- Linux-stable-mirror - lists.linaro.org

[PATCH 4.9.y] PCI: OF: Fix I/O space page leak

by Sergei Shtylyov

Commit a5fb9fb023a1435f2b42bccd7f547560f3a21dc3 upstream. When testing the R-Car PCIe driver on the Condor board, if the PCIe PHY driver was left disabled, the kernel crashed with this BUG: kernel BUG at lib/ioremap.c:72! Internal error: Oops - BUG: 0 [#1] PREEMPT SMP Modules linked in: CPU: 0 PID: 39 Comm: kworker/0:1 Not tainted 4.17.0-dirty #1092 Hardware name: Renesas Condor board based on r8a77980 (DT) Workqueue: events deferred_probe_work_func pstate: 80000005 (Nzcv daif -PAN -UAO) pc : ioremap_page_range+0x370/0x3c8 lr : ioremap_page_range+0x40/0x3c8 sp : ffff000008da39e0 x29: ffff000008da39e0 x28: 00e8000000000f07 x27: ffff7dfffee00000 x26: 0140000000000000 x25: ffff7dfffef00000 x24: 00000000000fe100 x23: ffff80007b906000 x22: ffff000008ab8000 x21: ffff000008bb1d58 x20: ffff7dfffef00000 x19: ffff800009c30fb8 x18: 0000000000000001 x17: 00000000000152d0 x16: 00000000014012d0 x15: 0000000000000000 x14: 0720072007200720 x13: 0720072007200720 x12: 0720072007200720 x11: 0720072007300730 x10: 00000000000000ae x9 : 0000000000000000 x8 : ffff7dffff000000 x7 : 0000000000000000 x6 : 0000000000000100 x5 : 0000000000000000 x4 : 000000007b906000 x3 : ffff80007c61a880 x2 : ffff7dfffeefffff x1 : 0000000040000000 x0 : 00e80000fe100f07 Process kworker/0:1 (pid: 39, stack limit = 0x (ptrval)) Call trace: ioremap_page_range+0x370/0x3c8 pci_remap_iospace+0x7c/0xac pci_parse_request_of_pci_ranges+0x13c/0x190 rcar_pcie_probe+0x4c/0xb04 platform_drv_probe+0x50/0xbc driver_probe_device+0x21c/0x308 __device_attach_driver+0x98/0xc8 bus_for_each_drv+0x54/0x94 __device_attach+0xc4/0x12c device_initial_probe+0x10/0x18 bus_probe_device+0x90/0x98 deferred_probe_work_func+0xb0/0x150 process_one_work+0x12c/0x29c worker_thread+0x200/0x3fc kthread+0x108/0x134 ret_from_fork+0x10/0x18 Code: f9004ba2 54000080 aa0003fb 17ffff48 (d4210000) It turned out that pci_remap_iospace() wasn't undone when the driver's probe failed, and since devm_phy_optional_get() returned -EPROBE_DEFER, the probe was retried, finally causing the BUG due to trying to remap already remapped pages. Introduce the devm_pci_remap_iospace() managed API and replace the pci_remap_iospace() call with it to fix the bug. Fixes: dbf9826d5797 ("PCI: generic: Convert to DT resource parsing API") Signed-off-by: Sergei Shtylyov <sergei.shtylyov(a)cogentembedded.com> [lorenzo.pieralisi(a)arm.com: split commit/updated the commit log] Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi(a)arm.com> Signed-off-by: Bjorn Helgaas <bhelgaas(a)google.com> Reviewed-by: Linus Walleij <linus.walleij(a)linaro.org> --- drivers/pci/host/pci-host-common.c | 2 - drivers/pci/host/pcie-rcar.c | 2 - drivers/pci/pci.c | 38 +++++++++++++++++++++++++++++++++++++ include/linux/pci.h | 2 + 4 files changed, 42 insertions(+), 2 deletions(-) Index: linux-stable/drivers/pci/host/pci-host-common.c =================================================================== --- linux-stable.orig/drivers/pci/host/pci-host-common.c +++ linux-stable/drivers/pci/host/pci-host-common.c @@ -45,7 +45,7 @@ static int gen_pci_parse_request_of_pci_ switch (resource_type(res)) { case IORESOURCE_IO: - err = pci_remap_iospace(res, iobase); + err = devm_pci_remap_iospace(dev, res, iobase); if (err) { dev_warn(dev, "error %d: failed to map resource %pR\n", err, res); Index: linux-stable/drivers/pci/host/pcie-rcar.c =================================================================== --- linux-stable.orig/drivers/pci/host/pcie-rcar.c +++ linux-stable/drivers/pci/host/pcie-rcar.c @@ -1102,7 +1102,7 @@ static int rcar_pcie_parse_request_of_pc struct resource *res = win->res; if (resource_type(res) == IORESOURCE_IO) { - err = pci_remap_iospace(res, iobase); + err = devm_pci_remap_iospace(dev, res, iobase); if (err) { dev_warn(dev, "error %d: failed to map resource %pR\n", err, res); Index: linux-stable/drivers/pci/pci.c =================================================================== --- linux-stable.orig/drivers/pci/pci.c +++ linux-stable/drivers/pci/pci.c @@ -3436,6 +3436,44 @@ char * __weak __init pcibios_setup(char return str; } +static void devm_pci_unmap_iospace(struct device *dev, void *ptr) +{ + struct resource **res = ptr; + + pci_unmap_iospace(*res); +} + +/** + * devm_pci_remap_iospace - Managed pci_remap_iospace() + * @dev: Generic device to remap IO address for + * @res: Resource describing the I/O space + * @phys_addr: physical address of range to be mapped + * + * Managed pci_remap_iospace(). Map is automatically unmapped on driver + * detach. + */ +int devm_pci_remap_iospace(struct device *dev, const struct resource *res, + phys_addr_t phys_addr) +{ + const struct resource **ptr; + int error; + + ptr = devres_alloc(devm_pci_unmap_iospace, sizeof(*ptr), GFP_KERNEL); + if (!ptr) + return -ENOMEM; + + error = pci_remap_iospace(res, phys_addr); + if (error) { + devres_free(ptr); + } else { + *ptr = res; + devres_add(dev, ptr); + } + + return error; +} +EXPORT_SYMBOL(devm_pci_remap_iospace); + /** * pcibios_set_master - enable PCI bus-mastering for device dev * @dev: the PCI device to enable Index: linux-stable/include/linux/pci.h =================================================================== --- linux-stable.orig/include/linux/pci.h +++ linux-stable/include/linux/pci.h @@ -1190,6 +1190,8 @@ int pci_register_io_range(phys_addr_t ad unsigned long pci_address_to_pio(phys_addr_t addr); phys_addr_t pci_pio_to_address(unsigned long pio); int pci_remap_iospace(const struct resource *res, phys_addr_t phys_addr); +int devm_pci_remap_iospace(struct device *dev, const struct resource *res, + phys_addr_t phys_addr); void pci_unmap_iospace(struct resource *res); static inline pci_bus_addr_t pci_bus_address(struct pci_dev *pdev, int bar)

7 years, 3 months

1
1
0 0

[PATCH v3 1/2] mm: migration: fix migration of huge PMD shared pages

by Mike Kravetz

The page migration code employs try_to_unmap() to try and unmap the source page. This is accomplished by using rmap_walk to find all vmas where the page is mapped. This search stops when page mapcount is zero. For shared PMD huge pages, the page map count is always 1 no matter the number of mappings. Shared mappings are tracked via the reference count of the PMD page. Therefore, try_to_unmap stops prematurely and does not completely unmap all mappings of the source page. This problem can result is data corruption as writes to the original source page can happen after contents of the page are copied to the target page. Hence, data is lost. This problem was originally seen as DB corruption of shared global areas after a huge page was soft offlined due to ECC memory errors. DB developers noticed they could reproduce the issue by (hotplug) offlining memory used to back huge pages. A simple testcase can reproduce the problem by creating a shared PMD mapping (note that this must be at least PUD_SIZE in size and PUD_SIZE aligned (1GB on x86)), and using migrate_pages() to migrate process pages between nodes while continually writing to the huge pages being migrated. To fix, have the try_to_unmap_one routine check for huge PMD sharing by calling huge_pmd_unshare for hugetlbfs huge pages. If it is a shared mapping it will be 'unshared' which removes the page table entry and drops the reference on the PMD page. After this, flush caches and TLB. mmu notifiers are called before locking page tables, but we can not be sure of PMD sharing until page tables are locked. Therefore, check for the possibility of PMD sharing before locking so that notifiers can prepare for the worst possible case. Fixes: 39dde65c9940 ("shared page table for hugetlb page") Cc: stable(a)vger.kernel.org Signed-off-by: Mike Kravetz <mike.kravetz(a)oracle.com> --- include/linux/hugetlb.h | 14 ++++++++++++++ mm/hugetlb.c | 40 +++++++++++++++++++++++++++++++++++++++ mm/rmap.c | 42 ++++++++++++++++++++++++++++++++++++++--- 3 files changed, 93 insertions(+), 3 deletions(-) diff --git a/include/linux/hugetlb.h b/include/linux/hugetlb.h index 36fa6a2a82e3..1c6cde68487f 100644 --- a/include/linux/hugetlb.h +++ b/include/linux/hugetlb.h @@ -140,6 +140,8 @@ pte_t *huge_pte_alloc(struct mm_struct *mm, pte_t *huge_pte_offset(struct mm_struct *mm, unsigned long addr, unsigned long sz); int huge_pmd_unshare(struct mm_struct *mm, unsigned long *addr, pte_t *ptep); +bool huge_pmd_sharing_possible(struct vm_area_struct *vma, + unsigned long *start, unsigned long *end); struct page *follow_huge_addr(struct mm_struct *mm, unsigned long address, int write); struct page *follow_huge_pd(struct vm_area_struct *vma, @@ -170,6 +172,18 @@ static inline unsigned long hugetlb_total_pages(void) return 0; } +static inline int huge_pmd_unshare(struct mm_struct *mm, unsigned long *addr, + pte_t *ptep) +{ + return 0; +} + +bool huge_pmd_sharing_possible(struct vm_area_struct *vma, + unsigned long *start, unsigned long *end) +{ + return false; +} + #define follow_hugetlb_page(m,v,p,vs,a,b,i,w,n) ({ BUG(); 0; }) #define follow_huge_addr(mm, addr, write) ERR_PTR(-EINVAL) #define copy_hugetlb_page_range(src, dst, vma) ({ BUG(); 0; }) diff --git a/mm/hugetlb.c b/mm/hugetlb.c index 3103099f64fd..fd155dc52117 100644 --- a/mm/hugetlb.c +++ b/mm/hugetlb.c @@ -4555,6 +4555,9 @@ static bool vma_shareable(struct vm_area_struct *vma, unsigned long addr) /* * check on proper vm_flags and page table alignment + * + * Note that this is the same check used in huge_pmd_sharing_possible. + * If you change one, consider changing both. */ if (vma->vm_flags & VM_MAYSHARE && vma->vm_start <= base && end <= vma->vm_end) @@ -4562,6 +4565,43 @@ static bool vma_shareable(struct vm_area_struct *vma, unsigned long addr) return false; } +/* + * Determine if start,end range within vma could be mapped by shared pmd. + * If yes, adjust start and end to cover range associated with possible + * shared pmd mappings. + */ +bool huge_pmd_sharing_possible(struct vm_area_struct *vma, + unsigned long *start, unsigned long *end) +{ + unsigned long check_addr = *start; + bool ret = false; + + if (!(vma->vm_flags & VM_MAYSHARE)) + return ret; + + for (check_addr = *start; check_addr < *end; check_addr += PUD_SIZE) { + unsigned long a_start = check_addr & PUD_MASK; + unsigned long a_end = a_start + PUD_SIZE; + + /* + * If sharing is possible, adjust start/end if necessary. + * + * Note that this is the same check used in vma_shareable. If + * you change one, consider changing both. + */ + if (vma->vm_start <= a_start && a_end <= vma->vm_end) { + if (a_start < *start) + *start = a_start; + if (a_end > *end) + *end = a_end; + + ret = true; + } + } + + return ret; +} + /* * Search for a shareable pmd page for hugetlb. In any case calls pmd_alloc() * and returns the corresponding pte. While this is not necessary for the diff --git a/mm/rmap.c b/mm/rmap.c index eb477809a5c0..8cf853a4b093 100644 --- a/mm/rmap.c +++ b/mm/rmap.c @@ -1362,11 +1362,21 @@ static bool try_to_unmap_one(struct page *page, struct vm_area_struct *vma, } /* - * We have to assume the worse case ie pmd for invalidation. Note that - * the page can not be free in this function as call of try_to_unmap() - * must hold a reference on the page. + * For THP, we have to assume the worse case ie pmd for invalidation. + * For hugetlb, it could be much worse if we need to do pud + * invalidation in the case of pmd sharing. + * + * Note that the page can not be free in this function as call of + * try_to_unmap() must hold a reference on the page. */ end = min(vma->vm_end, start + (PAGE_SIZE << compound_order(page))); + if (PageHuge(page)) { + /* + * If sharing is possible, start and end will be adjusted + * accordingly. + */ + (void)huge_pmd_sharing_possible(vma, &start, &end); + } mmu_notifier_invalidate_range_start(vma->vm_mm, start, end); while (page_vma_mapped_walk(&pvmw)) { @@ -1409,6 +1419,32 @@ static bool try_to_unmap_one(struct page *page, struct vm_area_struct *vma, subpage = page - page_to_pfn(page) + pte_pfn(*pvmw.pte); address = pvmw.address; + if (PageHuge(page)) { + if (huge_pmd_unshare(mm, &address, pvmw.pte)) { + /* + * huge_pmd_unshare unmapped an entire PMD + * page. There is no way of knowing exactly + * which PMDs may be cached for this mm, so + * we must flush them all. start/end were + * already adjusted above to cover this range. + */ + flush_cache_range(vma, start, end); + flush_tlb_range(vma, start, end); + mmu_notifier_invalidate_range(mm, start, end); + + /* + * The ref count of the PMD page was dropped + * which is part of the way map counting + * is done for shared PMDs. Return 'true' + * here. When there is no other sharing, + * huge_pmd_unshare returns false and we will + * unmap the actual page and drop map count + * to zero. + */ + page_vma_mapped_walk_done(&pvmw); + break; + } + } if (IS_ENABLED(CONFIG_MIGRATION) && (flags & TTU_MIGRATION) && -- 2.17.1

7 years, 3 months

4
16
0 0

please revert commit ce8556cca6 "kbuild: verify that $DEPMOD is installed" introduced in v4.18.4.

by H. Nikolaus Schaller

This patch requires that /sbin/depmod is installed and installable on the build host. But not all build hosts for cross compiling Linux are Linux systems and are able to provide a working port of depmod, especially at the file patch /sbin/depmod. I use, for example, a Darwin system to cross compile Linux and I run depmod -a on the embedded system once, after installing a new Linux kernel there. I have no problem with seeing a warning, but aborting the build process is IMHO a bad idea since the previous behaviour didn't harm many people as far as I see. Probably 99% of people compiling Linux kernels do that on Linux and 99% of those have depmod installed for optimal operation of their build host. So IMHO printing the warning is good enough. BR and thanks, Nikolaus Schaller

7 years, 3 months

6
10
0 0

[PATCH 1/5] drm/nouveau: Check backlight IDs are >= 0, not > 0

by Lyude Paul

Remember, ida IDs start at 0, not 1! Signed-off-by: Lyude Paul <lyude(a)redhat.com> Cc: stable(a)vger.kernel.org Cc: Jeffery Miller <jmiller(a)neverware.com> Cc: Karol Herbst <kherbst(a)redhat.com> --- drivers/gpu/drm/nouveau/nouveau_backlight.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/nouveau/nouveau_backlight.c b/drivers/gpu/drm/nouveau/nouveau_backlight.c index 408b955e5c39..6dd72bc32897 100644 --- a/drivers/gpu/drm/nouveau/nouveau_backlight.c +++ b/drivers/gpu/drm/nouveau/nouveau_backlight.c @@ -116,7 +116,7 @@ nv40_backlight_init(struct drm_connector *connector) &nv40_bl_ops, &props); if (IS_ERR(bd)) { - if (bl_connector.id > 0) + if (bl_connector.id >= 0) ida_simple_remove(&bl_ida, bl_connector.id); return PTR_ERR(bd); } @@ -249,7 +249,7 @@ nv50_backlight_init(struct drm_connector *connector) nv_encoder, ops, &props); if (IS_ERR(bd)) { - if (bl_connector.id > 0) + if (bl_connector.id >= 0) ida_simple_remove(&bl_ida, bl_connector.id); return PTR_ERR(bd); } -- 2.17.1

7 years, 3 months

2
2
0 0

Re: [PATCH 4.17 015/324] smb3: increase initial number of credits requested to allow write

by Greg Kroah-Hartman

On Thu, Aug 23, 2018 at 09:38:31AM -0400, Scott French wrote: > Please remove the stfrench @ gmail address. I am not Steve Now removed, sorry for the noise. greg k-h

7 years, 3 months

1
0
0 0

[PATCH] powerpc/nohash: fix pte_access_permitted()

by Christophe Leroy

Commit 5769beaf180a8 ("powerpc/mm: Add proper pte access check helper for other platforms") replaced generic pte_access_permitted() by an arch specific one. The generic one is defined as (pte_present(pte) && (!(write) || pte_write(pte))) The arch specific one is open coded checking that _PAGE_USER and _PAGE_WRITE (_PAGE_RW) flags are set, but lacking to check that _PAGE_RO and _PAGE_PRIVILEGED are unset, leading to a useless test on targets like the 8xx which defines _PAGE_RW and _PAGE_USER as 0. Commit 5fa5b16be5b31 ("powerpc/mm/hugetlb: Use pte_access_permitted for hugetlb access check") replaced some tests performed with pte helpers by a call to pte_access_permitted(), leading to the same issue. This patch rewrites powerpc/nohash pte_access_permitted() using pte helpers. Fixes: 5769beaf180a8 ("powerpc/mm: Add proper pte access check helper for other platforms") Fixes: 5fa5b16be5b31 ("powerpc/mm/hugetlb: Use pte_access_permitted for hugetlb access check") Cc: stable(a)vger.kernel.org # v4.15+ Signed-off-by: Christophe Leroy <christophe.leroy(a)c-s.fr> --- arch/powerpc/include/asm/nohash/pgtable.h | 9 +++------ 1 file changed, 3 insertions(+), 6 deletions(-) diff --git a/arch/powerpc/include/asm/nohash/pgtable.h b/arch/powerpc/include/asm/nohash/pgtable.h index 2160be2e4339..b321c82b3624 100644 --- a/arch/powerpc/include/asm/nohash/pgtable.h +++ b/arch/powerpc/include/asm/nohash/pgtable.h @@ -51,17 +51,14 @@ static inline int pte_present(pte_t pte) #define pte_access_permitted pte_access_permitted static inline bool pte_access_permitted(pte_t pte, bool write) { - unsigned long pteval = pte_val(pte); /* * A read-only access is controlled by _PAGE_USER bit. * We have _PAGE_READ set for WRITE and EXECUTE */ - unsigned long need_pte_bits = _PAGE_PRESENT | _PAGE_USER; - - if (write) - need_pte_bits |= _PAGE_WRITE; + if (!pte_present(pte) || !pte_user(pte) || !pte_read(pte)) + return false; - if ((pteval & need_pte_bits) != need_pte_bits) + if (write && !pte_write(pte)) return false; return true; -- 2.13.3

7 years, 3 months

4
3
0 0

request for 4.4-stable: e7c0b5991dd1b ("ovl: warn instead of error if d_type is not supported")

by SZ Lin (林上智)

Hi Greg, This patch is not marked for 4.4-stable, but it's already in 4.9 and 4.14 stable. Please apply to 4.4-stable. This patch turned the status from error to warning if d_type is not supported, and thus operation won't be interrupted. -- SZ Lin (林上智)

7 years, 3 months

1
0
0 0

request for 4.4-stable: 21765194cecf2 ("ovl: Do d_type check only if work dir creation was successful")

by SZ Lin (林上智)

Hi Greg, This patch is not marked for 4.4-stable, but it's already in 4.9 and 4.14 stable. Please apply to 4.4-stable. This patch fixed check machanism for d_type to avoid returning d_type not supported even if underlying filesystem might be supporting it. -- SZ Lin (林上智)

7 years, 3 months

1
0
0 0

request for 4.4-stable: 45aebeaf4f674 ("ovl: Ensure upper filesystem supports d_type")

by SZ Lin (林上智)

Hi Greg, This patch is not marked for 4.4-stable, but it's already in 4.9 and 4.14 stable. Please apply to 4.4-stable. This patch added a check mechanism for d_type in upper layer of overlayfs to avoid whiteouts issue. -- SZ Lin (林上智)

7 years, 3 months

1
0
0 0

[PATCH] mm: respect arch_dup_mmap() return value

by Nadav Amit

Commit d70f2a14b72a4 ("include/linux/sched/mm.h: uninline mmdrop_async(), etc") ignored the return value of arch_dup_mmap(). As a result, on x86, a failure to duplicate the LDT (e.g., due to memory allocation error), would leave the duplicated memory mapping in an inconsistent state. Fix by regarding the return value, as it was before the change. Fixes: d70f2a14b72a4 ("include/linux/sched/mm.h: uninline mmdrop_async(), etc") Cc: Andrew Morton <akpm(a)linux-foundation.org> Cc: stable(a)vger.kernel.org Signed-off-by: Nadav Amit <namit(a)vmware.com> --- kernel/fork.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/kernel/fork.c b/kernel/fork.c index 1b27babc4c78..4527d1d331de 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -549,8 +549,7 @@ static __latent_entropy int dup_mmap(struct mm_struct *mm, goto out; } /* a new mm has just been created */ - arch_dup_mmap(oldmm, mm); - retval = 0; + retval = arch_dup_mmap(oldmm, mm); out: up_write(&mm->mmap_sem); flush_tlb_mm(oldmm); -- 2.17.1

7 years, 3 months

2
1
0 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

Linux-stable-mirror