaas <helgaas@kernel.org> wrote: > From: Bjorn Helgaas <bhelgaas@google.com> > > Previously, rcar_pcie_init() used of_find_matching_node() to search the > entire device tree for compatible strings for which we need to install an > abort handler. If we found one, we got a device_node with refcount > incremented, but we discarded the pointer and never released that > reference. > > Extend the struct rcar_variant to indicate whether each variant requires an > abort handler. Install the handler in rcar_pcie_probe() when needed. > > Signed-off-by: Bjorn Helgaas <bhelgaas@google.com> > Cc: Liang He <windhl@126.com> > Cc: Geert Uytterhoeven <geert@linux-m68k.org> Thanks for your patch! > --- a/drivers/pci/controller/pcie-rcar-host.c > +++ b/drivers/pci/controller/pcie-rcar-host.c > @@ -964,12 +965,35 @@ static int rcar_pcie_parse_map_dma_ranges(struct rcar_pcie_host *host) > return err; > } > > +#ifdef CONFIG_ARM > +static int rcar_pcie_aarch32_abort_handler(unsigned long addr, > + unsigned int fsr, struct pt_regs *regs) > +{ > + return !fixup_exception(regs); > +} > +#endif > + > +static void rcar_pcie_hook_aborts(void) > +{ > +#ifdef CONFIG_ARM > +#ifdef CONFIG_ARM_LPAE > + hook_fault_code(17, rcar_pcie_aarch32_abort_handler, SIGBUS, 0, > + "asynchronous external abort"); > +#else > + hook_fault_code(22, rcar_pcie_aarch32_abort_handler, SIGBUS, 0, > + "imprecise external abort"); > +#endif > +#endif > +} > + > static const struct rcar_variant rcar_h1_data = { > .phy_init_fn = rcar_pcie_phy_init_h1, > + .hook_aborts = true, > }; > > static const struct rcar_variant rcar_gen2_data = { > .phy_init_fn = rcar_pcie_phy_init_gen2, > + .hook_aborts = true, > }; > > static const struct rcar_variant rcar_gen3_data = { > @@ -1035,6 +1059,9 @@ static int rcar_pcie_probe(struct platform_device *pdev) > goto err_clk_disable; > } > > + if (host->variant->hook_aborts) > + rcar_pcie_hook_aborts(); I was quite sure there was a good reason why this was not done in .probe() before... And indeed, the original submission[1] did have a comment explaining that: + /* + * Since probe() can be deferred we need to make sure that + * hook_fault_code is not called after __init memory is freed + * by kernel and since rcar_pcie_abort_handler() is a no-op, + * we can install the handler here without risking it + * accessing some uninitialized driver state. + */ No idea why it was removed in v2 and later, but the point is: hook_fault_code() is __init, so you cannot call it from a deferred probe. And you should have got a section mismatch warning ;-) [1] https://lore.kernel.org/all/20200912211853.15321-1-marek.vasut@gmail.com/ Gr{oetje,eeting}s, Geert -- Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that. -- Linus Torvalds ^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH 2/2] PCI: rcar: Resolve of_find_matching_node() reference leak 2022-06-22 7:45 ` Geert Uytterhoeven @ 2022-06-22 11:05 ` Bjorn Helgaas 0 siblings, 0 replies; 5+ messages in thread From: Bjorn Helgaas @ 2022-06-22 11:05 UTC (permalink / raw) To: Geert Uytterhoeven Cc: Marek Vasut, Yoshihiro Shimoda, Liang He, Lorenzo Pieralisi, Rob Herring, Krzysztof Wilczyński, linux-pci, Linux-Renesas, Linux Kernel Mailing List, Bjorn Helgaas On Wed, Jun 22, 2022 at 09:45:49AM +0200, Geert Uytterhoeven wrote: > Hi Bjorn, > > On Wed, Jun 22, 2022 at 4:57 AM Bjorn Helgaas <helgaas@kernel.org> wrote: > > From: Bjorn Helgaas <bhelgaas@google.com> > > > > Previously, rcar_pcie_init() used of_find_matching_node() to search the > > entire device tree for compatible strings for which we need to install an > > abort handler. If we found one, we got a device_node with refcount > > incremented, but we discarded the pointer and never released that > > reference. > > > > Extend the struct rcar_variant to indicate whether each variant requires an > > abort handler. Install the handler in rcar_pcie_probe() when needed. > > > > Signed-off-by: Bjorn Helgaas <bhelgaas@google.com> > > Cc: Liang He <windhl@126.com> > > Cc: Geert Uytterhoeven <geert@linux-m68k.org> > > Thanks for your patch! > > > --- a/drivers/pci/controller/pcie-rcar-host.c > > +++ b/drivers/pci/controller/pcie-rcar-host.c > > > @@ -964,12 +965,35 @@ static int rcar_pcie_parse_map_dma_ranges(struct rcar_pcie_host *host) > > return err; > > } > > > > +#ifdef CONFIG_ARM > > +static int rcar_pcie_aarch32_abort_handler(unsigned long addr, > > + unsigned int fsr, struct pt_regs *regs) > > +{ > > + return !fixup_exception(regs); > > +} > > +#endif > > + > > +static void rcar_pcie_hook_aborts(void) > > +{ > > +#ifdef CONFIG_ARM > > +#ifdef CONFIG_ARM_LPAE > > + hook_fault_code(17, rcar_pcie_aarch32_abort_handler, SIGBUS, 0, > > + "asynchronous external abort"); > > +#else > > + hook_fault_code(22, rcar_pcie_aarch32_abort_handler, SIGBUS, 0, > > + "imprecise external abort"); > > +#endif > > +#endif > > +} > > + > > static const struct rcar_variant rcar_h1_data = { > > .phy_init_fn = rcar_pcie_phy_init_h1, > > + .hook_aborts = true, > > }; > > > > static const struct rcar_variant rcar_gen2_data = { > > .phy_init_fn = rcar_pcie_phy_init_gen2, > > + .hook_aborts = true, > > }; > > > > static const struct rcar_variant rcar_gen3_data = { > > @@ -1035,6 +1059,9 @@ static int rcar_pcie_probe(struct platform_device *pdev) > > goto err_clk_disable; > > } > > > > + if (host->variant->hook_aborts) > > + rcar_pcie_hook_aborts(); > > I was quite sure there was a good reason why this was not done in > .probe() before... > > And indeed, the original submission[1] did have a comment explaining > that: > > + /* > + * Since probe() can be deferred we need to make sure that > + * hook_fault_code is not called after __init memory is freed > + * by kernel and since rcar_pcie_abort_handler() is a no-op, > + * we can install the handler here without risking it > + * accessing some uninitialized driver state. > + */ > > No idea why it was removed in v2 and later, but the point is: > hook_fault_code() is __init, so you cannot call it from a deferred > probe. > And you should have got a section mismatch warning ;-) Oooh, thanks for that! I missed the builtin_platform_driver_probe() vs builtin_platform_driver() difference that explains why doing this at probe-time works for pci-ixp4xx.c but not here. Bjorn ^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2022-06-22 11:05 UTC | newest] Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2022-06-22 2:57 [PATCH 0/2] PCI: rcar: Fix of_find_matching_node() reference leak Bjorn Helgaas 2022-06-22 2:57 ` [PATCH 1/2] PCI: rcar: Add dev struct for of_device_get_match_data() Bjorn Helgaas 2022-06-22 2:57 ` [PATCH 2/2] PCI: rcar: Resolve of_find_matching_node() reference leak Bjorn Helgaas 2022-06-22 7:45 ` Geert Uytterhoeven 2022-06-22 11:05 ` Bjorn Helgaas
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).