The patch titled
Subject: revert "ipc_schedule_free() can do vfree() now"
has been added to the -mm tree. Its filename is
revert-ipc_schedule_free-can-do-vfree-now.patch
Before you just go and hit "reply", please:
a) Consider who else should be cc'ed
b) Prefer to cc a suitable mailing list as well
c) Ideally: find the original patch on the mailing list and do a
reply-to-all to that, adding suitable additional cc's
*** Remember to use Documentation/SubmitChecklist when testing your code ***
The -mm tree is included into linux-next and is updated
there every 3-4 working days
------------------------------------------------------
From: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
Subject: revert "ipc_schedule_free() can do vfree() now"
This massively screws up pending IPC patches.
Cc: Al Viro <viro@xxxxxxxxxxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
ipc/util.c | 66 ++++++++++++++++++++++++++++++++++++++++++---------
1 file changed, 55 insertions(+), 11 deletions(-)
diff -puN ipc/util.c~revert-ipc_schedule_free-can-do-vfree-now ipc/util.c
--- a/ipc/util.c~revert-ipc_schedule_free-can-do-vfree-now
+++ a/ipc/util.c
@@ -468,9 +468,10 @@ void ipc_free(void* ptr, int size)
/*
* rcu allocations:
- * There are two headers that are prepended to the actual allocation:
+ * There are three headers that are prepended to the actual allocation:
* - during use: ipc_rcu_hdr.
* - during the rcu grace period: ipc_rcu_grace.
+ * - [only if vmalloc]: ipc_rcu_sched.
* Their lifetime doesn't overlap, thus the headers share the same memory.
* Unlike a normal union, they are right-aligned, thus some container_of
* forward/backward casting is necessary:
@@ -478,6 +479,7 @@ void ipc_free(void* ptr, int size)
struct ipc_rcu_hdr
{
int refcount;
+ int is_vmalloc;
void *data[0];
};
@@ -489,8 +491,25 @@ struct ipc_rcu_grace
void *data[0];
};
-#define HDRLEN (sizeof(struct ipc_rcu_grace) > sizeof(struct ipc_rcu_hdr) ? \
+struct ipc_rcu_sched
+{
+ struct work_struct work;
+ /* "void *" makes sure alignment of following data is sane. */
+ void *data[0];
+};
+
+#define HDRLEN_KMALLOC (sizeof(struct ipc_rcu_grace) > sizeof(struct ipc_rcu_hdr) ? \
sizeof(struct ipc_rcu_grace) : sizeof(struct ipc_rcu_hdr))
+#define HDRLEN_VMALLOC (sizeof(struct ipc_rcu_sched) > HDRLEN_KMALLOC ? \
+ sizeof(struct ipc_rcu_sched) : HDRLEN_KMALLOC)
+
+static inline int rcu_use_vmalloc(int size)
+{
+ /* Too big for a single page? */
+ if (HDRLEN_KMALLOC + size > PAGE_SIZE)
+ return 1;
+ return 0;
+}
/**
* ipc_rcu_alloc - allocate ipc and rcu space
@@ -503,16 +522,27 @@ struct ipc_rcu_grace
void* ipc_rcu_alloc(int size)
{
- size_t len = size + HDRLEN;
- void *out;
+ void* out;
/*
- * We prepend the allocation with the rcu struct
+ * We prepend the allocation with the rcu struct, and
+ * workqueue if necessary (for vmalloc).
*/
- out = (len > PAGE_SIZE) ? vmalloc(len) : kmalloc(len, GFP_KERNEL);
- if (out) {
- out += HDRLEN;
- container_of(out, struct ipc_rcu_hdr, data)->refcount = 1;
+ if (rcu_use_vmalloc(size)) {
+ out = vmalloc(HDRLEN_VMALLOC + size);
+ if (out) {
+ out += HDRLEN_VMALLOC;
+ container_of(out, struct ipc_rcu_hdr, data)->is_vmalloc = 1;
+ container_of(out, struct ipc_rcu_hdr, data)->refcount = 1;
+ }
+ } else {
+ out = kmalloc(HDRLEN_KMALLOC + size, GFP_KERNEL);
+ if (out) {
+ out += HDRLEN_KMALLOC;
+ container_of(out, struct ipc_rcu_hdr, data)->is_vmalloc = 0;
+ container_of(out, struct ipc_rcu_hdr, data)->refcount = 1;
+ }
}
+
return out;
}
@@ -521,15 +551,29 @@ void ipc_rcu_getref(void *ptr)
container_of(ptr, struct ipc_rcu_hdr, data)->refcount++;
}
+static void ipc_do_vfree(struct work_struct *work)
+{
+ vfree(container_of(work, struct ipc_rcu_sched, work));
+}
+
/**
* ipc_schedule_free - free ipc + rcu space
* @head: RCU callback structure for queued work
+ *
+ * Since RCU callback function is called in bh,
+ * we need to defer the vfree to schedule_work().
*/
static void ipc_schedule_free(struct rcu_head *head)
{
struct ipc_rcu_grace *grace;
+ struct ipc_rcu_sched *sched;
+
grace = container_of(head, struct ipc_rcu_grace, rcu);
- vfree(grace);
+ sched = container_of(&(grace->data[0]), struct ipc_rcu_sched,
+ data[0]);
+
+ INIT_WORK(&sched->work, ipc_do_vfree);
+ schedule_work(&sched->work);
}
void ipc_rcu_putref(void *ptr)
@@ -537,7 +581,7 @@ void ipc_rcu_putref(void *ptr)
if (--container_of(ptr, struct ipc_rcu_hdr, data)->refcount > 0)
return;
- if (is_vmalloc_addr(ptr)) {
+ if (container_of(ptr, struct ipc_rcu_hdr, data)->is_vmalloc) {
call_rcu(&container_of(ptr, struct ipc_rcu_grace, data)->rcu,
ipc_schedule_free);
} else {
_
Patches currently in -mm which might be from akpm@xxxxxxxxxxxxxxxxxxxx are
origin.patch
clear_refs-sanitize-accepted-commands-declaration.patch
linux-next.patch
arch-alpha-kernel-systblss-remove-debug-check.patch
i-need-old-gcc.patch
drivers-usb-storage-realtek_crc-fix-build.patch
revert-ipc-dont-allocate-a-copy-larger-than-max.patch
sound-soc-codecs-si476xc-dont-use-0bnnn.patch
drivers-video-implement-a-simple-framebuffer-driver.patch
inotify-invalid-mask-should-return-a-error-number-but-not-set-it-fix.patch
posix-timers-correctly-get-dying-task-time-sample-in-posix_cpu_timer_schedule.patch
mm.patch
xen-tmem-enable-xen-tmem-shim-to-be-built-loaded-as-a-module-fix.patch
staging-zcache-enable-ramster-to-be-built-loaded-as-a-module-fix.patch
selftest-add-simple-test-for-soft-dirty-bit.patch
include-linux-mmzoneh-cleanups.patch
include-linux-mmzoneh-cleanups-fix.patch
drop_caches-add-some-documentation-and-info-messsge.patch
memcg-debugging-facility-to-access-dangling-memcgs-fix.patch
kernel-smpc-cleanups.patch
lib-string_helpers-introduce-generic-string_unescape-fix.patch
kernel-timerc-ove-some-non-timer-related-syscalls-to-kernel-sysc-checkpatch-fixes.patch
epoll-trim-epitem-by-one-cache-line-on-x86_64-fix.patch
binfmt_elfc-use-get_random_int-to-fix-entropy-depleting.patch
nilfs2-fix-issue-with-flush-kernel-thread-after-remount-in-ro-mode-because-of-drivers-internal-error-or-metadata-corruption-fix.patch
hfsplus-fix-warnings-in-fs-hfsplus-bfindc-in-function-hfs_find_1st_rec_by_cnid-fix.patch
fat-additions-to-support-fat_fallocate-fix.patch
usermodehelper-export-_exec-and-_setup-functions-fix.patch
kexec-use-min_t-to-simplify-logic-fix.patch
revert-ipc_schedule_free-can-do-vfree-now.patch
ipc-introduce-obtaining-a-lockless-ipc-object-fix.patch
ipcsem-do-not-hold-ipc-lock-more-than-necessary-fix-checkpatch-fixes.patch
ipcsem-open-code-and-rename-sem_lock-fix.patch
ipcsem-fine-grained-locking-for-semtimedop-ipc-make-refcounter-atomic-fix.patch
ipc-sysv-shared-memory-limited-to-8tib-fix.patch
rapidio-make-enumeration-discovery-configurable-fix.patch
kernel-pidc-improve-flow-of-a-loop-inside-alloc_pidmap-fix.patch
pid_namespacec-h-simplify-defines-fix.patch
drivers-net-rename-random32-to-prandom_u32-fix.patch
gadget-remove-only-user-of-aio-retry-checkpatch-fixes.patch
aio-remove-retry-based-aio-checkpatch-fixes.patch
aio-add-kiocb_cancel.patch
aio-make-aio_put_req-lockless-checkpatch-fixes.patch
aio-refcounting-cleanup-checkpatch-fixes.patch
wait-add-wait_event_hrtimeout.patch
aio-make-aio_read_evt-more-efficient-convert-to-hrtimers-checkpatch-fixes.patch
aio-use-cancellation-list-lazily.patch
aio-give-shared-kioctx-fields-their-own-cachelines.patch
generic-dynamic-per-cpu-refcounting.patch
generic-dynamic-per-cpu-refcounting-checkpatch-fixes.patch
aio-dont-include-aioh-in-schedh.patch
aio-dont-include-aioh-in-schedh-fix.patch
aio-kill-ki_retry.patch
aio-kill-ki_retry-fix.patch
aio-kill-ki_retry-checkpatch-fixes.patch
block-prep-work-for-batch-completion-checkpatch-fixes.patch
block-prep-work-for-batch-completion-fix-2.patch
block-prep-work-for-batch-completion-fix-3.patch
block-prep-work-for-batch-completion-fix-3-fix.patch
block-prep-work-for-batch-completion-fix-4.patch
block-prep-work-for-batch-completion-fix-99.patch
block-aio-batch-completion-for-bios-kiocbs.patch
block-aio-batch-completion-for-bios-kiocbs-checkpatch-fixes.patch
block-aio-batch-completion-for-bios-kiocbs-fix.patch
lib-add-lz4-compressor-module-fix.patch
crypto-add-lz4-cryptographic-api-fix.patch
bpf-add-comments-explaining-the-schedule_work-operation.patch
debugging-keep-track-of-page-owners-fix-2-fix.patch
debugging-keep-track-of-page-owners-fix-2-fix-fix-fix.patch
journal_add_journal_head-debug.patch
kernel-forkc-export-kernel_thread-to-modules.patch
mutex-subsystem-synchro-test-module.patch
slab-leaks3-default-y.patch
put_bh-debug.patch
--
To unsubscribe from this list: send the line "unsubscribe mm-commits" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
