linux/fs/file.c

1308 lines
31 KiB
C
Raw Normal View History

License cleanup: add SPDX GPL-2.0 license identifier to files with no license Many source files in the tree are missing licensing information, which makes it harder for compliance tools to determine the correct license. By default all files without license information are under the default license of the kernel, which is GPL version 2. Update the files which contain no license information with the 'GPL-2.0' SPDX license identifier. The SPDX identifier is a legally binding shorthand, which can be used instead of the full boiler plate text. This patch is based on work done by Thomas Gleixner and Kate Stewart and Philippe Ombredanne. How this work was done: Patches were generated and checked against linux-4.14-rc6 for a subset of the use cases: - file had no licensing information it it. - file was a */uapi/* one with no licensing information in it, - file was a */uapi/* one with existing licensing information, Further patches will be generated in subsequent months to fix up cases where non-standard license headers were used, and references to license had to be inferred by heuristics based on keywords. The analysis to determine which SPDX License Identifier to be applied to a file was done in a spreadsheet of side by side results from of the output of two independent scanners (ScanCode & Windriver) producing SPDX tag:value files created by Philippe Ombredanne. Philippe prepared the base worksheet, and did an initial spot review of a few 1000 files. The 4.13 kernel was the starting point of the analysis with 60,537 files assessed. Kate Stewart did a file by file comparison of the scanner results in the spreadsheet to determine which SPDX license identifier(s) to be applied to the file. She confirmed any determination that was not immediately clear with lawyers working with the Linux Foundation. Criteria used to select files for SPDX license identifier tagging was: - Files considered eligible had to be source code files. - Make and config files were included as candidates if they contained >5 lines of source - File already had some variant of a license header in it (even if <5 lines). All documentation files were explicitly excluded. The following heuristics were used to determine which SPDX license identifiers to apply. - when both scanners couldn't find any license traces, file was considered to have no license information in it, and the top level COPYING file license applied. For non */uapi/* files that summary was: SPDX license identifier # files ---------------------------------------------------|------- GPL-2.0 11139 and resulted in the first patch in this series. If that file was a */uapi/* path one, it was "GPL-2.0 WITH Linux-syscall-note" otherwise it was "GPL-2.0". Results of that was: SPDX license identifier # files ---------------------------------------------------|------- GPL-2.0 WITH Linux-syscall-note 930 and resulted in the second patch in this series. - if a file had some form of licensing information in it, and was one of the */uapi/* ones, it was denoted with the Linux-syscall-note if any GPL family license was found in the file or had no licensing in it (per prior point). Results summary: SPDX license identifier # files ---------------------------------------------------|------ GPL-2.0 WITH Linux-syscall-note 270 GPL-2.0+ WITH Linux-syscall-note 169 ((GPL-2.0 WITH Linux-syscall-note) OR BSD-2-Clause) 21 ((GPL-2.0 WITH Linux-syscall-note) OR BSD-3-Clause) 17 LGPL-2.1+ WITH Linux-syscall-note 15 GPL-1.0+ WITH Linux-syscall-note 14 ((GPL-2.0+ WITH Linux-syscall-note) OR BSD-3-Clause) 5 LGPL-2.0+ WITH Linux-syscall-note 4 LGPL-2.1 WITH Linux-syscall-note 3 ((GPL-2.0 WITH Linux-syscall-note) OR MIT) 3 ((GPL-2.0 WITH Linux-syscall-note) AND MIT) 1 and that resulted in the third patch in this series. - when the two scanners agreed on the detected license(s), that became the concluded license(s). - when there was disagreement between the two scanners (one detected a license but the other didn't, or they both detected different licenses) a manual inspection of the file occurred. - In most cases a manual inspection of the information in the file resulted in a clear resolution of the license that should apply (and which scanner probably needed to revisit its heuristics). - When it was not immediately clear, the license identifier was confirmed with lawyers working with the Linux Foundation. - If there was any question as to the appropriate license identifier, the file was flagged for further research and to be revisited later in time. In total, over 70 hours of logged manual review was done on the spreadsheet to determine the SPDX license identifiers to apply to the source files by Kate, Philippe, Thomas and, in some cases, confirmation by lawyers working with the Linux Foundation. Kate also obtained a third independent scan of the 4.13 code base from FOSSology, and compared selected files where the other two scanners disagreed against that SPDX file, to see if there was new insights. The Windriver scanner is based on an older version of FOSSology in part, so they are related. Thomas did random spot checks in about 500 files from the spreadsheets for the uapi headers and agreed with SPDX license identifier in the files he inspected. For the non-uapi files Thomas did random spot checks in about 15000 files. In initial set of patches against 4.14-rc6, 3 files were found to have copy/paste license identifier errors, and have been fixed to reflect the correct identifier. Additionally Philippe spent 10 hours this week doing a detailed manual inspection and review of the 12,461 patched files from the initial patch version early this week with: - a full scancode scan run, collecting the matched texts, detected license ids and scores - reviewing anything where there was a license detected (about 500+ files) to ensure that the applied SPDX license was correct - reviewing anything where there was no detection but the patch license was not GPL-2.0 WITH Linux-syscall-note to ensure that the applied SPDX license was correct This produced a worksheet with 20 files needing minor correction. This worksheet was then exported into 3 different .csv files for the different types of files to be modified. These .csv files were then reviewed by Greg. Thomas wrote a script to parse the csv files and add the proper SPDX tag to the file, in the format that the file expected. This script was further refined by Greg based on the output to detect more types of files automatically and to distinguish between header and source .c files (which need different comment types.) Finally Greg ran the script using the .csv files to generate the patches. Reviewed-by: Kate Stewart <kstewart@linuxfoundation.org> Reviewed-by: Philippe Ombredanne <pombredanne@nexb.com> Reviewed-by: Thomas Gleixner <tglx@linutronix.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2017-11-01 17:07:57 +03:00
// SPDX-License-Identifier: GPL-2.0
/*
* linux/fs/file.c
*
* Copyright (C) 1998-1999, Stephen Tweedie and Bill Hawes
*
* Manage the dynamic fd arrays in the process files_struct.
*/
#include <linux/syscalls.h>
#include <linux/export.h>
#include <linux/fs.h>
open: add close_range() This adds the close_range() syscall. It allows to efficiently close a range of file descriptors up to all file descriptors of a calling task. I was contacted by FreeBSD as they wanted to have the same close_range() syscall as we proposed here. We've coordinated this and in the meantime, Kyle was fast enough to merge close_range() into FreeBSD already in April: https://reviews.freebsd.org/D21627 https://svnweb.freebsd.org/base?view=revision&revision=359836 and the current plan is to backport close_range() to FreeBSD 12.2 (cf. [2]) once its merged in Linux too. Python is in the process of switching to close_range() on FreeBSD and they are waiting on us to merge this to switch on Linux as well: https://bugs.python.org/issue38061 The syscall came up in a recent discussion around the new mount API and making new file descriptor types cloexec by default. During this discussion, Al suggested the close_range() syscall (cf. [1]). Note, a syscall in this manner has been requested by various people over time. First, it helps to close all file descriptors of an exec()ing task. This can be done safely via (quoting Al's example from [1] verbatim): /* that exec is sensitive */ unshare(CLONE_FILES); /* we don't want anything past stderr here */ close_range(3, ~0U); execve(....); The code snippet above is one way of working around the problem that file descriptors are not cloexec by default. This is aggravated by the fact that we can't just switch them over without massively regressing userspace. For a whole class of programs having an in-kernel method of closing all file descriptors is very helpful (e.g. demons, service managers, programming language standard libraries, container managers etc.). (Please note, unshare(CLONE_FILES) should only be needed if the calling task is multi-threaded and shares the file descriptor table with another thread in which case two threads could race with one thread allocating file descriptors and the other one closing them via close_range(). For the general case close_range() before the execve() is sufficient.) Second, it allows userspace to avoid implementing closing all file descriptors by parsing through /proc/<pid>/fd/* and calling close() on each file descriptor. From looking at various large(ish) userspace code bases this or similar patterns are very common in: - service managers (cf. [4]) - libcs (cf. [6]) - container runtimes (cf. [5]) - programming language runtimes/standard libraries - Python (cf. [2]) - Rust (cf. [7], [8]) As Dmitry pointed out there's even a long-standing glibc bug about missing kernel support for this task (cf. [3]). In addition, the syscall will also work for tasks that do not have procfs mounted and on kernels that do not have procfs support compiled in. In such situations the only way to make sure that all file descriptors are closed is to call close() on each file descriptor up to UINT_MAX or RLIMIT_NOFILE, OPEN_MAX trickery (cf. comment [8] on Rust). The performance is striking. For good measure, comparing the following simple close_all_fds() userspace implementation that is essentially just glibc's version in [6]: static int close_all_fds(void) { int dir_fd; DIR *dir; struct dirent *direntp; dir = opendir("/proc/self/fd"); if (!dir) return -1; dir_fd = dirfd(dir); while ((direntp = readdir(dir))) { int fd; if (strcmp(direntp->d_name, ".") == 0) continue; if (strcmp(direntp->d_name, "..") == 0) continue; fd = atoi(direntp->d_name); if (fd == dir_fd || fd == 0 || fd == 1 || fd == 2) continue; close(fd); } closedir(dir); return 0; } to close_range() yields: 1. closing 4 open files: - close_all_fds(): ~280 us - close_range(): ~24 us 2. closing 1000 open files: - close_all_fds(): ~5000 us - close_range(): ~800 us close_range() is designed to allow for some flexibility. Specifically, it does not simply always close all open file descriptors of a task. Instead, callers can specify an upper bound. This is e.g. useful for scenarios where specific file descriptors are created with well-known numbers that are supposed to be excluded from getting closed. For extra paranoia close_range() comes with a flags argument. This can e.g. be used to implement extension. Once can imagine userspace wanting to stop at the first error instead of ignoring errors under certain circumstances. There might be other valid ideas in the future. In any case, a flag argument doesn't hurt and keeps us on the safe side. From an implementation side this is kept rather dumb. It saw some input from David and Jann but all nonsense is obviously my own! - Errors to close file descriptors are currently ignored. (Could be changed by setting a flag in the future if needed.) - __close_range() is a rather simplistic wrapper around __close_fd(). My reasoning behind this is based on the nature of how __close_fd() needs to release an fd. But maybe I misunderstood specifics: We take the files_lock and rcu-dereference the fdtable of the calling task, we find the entry in the fdtable, get the file and need to release files_lock before calling filp_close(). In the meantime the fdtable might have been altered so we can't just retake the spinlock and keep the old rcu-reference of the fdtable around. Instead we need to grab a fresh reference to the fdtable. If my reasoning is correct then there's really no point in fancyfying __close_range(): We just need to rcu-dereference the fdtable of the calling task once to cap the max_fd value correctly and then go on calling __close_fd() in a loop. /* References */ [1]: https://lore.kernel.org/lkml/20190516165021.GD17978@ZenIV.linux.org.uk/ [2]: https://github.com/python/cpython/blob/9e4f2f3a6b8ee995c365e86d976937c141d867f8/Modules/_posixsubprocess.c#L220 [3]: https://sourceware.org/bugzilla/show_bug.cgi?id=10353#c7 [4]: https://github.com/systemd/systemd/blob/5238e9575906297608ff802a27e2ff9effa3b338/src/basic/fd-util.c#L217 [5]: https://github.com/lxc/lxc/blob/ddf4b77e11a4d08f09b7b9cd13e593f8c047edc5/src/lxc/start.c#L236 [6]: https://sourceware.org/git/?p=glibc.git;a=blob;f=sysdeps/unix/sysv/linux/grantpt.c;h=2030e07fa6e652aac32c775b8c6e005844c3c4eb;hb=HEAD#l17 Note that this is an internal implementation that is not exported. Currently, libc seems to not provide an exported version of this because of missing kernel support to do this. Note, in a recent patch series Florian made grantpt() a nop thereby removing the code referenced here. [7]: https://github.com/rust-lang/rust/issues/12148 [8]: https://github.com/rust-lang/rust/blob/5f47c0613ed4eb46fca3633c1297364c09e5e451/src/libstd/sys/unix/process2.rs#L303-L308 Rust's solution is slightly different but is equally unperformant. Rust calls getdtablesize() which is a glibc library function that simply returns the current RLIMIT_NOFILE or OPEN_MAX values. Rust then goes on to call close() on each fd. That's obviously overkill for most tasks. Rarely, tasks - especially non-demons - hit RLIMIT_NOFILE or OPEN_MAX. Let's be nice and assume an unprivileged user with RLIMIT_NOFILE set to 1024. Even in this case, there's a very high chance that in the common case Rust is calling the close() syscall 1021 times pointlessly if the task just has 0, 1, and 2 open. Suggested-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com> Cc: Arnd Bergmann <arnd@arndb.de> Cc: Kyle Evans <self@kyle-evans.net> Cc: Jann Horn <jannh@google.com> Cc: David Howells <dhowells@redhat.com> Cc: Dmitry V. Levin <ldv@altlinux.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Florian Weimer <fweimer@redhat.com> Cc: linux-api@vger.kernel.org
2019-05-24 12:30:34 +03:00
#include <linux/kernel.h>
#include <linux/mm.h>
#include <linux/sched/signal.h>
#include <linux/slab.h>
#include <linux/file.h>
#include <linux/fdtable.h>
#include <linux/bitops.h>
#include <linux/spinlock.h>
#include <linux/rcupdate.h>
#include <linux/close_range.h>
#include <net/sock.h>
#include "internal.h"
unsigned int sysctl_nr_open __read_mostly = 1024*1024;
unsigned int sysctl_nr_open_min = BITS_PER_LONG;
/* our min() is unusable in constant expressions ;-/ */
#define __const_min(x, y) ((x) < (y) ? (x) : (y))
unsigned int sysctl_nr_open_max =
__const_min(INT_MAX, ~(size_t)0/sizeof(void *)) & -BITS_PER_LONG;
static void __free_fdtable(struct fdtable *fdt)
{
kvfree(fdt->fd);
kvfree(fdt->open_fds);
kfree(fdt);
}
static void free_fdtable_rcu(struct rcu_head *rcu)
{
__free_fdtable(container_of(rcu, struct fdtable, rcu));
}
#define BITBIT_NR(nr) BITS_TO_LONGS(BITS_TO_LONGS(nr))
#define BITBIT_SIZE(nr) (BITBIT_NR(nr) * sizeof(long))
/*
* Copy 'count' fd bits from the old table to the new table and clear the extra
* space if any. This does not copy the file pointers. Called with the files
* spinlock held for write.
*/
static void copy_fd_bitmaps(struct fdtable *nfdt, struct fdtable *ofdt,
unsigned int count)
{
unsigned int cpy, set;
cpy = count / BITS_PER_BYTE;
set = (nfdt->max_fds - count) / BITS_PER_BYTE;
memcpy(nfdt->open_fds, ofdt->open_fds, cpy);
memset((char *)nfdt->open_fds + cpy, 0, set);
memcpy(nfdt->close_on_exec, ofdt->close_on_exec, cpy);
memset((char *)nfdt->close_on_exec + cpy, 0, set);
cpy = BITBIT_SIZE(count);
set = BITBIT_SIZE(nfdt->max_fds) - cpy;
memcpy(nfdt->full_fds_bits, ofdt->full_fds_bits, cpy);
memset((char *)nfdt->full_fds_bits + cpy, 0, set);
}
/*
* Copy all file descriptors from the old table to the new, expanded table and
* clear the extra space. Called with the files spinlock held for write.
*/
[PATCH] fdtable: Implement new pagesize-based fdtable allocator This patch provides an improved fdtable allocation scheme, useful for expanding fdtable file descriptor entries. The main focus is on the fdarray, as its memory usage grows 128 times faster than that of an fdset. The allocation algorithm sizes the fdarray in such a way that its memory usage increases in easy page-sized chunks. The overall algorithm expands the allowed size in powers of two, in order to amortize the cost of invoking vmalloc() for larger allocation sizes. Namely, the following sizes for the fdarray are considered, and the smallest that accommodates the requested fd count is chosen: pagesize / 4 pagesize / 2 pagesize <- memory allocator switch point pagesize * 2 pagesize * 4 ...etc... Unlike the current implementation, this allocation scheme does not require a loop to compute the optimal fdarray size, and can be done in efficient straightline code. Furthermore, since the fdarray overflows the pagesize boundary long before any of the fdsets do, it makes sense to optimize run-time by allocating both fdsets in a single swoop. Even together, they will still be, by far, smaller than the fdarray. The fdtable->open_fds is now used as the anchor for the fdset memory allocation. Signed-off-by: Vadim Lobanov <vlobanov@speakeasy.net> Cc: Christoph Hellwig <hch@lst.de> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Dipankar Sarma <dipankar@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 13:21:22 +03:00
static void copy_fdtable(struct fdtable *nfdt, struct fdtable *ofdt)
{
size_t cpy, set;
[PATCH] fdtable: Implement new pagesize-based fdtable allocator This patch provides an improved fdtable allocation scheme, useful for expanding fdtable file descriptor entries. The main focus is on the fdarray, as its memory usage grows 128 times faster than that of an fdset. The allocation algorithm sizes the fdarray in such a way that its memory usage increases in easy page-sized chunks. The overall algorithm expands the allowed size in powers of two, in order to amortize the cost of invoking vmalloc() for larger allocation sizes. Namely, the following sizes for the fdarray are considered, and the smallest that accommodates the requested fd count is chosen: pagesize / 4 pagesize / 2 pagesize <- memory allocator switch point pagesize * 2 pagesize * 4 ...etc... Unlike the current implementation, this allocation scheme does not require a loop to compute the optimal fdarray size, and can be done in efficient straightline code. Furthermore, since the fdarray overflows the pagesize boundary long before any of the fdsets do, it makes sense to optimize run-time by allocating both fdsets in a single swoop. Even together, they will still be, by far, smaller than the fdarray. The fdtable->open_fds is now used as the anchor for the fdset memory allocation. Signed-off-by: Vadim Lobanov <vlobanov@speakeasy.net> Cc: Christoph Hellwig <hch@lst.de> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Dipankar Sarma <dipankar@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 13:21:22 +03:00
BUG_ON(nfdt->max_fds < ofdt->max_fds);
cpy = ofdt->max_fds * sizeof(struct file *);
set = (nfdt->max_fds - ofdt->max_fds) * sizeof(struct file *);
memcpy(nfdt->fd, ofdt->fd, cpy);
memset((char *)nfdt->fd + cpy, 0, set);
[PATCH] fdtable: Implement new pagesize-based fdtable allocator This patch provides an improved fdtable allocation scheme, useful for expanding fdtable file descriptor entries. The main focus is on the fdarray, as its memory usage grows 128 times faster than that of an fdset. The allocation algorithm sizes the fdarray in such a way that its memory usage increases in easy page-sized chunks. The overall algorithm expands the allowed size in powers of two, in order to amortize the cost of invoking vmalloc() for larger allocation sizes. Namely, the following sizes for the fdarray are considered, and the smallest that accommodates the requested fd count is chosen: pagesize / 4 pagesize / 2 pagesize <- memory allocator switch point pagesize * 2 pagesize * 4 ...etc... Unlike the current implementation, this allocation scheme does not require a loop to compute the optimal fdarray size, and can be done in efficient straightline code. Furthermore, since the fdarray overflows the pagesize boundary long before any of the fdsets do, it makes sense to optimize run-time by allocating both fdsets in a single swoop. Even together, they will still be, by far, smaller than the fdarray. The fdtable->open_fds is now used as the anchor for the fdset memory allocation. Signed-off-by: Vadim Lobanov <vlobanov@speakeasy.net> Cc: Christoph Hellwig <hch@lst.de> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Dipankar Sarma <dipankar@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 13:21:22 +03:00
copy_fd_bitmaps(nfdt, ofdt, ofdt->max_fds);
}
[PATCH] fdtable: Implement new pagesize-based fdtable allocator This patch provides an improved fdtable allocation scheme, useful for expanding fdtable file descriptor entries. The main focus is on the fdarray, as its memory usage grows 128 times faster than that of an fdset. The allocation algorithm sizes the fdarray in such a way that its memory usage increases in easy page-sized chunks. The overall algorithm expands the allowed size in powers of two, in order to amortize the cost of invoking vmalloc() for larger allocation sizes. Namely, the following sizes for the fdarray are considered, and the smallest that accommodates the requested fd count is chosen: pagesize / 4 pagesize / 2 pagesize <- memory allocator switch point pagesize * 2 pagesize * 4 ...etc... Unlike the current implementation, this allocation scheme does not require a loop to compute the optimal fdarray size, and can be done in efficient straightline code. Furthermore, since the fdarray overflows the pagesize boundary long before any of the fdsets do, it makes sense to optimize run-time by allocating both fdsets in a single swoop. Even together, they will still be, by far, smaller than the fdarray. The fdtable->open_fds is now used as the anchor for the fdset memory allocation. Signed-off-by: Vadim Lobanov <vlobanov@speakeasy.net> Cc: Christoph Hellwig <hch@lst.de> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Dipankar Sarma <dipankar@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 13:21:22 +03:00
static struct fdtable * alloc_fdtable(unsigned int nr)
{
[PATCH] fdtable: Implement new pagesize-based fdtable allocator This patch provides an improved fdtable allocation scheme, useful for expanding fdtable file descriptor entries. The main focus is on the fdarray, as its memory usage grows 128 times faster than that of an fdset. The allocation algorithm sizes the fdarray in such a way that its memory usage increases in easy page-sized chunks. The overall algorithm expands the allowed size in powers of two, in order to amortize the cost of invoking vmalloc() for larger allocation sizes. Namely, the following sizes for the fdarray are considered, and the smallest that accommodates the requested fd count is chosen: pagesize / 4 pagesize / 2 pagesize <- memory allocator switch point pagesize * 2 pagesize * 4 ...etc... Unlike the current implementation, this allocation scheme does not require a loop to compute the optimal fdarray size, and can be done in efficient straightline code. Furthermore, since the fdarray overflows the pagesize boundary long before any of the fdsets do, it makes sense to optimize run-time by allocating both fdsets in a single swoop. Even together, they will still be, by far, smaller than the fdarray. The fdtable->open_fds is now used as the anchor for the fdset memory allocation. Signed-off-by: Vadim Lobanov <vlobanov@speakeasy.net> Cc: Christoph Hellwig <hch@lst.de> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Dipankar Sarma <dipankar@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 13:21:22 +03:00
struct fdtable *fdt;
void *data;
/*
[PATCH] fdtable: Implement new pagesize-based fdtable allocator This patch provides an improved fdtable allocation scheme, useful for expanding fdtable file descriptor entries. The main focus is on the fdarray, as its memory usage grows 128 times faster than that of an fdset. The allocation algorithm sizes the fdarray in such a way that its memory usage increases in easy page-sized chunks. The overall algorithm expands the allowed size in powers of two, in order to amortize the cost of invoking vmalloc() for larger allocation sizes. Namely, the following sizes for the fdarray are considered, and the smallest that accommodates the requested fd count is chosen: pagesize / 4 pagesize / 2 pagesize <- memory allocator switch point pagesize * 2 pagesize * 4 ...etc... Unlike the current implementation, this allocation scheme does not require a loop to compute the optimal fdarray size, and can be done in efficient straightline code. Furthermore, since the fdarray overflows the pagesize boundary long before any of the fdsets do, it makes sense to optimize run-time by allocating both fdsets in a single swoop. Even together, they will still be, by far, smaller than the fdarray. The fdtable->open_fds is now used as the anchor for the fdset memory allocation. Signed-off-by: Vadim Lobanov <vlobanov@speakeasy.net> Cc: Christoph Hellwig <hch@lst.de> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Dipankar Sarma <dipankar@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 13:21:22 +03:00
* Figure out how many fds we actually want to support in this fdtable.
* Allocation steps are keyed to the size of the fdarray, since it
* grows far faster than any of the other dynamic data. We try to fit
* the fdarray into comfortable page-tuned chunks: starting at 1024B
* and growing in powers of two from there on.
*/
[PATCH] fdtable: Implement new pagesize-based fdtable allocator This patch provides an improved fdtable allocation scheme, useful for expanding fdtable file descriptor entries. The main focus is on the fdarray, as its memory usage grows 128 times faster than that of an fdset. The allocation algorithm sizes the fdarray in such a way that its memory usage increases in easy page-sized chunks. The overall algorithm expands the allowed size in powers of two, in order to amortize the cost of invoking vmalloc() for larger allocation sizes. Namely, the following sizes for the fdarray are considered, and the smallest that accommodates the requested fd count is chosen: pagesize / 4 pagesize / 2 pagesize <- memory allocator switch point pagesize * 2 pagesize * 4 ...etc... Unlike the current implementation, this allocation scheme does not require a loop to compute the optimal fdarray size, and can be done in efficient straightline code. Furthermore, since the fdarray overflows the pagesize boundary long before any of the fdsets do, it makes sense to optimize run-time by allocating both fdsets in a single swoop. Even together, they will still be, by far, smaller than the fdarray. The fdtable->open_fds is now used as the anchor for the fdset memory allocation. Signed-off-by: Vadim Lobanov <vlobanov@speakeasy.net> Cc: Christoph Hellwig <hch@lst.de> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Dipankar Sarma <dipankar@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 13:21:22 +03:00
nr /= (1024 / sizeof(struct file *));
nr = roundup_pow_of_two(nr + 1);
nr *= (1024 / sizeof(struct file *));
/*
* Note that this can drive nr *below* what we had passed if sysctl_nr_open
* had been set lower between the check in expand_files() and here. Deal
* with that in caller, it's cheaper that way.
*
* We make sure that nr remains a multiple of BITS_PER_LONG - otherwise
* bitmaps handling below becomes unpleasant, to put it mildly...
*/
if (unlikely(nr > sysctl_nr_open))
nr = ((sysctl_nr_open - 1) | (BITS_PER_LONG - 1)) + 1;
2016-01-15 02:18:21 +03:00
fdt = kmalloc(sizeof(struct fdtable), GFP_KERNEL_ACCOUNT);
[PATCH] fdtable: Implement new pagesize-based fdtable allocator This patch provides an improved fdtable allocation scheme, useful for expanding fdtable file descriptor entries. The main focus is on the fdarray, as its memory usage grows 128 times faster than that of an fdset. The allocation algorithm sizes the fdarray in such a way that its memory usage increases in easy page-sized chunks. The overall algorithm expands the allowed size in powers of two, in order to amortize the cost of invoking vmalloc() for larger allocation sizes. Namely, the following sizes for the fdarray are considered, and the smallest that accommodates the requested fd count is chosen: pagesize / 4 pagesize / 2 pagesize <- memory allocator switch point pagesize * 2 pagesize * 4 ...etc... Unlike the current implementation, this allocation scheme does not require a loop to compute the optimal fdarray size, and can be done in efficient straightline code. Furthermore, since the fdarray overflows the pagesize boundary long before any of the fdsets do, it makes sense to optimize run-time by allocating both fdsets in a single swoop. Even together, they will still be, by far, smaller than the fdarray. The fdtable->open_fds is now used as the anchor for the fdset memory allocation. Signed-off-by: Vadim Lobanov <vlobanov@speakeasy.net> Cc: Christoph Hellwig <hch@lst.de> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Dipankar Sarma <dipankar@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 13:21:22 +03:00
if (!fdt)
goto out;
[PATCH] fdtable: Implement new pagesize-based fdtable allocator This patch provides an improved fdtable allocation scheme, useful for expanding fdtable file descriptor entries. The main focus is on the fdarray, as its memory usage grows 128 times faster than that of an fdset. The allocation algorithm sizes the fdarray in such a way that its memory usage increases in easy page-sized chunks. The overall algorithm expands the allowed size in powers of two, in order to amortize the cost of invoking vmalloc() for larger allocation sizes. Namely, the following sizes for the fdarray are considered, and the smallest that accommodates the requested fd count is chosen: pagesize / 4 pagesize / 2 pagesize <- memory allocator switch point pagesize * 2 pagesize * 4 ...etc... Unlike the current implementation, this allocation scheme does not require a loop to compute the optimal fdarray size, and can be done in efficient straightline code. Furthermore, since the fdarray overflows the pagesize boundary long before any of the fdsets do, it makes sense to optimize run-time by allocating both fdsets in a single swoop. Even together, they will still be, by far, smaller than the fdarray. The fdtable->open_fds is now used as the anchor for the fdset memory allocation. Signed-off-by: Vadim Lobanov <vlobanov@speakeasy.net> Cc: Christoph Hellwig <hch@lst.de> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Dipankar Sarma <dipankar@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 13:21:22 +03:00
fdt->max_fds = nr;
data = kvmalloc_array(nr, sizeof(struct file *), GFP_KERNEL_ACCOUNT);
[PATCH] fdtable: Implement new pagesize-based fdtable allocator This patch provides an improved fdtable allocation scheme, useful for expanding fdtable file descriptor entries. The main focus is on the fdarray, as its memory usage grows 128 times faster than that of an fdset. The allocation algorithm sizes the fdarray in such a way that its memory usage increases in easy page-sized chunks. The overall algorithm expands the allowed size in powers of two, in order to amortize the cost of invoking vmalloc() for larger allocation sizes. Namely, the following sizes for the fdarray are considered, and the smallest that accommodates the requested fd count is chosen: pagesize / 4 pagesize / 2 pagesize <- memory allocator switch point pagesize * 2 pagesize * 4 ...etc... Unlike the current implementation, this allocation scheme does not require a loop to compute the optimal fdarray size, and can be done in efficient straightline code. Furthermore, since the fdarray overflows the pagesize boundary long before any of the fdsets do, it makes sense to optimize run-time by allocating both fdsets in a single swoop. Even together, they will still be, by far, smaller than the fdarray. The fdtable->open_fds is now used as the anchor for the fdset memory allocation. Signed-off-by: Vadim Lobanov <vlobanov@speakeasy.net> Cc: Christoph Hellwig <hch@lst.de> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Dipankar Sarma <dipankar@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 13:21:22 +03:00
if (!data)
goto out_fdt;
fdt->fd = data;
data = kvmalloc(max_t(size_t,
2 * nr / BITS_PER_BYTE + BITBIT_SIZE(nr), L1_CACHE_BYTES),
GFP_KERNEL_ACCOUNT);
[PATCH] fdtable: Implement new pagesize-based fdtable allocator This patch provides an improved fdtable allocation scheme, useful for expanding fdtable file descriptor entries. The main focus is on the fdarray, as its memory usage grows 128 times faster than that of an fdset. The allocation algorithm sizes the fdarray in such a way that its memory usage increases in easy page-sized chunks. The overall algorithm expands the allowed size in powers of two, in order to amortize the cost of invoking vmalloc() for larger allocation sizes. Namely, the following sizes for the fdarray are considered, and the smallest that accommodates the requested fd count is chosen: pagesize / 4 pagesize / 2 pagesize <- memory allocator switch point pagesize * 2 pagesize * 4 ...etc... Unlike the current implementation, this allocation scheme does not require a loop to compute the optimal fdarray size, and can be done in efficient straightline code. Furthermore, since the fdarray overflows the pagesize boundary long before any of the fdsets do, it makes sense to optimize run-time by allocating both fdsets in a single swoop. Even together, they will still be, by far, smaller than the fdarray. The fdtable->open_fds is now used as the anchor for the fdset memory allocation. Signed-off-by: Vadim Lobanov <vlobanov@speakeasy.net> Cc: Christoph Hellwig <hch@lst.de> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Dipankar Sarma <dipankar@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 13:21:22 +03:00
if (!data)
goto out_arr;
fdt->open_fds = data;
[PATCH] fdtable: Implement new pagesize-based fdtable allocator This patch provides an improved fdtable allocation scheme, useful for expanding fdtable file descriptor entries. The main focus is on the fdarray, as its memory usage grows 128 times faster than that of an fdset. The allocation algorithm sizes the fdarray in such a way that its memory usage increases in easy page-sized chunks. The overall algorithm expands the allowed size in powers of two, in order to amortize the cost of invoking vmalloc() for larger allocation sizes. Namely, the following sizes for the fdarray are considered, and the smallest that accommodates the requested fd count is chosen: pagesize / 4 pagesize / 2 pagesize <- memory allocator switch point pagesize * 2 pagesize * 4 ...etc... Unlike the current implementation, this allocation scheme does not require a loop to compute the optimal fdarray size, and can be done in efficient straightline code. Furthermore, since the fdarray overflows the pagesize boundary long before any of the fdsets do, it makes sense to optimize run-time by allocating both fdsets in a single swoop. Even together, they will still be, by far, smaller than the fdarray. The fdtable->open_fds is now used as the anchor for the fdset memory allocation. Signed-off-by: Vadim Lobanov <vlobanov@speakeasy.net> Cc: Christoph Hellwig <hch@lst.de> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Dipankar Sarma <dipankar@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 13:21:22 +03:00
data += nr / BITS_PER_BYTE;
fdt->close_on_exec = data;
data += nr / BITS_PER_BYTE;
fdt->full_fds_bits = data;
[PATCH] fdtable: Implement new pagesize-based fdtable allocator This patch provides an improved fdtable allocation scheme, useful for expanding fdtable file descriptor entries. The main focus is on the fdarray, as its memory usage grows 128 times faster than that of an fdset. The allocation algorithm sizes the fdarray in such a way that its memory usage increases in easy page-sized chunks. The overall algorithm expands the allowed size in powers of two, in order to amortize the cost of invoking vmalloc() for larger allocation sizes. Namely, the following sizes for the fdarray are considered, and the smallest that accommodates the requested fd count is chosen: pagesize / 4 pagesize / 2 pagesize <- memory allocator switch point pagesize * 2 pagesize * 4 ...etc... Unlike the current implementation, this allocation scheme does not require a loop to compute the optimal fdarray size, and can be done in efficient straightline code. Furthermore, since the fdarray overflows the pagesize boundary long before any of the fdsets do, it makes sense to optimize run-time by allocating both fdsets in a single swoop. Even together, they will still be, by far, smaller than the fdarray. The fdtable->open_fds is now used as the anchor for the fdset memory allocation. Signed-off-by: Vadim Lobanov <vlobanov@speakeasy.net> Cc: Christoph Hellwig <hch@lst.de> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Dipankar Sarma <dipankar@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 13:21:22 +03:00
return fdt;
[PATCH] fdtable: Implement new pagesize-based fdtable allocator This patch provides an improved fdtable allocation scheme, useful for expanding fdtable file descriptor entries. The main focus is on the fdarray, as its memory usage grows 128 times faster than that of an fdset. The allocation algorithm sizes the fdarray in such a way that its memory usage increases in easy page-sized chunks. The overall algorithm expands the allowed size in powers of two, in order to amortize the cost of invoking vmalloc() for larger allocation sizes. Namely, the following sizes for the fdarray are considered, and the smallest that accommodates the requested fd count is chosen: pagesize / 4 pagesize / 2 pagesize <- memory allocator switch point pagesize * 2 pagesize * 4 ...etc... Unlike the current implementation, this allocation scheme does not require a loop to compute the optimal fdarray size, and can be done in efficient straightline code. Furthermore, since the fdarray overflows the pagesize boundary long before any of the fdsets do, it makes sense to optimize run-time by allocating both fdsets in a single swoop. Even together, they will still be, by far, smaller than the fdarray. The fdtable->open_fds is now used as the anchor for the fdset memory allocation. Signed-off-by: Vadim Lobanov <vlobanov@speakeasy.net> Cc: Christoph Hellwig <hch@lst.de> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Dipankar Sarma <dipankar@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 13:21:22 +03:00
out_arr:
kvfree(fdt->fd);
[PATCH] fdtable: Implement new pagesize-based fdtable allocator This patch provides an improved fdtable allocation scheme, useful for expanding fdtable file descriptor entries. The main focus is on the fdarray, as its memory usage grows 128 times faster than that of an fdset. The allocation algorithm sizes the fdarray in such a way that its memory usage increases in easy page-sized chunks. The overall algorithm expands the allowed size in powers of two, in order to amortize the cost of invoking vmalloc() for larger allocation sizes. Namely, the following sizes for the fdarray are considered, and the smallest that accommodates the requested fd count is chosen: pagesize / 4 pagesize / 2 pagesize <- memory allocator switch point pagesize * 2 pagesize * 4 ...etc... Unlike the current implementation, this allocation scheme does not require a loop to compute the optimal fdarray size, and can be done in efficient straightline code. Furthermore, since the fdarray overflows the pagesize boundary long before any of the fdsets do, it makes sense to optimize run-time by allocating both fdsets in a single swoop. Even together, they will still be, by far, smaller than the fdarray. The fdtable->open_fds is now used as the anchor for the fdset memory allocation. Signed-off-by: Vadim Lobanov <vlobanov@speakeasy.net> Cc: Christoph Hellwig <hch@lst.de> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Dipankar Sarma <dipankar@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 13:21:22 +03:00
out_fdt:
kfree(fdt);
[PATCH] fdtable: Implement new pagesize-based fdtable allocator This patch provides an improved fdtable allocation scheme, useful for expanding fdtable file descriptor entries. The main focus is on the fdarray, as its memory usage grows 128 times faster than that of an fdset. The allocation algorithm sizes the fdarray in such a way that its memory usage increases in easy page-sized chunks. The overall algorithm expands the allowed size in powers of two, in order to amortize the cost of invoking vmalloc() for larger allocation sizes. Namely, the following sizes for the fdarray are considered, and the smallest that accommodates the requested fd count is chosen: pagesize / 4 pagesize / 2 pagesize <- memory allocator switch point pagesize * 2 pagesize * 4 ...etc... Unlike the current implementation, this allocation scheme does not require a loop to compute the optimal fdarray size, and can be done in efficient straightline code. Furthermore, since the fdarray overflows the pagesize boundary long before any of the fdsets do, it makes sense to optimize run-time by allocating both fdsets in a single swoop. Even together, they will still be, by far, smaller than the fdarray. The fdtable->open_fds is now used as the anchor for the fdset memory allocation. Signed-off-by: Vadim Lobanov <vlobanov@speakeasy.net> Cc: Christoph Hellwig <hch@lst.de> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Dipankar Sarma <dipankar@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-10 13:21:22 +03:00
out:
return NULL;
}
/*
* Expand the file descriptor table.
* This function will allocate a new fdtable and both fd array and fdset, of
* the given size.
* Return <0 error code on error; 1 on successful completion.
* The files->file_lock should be held on entry, and will be held on exit.
*/
static int expand_fdtable(struct files_struct *files, unsigned int nr)
__releases(files->file_lock)
__acquires(files->file_lock)
{
struct fdtable *new_fdt, *cur_fdt;
spin_unlock(&files->file_lock);
new_fdt = alloc_fdtable(nr);
/* make sure all fd_install() have seen resize_in_progress
* or have finished their rcu_read_lock_sched() section.
*/
if (atomic_read(&files->count) > 1)
synchronize_rcu();
spin_lock(&files->file_lock);
if (!new_fdt)
return -ENOMEM;
/*
* extremely unlikely race - sysctl_nr_open decreased between the check in
* caller and alloc_fdtable(). Cheaper to catch it here...
*/
if (unlikely(new_fdt->max_fds <= nr)) {
__free_fdtable(new_fdt);
return -EMFILE;
}
cur_fdt = files_fdtable(files);
BUG_ON(nr < cur_fdt->max_fds);
copy_fdtable(new_fdt, cur_fdt);
rcu_assign_pointer(files->fdt, new_fdt);
if (cur_fdt != &files->fdtab)
call_rcu(&cur_fdt->rcu, free_fdtable_rcu);
/* coupled with smp_rmb() in fd_install() */
smp_wmb();
return 1;
}
/*
* Expand files.
* This function will expand the file structures, if the requested size exceeds
* the current capacity and there is room for expansion.
* Return <0 error code on error; 0 when nothing done; 1 when files were
* expanded and execution may have blocked.
* The files->file_lock should be held on entry, and will be held on exit.
*/
static int expand_files(struct files_struct *files, unsigned int nr)
__releases(files->file_lock)
__acquires(files->file_lock)
{
struct fdtable *fdt;
int expanded = 0;
repeat:
fdt = files_fdtable(files);
/* Do we need to expand? */
if (nr < fdt->max_fds)
return expanded;
/* Can we expand? */
if (nr >= sysctl_nr_open)
return -EMFILE;
if (unlikely(files->resize_in_progress)) {
spin_unlock(&files->file_lock);
expanded = 1;
wait_event(files->resize_wait, !files->resize_in_progress);
spin_lock(&files->file_lock);
goto repeat;
}
/* All good, so we try */
files->resize_in_progress = true;
expanded = expand_fdtable(files, nr);
files->resize_in_progress = false;
wake_up_all(&files->resize_wait);
return expanded;
}
static inline void __set_close_on_exec(unsigned int fd, struct fdtable *fdt)
{
__set_bit(fd, fdt->close_on_exec);
}
static inline void __clear_close_on_exec(unsigned int fd, struct fdtable *fdt)
{
if (test_bit(fd, fdt->close_on_exec))
__clear_bit(fd, fdt->close_on_exec);
}
static inline void __set_open_fd(unsigned int fd, struct fdtable *fdt)
{
__set_bit(fd, fdt->open_fds);
fd /= BITS_PER_LONG;
if (!~fdt->open_fds[fd])
__set_bit(fd, fdt->full_fds_bits);
}
static inline void __clear_open_fd(unsigned int fd, struct fdtable *fdt)
{
__clear_bit(fd, fdt->open_fds);
__clear_bit(fd / BITS_PER_LONG, fdt->full_fds_bits);
}
static unsigned int count_open_files(struct fdtable *fdt)
{
unsigned int size = fdt->max_fds;
unsigned int i;
/* Find the last open fd */
for (i = size / BITS_PER_LONG; i > 0; ) {
if (fdt->open_fds[--i])
break;
}
i = (i + 1) * BITS_PER_LONG;
return i;
}
static unsigned int sane_fdtable_size(struct fdtable *fdt, unsigned int max_fds)
{
unsigned int count;
count = count_open_files(fdt);
if (max_fds < NR_OPEN_DEFAULT)
max_fds = NR_OPEN_DEFAULT;
return min(count, max_fds);
}
/*
* Allocate a new files structure and copy contents from the
* passed in files structure.
* errorp will be valid only when the returned files_struct is NULL.
*/
struct files_struct *dup_fd(struct files_struct *oldf, unsigned int max_fds, int *errorp)
{
struct files_struct *newf;
struct file **old_fds, **new_fds;
unsigned int open_files, i;
struct fdtable *old_fdt, *new_fdt;
*errorp = -ENOMEM;
newf = kmem_cache_alloc(files_cachep, GFP_KERNEL);
if (!newf)
goto out;
atomic_set(&newf->count, 1);
spin_lock_init(&newf->file_lock);
newf->resize_in_progress = false;
init_waitqueue_head(&newf->resize_wait);
newf->next_fd = 0;
new_fdt = &newf->fdtab;
new_fdt->max_fds = NR_OPEN_DEFAULT;
new_fdt->close_on_exec = newf->close_on_exec_init;
new_fdt->open_fds = newf->open_fds_init;
new_fdt->full_fds_bits = newf->full_fds_bits_init;
new_fdt->fd = &newf->fd_array[0];
spin_lock(&oldf->file_lock);
old_fdt = files_fdtable(oldf);
open_files = sane_fdtable_size(old_fdt, max_fds);
/*
* Check whether we need to allocate a larger fd array and fd set.
*/
while (unlikely(open_files > new_fdt->max_fds)) {
spin_unlock(&oldf->file_lock);
if (new_fdt != &newf->fdtab)
__free_fdtable(new_fdt);
new_fdt = alloc_fdtable(open_files - 1);
if (!new_fdt) {
*errorp = -ENOMEM;
goto out_release;
}
/* beyond sysctl_nr_open; nothing to do */
if (unlikely(new_fdt->max_fds < open_files)) {
__free_fdtable(new_fdt);
*errorp = -EMFILE;
goto out_release;
}
/*
* Reacquire the oldf lock and a pointer to its fd table
* who knows it may have a new bigger fd table. We need
* the latest pointer.
*/
spin_lock(&oldf->file_lock);
old_fdt = files_fdtable(oldf);
open_files = sane_fdtable_size(old_fdt, max_fds);
}
copy_fd_bitmaps(new_fdt, old_fdt, open_files);
old_fds = old_fdt->fd;
new_fds = new_fdt->fd;
for (i = open_files; i != 0; i--) {
struct file *f = *old_fds++;
if (f) {
get_file(f);
} else {
/*
* The fd may be claimed in the fd bitmap but not yet
* instantiated in the files array if a sibling thread
* is partway through open(). So make sure that this
* fd is available to the new process.
*/
__clear_open_fd(open_files - i, new_fdt);
}
rcu_assign_pointer(*new_fds++, f);
}
spin_unlock(&oldf->file_lock);
/* clear the remainder */
memset(new_fds, 0, (new_fdt->max_fds - open_files) * sizeof(struct file *));
rcu_assign_pointer(newf->fdt, new_fdt);
return newf;
out_release:
kmem_cache_free(files_cachep, newf);
out:
return NULL;
}
static struct fdtable *close_files(struct files_struct * files)
{
/*
* It is safe to dereference the fd table without RCU or
* ->file_lock because this is the last reference to the
* files structure.
*/
struct fdtable *fdt = rcu_dereference_raw(files->fdt);
unsigned int i, j = 0;
for (;;) {
unsigned long set;
i = j * BITS_PER_LONG;
if (i >= fdt->max_fds)
break;
set = fdt->open_fds[j++];
while (set) {
if (set & 1) {
struct file * file = xchg(&fdt->fd[i], NULL);
if (file) {
filp_close(file, files);
cond_resched();
}
}
i++;
set >>= 1;
}
}
return fdt;
}
void put_files_struct(struct files_struct *files)
{
if (atomic_dec_and_test(&files->count)) {
struct fdtable *fdt = close_files(files);
/* free the arrays if they are not embedded */
if (fdt != &files->fdtab)
__free_fdtable(fdt);
kmem_cache_free(files_cachep, files);
}
}
void exit_files(struct task_struct *tsk)
{
struct files_struct * files = tsk->files;
if (files) {
task_lock(tsk);
tsk->files = NULL;
task_unlock(tsk);
put_files_struct(files);
}
}
struct files_struct init_files = {
.count = ATOMIC_INIT(1),
.fdt = &init_files.fdtab,
.fdtab = {
.max_fds = NR_OPEN_DEFAULT,
.fd = &init_files.fd_array[0],
.close_on_exec = init_files.close_on_exec_init,
.open_fds = init_files.open_fds_init,
.full_fds_bits = init_files.full_fds_bits_init,
},
.file_lock = __SPIN_LOCK_UNLOCKED(init_files.file_lock),
fs/file.c: initialize init_files.resize_wait (Taken from https://bugzilla.kernel.org/show_bug.cgi?id=200647) 'get_unused_fd_flags' in kthread cause kernel crash. It works fine on 4.1, but causes crash after get 64 fds. It also cause crash on ubuntu1404/1604/1804, centos7.5, and the crash messages are almost the same. The crash message on centos7.5 shows below: start fd 61 start fd 62 start fd 63 BUG: unable to handle kernel NULL pointer dereference at (null) IP: __wake_up_common+0x2e/0x90 PGD 0 Oops: 0000 [#1] SMP Modules linked in: test(OE) xt_CHECKSUM iptable_mangle ipt_MASQUERADE nf_nat_masquerade_ipv4 iptable_nat nf_nat_ipv4 nf_nat nf_conntrack_ipv4 nf_defrag_ipv4 xt_conntrack nf_conntrack ipt_REJECT nf_reject_ipv4 tun bridge stp llc ebtable_filter ebtables ip6table_filter ip6_tables iptable_filter devlink sunrpc kvm_intel kvm irqbypass crc32_pclmul ghash_clmulni_intel aesni_intel lrw gf128mul glue_helper ablk_helper cryptd sg ppdev pcspkr virtio_balloon parport_pc parport i2c_piix4 joydev ip_tables xfs libcrc32c sr_mod cdrom sd_mod crc_t10dif crct10dif_generic ata_generic pata_acpi virtio_scsi virtio_console virtio_net cirrus drm_kms_helper syscopyarea sysfillrect sysimgblt fb_sys_fops ttm crct10dif_pclmul crct10dif_common crc32c_intel drm ata_piix serio_raw libata virtio_pci virtio_ring i2c_core virtio floppy dm_mirror dm_region_hash dm_log dm_mod CPU: 2 PID: 1820 Comm: test_fd Kdump: loaded Tainted: G OE ------------ 3.10.0-862.3.3.el7.x86_64 #1 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.10.2-0-g5f4c7b1-prebuilt.qemu-project.org 04/01/2014 task: ffff8e92b9431fa0 ti: ffff8e94247a0000 task.ti: ffff8e94247a0000 RIP: 0010:__wake_up_common+0x2e/0x90 RSP: 0018:ffff8e94247a2d18 EFLAGS: 00010086 RAX: 0000000000000000 RBX: ffffffff9d09daa0 RCX: 0000000000000000 RDX: 0000000000000000 RSI: 0000000000000003 RDI: ffffffff9d09daa0 RBP: ffff8e94247a2d50 R08: 0000000000000000 R09: ffff8e92b95dfda8 R10: 0000000000000000 R11: 0000000000000000 R12: ffffffff9d09daa8 R13: 0000000000000003 R14: 0000000000000000 R15: 0000000000000003 FS: 0000000000000000(0000) GS:ffff8e9434e80000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 0000000000000000 CR3: 000000017c686000 CR4: 00000000000207e0 Call Trace: __wake_up+0x39/0x50 expand_files+0x131/0x250 __alloc_fd+0x47/0x170 get_unused_fd_flags+0x30/0x40 test_fd+0x12a/0x1c0 [test] kthread+0xd1/0xe0 ret_from_fork_nospec_begin+0x21/0x21 Code: 66 90 55 48 89 e5 41 57 41 89 f7 41 56 41 89 ce 41 55 41 54 49 89 fc 49 83 c4 08 53 48 83 ec 10 48 8b 47 08 89 55 cc 4c 89 45 d0 <48> 8b 08 49 39 c4 48 8d 78 e8 4c 8d 69 e8 75 08 eb 3b 4c 89 ef RIP __wake_up_common+0x2e/0x90 RSP <ffff8e94247a2d18> CR2: 0000000000000000 This issue exists since CentOS 7.5 3.10.0-862 and CentOS 7.4 (3.10.0-693.21.1 ) is ok. Root cause: the item 'resize_wait' is not initialized before being used. Reported-by: Richard Zhang <zhang.zijian@h3c.com> Reviewed-by: Andrew Morton <akpm@linux-foundation.org> Cc: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-03-06 02:41:56 +03:00
.resize_wait = __WAIT_QUEUE_HEAD_INITIALIZER(init_files.resize_wait),
};
static unsigned int find_next_fd(struct fdtable *fdt, unsigned int start)
{
unsigned int maxfd = fdt->max_fds;
unsigned int maxbit = maxfd / BITS_PER_LONG;
unsigned int bitbit = start / BITS_PER_LONG;
bitbit = find_next_zero_bit(fdt->full_fds_bits, maxbit, bitbit) * BITS_PER_LONG;
if (bitbit > maxfd)
return maxfd;
if (bitbit > start)
start = bitbit;
return find_next_zero_bit(fdt->open_fds, maxfd, start);
}
/*
* allocate a file descriptor, mark it busy.
*/
static int alloc_fd(unsigned start, unsigned end, unsigned flags)
{
struct files_struct *files = current->files;
unsigned int fd;
int error;
struct fdtable *fdt;
spin_lock(&files->file_lock);
repeat:
fdt = files_fdtable(files);
fd = start;
if (fd < files->next_fd)
fd = files->next_fd;
if (fd < fdt->max_fds)
fd = find_next_fd(fdt, fd);
/*
* N.B. For clone tasks sharing a files structure, this test
* will limit the total number of files that can be opened.
*/
error = -EMFILE;
if (fd >= end)
goto out;
error = expand_files(files, fd);
if (error < 0)
goto out;
/*
* If we needed to expand the fs array we
* might have blocked - try again.
*/
if (error)
goto repeat;
if (start <= files->next_fd)
files->next_fd = fd + 1;
__set_open_fd(fd, fdt);
if (flags & O_CLOEXEC)
__set_close_on_exec(fd, fdt);
else
__clear_close_on_exec(fd, fdt);
error = fd;
#if 1
/* Sanity check */
if (rcu_access_pointer(fdt->fd[fd]) != NULL) {
printk(KERN_WARNING "alloc_fd: slot %d not NULL!\n", fd);
rcu_assign_pointer(fdt->fd[fd], NULL);
}
#endif
out:
spin_unlock(&files->file_lock);
return error;
}
int __get_unused_fd_flags(unsigned flags, unsigned long nofile)
{
return alloc_fd(0, nofile, flags);
}
int get_unused_fd_flags(unsigned flags)
{
return __get_unused_fd_flags(flags, rlimit(RLIMIT_NOFILE));
}
EXPORT_SYMBOL(get_unused_fd_flags);
static void __put_unused_fd(struct files_struct *files, unsigned int fd)
{
struct fdtable *fdt = files_fdtable(files);
__clear_open_fd(fd, fdt);
if (fd < files->next_fd)
files->next_fd = fd;
}
void put_unused_fd(unsigned int fd)
{
struct files_struct *files = current->files;
spin_lock(&files->file_lock);
__put_unused_fd(files, fd);
spin_unlock(&files->file_lock);
}
EXPORT_SYMBOL(put_unused_fd);
/*
* Install a file pointer in the fd array.
*
* The VFS is full of places where we drop the files lock between
* setting the open_fds bitmap and installing the file in the file
* array. At any such point, we are vulnerable to a dup2() race
* installing a file in the array before us. We need to detect this and
* fput() the struct file we are about to overwrite in this case.
*
* It should never happen - if we allow dup2() do it, _really_ bad things
* will follow.
*
* This consumes the "file" refcount, so callers should treat it
* as if they had called fput(file).
*/
void fd_install(unsigned int fd, struct file *file)
{
struct files_struct *files = current->files;
struct fdtable *fdt;
rcu_read_lock_sched();
if (unlikely(files->resize_in_progress)) {
rcu_read_unlock_sched();
spin_lock(&files->file_lock);
fdt = files_fdtable(files);
BUG_ON(fdt->fd[fd] != NULL);
rcu_assign_pointer(fdt->fd[fd], file);
spin_unlock(&files->file_lock);
return;
}
/* coupled with smp_wmb() in expand_fdtable() */
smp_rmb();
fdt = rcu_dereference_sched(files->fdt);
BUG_ON(fdt->fd[fd] != NULL);
rcu_assign_pointer(fdt->fd[fd], file);
rcu_read_unlock_sched();
}
EXPORT_SYMBOL(fd_install);
/**
* pick_file - return file associatd with fd
* @files: file struct to retrieve file from
* @fd: file descriptor to retrieve file for
*
* If this functions returns an EINVAL error pointer the fd was beyond the
* current maximum number of file descriptors for that fdtable.
*
* Returns: The file associated with @fd, on error returns an error pointer.
*/
open: add close_range() This adds the close_range() syscall. It allows to efficiently close a range of file descriptors up to all file descriptors of a calling task. I was contacted by FreeBSD as they wanted to have the same close_range() syscall as we proposed here. We've coordinated this and in the meantime, Kyle was fast enough to merge close_range() into FreeBSD already in April: https://reviews.freebsd.org/D21627 https://svnweb.freebsd.org/base?view=revision&revision=359836 and the current plan is to backport close_range() to FreeBSD 12.2 (cf. [2]) once its merged in Linux too. Python is in the process of switching to close_range() on FreeBSD and they are waiting on us to merge this to switch on Linux as well: https://bugs.python.org/issue38061 The syscall came up in a recent discussion around the new mount API and making new file descriptor types cloexec by default. During this discussion, Al suggested the close_range() syscall (cf. [1]). Note, a syscall in this manner has been requested by various people over time. First, it helps to close all file descriptors of an exec()ing task. This can be done safely via (quoting Al's example from [1] verbatim): /* that exec is sensitive */ unshare(CLONE_FILES); /* we don't want anything past stderr here */ close_range(3, ~0U); execve(....); The code snippet above is one way of working around the problem that file descriptors are not cloexec by default. This is aggravated by the fact that we can't just switch them over without massively regressing userspace. For a whole class of programs having an in-kernel method of closing all file descriptors is very helpful (e.g. demons, service managers, programming language standard libraries, container managers etc.). (Please note, unshare(CLONE_FILES) should only be needed if the calling task is multi-threaded and shares the file descriptor table with another thread in which case two threads could race with one thread allocating file descriptors and the other one closing them via close_range(). For the general case close_range() before the execve() is sufficient.) Second, it allows userspace to avoid implementing closing all file descriptors by parsing through /proc/<pid>/fd/* and calling close() on each file descriptor. From looking at various large(ish) userspace code bases this or similar patterns are very common in: - service managers (cf. [4]) - libcs (cf. [6]) - container runtimes (cf. [5]) - programming language runtimes/standard libraries - Python (cf. [2]) - Rust (cf. [7], [8]) As Dmitry pointed out there's even a long-standing glibc bug about missing kernel support for this task (cf. [3]). In addition, the syscall will also work for tasks that do not have procfs mounted and on kernels that do not have procfs support compiled in. In such situations the only way to make sure that all file descriptors are closed is to call close() on each file descriptor up to UINT_MAX or RLIMIT_NOFILE, OPEN_MAX trickery (cf. comment [8] on Rust). The performance is striking. For good measure, comparing the following simple close_all_fds() userspace implementation that is essentially just glibc's version in [6]: static int close_all_fds(void) { int dir_fd; DIR *dir; struct dirent *direntp; dir = opendir("/proc/self/fd"); if (!dir) return -1; dir_fd = dirfd(dir); while ((direntp = readdir(dir))) { int fd; if (strcmp(direntp->d_name, ".") == 0) continue; if (strcmp(direntp->d_name, "..") == 0) continue; fd = atoi(direntp->d_name); if (fd == dir_fd || fd == 0 || fd == 1 || fd == 2) continue; close(fd); } closedir(dir); return 0; } to close_range() yields: 1. closing 4 open files: - close_all_fds(): ~280 us - close_range(): ~24 us 2. closing 1000 open files: - close_all_fds(): ~5000 us - close_range(): ~800 us close_range() is designed to allow for some flexibility. Specifically, it does not simply always close all open file descriptors of a task. Instead, callers can specify an upper bound. This is e.g. useful for scenarios where specific file descriptors are created with well-known numbers that are supposed to be excluded from getting closed. For extra paranoia close_range() comes with a flags argument. This can e.g. be used to implement extension. Once can imagine userspace wanting to stop at the first error instead of ignoring errors under certain circumstances. There might be other valid ideas in the future. In any case, a flag argument doesn't hurt and keeps us on the safe side. From an implementation side this is kept rather dumb. It saw some input from David and Jann but all nonsense is obviously my own! - Errors to close file descriptors are currently ignored. (Could be changed by setting a flag in the future if needed.) - __close_range() is a rather simplistic wrapper around __close_fd(). My reasoning behind this is based on the nature of how __close_fd() needs to release an fd. But maybe I misunderstood specifics: We take the files_lock and rcu-dereference the fdtable of the calling task, we find the entry in the fdtable, get the file and need to release files_lock before calling filp_close(). In the meantime the fdtable might have been altered so we can't just retake the spinlock and keep the old rcu-reference of the fdtable around. Instead we need to grab a fresh reference to the fdtable. If my reasoning is correct then there's really no point in fancyfying __close_range(): We just need to rcu-dereference the fdtable of the calling task once to cap the max_fd value correctly and then go on calling __close_fd() in a loop. /* References */ [1]: https://lore.kernel.org/lkml/20190516165021.GD17978@ZenIV.linux.org.uk/ [2]: https://github.com/python/cpython/blob/9e4f2f3a6b8ee995c365e86d976937c141d867f8/Modules/_posixsubprocess.c#L220 [3]: https://sourceware.org/bugzilla/show_bug.cgi?id=10353#c7 [4]: https://github.com/systemd/systemd/blob/5238e9575906297608ff802a27e2ff9effa3b338/src/basic/fd-util.c#L217 [5]: https://github.com/lxc/lxc/blob/ddf4b77e11a4d08f09b7b9cd13e593f8c047edc5/src/lxc/start.c#L236 [6]: https://sourceware.org/git/?p=glibc.git;a=blob;f=sysdeps/unix/sysv/linux/grantpt.c;h=2030e07fa6e652aac32c775b8c6e005844c3c4eb;hb=HEAD#l17 Note that this is an internal implementation that is not exported. Currently, libc seems to not provide an exported version of this because of missing kernel support to do this. Note, in a recent patch series Florian made grantpt() a nop thereby removing the code referenced here. [7]: https://github.com/rust-lang/rust/issues/12148 [8]: https://github.com/rust-lang/rust/blob/5f47c0613ed4eb46fca3633c1297364c09e5e451/src/libstd/sys/unix/process2.rs#L303-L308 Rust's solution is slightly different but is equally unperformant. Rust calls getdtablesize() which is a glibc library function that simply returns the current RLIMIT_NOFILE or OPEN_MAX values. Rust then goes on to call close() on each fd. That's obviously overkill for most tasks. Rarely, tasks - especially non-demons - hit RLIMIT_NOFILE or OPEN_MAX. Let's be nice and assume an unprivileged user with RLIMIT_NOFILE set to 1024. Even in this case, there's a very high chance that in the common case Rust is calling the close() syscall 1021 times pointlessly if the task just has 0, 1, and 2 open. Suggested-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com> Cc: Arnd Bergmann <arnd@arndb.de> Cc: Kyle Evans <self@kyle-evans.net> Cc: Jann Horn <jannh@google.com> Cc: David Howells <dhowells@redhat.com> Cc: Dmitry V. Levin <ldv@altlinux.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Florian Weimer <fweimer@redhat.com> Cc: linux-api@vger.kernel.org
2019-05-24 12:30:34 +03:00
static struct file *pick_file(struct files_struct *files, unsigned fd)
{
struct file *file;
struct fdtable *fdt;
spin_lock(&files->file_lock);
fdt = files_fdtable(files);
if (fd >= fdt->max_fds) {
file = ERR_PTR(-EINVAL);
goto out_unlock;
}
file = fdt->fd[fd];
if (!file) {
file = ERR_PTR(-EBADF);
goto out_unlock;
}
rcu_assign_pointer(fdt->fd[fd], NULL);
__put_unused_fd(files, fd);
out_unlock:
spin_unlock(&files->file_lock);
open: add close_range() This adds the close_range() syscall. It allows to efficiently close a range of file descriptors up to all file descriptors of a calling task. I was contacted by FreeBSD as they wanted to have the same close_range() syscall as we proposed here. We've coordinated this and in the meantime, Kyle was fast enough to merge close_range() into FreeBSD already in April: https://reviews.freebsd.org/D21627 https://svnweb.freebsd.org/base?view=revision&revision=359836 and the current plan is to backport close_range() to FreeBSD 12.2 (cf. [2]) once its merged in Linux too. Python is in the process of switching to close_range() on FreeBSD and they are waiting on us to merge this to switch on Linux as well: https://bugs.python.org/issue38061 The syscall came up in a recent discussion around the new mount API and making new file descriptor types cloexec by default. During this discussion, Al suggested the close_range() syscall (cf. [1]). Note, a syscall in this manner has been requested by various people over time. First, it helps to close all file descriptors of an exec()ing task. This can be done safely via (quoting Al's example from [1] verbatim): /* that exec is sensitive */ unshare(CLONE_FILES); /* we don't want anything past stderr here */ close_range(3, ~0U); execve(....); The code snippet above is one way of working around the problem that file descriptors are not cloexec by default. This is aggravated by the fact that we can't just switch them over without massively regressing userspace. For a whole class of programs having an in-kernel method of closing all file descriptors is very helpful (e.g. demons, service managers, programming language standard libraries, container managers etc.). (Please note, unshare(CLONE_FILES) should only be needed if the calling task is multi-threaded and shares the file descriptor table with another thread in which case two threads could race with one thread allocating file descriptors and the other one closing them via close_range(). For the general case close_range() before the execve() is sufficient.) Second, it allows userspace to avoid implementing closing all file descriptors by parsing through /proc/<pid>/fd/* and calling close() on each file descriptor. From looking at various large(ish) userspace code bases this or similar patterns are very common in: - service managers (cf. [4]) - libcs (cf. [6]) - container runtimes (cf. [5]) - programming language runtimes/standard libraries - Python (cf. [2]) - Rust (cf. [7], [8]) As Dmitry pointed out there's even a long-standing glibc bug about missing kernel support for this task (cf. [3]). In addition, the syscall will also work for tasks that do not have procfs mounted and on kernels that do not have procfs support compiled in. In such situations the only way to make sure that all file descriptors are closed is to call close() on each file descriptor up to UINT_MAX or RLIMIT_NOFILE, OPEN_MAX trickery (cf. comment [8] on Rust). The performance is striking. For good measure, comparing the following simple close_all_fds() userspace implementation that is essentially just glibc's version in [6]: static int close_all_fds(void) { int dir_fd; DIR *dir; struct dirent *direntp; dir = opendir("/proc/self/fd"); if (!dir) return -1; dir_fd = dirfd(dir); while ((direntp = readdir(dir))) { int fd; if (strcmp(direntp->d_name, ".") == 0) continue; if (strcmp(direntp->d_name, "..") == 0) continue; fd = atoi(direntp->d_name); if (fd == dir_fd || fd == 0 || fd == 1 || fd == 2) continue; close(fd); } closedir(dir); return 0; } to close_range() yields: 1. closing 4 open files: - close_all_fds(): ~280 us - close_range(): ~24 us 2. closing 1000 open files: - close_all_fds(): ~5000 us - close_range(): ~800 us close_range() is designed to allow for some flexibility. Specifically, it does not simply always close all open file descriptors of a task. Instead, callers can specify an upper bound. This is e.g. useful for scenarios where specific file descriptors are created with well-known numbers that are supposed to be excluded from getting closed. For extra paranoia close_range() comes with a flags argument. This can e.g. be used to implement extension. Once can imagine userspace wanting to stop at the first error instead of ignoring errors under certain circumstances. There might be other valid ideas in the future. In any case, a flag argument doesn't hurt and keeps us on the safe side. From an implementation side this is kept rather dumb. It saw some input from David and Jann but all nonsense is obviously my own! - Errors to close file descriptors are currently ignored. (Could be changed by setting a flag in the future if needed.) - __close_range() is a rather simplistic wrapper around __close_fd(). My reasoning behind this is based on the nature of how __close_fd() needs to release an fd. But maybe I misunderstood specifics: We take the files_lock and rcu-dereference the fdtable of the calling task, we find the entry in the fdtable, get the file and need to release files_lock before calling filp_close(). In the meantime the fdtable might have been altered so we can't just retake the spinlock and keep the old rcu-reference of the fdtable around. Instead we need to grab a fresh reference to the fdtable. If my reasoning is correct then there's really no point in fancyfying __close_range(): We just need to rcu-dereference the fdtable of the calling task once to cap the max_fd value correctly and then go on calling __close_fd() in a loop. /* References */ [1]: https://lore.kernel.org/lkml/20190516165021.GD17978@ZenIV.linux.org.uk/ [2]: https://github.com/python/cpython/blob/9e4f2f3a6b8ee995c365e86d976937c141d867f8/Modules/_posixsubprocess.c#L220 [3]: https://sourceware.org/bugzilla/show_bug.cgi?id=10353#c7 [4]: https://github.com/systemd/systemd/blob/5238e9575906297608ff802a27e2ff9effa3b338/src/basic/fd-util.c#L217 [5]: https://github.com/lxc/lxc/blob/ddf4b77e11a4d08f09b7b9cd13e593f8c047edc5/src/lxc/start.c#L236 [6]: https://sourceware.org/git/?p=glibc.git;a=blob;f=sysdeps/unix/sysv/linux/grantpt.c;h=2030e07fa6e652aac32c775b8c6e005844c3c4eb;hb=HEAD#l17 Note that this is an internal implementation that is not exported. Currently, libc seems to not provide an exported version of this because of missing kernel support to do this. Note, in a recent patch series Florian made grantpt() a nop thereby removing the code referenced here. [7]: https://github.com/rust-lang/rust/issues/12148 [8]: https://github.com/rust-lang/rust/blob/5f47c0613ed4eb46fca3633c1297364c09e5e451/src/libstd/sys/unix/process2.rs#L303-L308 Rust's solution is slightly different but is equally unperformant. Rust calls getdtablesize() which is a glibc library function that simply returns the current RLIMIT_NOFILE or OPEN_MAX values. Rust then goes on to call close() on each fd. That's obviously overkill for most tasks. Rarely, tasks - especially non-demons - hit RLIMIT_NOFILE or OPEN_MAX. Let's be nice and assume an unprivileged user with RLIMIT_NOFILE set to 1024. Even in this case, there's a very high chance that in the common case Rust is calling the close() syscall 1021 times pointlessly if the task just has 0, 1, and 2 open. Suggested-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com> Cc: Arnd Bergmann <arnd@arndb.de> Cc: Kyle Evans <self@kyle-evans.net> Cc: Jann Horn <jannh@google.com> Cc: David Howells <dhowells@redhat.com> Cc: Dmitry V. Levin <ldv@altlinux.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Florian Weimer <fweimer@redhat.com> Cc: linux-api@vger.kernel.org
2019-05-24 12:30:34 +03:00
return file;
}
int close_fd(unsigned fd)
open: add close_range() This adds the close_range() syscall. It allows to efficiently close a range of file descriptors up to all file descriptors of a calling task. I was contacted by FreeBSD as they wanted to have the same close_range() syscall as we proposed here. We've coordinated this and in the meantime, Kyle was fast enough to merge close_range() into FreeBSD already in April: https://reviews.freebsd.org/D21627 https://svnweb.freebsd.org/base?view=revision&revision=359836 and the current plan is to backport close_range() to FreeBSD 12.2 (cf. [2]) once its merged in Linux too. Python is in the process of switching to close_range() on FreeBSD and they are waiting on us to merge this to switch on Linux as well: https://bugs.python.org/issue38061 The syscall came up in a recent discussion around the new mount API and making new file descriptor types cloexec by default. During this discussion, Al suggested the close_range() syscall (cf. [1]). Note, a syscall in this manner has been requested by various people over time. First, it helps to close all file descriptors of an exec()ing task. This can be done safely via (quoting Al's example from [1] verbatim): /* that exec is sensitive */ unshare(CLONE_FILES); /* we don't want anything past stderr here */ close_range(3, ~0U); execve(....); The code snippet above is one way of working around the problem that file descriptors are not cloexec by default. This is aggravated by the fact that we can't just switch them over without massively regressing userspace. For a whole class of programs having an in-kernel method of closing all file descriptors is very helpful (e.g. demons, service managers, programming language standard libraries, container managers etc.). (Please note, unshare(CLONE_FILES) should only be needed if the calling task is multi-threaded and shares the file descriptor table with another thread in which case two threads could race with one thread allocating file descriptors and the other one closing them via close_range(). For the general case close_range() before the execve() is sufficient.) Second, it allows userspace to avoid implementing closing all file descriptors by parsing through /proc/<pid>/fd/* and calling close() on each file descriptor. From looking at various large(ish) userspace code bases this or similar patterns are very common in: - service managers (cf. [4]) - libcs (cf. [6]) - container runtimes (cf. [5]) - programming language runtimes/standard libraries - Python (cf. [2]) - Rust (cf. [7], [8]) As Dmitry pointed out there's even a long-standing glibc bug about missing kernel support for this task (cf. [3]). In addition, the syscall will also work for tasks that do not have procfs mounted and on kernels that do not have procfs support compiled in. In such situations the only way to make sure that all file descriptors are closed is to call close() on each file descriptor up to UINT_MAX or RLIMIT_NOFILE, OPEN_MAX trickery (cf. comment [8] on Rust). The performance is striking. For good measure, comparing the following simple close_all_fds() userspace implementation that is essentially just glibc's version in [6]: static int close_all_fds(void) { int dir_fd; DIR *dir; struct dirent *direntp; dir = opendir("/proc/self/fd"); if (!dir) return -1; dir_fd = dirfd(dir); while ((direntp = readdir(dir))) { int fd; if (strcmp(direntp->d_name, ".") == 0) continue; if (strcmp(direntp->d_name, "..") == 0) continue; fd = atoi(direntp->d_name); if (fd == dir_fd || fd == 0 || fd == 1 || fd == 2) continue; close(fd); } closedir(dir); return 0; } to close_range() yields: 1. closing 4 open files: - close_all_fds(): ~280 us - close_range(): ~24 us 2. closing 1000 open files: - close_all_fds(): ~5000 us - close_range(): ~800 us close_range() is designed to allow for some flexibility. Specifically, it does not simply always close all open file descriptors of a task. Instead, callers can specify an upper bound. This is e.g. useful for scenarios where specific file descriptors are created with well-known numbers that are supposed to be excluded from getting closed. For extra paranoia close_range() comes with a flags argument. This can e.g. be used to implement extension. Once can imagine userspace wanting to stop at the first error instead of ignoring errors under certain circumstances. There might be other valid ideas in the future. In any case, a flag argument doesn't hurt and keeps us on the safe side. From an implementation side this is kept rather dumb. It saw some input from David and Jann but all nonsense is obviously my own! - Errors to close file descriptors are currently ignored. (Could be changed by setting a flag in the future if needed.) - __close_range() is a rather simplistic wrapper around __close_fd(). My reasoning behind this is based on the nature of how __close_fd() needs to release an fd. But maybe I misunderstood specifics: We take the files_lock and rcu-dereference the fdtable of the calling task, we find the entry in the fdtable, get the file and need to release files_lock before calling filp_close(). In the meantime the fdtable might have been altered so we can't just retake the spinlock and keep the old rcu-reference of the fdtable around. Instead we need to grab a fresh reference to the fdtable. If my reasoning is correct then there's really no point in fancyfying __close_range(): We just need to rcu-dereference the fdtable of the calling task once to cap the max_fd value correctly and then go on calling __close_fd() in a loop. /* References */ [1]: https://lore.kernel.org/lkml/20190516165021.GD17978@ZenIV.linux.org.uk/ [2]: https://github.com/python/cpython/blob/9e4f2f3a6b8ee995c365e86d976937c141d867f8/Modules/_posixsubprocess.c#L220 [3]: https://sourceware.org/bugzilla/show_bug.cgi?id=10353#c7 [4]: https://github.com/systemd/systemd/blob/5238e9575906297608ff802a27e2ff9effa3b338/src/basic/fd-util.c#L217 [5]: https://github.com/lxc/lxc/blob/ddf4b77e11a4d08f09b7b9cd13e593f8c047edc5/src/lxc/start.c#L236 [6]: https://sourceware.org/git/?p=glibc.git;a=blob;f=sysdeps/unix/sysv/linux/grantpt.c;h=2030e07fa6e652aac32c775b8c6e005844c3c4eb;hb=HEAD#l17 Note that this is an internal implementation that is not exported. Currently, libc seems to not provide an exported version of this because of missing kernel support to do this. Note, in a recent patch series Florian made grantpt() a nop thereby removing the code referenced here. [7]: https://github.com/rust-lang/rust/issues/12148 [8]: https://github.com/rust-lang/rust/blob/5f47c0613ed4eb46fca3633c1297364c09e5e451/src/libstd/sys/unix/process2.rs#L303-L308 Rust's solution is slightly different but is equally unperformant. Rust calls getdtablesize() which is a glibc library function that simply returns the current RLIMIT_NOFILE or OPEN_MAX values. Rust then goes on to call close() on each fd. That's obviously overkill for most tasks. Rarely, tasks - especially non-demons - hit RLIMIT_NOFILE or OPEN_MAX. Let's be nice and assume an unprivileged user with RLIMIT_NOFILE set to 1024. Even in this case, there's a very high chance that in the common case Rust is calling the close() syscall 1021 times pointlessly if the task just has 0, 1, and 2 open. Suggested-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com> Cc: Arnd Bergmann <arnd@arndb.de> Cc: Kyle Evans <self@kyle-evans.net> Cc: Jann Horn <jannh@google.com> Cc: David Howells <dhowells@redhat.com> Cc: Dmitry V. Levin <ldv@altlinux.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Florian Weimer <fweimer@redhat.com> Cc: linux-api@vger.kernel.org
2019-05-24 12:30:34 +03:00
{
struct files_struct *files = current->files;
open: add close_range() This adds the close_range() syscall. It allows to efficiently close a range of file descriptors up to all file descriptors of a calling task. I was contacted by FreeBSD as they wanted to have the same close_range() syscall as we proposed here. We've coordinated this and in the meantime, Kyle was fast enough to merge close_range() into FreeBSD already in April: https://reviews.freebsd.org/D21627 https://svnweb.freebsd.org/base?view=revision&revision=359836 and the current plan is to backport close_range() to FreeBSD 12.2 (cf. [2]) once its merged in Linux too. Python is in the process of switching to close_range() on FreeBSD and they are waiting on us to merge this to switch on Linux as well: https://bugs.python.org/issue38061 The syscall came up in a recent discussion around the new mount API and making new file descriptor types cloexec by default. During this discussion, Al suggested the close_range() syscall (cf. [1]). Note, a syscall in this manner has been requested by various people over time. First, it helps to close all file descriptors of an exec()ing task. This can be done safely via (quoting Al's example from [1] verbatim): /* that exec is sensitive */ unshare(CLONE_FILES); /* we don't want anything past stderr here */ close_range(3, ~0U); execve(....); The code snippet above is one way of working around the problem that file descriptors are not cloexec by default. This is aggravated by the fact that we can't just switch them over without massively regressing userspace. For a whole class of programs having an in-kernel method of closing all file descriptors is very helpful (e.g. demons, service managers, programming language standard libraries, container managers etc.). (Please note, unshare(CLONE_FILES) should only be needed if the calling task is multi-threaded and shares the file descriptor table with another thread in which case two threads could race with one thread allocating file descriptors and the other one closing them via close_range(). For the general case close_range() before the execve() is sufficient.) Second, it allows userspace to avoid implementing closing all file descriptors by parsing through /proc/<pid>/fd/* and calling close() on each file descriptor. From looking at various large(ish) userspace code bases this or similar patterns are very common in: - service managers (cf. [4]) - libcs (cf. [6]) - container runtimes (cf. [5]) - programming language runtimes/standard libraries - Python (cf. [2]) - Rust (cf. [7], [8]) As Dmitry pointed out there's even a long-standing glibc bug about missing kernel support for this task (cf. [3]). In addition, the syscall will also work for tasks that do not have procfs mounted and on kernels that do not have procfs support compiled in. In such situations the only way to make sure that all file descriptors are closed is to call close() on each file descriptor up to UINT_MAX or RLIMIT_NOFILE, OPEN_MAX trickery (cf. comment [8] on Rust). The performance is striking. For good measure, comparing the following simple close_all_fds() userspace implementation that is essentially just glibc's version in [6]: static int close_all_fds(void) { int dir_fd; DIR *dir; struct dirent *direntp; dir = opendir("/proc/self/fd"); if (!dir) return -1; dir_fd = dirfd(dir); while ((direntp = readdir(dir))) { int fd; if (strcmp(direntp->d_name, ".") == 0) continue; if (strcmp(direntp->d_name, "..") == 0) continue; fd = atoi(direntp->d_name); if (fd == dir_fd || fd == 0 || fd == 1 || fd == 2) continue; close(fd); } closedir(dir); return 0; } to close_range() yields: 1. closing 4 open files: - close_all_fds(): ~280 us - close_range(): ~24 us 2. closing 1000 open files: - close_all_fds(): ~5000 us - close_range(): ~800 us close_range() is designed to allow for some flexibility. Specifically, it does not simply always close all open file descriptors of a task. Instead, callers can specify an upper bound. This is e.g. useful for scenarios where specific file descriptors are created with well-known numbers that are supposed to be excluded from getting closed. For extra paranoia close_range() comes with a flags argument. This can e.g. be used to implement extension. Once can imagine userspace wanting to stop at the first error instead of ignoring errors under certain circumstances. There might be other valid ideas in the future. In any case, a flag argument doesn't hurt and keeps us on the safe side. From an implementation side this is kept rather dumb. It saw some input from David and Jann but all nonsense is obviously my own! - Errors to close file descriptors are currently ignored. (Could be changed by setting a flag in the future if needed.) - __close_range() is a rather simplistic wrapper around __close_fd(). My reasoning behind this is based on the nature of how __close_fd() needs to release an fd. But maybe I misunderstood specifics: We take the files_lock and rcu-dereference the fdtable of the calling task, we find the entry in the fdtable, get the file and need to release files_lock before calling filp_close(). In the meantime the fdtable might have been altered so we can't just retake the spinlock and keep the old rcu-reference of the fdtable around. Instead we need to grab a fresh reference to the fdtable. If my reasoning is correct then there's really no point in fancyfying __close_range(): We just need to rcu-dereference the fdtable of the calling task once to cap the max_fd value correctly and then go on calling __close_fd() in a loop. /* References */ [1]: https://lore.kernel.org/lkml/20190516165021.GD17978@ZenIV.linux.org.uk/ [2]: https://github.com/python/cpython/blob/9e4f2f3a6b8ee995c365e86d976937c141d867f8/Modules/_posixsubprocess.c#L220 [3]: https://sourceware.org/bugzilla/show_bug.cgi?id=10353#c7 [4]: https://github.com/systemd/systemd/blob/5238e9575906297608ff802a27e2ff9effa3b338/src/basic/fd-util.c#L217 [5]: https://github.com/lxc/lxc/blob/ddf4b77e11a4d08f09b7b9cd13e593f8c047edc5/src/lxc/start.c#L236 [6]: https://sourceware.org/git/?p=glibc.git;a=blob;f=sysdeps/unix/sysv/linux/grantpt.c;h=2030e07fa6e652aac32c775b8c6e005844c3c4eb;hb=HEAD#l17 Note that this is an internal implementation that is not exported. Currently, libc seems to not provide an exported version of this because of missing kernel support to do this. Note, in a recent patch series Florian made grantpt() a nop thereby removing the code referenced here. [7]: https://github.com/rust-lang/rust/issues/12148 [8]: https://github.com/rust-lang/rust/blob/5f47c0613ed4eb46fca3633c1297364c09e5e451/src/libstd/sys/unix/process2.rs#L303-L308 Rust's solution is slightly different but is equally unperformant. Rust calls getdtablesize() which is a glibc library function that simply returns the current RLIMIT_NOFILE or OPEN_MAX values. Rust then goes on to call close() on each fd. That's obviously overkill for most tasks. Rarely, tasks - especially non-demons - hit RLIMIT_NOFILE or OPEN_MAX. Let's be nice and assume an unprivileged user with RLIMIT_NOFILE set to 1024. Even in this case, there's a very high chance that in the common case Rust is calling the close() syscall 1021 times pointlessly if the task just has 0, 1, and 2 open. Suggested-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com> Cc: Arnd Bergmann <arnd@arndb.de> Cc: Kyle Evans <self@kyle-evans.net> Cc: Jann Horn <jannh@google.com> Cc: David Howells <dhowells@redhat.com> Cc: Dmitry V. Levin <ldv@altlinux.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Florian Weimer <fweimer@redhat.com> Cc: linux-api@vger.kernel.org
2019-05-24 12:30:34 +03:00
struct file *file;
file = pick_file(files, fd);
if (IS_ERR(file))
open: add close_range() This adds the close_range() syscall. It allows to efficiently close a range of file descriptors up to all file descriptors of a calling task. I was contacted by FreeBSD as they wanted to have the same close_range() syscall as we proposed here. We've coordinated this and in the meantime, Kyle was fast enough to merge close_range() into FreeBSD already in April: https://reviews.freebsd.org/D21627 https://svnweb.freebsd.org/base?view=revision&revision=359836 and the current plan is to backport close_range() to FreeBSD 12.2 (cf. [2]) once its merged in Linux too. Python is in the process of switching to close_range() on FreeBSD and they are waiting on us to merge this to switch on Linux as well: https://bugs.python.org/issue38061 The syscall came up in a recent discussion around the new mount API and making new file descriptor types cloexec by default. During this discussion, Al suggested the close_range() syscall (cf. [1]). Note, a syscall in this manner has been requested by various people over time. First, it helps to close all file descriptors of an exec()ing task. This can be done safely via (quoting Al's example from [1] verbatim): /* that exec is sensitive */ unshare(CLONE_FILES); /* we don't want anything past stderr here */ close_range(3, ~0U); execve(....); The code snippet above is one way of working around the problem that file descriptors are not cloexec by default. This is aggravated by the fact that we can't just switch them over without massively regressing userspace. For a whole class of programs having an in-kernel method of closing all file descriptors is very helpful (e.g. demons, service managers, programming language standard libraries, container managers etc.). (Please note, unshare(CLONE_FILES) should only be needed if the calling task is multi-threaded and shares the file descriptor table with another thread in which case two threads could race with one thread allocating file descriptors and the other one closing them via close_range(). For the general case close_range() before the execve() is sufficient.) Second, it allows userspace to avoid implementing closing all file descriptors by parsing through /proc/<pid>/fd/* and calling close() on each file descriptor. From looking at various large(ish) userspace code bases this or similar patterns are very common in: - service managers (cf. [4]) - libcs (cf. [6]) - container runtimes (cf. [5]) - programming language runtimes/standard libraries - Python (cf. [2]) - Rust (cf. [7], [8]) As Dmitry pointed out there's even a long-standing glibc bug about missing kernel support for this task (cf. [3]). In addition, the syscall will also work for tasks that do not have procfs mounted and on kernels that do not have procfs support compiled in. In such situations the only way to make sure that all file descriptors are closed is to call close() on each file descriptor up to UINT_MAX or RLIMIT_NOFILE, OPEN_MAX trickery (cf. comment [8] on Rust). The performance is striking. For good measure, comparing the following simple close_all_fds() userspace implementation that is essentially just glibc's version in [6]: static int close_all_fds(void) { int dir_fd; DIR *dir; struct dirent *direntp; dir = opendir("/proc/self/fd"); if (!dir) return -1; dir_fd = dirfd(dir); while ((direntp = readdir(dir))) { int fd; if (strcmp(direntp->d_name, ".") == 0) continue; if (strcmp(direntp->d_name, "..") == 0) continue; fd = atoi(direntp->d_name); if (fd == dir_fd || fd == 0 || fd == 1 || fd == 2) continue; close(fd); } closedir(dir); return 0; } to close_range() yields: 1. closing 4 open files: - close_all_fds(): ~280 us - close_range(): ~24 us 2. closing 1000 open files: - close_all_fds(): ~5000 us - close_range(): ~800 us close_range() is designed to allow for some flexibility. Specifically, it does not simply always close all open file descriptors of a task. Instead, callers can specify an upper bound. This is e.g. useful for scenarios where specific file descriptors are created with well-known numbers that are supposed to be excluded from getting closed. For extra paranoia close_range() comes with a flags argument. This can e.g. be used to implement extension. Once can imagine userspace wanting to stop at the first error instead of ignoring errors under certain circumstances. There might be other valid ideas in the future. In any case, a flag argument doesn't hurt and keeps us on the safe side. From an implementation side this is kept rather dumb. It saw some input from David and Jann but all nonsense is obviously my own! - Errors to close file descriptors are currently ignored. (Could be changed by setting a flag in the future if needed.) - __close_range() is a rather simplistic wrapper around __close_fd(). My reasoning behind this is based on the nature of how __close_fd() needs to release an fd. But maybe I misunderstood specifics: We take the files_lock and rcu-dereference the fdtable of the calling task, we find the entry in the fdtable, get the file and need to release files_lock before calling filp_close(). In the meantime the fdtable might have been altered so we can't just retake the spinlock and keep the old rcu-reference of the fdtable around. Instead we need to grab a fresh reference to the fdtable. If my reasoning is correct then there's really no point in fancyfying __close_range(): We just need to rcu-dereference the fdtable of the calling task once to cap the max_fd value correctly and then go on calling __close_fd() in a loop. /* References */ [1]: https://lore.kernel.org/lkml/20190516165021.GD17978@ZenIV.linux.org.uk/ [2]: https://github.com/python/cpython/blob/9e4f2f3a6b8ee995c365e86d976937c141d867f8/Modules/_posixsubprocess.c#L220 [3]: https://sourceware.org/bugzilla/show_bug.cgi?id=10353#c7 [4]: https://github.com/systemd/systemd/blob/5238e9575906297608ff802a27e2ff9effa3b338/src/basic/fd-util.c#L217 [5]: https://github.com/lxc/lxc/blob/ddf4b77e11a4d08f09b7b9cd13e593f8c047edc5/src/lxc/start.c#L236 [6]: https://sourceware.org/git/?p=glibc.git;a=blob;f=sysdeps/unix/sysv/linux/grantpt.c;h=2030e07fa6e652aac32c775b8c6e005844c3c4eb;hb=HEAD#l17 Note that this is an internal implementation that is not exported. Currently, libc seems to not provide an exported version of this because of missing kernel support to do this. Note, in a recent patch series Florian made grantpt() a nop thereby removing the code referenced here. [7]: https://github.com/rust-lang/rust/issues/12148 [8]: https://github.com/rust-lang/rust/blob/5f47c0613ed4eb46fca3633c1297364c09e5e451/src/libstd/sys/unix/process2.rs#L303-L308 Rust's solution is slightly different but is equally unperformant. Rust calls getdtablesize() which is a glibc library function that simply returns the current RLIMIT_NOFILE or OPEN_MAX values. Rust then goes on to call close() on each fd. That's obviously overkill for most tasks. Rarely, tasks - especially non-demons - hit RLIMIT_NOFILE or OPEN_MAX. Let's be nice and assume an unprivileged user with RLIMIT_NOFILE set to 1024. Even in this case, there's a very high chance that in the common case Rust is calling the close() syscall 1021 times pointlessly if the task just has 0, 1, and 2 open. Suggested-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com> Cc: Arnd Bergmann <arnd@arndb.de> Cc: Kyle Evans <self@kyle-evans.net> Cc: Jann Horn <jannh@google.com> Cc: David Howells <dhowells@redhat.com> Cc: Dmitry V. Levin <ldv@altlinux.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Florian Weimer <fweimer@redhat.com> Cc: linux-api@vger.kernel.org
2019-05-24 12:30:34 +03:00
return -EBADF;
return filp_close(file, files);
}
EXPORT_SYMBOL(close_fd); /* for ksys_close() */
/**
* last_fd - return last valid index into fd table
* @cur_fds: files struct
*
* Context: Either rcu read lock or files_lock must be held.
*
* Returns: Last valid index into fdtable.
*/
static inline unsigned last_fd(struct fdtable *fdt)
{
return fdt->max_fds - 1;
}
static inline void __range_cloexec(struct files_struct *cur_fds,
unsigned int fd, unsigned int max_fd)
{
struct fdtable *fdt;
/* make sure we're using the correct maximum value */
spin_lock(&cur_fds->file_lock);
fdt = files_fdtable(cur_fds);
max_fd = min(last_fd(fdt), max_fd);
if (fd <= max_fd)
bitmap_set(fdt->close_on_exec, fd, max_fd - fd + 1);
spin_unlock(&cur_fds->file_lock);
}
static inline void __range_close(struct files_struct *cur_fds, unsigned int fd,
unsigned int max_fd)
{
while (fd <= max_fd) {
struct file *file;
file = pick_file(cur_fds, fd++);
if (!IS_ERR(file)) {
/* found a valid file to close */
filp_close(file, cur_fds);
cond_resched();
continue;
}
/* beyond the last fd in that table */
if (PTR_ERR(file) == -EINVAL)
return;
}
}
open: add close_range() This adds the close_range() syscall. It allows to efficiently close a range of file descriptors up to all file descriptors of a calling task. I was contacted by FreeBSD as they wanted to have the same close_range() syscall as we proposed here. We've coordinated this and in the meantime, Kyle was fast enough to merge close_range() into FreeBSD already in April: https://reviews.freebsd.org/D21627 https://svnweb.freebsd.org/base?view=revision&revision=359836 and the current plan is to backport close_range() to FreeBSD 12.2 (cf. [2]) once its merged in Linux too. Python is in the process of switching to close_range() on FreeBSD and they are waiting on us to merge this to switch on Linux as well: https://bugs.python.org/issue38061 The syscall came up in a recent discussion around the new mount API and making new file descriptor types cloexec by default. During this discussion, Al suggested the close_range() syscall (cf. [1]). Note, a syscall in this manner has been requested by various people over time. First, it helps to close all file descriptors of an exec()ing task. This can be done safely via (quoting Al's example from [1] verbatim): /* that exec is sensitive */ unshare(CLONE_FILES); /* we don't want anything past stderr here */ close_range(3, ~0U); execve(....); The code snippet above is one way of working around the problem that file descriptors are not cloexec by default. This is aggravated by the fact that we can't just switch them over without massively regressing userspace. For a whole class of programs having an in-kernel method of closing all file descriptors is very helpful (e.g. demons, service managers, programming language standard libraries, container managers etc.). (Please note, unshare(CLONE_FILES) should only be needed if the calling task is multi-threaded and shares the file descriptor table with another thread in which case two threads could race with one thread allocating file descriptors and the other one closing them via close_range(). For the general case close_range() before the execve() is sufficient.) Second, it allows userspace to avoid implementing closing all file descriptors by parsing through /proc/<pid>/fd/* and calling close() on each file descriptor. From looking at various large(ish) userspace code bases this or similar patterns are very common in: - service managers (cf. [4]) - libcs (cf. [6]) - container runtimes (cf. [5]) - programming language runtimes/standard libraries - Python (cf. [2]) - Rust (cf. [7], [8]) As Dmitry pointed out there's even a long-standing glibc bug about missing kernel support for this task (cf. [3]). In addition, the syscall will also work for tasks that do not have procfs mounted and on kernels that do not have procfs support compiled in. In such situations the only way to make sure that all file descriptors are closed is to call close() on each file descriptor up to UINT_MAX or RLIMIT_NOFILE, OPEN_MAX trickery (cf. comment [8] on Rust). The performance is striking. For good measure, comparing the following simple close_all_fds() userspace implementation that is essentially just glibc's version in [6]: static int close_all_fds(void) { int dir_fd; DIR *dir; struct dirent *direntp; dir = opendir("/proc/self/fd"); if (!dir) return -1; dir_fd = dirfd(dir); while ((direntp = readdir(dir))) { int fd; if (strcmp(direntp->d_name, ".") == 0) continue; if (strcmp(direntp->d_name, "..") == 0) continue; fd = atoi(direntp->d_name); if (fd == dir_fd || fd == 0 || fd == 1 || fd == 2) continue; close(fd); } closedir(dir); return 0; } to close_range() yields: 1. closing 4 open files: - close_all_fds(): ~280 us - close_range(): ~24 us 2. closing 1000 open files: - close_all_fds(): ~5000 us - close_range(): ~800 us close_range() is designed to allow for some flexibility. Specifically, it does not simply always close all open file descriptors of a task. Instead, callers can specify an upper bound. This is e.g. useful for scenarios where specific file descriptors are created with well-known numbers that are supposed to be excluded from getting closed. For extra paranoia close_range() comes with a flags argument. This can e.g. be used to implement extension. Once can imagine userspace wanting to stop at the first error instead of ignoring errors under certain circumstances. There might be other valid ideas in the future. In any case, a flag argument doesn't hurt and keeps us on the safe side. From an implementation side this is kept rather dumb. It saw some input from David and Jann but all nonsense is obviously my own! - Errors to close file descriptors are currently ignored. (Could be changed by setting a flag in the future if needed.) - __close_range() is a rather simplistic wrapper around __close_fd(). My reasoning behind this is based on the nature of how __close_fd() needs to release an fd. But maybe I misunderstood specifics: We take the files_lock and rcu-dereference the fdtable of the calling task, we find the entry in the fdtable, get the file and need to release files_lock before calling filp_close(). In the meantime the fdtable might have been altered so we can't just retake the spinlock and keep the old rcu-reference of the fdtable around. Instead we need to grab a fresh reference to the fdtable. If my reasoning is correct then there's really no point in fancyfying __close_range(): We just need to rcu-dereference the fdtable of the calling task once to cap the max_fd value correctly and then go on calling __close_fd() in a loop. /* References */ [1]: https://lore.kernel.org/lkml/20190516165021.GD17978@ZenIV.linux.org.uk/ [2]: https://github.com/python/cpython/blob/9e4f2f3a6b8ee995c365e86d976937c141d867f8/Modules/_posixsubprocess.c#L220 [3]: https://sourceware.org/bugzilla/show_bug.cgi?id=10353#c7 [4]: https://github.com/systemd/systemd/blob/5238e9575906297608ff802a27e2ff9effa3b338/src/basic/fd-util.c#L217 [5]: https://github.com/lxc/lxc/blob/ddf4b77e11a4d08f09b7b9cd13e593f8c047edc5/src/lxc/start.c#L236 [6]: https://sourceware.org/git/?p=glibc.git;a=blob;f=sysdeps/unix/sysv/linux/grantpt.c;h=2030e07fa6e652aac32c775b8c6e005844c3c4eb;hb=HEAD#l17 Note that this is an internal implementation that is not exported. Currently, libc seems to not provide an exported version of this because of missing kernel support to do this. Note, in a recent patch series Florian made grantpt() a nop thereby removing the code referenced here. [7]: https://github.com/rust-lang/rust/issues/12148 [8]: https://github.com/rust-lang/rust/blob/5f47c0613ed4eb46fca3633c1297364c09e5e451/src/libstd/sys/unix/process2.rs#L303-L308 Rust's solution is slightly different but is equally unperformant. Rust calls getdtablesize() which is a glibc library function that simply returns the current RLIMIT_NOFILE or OPEN_MAX values. Rust then goes on to call close() on each fd. That's obviously overkill for most tasks. Rarely, tasks - especially non-demons - hit RLIMIT_NOFILE or OPEN_MAX. Let's be nice and assume an unprivileged user with RLIMIT_NOFILE set to 1024. Even in this case, there's a very high chance that in the common case Rust is calling the close() syscall 1021 times pointlessly if the task just has 0, 1, and 2 open. Suggested-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com> Cc: Arnd Bergmann <arnd@arndb.de> Cc: Kyle Evans <self@kyle-evans.net> Cc: Jann Horn <jannh@google.com> Cc: David Howells <dhowells@redhat.com> Cc: Dmitry V. Levin <ldv@altlinux.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Florian Weimer <fweimer@redhat.com> Cc: linux-api@vger.kernel.org
2019-05-24 12:30:34 +03:00
/**
* __close_range() - Close all file descriptors in a given range.
*
* @fd: starting file descriptor to close
* @max_fd: last file descriptor to close
*
* This closes a range of file descriptors. All file descriptors
* from @fd up to and including @max_fd are closed.
*/
int __close_range(unsigned fd, unsigned max_fd, unsigned int flags)
open: add close_range() This adds the close_range() syscall. It allows to efficiently close a range of file descriptors up to all file descriptors of a calling task. I was contacted by FreeBSD as they wanted to have the same close_range() syscall as we proposed here. We've coordinated this and in the meantime, Kyle was fast enough to merge close_range() into FreeBSD already in April: https://reviews.freebsd.org/D21627 https://svnweb.freebsd.org/base?view=revision&revision=359836 and the current plan is to backport close_range() to FreeBSD 12.2 (cf. [2]) once its merged in Linux too. Python is in the process of switching to close_range() on FreeBSD and they are waiting on us to merge this to switch on Linux as well: https://bugs.python.org/issue38061 The syscall came up in a recent discussion around the new mount API and making new file descriptor types cloexec by default. During this discussion, Al suggested the close_range() syscall (cf. [1]). Note, a syscall in this manner has been requested by various people over time. First, it helps to close all file descriptors of an exec()ing task. This can be done safely via (quoting Al's example from [1] verbatim): /* that exec is sensitive */ unshare(CLONE_FILES); /* we don't want anything past stderr here */ close_range(3, ~0U); execve(....); The code snippet above is one way of working around the problem that file descriptors are not cloexec by default. This is aggravated by the fact that we can't just switch them over without massively regressing userspace. For a whole class of programs having an in-kernel method of closing all file descriptors is very helpful (e.g. demons, service managers, programming language standard libraries, container managers etc.). (Please note, unshare(CLONE_FILES) should only be needed if the calling task is multi-threaded and shares the file descriptor table with another thread in which case two threads could race with one thread allocating file descriptors and the other one closing them via close_range(). For the general case close_range() before the execve() is sufficient.) Second, it allows userspace to avoid implementing closing all file descriptors by parsing through /proc/<pid>/fd/* and calling close() on each file descriptor. From looking at various large(ish) userspace code bases this or similar patterns are very common in: - service managers (cf. [4]) - libcs (cf. [6]) - container runtimes (cf. [5]) - programming language runtimes/standard libraries - Python (cf. [2]) - Rust (cf. [7], [8]) As Dmitry pointed out there's even a long-standing glibc bug about missing kernel support for this task (cf. [3]). In addition, the syscall will also work for tasks that do not have procfs mounted and on kernels that do not have procfs support compiled in. In such situations the only way to make sure that all file descriptors are closed is to call close() on each file descriptor up to UINT_MAX or RLIMIT_NOFILE, OPEN_MAX trickery (cf. comment [8] on Rust). The performance is striking. For good measure, comparing the following simple close_all_fds() userspace implementation that is essentially just glibc's version in [6]: static int close_all_fds(void) { int dir_fd; DIR *dir; struct dirent *direntp; dir = opendir("/proc/self/fd"); if (!dir) return -1; dir_fd = dirfd(dir); while ((direntp = readdir(dir))) { int fd; if (strcmp(direntp->d_name, ".") == 0) continue; if (strcmp(direntp->d_name, "..") == 0) continue; fd = atoi(direntp->d_name); if (fd == dir_fd || fd == 0 || fd == 1 || fd == 2) continue; close(fd); } closedir(dir); return 0; } to close_range() yields: 1. closing 4 open files: - close_all_fds(): ~280 us - close_range(): ~24 us 2. closing 1000 open files: - close_all_fds(): ~5000 us - close_range(): ~800 us close_range() is designed to allow for some flexibility. Specifically, it does not simply always close all open file descriptors of a task. Instead, callers can specify an upper bound. This is e.g. useful for scenarios where specific file descriptors are created with well-known numbers that are supposed to be excluded from getting closed. For extra paranoia close_range() comes with a flags argument. This can e.g. be used to implement extension. Once can imagine userspace wanting to stop at the first error instead of ignoring errors under certain circumstances. There might be other valid ideas in the future. In any case, a flag argument doesn't hurt and keeps us on the safe side. From an implementation side this is kept rather dumb. It saw some input from David and Jann but all nonsense is obviously my own! - Errors to close file descriptors are currently ignored. (Could be changed by setting a flag in the future if needed.) - __close_range() is a rather simplistic wrapper around __close_fd(). My reasoning behind this is based on the nature of how __close_fd() needs to release an fd. But maybe I misunderstood specifics: We take the files_lock and rcu-dereference the fdtable of the calling task, we find the entry in the fdtable, get the file and need to release files_lock before calling filp_close(). In the meantime the fdtable might have been altered so we can't just retake the spinlock and keep the old rcu-reference of the fdtable around. Instead we need to grab a fresh reference to the fdtable. If my reasoning is correct then there's really no point in fancyfying __close_range(): We just need to rcu-dereference the fdtable of the calling task once to cap the max_fd value correctly and then go on calling __close_fd() in a loop. /* References */ [1]: https://lore.kernel.org/lkml/20190516165021.GD17978@ZenIV.linux.org.uk/ [2]: https://github.com/python/cpython/blob/9e4f2f3a6b8ee995c365e86d976937c141d867f8/Modules/_posixsubprocess.c#L220 [3]: https://sourceware.org/bugzilla/show_bug.cgi?id=10353#c7 [4]: https://github.com/systemd/systemd/blob/5238e9575906297608ff802a27e2ff9effa3b338/src/basic/fd-util.c#L217 [5]: https://github.com/lxc/lxc/blob/ddf4b77e11a4d08f09b7b9cd13e593f8c047edc5/src/lxc/start.c#L236 [6]: https://sourceware.org/git/?p=glibc.git;a=blob;f=sysdeps/unix/sysv/linux/grantpt.c;h=2030e07fa6e652aac32c775b8c6e005844c3c4eb;hb=HEAD#l17 Note that this is an internal implementation that is not exported. Currently, libc seems to not provide an exported version of this because of missing kernel support to do this. Note, in a recent patch series Florian made grantpt() a nop thereby removing the code referenced here. [7]: https://github.com/rust-lang/rust/issues/12148 [8]: https://github.com/rust-lang/rust/blob/5f47c0613ed4eb46fca3633c1297364c09e5e451/src/libstd/sys/unix/process2.rs#L303-L308 Rust's solution is slightly different but is equally unperformant. Rust calls getdtablesize() which is a glibc library function that simply returns the current RLIMIT_NOFILE or OPEN_MAX values. Rust then goes on to call close() on each fd. That's obviously overkill for most tasks. Rarely, tasks - especially non-demons - hit RLIMIT_NOFILE or OPEN_MAX. Let's be nice and assume an unprivileged user with RLIMIT_NOFILE set to 1024. Even in this case, there's a very high chance that in the common case Rust is calling the close() syscall 1021 times pointlessly if the task just has 0, 1, and 2 open. Suggested-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com> Cc: Arnd Bergmann <arnd@arndb.de> Cc: Kyle Evans <self@kyle-evans.net> Cc: Jann Horn <jannh@google.com> Cc: David Howells <dhowells@redhat.com> Cc: Dmitry V. Levin <ldv@altlinux.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Florian Weimer <fweimer@redhat.com> Cc: linux-api@vger.kernel.org
2019-05-24 12:30:34 +03:00
{
struct task_struct *me = current;
struct files_struct *cur_fds = me->files, *fds = NULL;
if (flags & ~(CLOSE_RANGE_UNSHARE | CLOSE_RANGE_CLOEXEC))
return -EINVAL;
open: add close_range() This adds the close_range() syscall. It allows to efficiently close a range of file descriptors up to all file descriptors of a calling task. I was contacted by FreeBSD as they wanted to have the same close_range() syscall as we proposed here. We've coordinated this and in the meantime, Kyle was fast enough to merge close_range() into FreeBSD already in April: https://reviews.freebsd.org/D21627 https://svnweb.freebsd.org/base?view=revision&revision=359836 and the current plan is to backport close_range() to FreeBSD 12.2 (cf. [2]) once its merged in Linux too. Python is in the process of switching to close_range() on FreeBSD and they are waiting on us to merge this to switch on Linux as well: https://bugs.python.org/issue38061 The syscall came up in a recent discussion around the new mount API and making new file descriptor types cloexec by default. During this discussion, Al suggested the close_range() syscall (cf. [1]). Note, a syscall in this manner has been requested by various people over time. First, it helps to close all file descriptors of an exec()ing task. This can be done safely via (quoting Al's example from [1] verbatim): /* that exec is sensitive */ unshare(CLONE_FILES); /* we don't want anything past stderr here */ close_range(3, ~0U); execve(....); The code snippet above is one way of working around the problem that file descriptors are not cloexec by default. This is aggravated by the fact that we can't just switch them over without massively regressing userspace. For a whole class of programs having an in-kernel method of closing all file descriptors is very helpful (e.g. demons, service managers, programming language standard libraries, container managers etc.). (Please note, unshare(CLONE_FILES) should only be needed if the calling task is multi-threaded and shares the file descriptor table with another thread in which case two threads could race with one thread allocating file descriptors and the other one closing them via close_range(). For the general case close_range() before the execve() is sufficient.) Second, it allows userspace to avoid implementing closing all file descriptors by parsing through /proc/<pid>/fd/* and calling close() on each file descriptor. From looking at various large(ish) userspace code bases this or similar patterns are very common in: - service managers (cf. [4]) - libcs (cf. [6]) - container runtimes (cf. [5]) - programming language runtimes/standard libraries - Python (cf. [2]) - Rust (cf. [7], [8]) As Dmitry pointed out there's even a long-standing glibc bug about missing kernel support for this task (cf. [3]). In addition, the syscall will also work for tasks that do not have procfs mounted and on kernels that do not have procfs support compiled in. In such situations the only way to make sure that all file descriptors are closed is to call close() on each file descriptor up to UINT_MAX or RLIMIT_NOFILE, OPEN_MAX trickery (cf. comment [8] on Rust). The performance is striking. For good measure, comparing the following simple close_all_fds() userspace implementation that is essentially just glibc's version in [6]: static int close_all_fds(void) { int dir_fd; DIR *dir; struct dirent *direntp; dir = opendir("/proc/self/fd"); if (!dir) return -1; dir_fd = dirfd(dir); while ((direntp = readdir(dir))) { int fd; if (strcmp(direntp->d_name, ".") == 0) continue; if (strcmp(direntp->d_name, "..") == 0) continue; fd = atoi(direntp->d_name); if (fd == dir_fd || fd == 0 || fd == 1 || fd == 2) continue; close(fd); } closedir(dir); return 0; } to close_range() yields: 1. closing 4 open files: - close_all_fds(): ~280 us - close_range(): ~24 us 2. closing 1000 open files: - close_all_fds(): ~5000 us - close_range(): ~800 us close_range() is designed to allow for some flexibility. Specifically, it does not simply always close all open file descriptors of a task. Instead, callers can specify an upper bound. This is e.g. useful for scenarios where specific file descriptors are created with well-known numbers that are supposed to be excluded from getting closed. For extra paranoia close_range() comes with a flags argument. This can e.g. be used to implement extension. Once can imagine userspace wanting to stop at the first error instead of ignoring errors under certain circumstances. There might be other valid ideas in the future. In any case, a flag argument doesn't hurt and keeps us on the safe side. From an implementation side this is kept rather dumb. It saw some input from David and Jann but all nonsense is obviously my own! - Errors to close file descriptors are currently ignored. (Could be changed by setting a flag in the future if needed.) - __close_range() is a rather simplistic wrapper around __close_fd(). My reasoning behind this is based on the nature of how __close_fd() needs to release an fd. But maybe I misunderstood specifics: We take the files_lock and rcu-dereference the fdtable of the calling task, we find the entry in the fdtable, get the file and need to release files_lock before calling filp_close(). In the meantime the fdtable might have been altered so we can't just retake the spinlock and keep the old rcu-reference of the fdtable around. Instead we need to grab a fresh reference to the fdtable. If my reasoning is correct then there's really no point in fancyfying __close_range(): We just need to rcu-dereference the fdtable of the calling task once to cap the max_fd value correctly and then go on calling __close_fd() in a loop. /* References */ [1]: https://lore.kernel.org/lkml/20190516165021.GD17978@ZenIV.linux.org.uk/ [2]: https://github.com/python/cpython/blob/9e4f2f3a6b8ee995c365e86d976937c141d867f8/Modules/_posixsubprocess.c#L220 [3]: https://sourceware.org/bugzilla/show_bug.cgi?id=10353#c7 [4]: https://github.com/systemd/systemd/blob/5238e9575906297608ff802a27e2ff9effa3b338/src/basic/fd-util.c#L217 [5]: https://github.com/lxc/lxc/blob/ddf4b77e11a4d08f09b7b9cd13e593f8c047edc5/src/lxc/start.c#L236 [6]: https://sourceware.org/git/?p=glibc.git;a=blob;f=sysdeps/unix/sysv/linux/grantpt.c;h=2030e07fa6e652aac32c775b8c6e005844c3c4eb;hb=HEAD#l17 Note that this is an internal implementation that is not exported. Currently, libc seems to not provide an exported version of this because of missing kernel support to do this. Note, in a recent patch series Florian made grantpt() a nop thereby removing the code referenced here. [7]: https://github.com/rust-lang/rust/issues/12148 [8]: https://github.com/rust-lang/rust/blob/5f47c0613ed4eb46fca3633c1297364c09e5e451/src/libstd/sys/unix/process2.rs#L303-L308 Rust's solution is slightly different but is equally unperformant. Rust calls getdtablesize() which is a glibc library function that simply returns the current RLIMIT_NOFILE or OPEN_MAX values. Rust then goes on to call close() on each fd. That's obviously overkill for most tasks. Rarely, tasks - especially non-demons - hit RLIMIT_NOFILE or OPEN_MAX. Let's be nice and assume an unprivileged user with RLIMIT_NOFILE set to 1024. Even in this case, there's a very high chance that in the common case Rust is calling the close() syscall 1021 times pointlessly if the task just has 0, 1, and 2 open. Suggested-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com> Cc: Arnd Bergmann <arnd@arndb.de> Cc: Kyle Evans <self@kyle-evans.net> Cc: Jann Horn <jannh@google.com> Cc: David Howells <dhowells@redhat.com> Cc: Dmitry V. Levin <ldv@altlinux.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Florian Weimer <fweimer@redhat.com> Cc: linux-api@vger.kernel.org
2019-05-24 12:30:34 +03:00
if (fd > max_fd)
return -EINVAL;
if (flags & CLOSE_RANGE_UNSHARE) {
int ret;
unsigned int max_unshare_fds = NR_OPEN_MAX;
/*
* If the caller requested all fds to be made cloexec we always
* copy all of the file descriptors since they still want to
* use them.
*/
if (!(flags & CLOSE_RANGE_CLOEXEC)) {
/*
* If the requested range is greater than the current
* maximum, we're closing everything so only copy all
* file descriptors beneath the lowest file descriptor.
*/
rcu_read_lock();
if (max_fd >= last_fd(files_fdtable(cur_fds)))
max_unshare_fds = fd;
rcu_read_unlock();
}
ret = unshare_fd(CLONE_FILES, max_unshare_fds, &fds);
if (ret)
return ret;
/*
* We used to share our file descriptor table, and have now
* created a private one, make sure we're using it below.
*/
if (fds)
swap(cur_fds, fds);
}
if (flags & CLOSE_RANGE_CLOEXEC)
__range_cloexec(cur_fds, fd, max_fd);
else
__range_close(cur_fds, fd, max_fd);
open: add close_range() This adds the close_range() syscall. It allows to efficiently close a range of file descriptors up to all file descriptors of a calling task. I was contacted by FreeBSD as they wanted to have the same close_range() syscall as we proposed here. We've coordinated this and in the meantime, Kyle was fast enough to merge close_range() into FreeBSD already in April: https://reviews.freebsd.org/D21627 https://svnweb.freebsd.org/base?view=revision&revision=359836 and the current plan is to backport close_range() to FreeBSD 12.2 (cf. [2]) once its merged in Linux too. Python is in the process of switching to close_range() on FreeBSD and they are waiting on us to merge this to switch on Linux as well: https://bugs.python.org/issue38061 The syscall came up in a recent discussion around the new mount API and making new file descriptor types cloexec by default. During this discussion, Al suggested the close_range() syscall (cf. [1]). Note, a syscall in this manner has been requested by various people over time. First, it helps to close all file descriptors of an exec()ing task. This can be done safely via (quoting Al's example from [1] verbatim): /* that exec is sensitive */ unshare(CLONE_FILES); /* we don't want anything past stderr here */ close_range(3, ~0U); execve(....); The code snippet above is one way of working around the problem that file descriptors are not cloexec by default. This is aggravated by the fact that we can't just switch them over without massively regressing userspace. For a whole class of programs having an in-kernel method of closing all file descriptors is very helpful (e.g. demons, service managers, programming language standard libraries, container managers etc.). (Please note, unshare(CLONE_FILES) should only be needed if the calling task is multi-threaded and shares the file descriptor table with another thread in which case two threads could race with one thread allocating file descriptors and the other one closing them via close_range(). For the general case close_range() before the execve() is sufficient.) Second, it allows userspace to avoid implementing closing all file descriptors by parsing through /proc/<pid>/fd/* and calling close() on each file descriptor. From looking at various large(ish) userspace code bases this or similar patterns are very common in: - service managers (cf. [4]) - libcs (cf. [6]) - container runtimes (cf. [5]) - programming language runtimes/standard libraries - Python (cf. [2]) - Rust (cf. [7], [8]) As Dmitry pointed out there's even a long-standing glibc bug about missing kernel support for this task (cf. [3]). In addition, the syscall will also work for tasks that do not have procfs mounted and on kernels that do not have procfs support compiled in. In such situations the only way to make sure that all file descriptors are closed is to call close() on each file descriptor up to UINT_MAX or RLIMIT_NOFILE, OPEN_MAX trickery (cf. comment [8] on Rust). The performance is striking. For good measure, comparing the following simple close_all_fds() userspace implementation that is essentially just glibc's version in [6]: static int close_all_fds(void) { int dir_fd; DIR *dir; struct dirent *direntp; dir = opendir("/proc/self/fd"); if (!dir) return -1; dir_fd = dirfd(dir); while ((direntp = readdir(dir))) { int fd; if (strcmp(direntp->d_name, ".") == 0) continue; if (strcmp(direntp->d_name, "..") == 0) continue; fd = atoi(direntp->d_name); if (fd == dir_fd || fd == 0 || fd == 1 || fd == 2) continue; close(fd); } closedir(dir); return 0; } to close_range() yields: 1. closing 4 open files: - close_all_fds(): ~280 us - close_range(): ~24 us 2. closing 1000 open files: - close_all_fds(): ~5000 us - close_range(): ~800 us close_range() is designed to allow for some flexibility. Specifically, it does not simply always close all open file descriptors of a task. Instead, callers can specify an upper bound. This is e.g. useful for scenarios where specific file descriptors are created with well-known numbers that are supposed to be excluded from getting closed. For extra paranoia close_range() comes with a flags argument. This can e.g. be used to implement extension. Once can imagine userspace wanting to stop at the first error instead of ignoring errors under certain circumstances. There might be other valid ideas in the future. In any case, a flag argument doesn't hurt and keeps us on the safe side. From an implementation side this is kept rather dumb. It saw some input from David and Jann but all nonsense is obviously my own! - Errors to close file descriptors are currently ignored. (Could be changed by setting a flag in the future if needed.) - __close_range() is a rather simplistic wrapper around __close_fd(). My reasoning behind this is based on the nature of how __close_fd() needs to release an fd. But maybe I misunderstood specifics: We take the files_lock and rcu-dereference the fdtable of the calling task, we find the entry in the fdtable, get the file and need to release files_lock before calling filp_close(). In the meantime the fdtable might have been altered so we can't just retake the spinlock and keep the old rcu-reference of the fdtable around. Instead we need to grab a fresh reference to the fdtable. If my reasoning is correct then there's really no point in fancyfying __close_range(): We just need to rcu-dereference the fdtable of the calling task once to cap the max_fd value correctly and then go on calling __close_fd() in a loop. /* References */ [1]: https://lore.kernel.org/lkml/20190516165021.GD17978@ZenIV.linux.org.uk/ [2]: https://github.com/python/cpython/blob/9e4f2f3a6b8ee995c365e86d976937c141d867f8/Modules/_posixsubprocess.c#L220 [3]: https://sourceware.org/bugzilla/show_bug.cgi?id=10353#c7 [4]: https://github.com/systemd/systemd/blob/5238e9575906297608ff802a27e2ff9effa3b338/src/basic/fd-util.c#L217 [5]: https://github.com/lxc/lxc/blob/ddf4b77e11a4d08f09b7b9cd13e593f8c047edc5/src/lxc/start.c#L236 [6]: https://sourceware.org/git/?p=glibc.git;a=blob;f=sysdeps/unix/sysv/linux/grantpt.c;h=2030e07fa6e652aac32c775b8c6e005844c3c4eb;hb=HEAD#l17 Note that this is an internal implementation that is not exported. Currently, libc seems to not provide an exported version of this because of missing kernel support to do this. Note, in a recent patch series Florian made grantpt() a nop thereby removing the code referenced here. [7]: https://github.com/rust-lang/rust/issues/12148 [8]: https://github.com/rust-lang/rust/blob/5f47c0613ed4eb46fca3633c1297364c09e5e451/src/libstd/sys/unix/process2.rs#L303-L308 Rust's solution is slightly different but is equally unperformant. Rust calls getdtablesize() which is a glibc library function that simply returns the current RLIMIT_NOFILE or OPEN_MAX values. Rust then goes on to call close() on each fd. That's obviously overkill for most tasks. Rarely, tasks - especially non-demons - hit RLIMIT_NOFILE or OPEN_MAX. Let's be nice and assume an unprivileged user with RLIMIT_NOFILE set to 1024. Even in this case, there's a very high chance that in the common case Rust is calling the close() syscall 1021 times pointlessly if the task just has 0, 1, and 2 open. Suggested-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com> Cc: Arnd Bergmann <arnd@arndb.de> Cc: Kyle Evans <self@kyle-evans.net> Cc: Jann Horn <jannh@google.com> Cc: David Howells <dhowells@redhat.com> Cc: Dmitry V. Levin <ldv@altlinux.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Florian Weimer <fweimer@redhat.com> Cc: linux-api@vger.kernel.org
2019-05-24 12:30:34 +03:00
if (fds) {
/*
* We're done closing the files we were supposed to. Time to install
* the new file descriptor table and drop the old one.
*/
task_lock(me);
me->files = cur_fds;
task_unlock(me);
put_files_struct(fds);
}
open: add close_range() This adds the close_range() syscall. It allows to efficiently close a range of file descriptors up to all file descriptors of a calling task. I was contacted by FreeBSD as they wanted to have the same close_range() syscall as we proposed here. We've coordinated this and in the meantime, Kyle was fast enough to merge close_range() into FreeBSD already in April: https://reviews.freebsd.org/D21627 https://svnweb.freebsd.org/base?view=revision&revision=359836 and the current plan is to backport close_range() to FreeBSD 12.2 (cf. [2]) once its merged in Linux too. Python is in the process of switching to close_range() on FreeBSD and they are waiting on us to merge this to switch on Linux as well: https://bugs.python.org/issue38061 The syscall came up in a recent discussion around the new mount API and making new file descriptor types cloexec by default. During this discussion, Al suggested the close_range() syscall (cf. [1]). Note, a syscall in this manner has been requested by various people over time. First, it helps to close all file descriptors of an exec()ing task. This can be done safely via (quoting Al's example from [1] verbatim): /* that exec is sensitive */ unshare(CLONE_FILES); /* we don't want anything past stderr here */ close_range(3, ~0U); execve(....); The code snippet above is one way of working around the problem that file descriptors are not cloexec by default. This is aggravated by the fact that we can't just switch them over without massively regressing userspace. For a whole class of programs having an in-kernel method of closing all file descriptors is very helpful (e.g. demons, service managers, programming language standard libraries, container managers etc.). (Please note, unshare(CLONE_FILES) should only be needed if the calling task is multi-threaded and shares the file descriptor table with another thread in which case two threads could race with one thread allocating file descriptors and the other one closing them via close_range(). For the general case close_range() before the execve() is sufficient.) Second, it allows userspace to avoid implementing closing all file descriptors by parsing through /proc/<pid>/fd/* and calling close() on each file descriptor. From looking at various large(ish) userspace code bases this or similar patterns are very common in: - service managers (cf. [4]) - libcs (cf. [6]) - container runtimes (cf. [5]) - programming language runtimes/standard libraries - Python (cf. [2]) - Rust (cf. [7], [8]) As Dmitry pointed out there's even a long-standing glibc bug about missing kernel support for this task (cf. [3]). In addition, the syscall will also work for tasks that do not have procfs mounted and on kernels that do not have procfs support compiled in. In such situations the only way to make sure that all file descriptors are closed is to call close() on each file descriptor up to UINT_MAX or RLIMIT_NOFILE, OPEN_MAX trickery (cf. comment [8] on Rust). The performance is striking. For good measure, comparing the following simple close_all_fds() userspace implementation that is essentially just glibc's version in [6]: static int close_all_fds(void) { int dir_fd; DIR *dir; struct dirent *direntp; dir = opendir("/proc/self/fd"); if (!dir) return -1; dir_fd = dirfd(dir); while ((direntp = readdir(dir))) { int fd; if (strcmp(direntp->d_name, ".") == 0) continue; if (strcmp(direntp->d_name, "..") == 0) continue; fd = atoi(direntp->d_name); if (fd == dir_fd || fd == 0 || fd == 1 || fd == 2) continue; close(fd); } closedir(dir); return 0; } to close_range() yields: 1. closing 4 open files: - close_all_fds(): ~280 us - close_range(): ~24 us 2. closing 1000 open files: - close_all_fds(): ~5000 us - close_range(): ~800 us close_range() is designed to allow for some flexibility. Specifically, it does not simply always close all open file descriptors of a task. Instead, callers can specify an upper bound. This is e.g. useful for scenarios where specific file descriptors are created with well-known numbers that are supposed to be excluded from getting closed. For extra paranoia close_range() comes with a flags argument. This can e.g. be used to implement extension. Once can imagine userspace wanting to stop at the first error instead of ignoring errors under certain circumstances. There might be other valid ideas in the future. In any case, a flag argument doesn't hurt and keeps us on the safe side. From an implementation side this is kept rather dumb. It saw some input from David and Jann but all nonsense is obviously my own! - Errors to close file descriptors are currently ignored. (Could be changed by setting a flag in the future if needed.) - __close_range() is a rather simplistic wrapper around __close_fd(). My reasoning behind this is based on the nature of how __close_fd() needs to release an fd. But maybe I misunderstood specifics: We take the files_lock and rcu-dereference the fdtable of the calling task, we find the entry in the fdtable, get the file and need to release files_lock before calling filp_close(). In the meantime the fdtable might have been altered so we can't just retake the spinlock and keep the old rcu-reference of the fdtable around. Instead we need to grab a fresh reference to the fdtable. If my reasoning is correct then there's really no point in fancyfying __close_range(): We just need to rcu-dereference the fdtable of the calling task once to cap the max_fd value correctly and then go on calling __close_fd() in a loop. /* References */ [1]: https://lore.kernel.org/lkml/20190516165021.GD17978@ZenIV.linux.org.uk/ [2]: https://github.com/python/cpython/blob/9e4f2f3a6b8ee995c365e86d976937c141d867f8/Modules/_posixsubprocess.c#L220 [3]: https://sourceware.org/bugzilla/show_bug.cgi?id=10353#c7 [4]: https://github.com/systemd/systemd/blob/5238e9575906297608ff802a27e2ff9effa3b338/src/basic/fd-util.c#L217 [5]: https://github.com/lxc/lxc/blob/ddf4b77e11a4d08f09b7b9cd13e593f8c047edc5/src/lxc/start.c#L236 [6]: https://sourceware.org/git/?p=glibc.git;a=blob;f=sysdeps/unix/sysv/linux/grantpt.c;h=2030e07fa6e652aac32c775b8c6e005844c3c4eb;hb=HEAD#l17 Note that this is an internal implementation that is not exported. Currently, libc seems to not provide an exported version of this because of missing kernel support to do this. Note, in a recent patch series Florian made grantpt() a nop thereby removing the code referenced here. [7]: https://github.com/rust-lang/rust/issues/12148 [8]: https://github.com/rust-lang/rust/blob/5f47c0613ed4eb46fca3633c1297364c09e5e451/src/libstd/sys/unix/process2.rs#L303-L308 Rust's solution is slightly different but is equally unperformant. Rust calls getdtablesize() which is a glibc library function that simply returns the current RLIMIT_NOFILE or OPEN_MAX values. Rust then goes on to call close() on each fd. That's obviously overkill for most tasks. Rarely, tasks - especially non-demons - hit RLIMIT_NOFILE or OPEN_MAX. Let's be nice and assume an unprivileged user with RLIMIT_NOFILE set to 1024. Even in this case, there's a very high chance that in the common case Rust is calling the close() syscall 1021 times pointlessly if the task just has 0, 1, and 2 open. Suggested-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com> Cc: Arnd Bergmann <arnd@arndb.de> Cc: Kyle Evans <self@kyle-evans.net> Cc: Jann Horn <jannh@google.com> Cc: David Howells <dhowells@redhat.com> Cc: Dmitry V. Levin <ldv@altlinux.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Florian Weimer <fweimer@redhat.com> Cc: linux-api@vger.kernel.org
2019-05-24 12:30:34 +03:00
return 0;
}
/*
* See close_fd_get_file() below, this variant assumes current->files->file_lock
* is held.
*/
int __close_fd_get_file(unsigned int fd, struct file **res)
{
struct files_struct *files = current->files;
struct file *file;
struct fdtable *fdt;
fdt = files_fdtable(files);
if (fd >= fdt->max_fds)
goto out_err;
file = fdt->fd[fd];
if (!file)
goto out_err;
rcu_assign_pointer(fdt->fd[fd], NULL);
__put_unused_fd(files, fd);
get_file(file);
*res = file;
return 0;
out_err:
*res = NULL;
return -ENOENT;
}
/*
* variant of close_fd that gets a ref on the file for later fput.
* The caller must ensure that filp_close() called on the file, and then
* an fput().
*/
int close_fd_get_file(unsigned int fd, struct file **res)
{
struct files_struct *files = current->files;
int ret;
spin_lock(&files->file_lock);
ret = __close_fd_get_file(fd, res);
spin_unlock(&files->file_lock);
return ret;
}
void do_close_on_exec(struct files_struct *files)
{
unsigned i;
struct fdtable *fdt;
/* exec unshares first */
spin_lock(&files->file_lock);
for (i = 0; ; i++) {
unsigned long set;
unsigned fd = i * BITS_PER_LONG;
fdt = files_fdtable(files);
if (fd >= fdt->max_fds)
break;
set = fdt->close_on_exec[i];
if (!set)
continue;
fdt->close_on_exec[i] = 0;
for ( ; set ; fd++, set >>= 1) {
struct file *file;
if (!(set & 1))
continue;
file = fdt->fd[fd];
if (!file)
continue;
rcu_assign_pointer(fdt->fd[fd], NULL);
__put_unused_fd(files, fd);
spin_unlock(&files->file_lock);
filp_close(file, files);
cond_resched();
spin_lock(&files->file_lock);
}
}
spin_unlock(&files->file_lock);
}
fget: clarify and improve __fget_files() implementation Commit 054aa8d439b9 ("fget: check that the fd still exists after getting a ref to it") fixed a race with getting a reference to a file just as it was being closed. It was a fairly minimal patch, and I didn't think re-checking the file pointer lookup would be a measurable overhead, since it was all right there and cached. But I was wrong, as pointed out by the kernel test robot. The 'poll2' case of the will-it-scale.per_thread_ops benchmark regressed quite noticeably. Admittedly it seems to be a very artificial test: doing "poll()" system calls on regular files in a very tight loop in multiple threads. That means that basically all the time is spent just looking up file descriptors without ever doing anything useful with them (not that doing 'poll()' on a regular file is useful to begin with). And as a result it shows the extra "re-check fd" cost as a sore thumb. Happily, the regression is fixable by just writing the code to loook up the fd to be better and clearer. There's still a cost to verify the file pointer, but now it's basically in the noise even for that benchmark that does nothing else - and the code is more understandable and has better comments too. [ Side note: this patch is also a classic case of one that looks very messy with the default greedy Myers diff - it's much more legible with either the patience of histogram diff algorithm ] Link: https://lore.kernel.org/lkml/20211210053743.GA36420@xsang-OptiPlex-9020/ Link: https://lore.kernel.org/lkml/20211213083154.GA20853@linux.intel.com/ Reported-by: kernel test robot <oliver.sang@intel.com> Tested-by: Carel Si <beibei.si@intel.com> Cc: Jann Horn <jannh@google.com> Cc: Miklos Szeredi <mszeredi@redhat.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2021-12-11 01:00:15 +03:00
static inline struct file *__fget_files_rcu(struct files_struct *files,
unsigned int fd, fmode_t mask, unsigned int refs)
{
fget: clarify and improve __fget_files() implementation Commit 054aa8d439b9 ("fget: check that the fd still exists after getting a ref to it") fixed a race with getting a reference to a file just as it was being closed. It was a fairly minimal patch, and I didn't think re-checking the file pointer lookup would be a measurable overhead, since it was all right there and cached. But I was wrong, as pointed out by the kernel test robot. The 'poll2' case of the will-it-scale.per_thread_ops benchmark regressed quite noticeably. Admittedly it seems to be a very artificial test: doing "poll()" system calls on regular files in a very tight loop in multiple threads. That means that basically all the time is spent just looking up file descriptors without ever doing anything useful with them (not that doing 'poll()' on a regular file is useful to begin with). And as a result it shows the extra "re-check fd" cost as a sore thumb. Happily, the regression is fixable by just writing the code to loook up the fd to be better and clearer. There's still a cost to verify the file pointer, but now it's basically in the noise even for that benchmark that does nothing else - and the code is more understandable and has better comments too. [ Side note: this patch is also a classic case of one that looks very messy with the default greedy Myers diff - it's much more legible with either the patience of histogram diff algorithm ] Link: https://lore.kernel.org/lkml/20211210053743.GA36420@xsang-OptiPlex-9020/ Link: https://lore.kernel.org/lkml/20211213083154.GA20853@linux.intel.com/ Reported-by: kernel test robot <oliver.sang@intel.com> Tested-by: Carel Si <beibei.si@intel.com> Cc: Jann Horn <jannh@google.com> Cc: Miklos Szeredi <mszeredi@redhat.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2021-12-11 01:00:15 +03:00
for (;;) {
struct file *file;
struct fdtable *fdt = rcu_dereference_raw(files->fdt);
struct file __rcu **fdentry;
fget: clarify and improve __fget_files() implementation Commit 054aa8d439b9 ("fget: check that the fd still exists after getting a ref to it") fixed a race with getting a reference to a file just as it was being closed. It was a fairly minimal patch, and I didn't think re-checking the file pointer lookup would be a measurable overhead, since it was all right there and cached. But I was wrong, as pointed out by the kernel test robot. The 'poll2' case of the will-it-scale.per_thread_ops benchmark regressed quite noticeably. Admittedly it seems to be a very artificial test: doing "poll()" system calls on regular files in a very tight loop in multiple threads. That means that basically all the time is spent just looking up file descriptors without ever doing anything useful with them (not that doing 'poll()' on a regular file is useful to begin with). And as a result it shows the extra "re-check fd" cost as a sore thumb. Happily, the regression is fixable by just writing the code to loook up the fd to be better and clearer. There's still a cost to verify the file pointer, but now it's basically in the noise even for that benchmark that does nothing else - and the code is more understandable and has better comments too. [ Side note: this patch is also a classic case of one that looks very messy with the default greedy Myers diff - it's much more legible with either the patience of histogram diff algorithm ] Link: https://lore.kernel.org/lkml/20211210053743.GA36420@xsang-OptiPlex-9020/ Link: https://lore.kernel.org/lkml/20211213083154.GA20853@linux.intel.com/ Reported-by: kernel test robot <oliver.sang@intel.com> Tested-by: Carel Si <beibei.si@intel.com> Cc: Jann Horn <jannh@google.com> Cc: Miklos Szeredi <mszeredi@redhat.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2021-12-11 01:00:15 +03:00
if (unlikely(fd >= fdt->max_fds))
return NULL;
fdentry = fdt->fd + array_index_nospec(fd, fdt->max_fds);
file = rcu_dereference_raw(*fdentry);
if (unlikely(!file))
return NULL;
if (unlikely(file->f_mode & mask))
return NULL;
/*
* Ok, we have a file pointer. However, because we do
* this all locklessly under RCU, we may be racing with
* that file being closed.
*
* Such a race can take two forms:
*
* (a) the file ref already went down to zero,
* and get_file_rcu_many() fails. Just try
* again:
*/
if (unlikely(!get_file_rcu_many(file, refs)))
continue;
/*
* (b) the file table entry has changed under us.
* Note that we don't need to re-check the 'fdt->fd'
* pointer having changed, because it always goes
* hand-in-hand with 'fdt'.
*
* If so, we need to put our refs and try again.
*/
fget: clarify and improve __fget_files() implementation Commit 054aa8d439b9 ("fget: check that the fd still exists after getting a ref to it") fixed a race with getting a reference to a file just as it was being closed. It was a fairly minimal patch, and I didn't think re-checking the file pointer lookup would be a measurable overhead, since it was all right there and cached. But I was wrong, as pointed out by the kernel test robot. The 'poll2' case of the will-it-scale.per_thread_ops benchmark regressed quite noticeably. Admittedly it seems to be a very artificial test: doing "poll()" system calls on regular files in a very tight loop in multiple threads. That means that basically all the time is spent just looking up file descriptors without ever doing anything useful with them (not that doing 'poll()' on a regular file is useful to begin with). And as a result it shows the extra "re-check fd" cost as a sore thumb. Happily, the regression is fixable by just writing the code to loook up the fd to be better and clearer. There's still a cost to verify the file pointer, but now it's basically in the noise even for that benchmark that does nothing else - and the code is more understandable and has better comments too. [ Side note: this patch is also a classic case of one that looks very messy with the default greedy Myers diff - it's much more legible with either the patience of histogram diff algorithm ] Link: https://lore.kernel.org/lkml/20211210053743.GA36420@xsang-OptiPlex-9020/ Link: https://lore.kernel.org/lkml/20211213083154.GA20853@linux.intel.com/ Reported-by: kernel test robot <oliver.sang@intel.com> Tested-by: Carel Si <beibei.si@intel.com> Cc: Jann Horn <jannh@google.com> Cc: Miklos Szeredi <mszeredi@redhat.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2021-12-11 01:00:15 +03:00
if (unlikely(rcu_dereference_raw(files->fdt) != fdt) ||
unlikely(rcu_dereference_raw(*fdentry) != file)) {
fget: check that the fd still exists after getting a ref to it Jann Horn points out that there is another possible race wrt Unix domain socket garbage collection, somewhat reminiscent of the one fixed in commit cbcf01128d0a ("af_unix: fix garbage collect vs MSG_PEEK"). See the extended comment about the garbage collection requirements added to unix_peek_fds() by that commit for details. The race comes from how we can locklessly look up a file descriptor just as it is in the process of being closed, and with the right artificial timing (Jann added a few strategic 'mdelay(500)' calls to do that), the Unix domain socket garbage collector could see the reference count decrement of the close() happen before fget() took its reference to the file and the file was attached onto a new file descriptor. This is all (intentionally) correct on the 'struct file *' side, with RCU lookups and lockless reference counting very much part of the design. Getting that reference count out of order isn't a problem per se. But the garbage collector can get confused by seeing this situation of having seen a file not having any remaining external references and then seeing it being attached to an fd. In commit cbcf01128d0a ("af_unix: fix garbage collect vs MSG_PEEK") the fix was to serialize the file descriptor install with the garbage collector by taking and releasing the unix_gc_lock. That's not really an option here, but since this all happens when we are in the process of looking up a file descriptor, we can instead simply just re-check that the file hasn't been closed in the meantime, and just re-do the lookup if we raced with a concurrent close() of the same file descriptor. Reported-and-tested-by: Jann Horn <jannh@google.com> Acked-by: Miklos Szeredi <mszeredi@redhat.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2021-12-01 21:06:14 +03:00
fput_many(file, refs);
fget: clarify and improve __fget_files() implementation Commit 054aa8d439b9 ("fget: check that the fd still exists after getting a ref to it") fixed a race with getting a reference to a file just as it was being closed. It was a fairly minimal patch, and I didn't think re-checking the file pointer lookup would be a measurable overhead, since it was all right there and cached. But I was wrong, as pointed out by the kernel test robot. The 'poll2' case of the will-it-scale.per_thread_ops benchmark regressed quite noticeably. Admittedly it seems to be a very artificial test: doing "poll()" system calls on regular files in a very tight loop in multiple threads. That means that basically all the time is spent just looking up file descriptors without ever doing anything useful with them (not that doing 'poll()' on a regular file is useful to begin with). And as a result it shows the extra "re-check fd" cost as a sore thumb. Happily, the regression is fixable by just writing the code to loook up the fd to be better and clearer. There's still a cost to verify the file pointer, but now it's basically in the noise even for that benchmark that does nothing else - and the code is more understandable and has better comments too. [ Side note: this patch is also a classic case of one that looks very messy with the default greedy Myers diff - it's much more legible with either the patience of histogram diff algorithm ] Link: https://lore.kernel.org/lkml/20211210053743.GA36420@xsang-OptiPlex-9020/ Link: https://lore.kernel.org/lkml/20211213083154.GA20853@linux.intel.com/ Reported-by: kernel test robot <oliver.sang@intel.com> Tested-by: Carel Si <beibei.si@intel.com> Cc: Jann Horn <jannh@google.com> Cc: Miklos Szeredi <mszeredi@redhat.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2021-12-11 01:00:15 +03:00
continue;
fget: check that the fd still exists after getting a ref to it Jann Horn points out that there is another possible race wrt Unix domain socket garbage collection, somewhat reminiscent of the one fixed in commit cbcf01128d0a ("af_unix: fix garbage collect vs MSG_PEEK"). See the extended comment about the garbage collection requirements added to unix_peek_fds() by that commit for details. The race comes from how we can locklessly look up a file descriptor just as it is in the process of being closed, and with the right artificial timing (Jann added a few strategic 'mdelay(500)' calls to do that), the Unix domain socket garbage collector could see the reference count decrement of the close() happen before fget() took its reference to the file and the file was attached onto a new file descriptor. This is all (intentionally) correct on the 'struct file *' side, with RCU lookups and lockless reference counting very much part of the design. Getting that reference count out of order isn't a problem per se. But the garbage collector can get confused by seeing this situation of having seen a file not having any remaining external references and then seeing it being attached to an fd. In commit cbcf01128d0a ("af_unix: fix garbage collect vs MSG_PEEK") the fix was to serialize the file descriptor install with the garbage collector by taking and releasing the unix_gc_lock. That's not really an option here, but since this all happens when we are in the process of looking up a file descriptor, we can instead simply just re-check that the file hasn't been closed in the meantime, and just re-do the lookup if we raced with a concurrent close() of the same file descriptor. Reported-and-tested-by: Jann Horn <jannh@google.com> Acked-by: Miklos Szeredi <mszeredi@redhat.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2021-12-01 21:06:14 +03:00
}
fget: clarify and improve __fget_files() implementation Commit 054aa8d439b9 ("fget: check that the fd still exists after getting a ref to it") fixed a race with getting a reference to a file just as it was being closed. It was a fairly minimal patch, and I didn't think re-checking the file pointer lookup would be a measurable overhead, since it was all right there and cached. But I was wrong, as pointed out by the kernel test robot. The 'poll2' case of the will-it-scale.per_thread_ops benchmark regressed quite noticeably. Admittedly it seems to be a very artificial test: doing "poll()" system calls on regular files in a very tight loop in multiple threads. That means that basically all the time is spent just looking up file descriptors without ever doing anything useful with them (not that doing 'poll()' on a regular file is useful to begin with). And as a result it shows the extra "re-check fd" cost as a sore thumb. Happily, the regression is fixable by just writing the code to loook up the fd to be better and clearer. There's still a cost to verify the file pointer, but now it's basically in the noise even for that benchmark that does nothing else - and the code is more understandable and has better comments too. [ Side note: this patch is also a classic case of one that looks very messy with the default greedy Myers diff - it's much more legible with either the patience of histogram diff algorithm ] Link: https://lore.kernel.org/lkml/20211210053743.GA36420@xsang-OptiPlex-9020/ Link: https://lore.kernel.org/lkml/20211213083154.GA20853@linux.intel.com/ Reported-by: kernel test robot <oliver.sang@intel.com> Tested-by: Carel Si <beibei.si@intel.com> Cc: Jann Horn <jannh@google.com> Cc: Miklos Szeredi <mszeredi@redhat.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2021-12-11 01:00:15 +03:00
/*
* Ok, we have a ref to the file, and checked that it
* still exists.
*/
return file;
}
fget: clarify and improve __fget_files() implementation Commit 054aa8d439b9 ("fget: check that the fd still exists after getting a ref to it") fixed a race with getting a reference to a file just as it was being closed. It was a fairly minimal patch, and I didn't think re-checking the file pointer lookup would be a measurable overhead, since it was all right there and cached. But I was wrong, as pointed out by the kernel test robot. The 'poll2' case of the will-it-scale.per_thread_ops benchmark regressed quite noticeably. Admittedly it seems to be a very artificial test: doing "poll()" system calls on regular files in a very tight loop in multiple threads. That means that basically all the time is spent just looking up file descriptors without ever doing anything useful with them (not that doing 'poll()' on a regular file is useful to begin with). And as a result it shows the extra "re-check fd" cost as a sore thumb. Happily, the regression is fixable by just writing the code to loook up the fd to be better and clearer. There's still a cost to verify the file pointer, but now it's basically in the noise even for that benchmark that does nothing else - and the code is more understandable and has better comments too. [ Side note: this patch is also a classic case of one that looks very messy with the default greedy Myers diff - it's much more legible with either the patience of histogram diff algorithm ] Link: https://lore.kernel.org/lkml/20211210053743.GA36420@xsang-OptiPlex-9020/ Link: https://lore.kernel.org/lkml/20211213083154.GA20853@linux.intel.com/ Reported-by: kernel test robot <oliver.sang@intel.com> Tested-by: Carel Si <beibei.si@intel.com> Cc: Jann Horn <jannh@google.com> Cc: Miklos Szeredi <mszeredi@redhat.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2021-12-11 01:00:15 +03:00
}
static struct file *__fget_files(struct files_struct *files, unsigned int fd,
fmode_t mask, unsigned int refs)
{
struct file *file;
rcu_read_lock();
file = __fget_files_rcu(files, fd, mask, refs);
rcu_read_unlock();
return file;
}
static inline struct file *__fget(unsigned int fd, fmode_t mask,
unsigned int refs)
{
return __fget_files(current->files, fd, mask, refs);
}
struct file *fget_many(unsigned int fd, unsigned int refs)
{
return __fget(fd, FMODE_PATH, refs);
}
struct file *fget(unsigned int fd)
{
return __fget(fd, FMODE_PATH, 1);
}
EXPORT_SYMBOL(fget);
struct file *fget_raw(unsigned int fd)
{
return __fget(fd, 0, 1);
}
EXPORT_SYMBOL(fget_raw);
struct file *fget_task(struct task_struct *task, unsigned int fd)
{
struct file *file = NULL;
task_lock(task);
if (task->files)
file = __fget_files(task->files, fd, 0, 1);
task_unlock(task);
return file;
}
struct file *task_lookup_fd_rcu(struct task_struct *task, unsigned int fd)
{
/* Must be called with rcu_read_lock held */
struct files_struct *files;
struct file *file = NULL;
task_lock(task);
files = task->files;
if (files)
file = files_lookup_fd_rcu(files, fd);
task_unlock(task);
return file;
}
struct file *task_lookup_next_fd_rcu(struct task_struct *task, unsigned int *ret_fd)
{
/* Must be called with rcu_read_lock held */
struct files_struct *files;
unsigned int fd = *ret_fd;
struct file *file = NULL;
task_lock(task);
files = task->files;
if (files) {
for (; fd < files_fdtable(files)->max_fds; fd++) {
file = files_lookup_fd_rcu(files, fd);
if (file)
break;
}
}
task_unlock(task);
*ret_fd = fd;
return file;
}
/*
* Lightweight file lookup - no refcnt increment if fd table isn't shared.
*
* You can use this instead of fget if you satisfy all of the following
* conditions:
* 1) You must call fput_light before exiting the syscall and returning control
* to userspace (i.e. you cannot remember the returned struct file * after
* returning to userspace).
* 2) You must not call filp_close on the returned struct file * in between
* calls to fget_light and fput_light.
* 3) You must not clone the current task in between the calls to fget_light
* and fput_light.
*
* The fput_needed flag returned by fget_light should be passed to the
* corresponding fput_light.
*/
static unsigned long __fget_light(unsigned int fd, fmode_t mask)
{
struct files_struct *files = current->files;
struct file *file;
if (atomic_read(&files->count) == 1) {
file = files_lookup_fd_raw(files, fd);
if (!file || unlikely(file->f_mode & mask))
return 0;
return (unsigned long)file;
} else {
file = __fget(fd, mask, 1);
if (!file)
return 0;
return FDPUT_FPUT | (unsigned long)file;
}
}
unsigned long __fdget(unsigned int fd)
{
return __fget_light(fd, FMODE_PATH);
}
EXPORT_SYMBOL(__fdget);
unsigned long __fdget_raw(unsigned int fd)
{
return __fget_light(fd, 0);
}
unsigned long __fdget_pos(unsigned int fd)
{
unsigned long v = __fdget(fd);
struct file *file = (struct file *)(v & ~3);
Revert "vfs: properly and reliably lock f_pos in fdget_pos()" This reverts commit 0be0ee71816b2b6725e2b4f32ad6726c9d729777. I was hoping it would be benign to switch over entirely to FMODE_STREAM, and we'd have just a couple of small fixups we'd need, but it looks like we're not quite there yet. While it worked fine on both my desktop and laptop, they are fairly similar in other respects, and run mostly the same loads. Kenneth Crudup reports that it seems to break both his vmware installation and the KDE upower service. In both cases apparently leading to timeouts due to waitinmg for the f_pos lock. There are a number of character devices in particular that definitely want stream-like behavior, but that currently don't get marked as streams, and as a result get the exclusion between concurrent read()/write() on the same file descriptor. Which doesn't work well for them. The most obvious example if this is /dev/console and /dev/tty, which use console_fops and tty_fops respectively (and ptmx_fops for the pty master side). It may be that it's just this that causes problems, but we clearly weren't ready yet. Because there's a number of other likely common cases that don't have llseek implementations and would seem to act as stream devices: /dev/fuse (fuse_dev_operations) /dev/mcelog (mce_chrdev_ops) /dev/mei0 (mei_fops) /dev/net/tun (tun_fops) /dev/nvme0 (nvme_dev_fops) /dev/tpm0 (tpm_fops) /proc/self/ns/mnt (ns_file_operations) /dev/snd/pcm* (snd_pcm_f_ops[]) and while some of these could be trivially automatically detected by the vfs layer when the character device is opened by just noticing that they have no read or write operations either, it often isn't that obvious. Some character devices most definitely do use the file position, even if they don't allow seeking: the firmware update code, for example, uses simple_read_from_buffer() that does use f_pos, but doesn't allow seeking back and forth. We'll revisit this when there's a better way to detect the problem and fix it (possibly with a coccinelle script to do more of the FMODE_STREAM annotations). Reported-by: Kenneth R. Crudup <kenny@panix.com> Cc: Kirill Smelkov <kirr@nexedi.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-11-26 22:34:06 +03:00
if (file && (file->f_mode & FMODE_ATOMIC_POS)) {
if (file_count(file) > 1) {
v |= FDPUT_POS_UNLOCK;
mutex_lock(&file->f_pos_lock);
}
}
return v;
}
void __f_unlock_pos(struct file *f)
{
mutex_unlock(&f->f_pos_lock);
}
/*
* We only lock f_pos if we have threads or if the file might be
* shared with another process. In both cases we'll have an elevated
* file count (done either by fdget() or by fork()).
*/
void set_close_on_exec(unsigned int fd, int flag)
{
struct files_struct *files = current->files;
struct fdtable *fdt;
spin_lock(&files->file_lock);
fdt = files_fdtable(files);
if (flag)
__set_close_on_exec(fd, fdt);
else
__clear_close_on_exec(fd, fdt);
spin_unlock(&files->file_lock);
}
bool get_close_on_exec(unsigned int fd)
{
struct files_struct *files = current->files;
struct fdtable *fdt;
bool res;
rcu_read_lock();
fdt = files_fdtable(files);
res = close_on_exec(fd, fdt);
rcu_read_unlock();
return res;
}
static int do_dup2(struct files_struct *files,
struct file *file, unsigned fd, unsigned flags)
__releases(&files->file_lock)
{
struct file *tofree;
struct fdtable *fdt;
/*
* We need to detect attempts to do dup2() over allocated but still
* not finished descriptor. NB: OpenBSD avoids that at the price of
* extra work in their equivalent of fget() - they insert struct
* file immediately after grabbing descriptor, mark it larval if
* more work (e.g. actual opening) is needed and make sure that
* fget() treats larval files as absent. Potentially interesting,
* but while extra work in fget() is trivial, locking implications
* and amount of surgery on open()-related paths in VFS are not.
* FreeBSD fails with -EBADF in the same situation, NetBSD "solution"
* deadlocks in rather amusing ways, AFAICS. All of that is out of
* scope of POSIX or SUS, since neither considers shared descriptor
* tables and this condition does not arise without those.
*/
fdt = files_fdtable(files);
tofree = fdt->fd[fd];
if (!tofree && fd_is_open(fd, fdt))
goto Ebusy;
get_file(file);
rcu_assign_pointer(fdt->fd[fd], file);
__set_open_fd(fd, fdt);
if (flags & O_CLOEXEC)
__set_close_on_exec(fd, fdt);
else
__clear_close_on_exec(fd, fdt);
spin_unlock(&files->file_lock);
if (tofree)
filp_close(tofree, files);
return fd;
Ebusy:
spin_unlock(&files->file_lock);
return -EBUSY;
}
int replace_fd(unsigned fd, struct file *file, unsigned flags)
{
int err;
struct files_struct *files = current->files;
if (!file)
return close_fd(fd);
if (fd >= rlimit(RLIMIT_NOFILE))
return -EBADF;
spin_lock(&files->file_lock);
err = expand_files(files, fd);
if (unlikely(err < 0))
goto out_unlock;
return do_dup2(files, file, fd, flags);
out_unlock:
spin_unlock(&files->file_lock);
return err;
}
/**
* __receive_fd() - Install received file into file descriptor table
* @file: struct file that was received from another process
* @ufd: __user pointer to write new fd number to
* @o_flags: the O_* flags to apply to the new fd entry
*
* Installs a received file into the file descriptor table, with appropriate
* checks and count updates. Optionally writes the fd number to userspace, if
* @ufd is non-NULL.
*
* This helper handles its own reference counting of the incoming
* struct file.
*
* Returns newly install fd or -ve on error.
*/
int __receive_fd(struct file *file, int __user *ufd, unsigned int o_flags)
{
int new_fd;
int error;
error = security_file_receive(file);
if (error)
return error;
new_fd = get_unused_fd_flags(o_flags);
if (new_fd < 0)
return new_fd;
if (ufd) {
error = put_user(new_fd, ufd);
if (error) {
put_unused_fd(new_fd);
return error;
}
}
fd_install(new_fd, get_file(file));
__receive_sock(file);
return new_fd;
}
int receive_fd_replace(int new_fd, struct file *file, unsigned int o_flags)
{
int error;
error = security_file_receive(file);
if (error)
return error;
error = replace_fd(new_fd, file, o_flags);
if (error)
return error;
__receive_sock(file);
return new_fd;
}
int receive_fd(struct file *file, unsigned int o_flags)
{
return __receive_fd(file, NULL, o_flags);
}
EXPORT_SYMBOL_GPL(receive_fd);
static int ksys_dup3(unsigned int oldfd, unsigned int newfd, int flags)
{
int err = -EBADF;
struct file *file;
struct files_struct *files = current->files;
if ((flags & ~O_CLOEXEC) != 0)
return -EINVAL;
if (unlikely(oldfd == newfd))
return -EINVAL;
if (newfd >= rlimit(RLIMIT_NOFILE))
return -EBADF;
spin_lock(&files->file_lock);
err = expand_files(files, newfd);
file = files_lookup_fd_locked(files, oldfd);
if (unlikely(!file))
goto Ebadf;
if (unlikely(err < 0)) {
if (err == -EMFILE)
goto Ebadf;
goto out_unlock;
}
return do_dup2(files, file, newfd, flags);
Ebadf:
err = -EBADF;
out_unlock:
spin_unlock(&files->file_lock);
return err;
}
SYSCALL_DEFINE3(dup3, unsigned int, oldfd, unsigned int, newfd, int, flags)
{
return ksys_dup3(oldfd, newfd, flags);
}
SYSCALL_DEFINE2(dup2, unsigned int, oldfd, unsigned int, newfd)
{
if (unlikely(newfd == oldfd)) { /* corner case */
struct files_struct *files = current->files;
int retval = oldfd;
rcu_read_lock();
if (!files_lookup_fd_rcu(files, oldfd))
retval = -EBADF;
rcu_read_unlock();
return retval;
}
return ksys_dup3(oldfd, newfd, 0);
}
SYSCALL_DEFINE1(dup, unsigned int, fildes)
{
int ret = -EBADF;
struct file *file = fget_raw(fildes);
if (file) {
ret = get_unused_fd_flags(0);
if (ret >= 0)
fd_install(ret, file);
else
fput(file);
}
return ret;
}
int f_dupfd(unsigned int from, struct file *file, unsigned flags)
{
unsigned long nofile = rlimit(RLIMIT_NOFILE);
int err;
if (from >= nofile)
return -EINVAL;
err = alloc_fd(from, nofile, flags);
if (err >= 0) {
get_file(file);
fd_install(err, file);
}
return err;
}
int iterate_fd(struct files_struct *files, unsigned n,
int (*f)(const void *, struct file *, unsigned),
const void *p)
{
struct fdtable *fdt;
int res = 0;
if (!files)
return 0;
spin_lock(&files->file_lock);
for (fdt = files_fdtable(files); n < fdt->max_fds; n++) {
struct file *file;
file = rcu_dereference_check_fdtable(files, fdt->fd[n]);
if (!file)
continue;
res = f(p, file, n);
if (res)
break;
}
spin_unlock(&files->file_lock);
return res;
}
EXPORT_SYMBOL(iterate_fd);