* [PATCH 1/2] x86/PCI: add kernel option and taint it when we add a 64bit window @ 2018-01-10 12:25 Christian König 2018-01-10 12:25 ` [PATCH 2/2] x86/PCI: limit the size of the 64bit window to 256GB v2 Christian König 2018-01-10 20:10 ` [PATCH 1/2] x86/PCI: add kernel option and taint it when we add a 64bit window Randy Dunlap 0 siblings, 2 replies; 6+ messages in thread From: Christian König @ 2018-01-10 12:25 UTC (permalink / raw) To: bhelgaas; +Cc: linux-kernel Only try to enable a 64bit window on AMD CPUs when pci=big_root_window is specified and taint the kernel when we add the window. Signed-off-by: Christian König <christian.koenig@amd.com> --- arch/x86/include/asm/pci_x86.h | 1 + arch/x86/pci/common.c | 5 +++++ arch/x86/pci/fixup.c | 4 ++++ 3 files changed, 10 insertions(+) diff --git a/arch/x86/include/asm/pci_x86.h b/arch/x86/include/asm/pci_x86.h index 7a5d6695abd3..eb66fa9cd0fc 100644 --- a/arch/x86/include/asm/pci_x86.h +++ b/arch/x86/include/asm/pci_x86.h @@ -38,6 +38,7 @@ do { \ #define PCI_NOASSIGN_ROMS 0x80000 #define PCI_ROOT_NO_CRS 0x100000 #define PCI_NOASSIGN_BARS 0x200000 +#define PCI_BIG_ROOT_WINDOW 0x400000 extern unsigned int pci_probe; extern unsigned long pirq_table_addr; diff --git a/arch/x86/pci/common.c b/arch/x86/pci/common.c index 7a5350d08cef..563049c483a1 100644 --- a/arch/x86/pci/common.c +++ b/arch/x86/pci/common.c @@ -594,6 +594,11 @@ char *__init pcibios_setup(char *str) } else if (!strcmp(str, "nocrs")) { pci_probe |= PCI_ROOT_NO_CRS; return NULL; +#ifdef CONFIG_PHYS_ADDR_T_64BIT + } else if (!strcmp(str, "big_root_window")) { + pci_probe |= PCI_BIG_ROOT_WINDOW; + return NULL; +#endif } else if (!strcmp(str, "earlydump")) { pci_early_dump_regs = 1; return NULL; diff --git a/arch/x86/pci/fixup.c b/arch/x86/pci/fixup.c index e663d6bf1328..a91280da2ea1 100644 --- a/arch/x86/pci/fixup.c +++ b/arch/x86/pci/fixup.c @@ -667,6 +667,9 @@ static void pci_amd_enable_64bit_bar(struct pci_dev *dev) struct resource *res, *conflict; struct pci_dev *other; + if (!(pci_probe & PCI_BIG_ROOT_WINDOW)) + return; + /* Check that we are the only device of that type */ other = pci_get_device(dev->vendor, dev->device, NULL); if (other != dev || @@ -715,6 +718,7 @@ static void pci_amd_enable_64bit_bar(struct pci_dev *dev) } dev_info(&dev->dev, "adding root bus resource %pR\n", res); + add_taint(TAINT_FIRMWARE_WORKAROUND, LOCKDEP_STILL_OK); base = ((res->start >> 8) & AMD_141b_MMIO_BASE_MMIOBASE_MASK) | AMD_141b_MMIO_BASE_RE_MASK | AMD_141b_MMIO_BASE_WE_MASK; -- 2.11.0 ^ permalink raw reply [flat|nested] 6+ messages in thread
* [PATCH 2/2] x86/PCI: limit the size of the 64bit window to 256GB v2 2018-01-10 12:25 [PATCH 1/2] x86/PCI: add kernel option and taint it when we add a 64bit window Christian König @ 2018-01-10 12:25 ` Christian König 2018-01-10 18:02 ` Bjorn Helgaas 2018-01-10 20:10 ` [PATCH 1/2] x86/PCI: add kernel option and taint it when we add a 64bit window Randy Dunlap 1 sibling, 1 reply; 6+ messages in thread From: Christian König @ 2018-01-10 12:25 UTC (permalink / raw) To: bhelgaas; +Cc: linux-kernel Avoid problems with BIOS implementations which don't report all used resources to the OS by only allocating a 256GB window directly below the hardware limit. v2: cleanup code a bit more, update comment and explain the hw limit Signed-off-by: Christian König <christian.koenig@amd.com> --- arch/x86/pci/fixup.c | 19 +++++++++---------- 1 file changed, 9 insertions(+), 10 deletions(-) diff --git a/arch/x86/pci/fixup.c b/arch/x86/pci/fixup.c index a91280da2ea1..9c1c98d7e3a7 100644 --- a/arch/x86/pci/fixup.c +++ b/arch/x86/pci/fixup.c @@ -662,10 +662,11 @@ DECLARE_PCI_FIXUP_EARLY(PCI_VENDOR_ID_INTEL, 0x2033, quirk_no_aersid); */ static void pci_amd_enable_64bit_bar(struct pci_dev *dev) { - unsigned i; u32 base, limit, high; - struct resource *res, *conflict; struct pci_dev *other; + struct resource *res; + unsigned i; + int r; if (!(pci_probe & PCI_BIG_ROOT_WINDOW)) return; @@ -702,19 +703,17 @@ static void pci_amd_enable_64bit_bar(struct pci_dev *dev) if (!res) return; + /* Allocate a 256GB window directly below the 0xfd00000000 hw limit */ res->name = "PCI Bus 0000:00"; res->flags = IORESOURCE_PREFETCH | IORESOURCE_MEM | IORESOURCE_MEM_64 | IORESOURCE_WINDOW; - res->start = 0x100000000ull; + res->start = 0xbd00000000ull; res->end = 0xfd00000000ull - 1; - /* Just grab the free area behind system memory for this */ - while ((conflict = request_resource_conflict(&iomem_resource, res))) { - if (conflict->end >= res->end) { - kfree(res); - return; - } - res->start = conflict->end + 1; + r = request_resource(&iomem_resource, res); + if (r) { + kfree(res); + return; } dev_info(&dev->dev, "adding root bus resource %pR\n", res); -- 2.11.0 ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH 2/2] x86/PCI: limit the size of the 64bit window to 256GB v2 2018-01-10 12:25 ` [PATCH 2/2] x86/PCI: limit the size of the 64bit window to 256GB v2 Christian König @ 2018-01-10 18:02 ` Bjorn Helgaas 2018-01-10 18:14 ` Christian König 0 siblings, 1 reply; 6+ messages in thread From: Bjorn Helgaas @ 2018-01-10 18:02 UTC (permalink / raw) To: Christian König Cc: bhelgaas, linux-kernel, linux-pci, Linus Torvalds, Aaro Koskinen, Andy Shevchenko, Boris Ostrovsky, Juergen Gross, Alex Deucher, David Airlie [+cc linux-pci, previous cc list] On Wed, Jan 10, 2018 at 01:25:51PM +0100, Christian König wrote: > Avoid problems with BIOS implementations which don't report all used > resources to the OS by only allocating a 256GB window directly below the > hardware limit. > > v2: cleanup code a bit more, update comment and explain the hw limit Fixes a silent reboot loop reported by Aaro Koskinen <aaro.koskinen@iki.fi> on an AMD-based MSI MS-7699/760GA-P43(FX) system. Link: https://lkml.kernel.org/r/20180105220412.fzpwqe4zljdawr36@darkstar.musicnaut.iki.fi Reported-by: Aaro Koskinen <aaro.koskinen@iki.fi> > Signed-off-by: Christian König <christian.koenig@amd.com> > --- > arch/x86/pci/fixup.c | 19 +++++++++---------- > 1 file changed, 9 insertions(+), 10 deletions(-) > > diff --git a/arch/x86/pci/fixup.c b/arch/x86/pci/fixup.c > index a91280da2ea1..9c1c98d7e3a7 100644 > --- a/arch/x86/pci/fixup.c > +++ b/arch/x86/pci/fixup.c > @@ -662,10 +662,11 @@ DECLARE_PCI_FIXUP_EARLY(PCI_VENDOR_ID_INTEL, 0x2033, quirk_no_aersid); > */ > static void pci_amd_enable_64bit_bar(struct pci_dev *dev) > { > - unsigned i; > u32 base, limit, high; > - struct resource *res, *conflict; > struct pci_dev *other; > + struct resource *res; > + unsigned i; > + int r; > > if (!(pci_probe & PCI_BIG_ROOT_WINDOW)) > return; > @@ -702,19 +703,17 @@ static void pci_amd_enable_64bit_bar(struct pci_dev *dev) > if (!res) > return; > > + /* Allocate a 256GB window directly below the 0xfd00000000 hw limit */ Can you include a citation for the hardware limit in the changelog? I'll try to dig it up myself if you don't beat me to it. > res->name = "PCI Bus 0000:00"; > res->flags = IORESOURCE_PREFETCH | IORESOURCE_MEM | > IORESOURCE_MEM_64 | IORESOURCE_WINDOW; > - res->start = 0x100000000ull; > + res->start = 0xbd00000000ull; > res->end = 0xfd00000000ull - 1; > > - /* Just grab the free area behind system memory for this */ > - while ((conflict = request_resource_conflict(&iomem_resource, res))) { > - if (conflict->end >= res->end) { > - kfree(res); > - return; > - } > - res->start = conflict->end + 1; > + r = request_resource(&iomem_resource, res); > + if (r) { > + kfree(res); > + return; > } > > dev_info(&dev->dev, "adding root bus resource %pR\n", res); > -- > 2.11.0 > ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH 2/2] x86/PCI: limit the size of the 64bit window to 256GB v2 2018-01-10 18:02 ` Bjorn Helgaas @ 2018-01-10 18:14 ` Christian König 2018-01-10 20:37 ` Bjorn Helgaas 0 siblings, 1 reply; 6+ messages in thread From: Christian König @ 2018-01-10 18:14 UTC (permalink / raw) To: Bjorn Helgaas Cc: bhelgaas, linux-kernel, linux-pci, Linus Torvalds, Aaro Koskinen, Andy Shevchenko, Boris Ostrovsky, Juergen Gross, Alex Deucher, David Airlie Am 10.01.2018 um 19:02 schrieb Bjorn Helgaas: > [+cc linux-pci, previous cc list] > > On Wed, Jan 10, 2018 at 01:25:51PM +0100, Christian König wrote: >> Avoid problems with BIOS implementations which don't report all used >> resources to the OS by only allocating a 256GB window directly below the >> hardware limit. >> >> v2: cleanup code a bit more, update comment and explain the hw limit > Fixes a silent reboot loop reported by Aaro Koskinen > <aaro.koskinen@iki.fi> on an AMD-based MSI MS-7699/760GA-P43(FX) > system. > > Link: https://lkml.kernel.org/r/20180105220412.fzpwqe4zljdawr36@darkstar.musicnaut.iki.fi > Reported-by: Aaro Koskinen <aaro.koskinen@iki.fi> > >> Signed-off-by: Christian König <christian.koenig@amd.com> >> --- >> arch/x86/pci/fixup.c | 19 +++++++++---------- >> 1 file changed, 9 insertions(+), 10 deletions(-) >> >> diff --git a/arch/x86/pci/fixup.c b/arch/x86/pci/fixup.c >> index a91280da2ea1..9c1c98d7e3a7 100644 >> --- a/arch/x86/pci/fixup.c >> +++ b/arch/x86/pci/fixup.c >> @@ -662,10 +662,11 @@ DECLARE_PCI_FIXUP_EARLY(PCI_VENDOR_ID_INTEL, 0x2033, quirk_no_aersid); >> */ >> static void pci_amd_enable_64bit_bar(struct pci_dev *dev) >> { >> - unsigned i; >> u32 base, limit, high; >> - struct resource *res, *conflict; >> struct pci_dev *other; >> + struct resource *res; >> + unsigned i; >> + int r; >> >> if (!(pci_probe & PCI_BIG_ROOT_WINDOW)) >> return; >> @@ -702,19 +703,17 @@ static void pci_amd_enable_64bit_bar(struct pci_dev *dev) >> if (!res) >> return; >> >> + /* Allocate a 256GB window directly below the 0xfd00000000 hw limit */ > Can you include a citation for the hardware limit in the changelog? > I'll try to dig it up myself if you don't beat me to it. Sure, no problem. The first one in the series is ok as it is? Christian. > >> res->name = "PCI Bus 0000:00"; >> res->flags = IORESOURCE_PREFETCH | IORESOURCE_MEM | >> IORESOURCE_MEM_64 | IORESOURCE_WINDOW; >> - res->start = 0x100000000ull; >> + res->start = 0xbd00000000ull; >> res->end = 0xfd00000000ull - 1; >> >> - /* Just grab the free area behind system memory for this */ >> - while ((conflict = request_resource_conflict(&iomem_resource, res))) { >> - if (conflict->end >= res->end) { >> - kfree(res); >> - return; >> - } >> - res->start = conflict->end + 1; >> + r = request_resource(&iomem_resource, res); >> + if (r) { >> + kfree(res); >> + return; >> } >> >> dev_info(&dev->dev, "adding root bus resource %pR\n", res); >> -- >> 2.11.0 >> ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH 2/2] x86/PCI: limit the size of the 64bit window to 256GB v2 2018-01-10 18:14 ` Christian König @ 2018-01-10 20:37 ` Bjorn Helgaas 0 siblings, 0 replies; 6+ messages in thread From: Bjorn Helgaas @ 2018-01-10 20:37 UTC (permalink / raw) To: christian.koenig Cc: bhelgaas, linux-kernel, linux-pci, Linus Torvalds, Aaro Koskinen, Andy Shevchenko, Boris Ostrovsky, Juergen Gross, Alex Deucher, David Airlie On Wed, Jan 10, 2018 at 07:14:56PM +0100, Christian König wrote: > Am 10.01.2018 um 19:02 schrieb Bjorn Helgaas: > >[+cc linux-pci, previous cc list] > > > >On Wed, Jan 10, 2018 at 01:25:51PM +0100, Christian König wrote: > >>Avoid problems with BIOS implementations which don't report all used > >>resources to the OS by only allocating a 256GB window directly below the > >>hardware limit. > >> > >>v2: cleanup code a bit more, update comment and explain the hw limit > >Fixes a silent reboot loop reported by Aaro Koskinen > ><aaro.koskinen@iki.fi> on an AMD-based MSI MS-7699/760GA-P43(FX) > >system. > > > >Link: https://lkml.kernel.org/r/20180105220412.fzpwqe4zljdawr36@darkstar.musicnaut.iki.fi > >Reported-by: Aaro Koskinen <aaro.koskinen@iki.fi> > > > >>Signed-off-by: Christian König <christian.koenig@amd.com> > >>--- > >> arch/x86/pci/fixup.c | 19 +++++++++---------- > >> 1 file changed, 9 insertions(+), 10 deletions(-) > >> > >>diff --git a/arch/x86/pci/fixup.c b/arch/x86/pci/fixup.c > >>index a91280da2ea1..9c1c98d7e3a7 100644 > >>--- a/arch/x86/pci/fixup.c > >>+++ b/arch/x86/pci/fixup.c > >>@@ -662,10 +662,11 @@ DECLARE_PCI_FIXUP_EARLY(PCI_VENDOR_ID_INTEL, 0x2033, quirk_no_aersid); > >> */ > >> static void pci_amd_enable_64bit_bar(struct pci_dev *dev) > >> { > >>- unsigned i; > >> u32 base, limit, high; > >>- struct resource *res, *conflict; > >> struct pci_dev *other; > >>+ struct resource *res; > >>+ unsigned i; > >>+ int r; > >> if (!(pci_probe & PCI_BIG_ROOT_WINDOW)) > >> return; > >>@@ -702,19 +703,17 @@ static void pci_amd_enable_64bit_bar(struct pci_dev *dev) > >> if (!res) > >> return; > >>+ /* Allocate a 256GB window directly below the 0xfd00000000 hw limit */ > >Can you include a citation for the hardware limit in the changelog? > >I'll try to dig it up myself if you don't beat me to it. > > Sure, no problem. The first one in the series is ok as it is? Yeah, I think so, although it would be good to add doc as Randy suggests. Can you cc the folks involved (in particular, linux-pci) next time, too? > >> res->name = "PCI Bus 0000:00"; > >> res->flags = IORESOURCE_PREFETCH | IORESOURCE_MEM | > >> IORESOURCE_MEM_64 | IORESOURCE_WINDOW; > >>- res->start = 0x100000000ull; > >>+ res->start = 0xbd00000000ull; > >> res->end = 0xfd00000000ull - 1; > >>- /* Just grab the free area behind system memory for this */ > >>- while ((conflict = request_resource_conflict(&iomem_resource, res))) { > >>- if (conflict->end >= res->end) { > >>- kfree(res); > >>- return; > >>- } > >>- res->start = conflict->end + 1; > >>+ r = request_resource(&iomem_resource, res); > >>+ if (r) { > >>+ kfree(res); > >>+ return; > >> } > >> dev_info(&dev->dev, "adding root bus resource %pR\n", res); > >>-- > >>2.11.0 > >> > ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH 1/2] x86/PCI: add kernel option and taint it when we add a 64bit window 2018-01-10 12:25 [PATCH 1/2] x86/PCI: add kernel option and taint it when we add a 64bit window Christian König 2018-01-10 12:25 ` [PATCH 2/2] x86/PCI: limit the size of the 64bit window to 256GB v2 Christian König @ 2018-01-10 20:10 ` Randy Dunlap 1 sibling, 0 replies; 6+ messages in thread From: Randy Dunlap @ 2018-01-10 20:10 UTC (permalink / raw) To: Christian König, bhelgaas; +Cc: linux-kernel On 01/10/18 04:25, Christian König wrote: > Only try to enable a 64bit window on AMD CPUs when pci=big_root_window is > specified and taint the kernel when we add the window. > > Signed-off-by: Christian König <christian.koenig@amd.com> > --- > arch/x86/include/asm/pci_x86.h | 1 + > arch/x86/pci/common.c | 5 +++++ > arch/x86/pci/fixup.c | 4 ++++ > 3 files changed, 10 insertions(+) > Please update Documentation/admin-guide/kernel-parameters.txt with this new option of "pci=". -- ~Randy ^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2018-01-10 20:37 UTC | newest] Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2018-01-10 12:25 [PATCH 1/2] x86/PCI: add kernel option and taint it when we add a 64bit window Christian König 2018-01-10 12:25 ` [PATCH 2/2] x86/PCI: limit the size of the 64bit window to 256GB v2 Christian König 2018-01-10 18:02 ` Bjorn Helgaas 2018-01-10 18:14 ` Christian König 2018-01-10 20:37 ` Bjorn Helgaas 2018-01-10 20:10 ` [PATCH 1/2] x86/PCI: add kernel option and taint it when we add a 64bit window Randy Dunlap
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®