From: "Mike Rapoport (IBM)" <rppt@xxxxxxxxxx> It is enough to use a file name to cross-reference another rst document. Jon says: The right things will happen in the HTML output, readers of the plain-text will know immediately where to go, and we don't have to add the label clutter. Drop reference markup and unnecessary labels and use plain file names. Signed-off-by: Mike Rapoport (IBM) <rppt@xxxxxxxxxx> --- Documentation/admin-guide/mm/concepts.rst | 13 +++++-------- Documentation/admin-guide/mm/hugetlbpage.rst | 4 +--- Documentation/admin-guide/mm/idle_page_tracking.rst | 7 ++----- Documentation/admin-guide/mm/index.rst | 3 +-- Documentation/admin-guide/mm/ksm.rst | 2 -- Documentation/admin-guide/mm/memory-hotplug.rst | 2 -- Documentation/admin-guide/mm/numa_memory_policy.rst | 2 -- Documentation/admin-guide/mm/numaperf.rst | 2 -- Documentation/admin-guide/mm/pagemap.rst | 11 ++++------- Documentation/admin-guide/mm/shrinker_debugfs.rst | 2 -- Documentation/admin-guide/mm/soft-dirty.rst | 2 -- Documentation/admin-guide/mm/swap_numa.rst | 2 -- Documentation/admin-guide/mm/transhuge.rst | 2 -- Documentation/admin-guide/mm/userfaultfd.rst | 2 -- Documentation/admin-guide/mm/zswap.rst | 2 -- .../translations/zh_CN/admin-guide/mm/index.rst | 2 +- 16 files changed, 14 insertions(+), 46 deletions(-) diff --git a/Documentation/admin-guide/mm/concepts.rst b/Documentation/admin-guide/mm/concepts.rst index c79f1e336222..e796b0a7e4a5 100644 --- a/Documentation/admin-guide/mm/concepts.rst +++ b/Documentation/admin-guide/mm/concepts.rst @@ -1,5 +1,3 @@ -.. _mm_concepts: - ================= Concepts overview ================= @@ -86,16 +84,15 @@ memory with the huge pages. The first one is `HugeTLB filesystem`, or hugetlbfs. It is a pseudo filesystem that uses RAM as its backing store. For the files created in this filesystem the data resides in the memory and mapped using huge pages. The hugetlbfs is described at -:ref:`Documentation/admin-guide/mm/hugetlbpage.rst <hugetlbpage>`. +Documentation/admin-guide/mm/hugetlbpage.rst. Another, more recent, mechanism that enables use of the huge pages is called `Transparent HugePages`, or THP. Unlike the hugetlbfs that requires users and/or system administrators to configure what parts of the system memory should and can be mapped by the huge pages, THP manages such mappings transparently to the user and hence the -name. See -:ref:`Documentation/admin-guide/mm/transhuge.rst <admin_guide_transhuge>` -for more details about THP. +name. See Documentation/admin-guide/mm/transhuge.rst for more details +about THP. Zones ===== @@ -125,8 +122,8 @@ processor. Each bank is referred to as a `node` and for each node Linux constructs an independent memory management subsystem. A node has its own set of zones, lists of free and used pages and various statistics counters. You can find more details about NUMA in -:ref:`Documentation/mm/numa.rst <numa>` and in -:ref:`Documentation/admin-guide/mm/numa_memory_policy.rst <numa_memory_policy>`. +Documentation/mm/numa.rst` and in +Documentation/admin-guide/mm/numa_memory_policy.rst. Page cache ========== diff --git a/Documentation/admin-guide/mm/hugetlbpage.rst b/Documentation/admin-guide/mm/hugetlbpage.rst index 19f27c0d92e0..5f9121b287d9 100644 --- a/Documentation/admin-guide/mm/hugetlbpage.rst +++ b/Documentation/admin-guide/mm/hugetlbpage.rst @@ -1,5 +1,3 @@ -.. _hugetlbpage: - ============= HugeTLB Pages ============= @@ -313,7 +311,7 @@ memory policy mode--bind, preferred, local or interleave--may be used. The resulting effect on persistent huge page allocation is as follows: #. Regardless of mempolicy mode [see - :ref:`Documentation/admin-guide/mm/numa_memory_policy.rst <numa_memory_policy>`], + Documentation/admin-guide/mm/numa_memory_policy.rst], persistent huge pages will be distributed across the node or nodes specified in the mempolicy as if "interleave" had been specified. However, if a node in the policy does not contain sufficient contiguous diff --git a/Documentation/admin-guide/mm/idle_page_tracking.rst b/Documentation/admin-guide/mm/idle_page_tracking.rst index df9394fb39c2..b5a285bd73fd 100644 --- a/Documentation/admin-guide/mm/idle_page_tracking.rst +++ b/Documentation/admin-guide/mm/idle_page_tracking.rst @@ -1,5 +1,3 @@ -.. _idle_page_tracking: - ================== Idle Page Tracking ================== @@ -70,9 +68,8 @@ If the tool is run initially with the appropriate option, it will mark all the queried pages as idle. Subsequent runs of the tool can then show which pages have their idle flag cleared in the interim. -See :ref:`Documentation/admin-guide/mm/pagemap.rst <pagemap>` for more -information about ``/proc/pid/pagemap``, ``/proc/kpageflags``, and -``/proc/kpagecgroup``. +See Documentation/admin-guide/mm/pagemap.rst for more information about +``/proc/pid/pagemap``, ``/proc/kpageflags``, and ``/proc/kpagecgroup``. .. _impl_details: diff --git a/Documentation/admin-guide/mm/index.rst b/Documentation/admin-guide/mm/index.rst index d1064e0ba34a..1f883abf3f00 100644 --- a/Documentation/admin-guide/mm/index.rst +++ b/Documentation/admin-guide/mm/index.rst @@ -16,8 +16,7 @@ are described in Documentation/admin-guide/sysctl/vm.rst and in `man 5 proc`_. .. _man 5 proc: http://man7.org/linux/man-pages/man5/proc.5.html Linux memory management has its own jargon and if you are not yet -familiar with it, consider reading -:ref:`Documentation/admin-guide/mm/concepts.rst <mm_concepts>`. +familiar with it, consider reading Documentation/admin-guide/mm/concepts.rst. Here we document in detail how to interact with various mechanisms in the Linux memory management. diff --git a/Documentation/admin-guide/mm/ksm.rst b/Documentation/admin-guide/mm/ksm.rst index fb6ba2002a4b..eed51a910c94 100644 --- a/Documentation/admin-guide/mm/ksm.rst +++ b/Documentation/admin-guide/mm/ksm.rst @@ -1,5 +1,3 @@ -.. _admin_guide_ksm: - ======================= Kernel Samepage Merging ======================= diff --git a/Documentation/admin-guide/mm/memory-hotplug.rst b/Documentation/admin-guide/mm/memory-hotplug.rst index a3c9e8ad8fa0..1b02fe5807cc 100644 --- a/Documentation/admin-guide/mm/memory-hotplug.rst +++ b/Documentation/admin-guide/mm/memory-hotplug.rst @@ -1,5 +1,3 @@ -.. _admin_guide_memory_hotplug: - ================== Memory Hot(Un)Plug ================== diff --git a/Documentation/admin-guide/mm/numa_memory_policy.rst b/Documentation/admin-guide/mm/numa_memory_policy.rst index 5a6afecbb0d0..f5cc2a368dac 100644 --- a/Documentation/admin-guide/mm/numa_memory_policy.rst +++ b/Documentation/admin-guide/mm/numa_memory_policy.rst @@ -1,5 +1,3 @@ -.. _numa_memory_policy: - ================== NUMA Memory Policy ================== diff --git a/Documentation/admin-guide/mm/numaperf.rst b/Documentation/admin-guide/mm/numaperf.rst index 166697325947..24e63e740420 100644 --- a/Documentation/admin-guide/mm/numaperf.rst +++ b/Documentation/admin-guide/mm/numaperf.rst @@ -1,5 +1,3 @@ -.. _numaperf: - ============= NUMA Locality ============= diff --git a/Documentation/admin-guide/mm/pagemap.rst b/Documentation/admin-guide/mm/pagemap.rst index 6e2e416af783..1a22674ab18e 100644 --- a/Documentation/admin-guide/mm/pagemap.rst +++ b/Documentation/admin-guide/mm/pagemap.rst @@ -1,5 +1,3 @@ -.. _pagemap: - ============================= Examining Process Page Tables ============================= @@ -19,10 +17,10 @@ There are four components to pagemap: * Bits 0-4 swap type if swapped * Bits 5-54 swap offset if swapped * Bit 55 pte is soft-dirty (see - :ref:`Documentation/admin-guide/mm/soft-dirty.rst <soft_dirty>`) + Documentation/admin-guide/mm/soft-dirty.rst) * Bit 56 page exclusively mapped (since 4.2) * Bit 57 pte is uffd-wp write-protected (since 5.13) (see - :ref:`Documentation/admin-guide/mm/userfaultfd.rst <userfaultfd>`) + Documentation/admin-guide/mm/userfaultfd.rst) * Bits 58-60 zero * Bit 61 page is file-page or shared-anon (since 3.5) * Bit 62 page swapped @@ -105,8 +103,7 @@ Short descriptions to the page flags A compound page with order N consists of 2^N physically contiguous pages. A compound page with order 2 takes the form of "HTTT", where H donates its head page and T donates its tail page(s). The major consumers of compound - pages are hugeTLB pages - (:ref:`Documentation/admin-guide/mm/hugetlbpage.rst <hugetlbpage>`), + pages are hugeTLB pages (Documentation/admin-guide/mm/hugetlbpage.rst), the SLUB etc. memory allocators and various device drivers. However in this interface, only huge/giga pages are made visible to end users. @@ -128,7 +125,7 @@ Short descriptions to the page flags Zero page for pfn_zero or huge_zero page. 25 - IDLE The page has not been accessed since it was marked idle (see - :ref:`Documentation/admin-guide/mm/idle_page_tracking.rst <idle_page_tracking>`). + Documentation/admin-guide/mm/idle_page_tracking.rst). Note that this flag may be stale in case the page was accessed via a PTE. To make sure the flag is up-to-date one has to read ``/sys/kernel/mm/page_idle/bitmap`` first. diff --git a/Documentation/admin-guide/mm/shrinker_debugfs.rst b/Documentation/admin-guide/mm/shrinker_debugfs.rst index 3887f0b294fe..c582033bd113 100644 --- a/Documentation/admin-guide/mm/shrinker_debugfs.rst +++ b/Documentation/admin-guide/mm/shrinker_debugfs.rst @@ -1,5 +1,3 @@ -.. _shrinker_debugfs: - ========================== Shrinker Debugfs Interface ========================== diff --git a/Documentation/admin-guide/mm/soft-dirty.rst b/Documentation/admin-guide/mm/soft-dirty.rst index cb0cfd6672fa..aeea936caa44 100644 --- a/Documentation/admin-guide/mm/soft-dirty.rst +++ b/Documentation/admin-guide/mm/soft-dirty.rst @@ -1,5 +1,3 @@ -.. _soft_dirty: - =============== Soft-Dirty PTEs =============== diff --git a/Documentation/admin-guide/mm/swap_numa.rst b/Documentation/admin-guide/mm/swap_numa.rst index e0466f2db8fa..2e630627bcee 100644 --- a/Documentation/admin-guide/mm/swap_numa.rst +++ b/Documentation/admin-guide/mm/swap_numa.rst @@ -1,5 +1,3 @@ -.. _swap_numa: - =========================================== Automatically bind swap device to numa node =========================================== diff --git a/Documentation/admin-guide/mm/transhuge.rst b/Documentation/admin-guide/mm/transhuge.rst index 8ee78ec232eb..b0cc8243e093 100644 --- a/Documentation/admin-guide/mm/transhuge.rst +++ b/Documentation/admin-guide/mm/transhuge.rst @@ -1,5 +1,3 @@ -.. _admin_guide_transhuge: - ============================ Transparent Hugepage Support ============================ diff --git a/Documentation/admin-guide/mm/userfaultfd.rst b/Documentation/admin-guide/mm/userfaultfd.rst index 83f31919ebb3..7dc823b56ca4 100644 --- a/Documentation/admin-guide/mm/userfaultfd.rst +++ b/Documentation/admin-guide/mm/userfaultfd.rst @@ -1,5 +1,3 @@ -.. _userfaultfd: - =========== Userfaultfd =========== diff --git a/Documentation/admin-guide/mm/zswap.rst b/Documentation/admin-guide/mm/zswap.rst index f67de481c7f6..fa5f6c9c78f3 100644 --- a/Documentation/admin-guide/mm/zswap.rst +++ b/Documentation/admin-guide/mm/zswap.rst @@ -1,5 +1,3 @@ -.. _zswap: - ===== zswap ===== diff --git a/Documentation/translations/zh_CN/admin-guide/mm/index.rst b/Documentation/translations/zh_CN/admin-guide/mm/index.rst index 702271c5b683..a8fd2c4a8796 100644 --- a/Documentation/translations/zh_CN/admin-guide/mm/index.rst +++ b/Documentation/translations/zh_CN/admin-guide/mm/index.rst @@ -22,7 +22,7 @@ Linux内存管理是一个具有许多可配置设置的复杂系统, 且这些 .. _man 5 proc: http://man7.org/linux/man-pages/man5/proc.5.html Linux内存管理有它自己的术语,如果你还不熟悉它,请考虑阅读下面参考: -:ref:`Documentation/admin-guide/mm/concepts.rst <mm_concepts>`. +Documentation/admin-guide/mm/concepts.rst. 在此目录下,我们详细描述了如何与Linux内存管理中的各种机制交互。 -- 2.35.1