From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from smtp.kernel.org (aws-us-west-2-korg-mail-1.web.codeaurora.org [10.30.226.201]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id D30293033FE for ; Tue, 10 Mar 2026 14:35:54 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=10.30.226.201 ARC-Seal:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1773153354; cv=none; b=jJ6xFDNJni1t4n8o11r6YNIPXdrwzu71W1MTfQfTE+KeCH6De9cLcyjrlhqrwCo/wthJqWpG7Do2QJXTofOYoauPh8ZJYy04TlwQWbcV1sSvyu3j8BPKNCiDmPUyJuJkeWBcsNDvrrFQmcXhkjfguq1zVbgjdCZaVjS4PYgbPko= ARC-Message-Signature:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1773153354; c=relaxed/simple; bh=pxqba9wmZxQ1nmSRgyrt1Z9u0BBfiTIH8Z8HAWwBMIY=; h=Message-ID:Date:From:To:Cc:Subject:References:MIME-Version: Content-Type; b=A2XMs3KFPnRexeC9wvyiReDWBmBFFAd5fTfPOJBnz0BogIRsBkYtW3paipRT9xx5fL7AQ2FFvYkR7jRVFzTlkZbWMvjxJ9zrTjoOb44iMdUb1zBhEeHDg/su0ihwrpjgeN1J16/HFZCmzpEQphh/d22WtnB5L9f6YE9fl1FE6po= ARC-Authentication-Results:i=1; smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=kernel.org header.i=@kernel.org header.b=e9Ys+0kp; arc=none smtp.client-ip=10.30.226.201 Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=kernel.org header.i=@kernel.org header.b="e9Ys+0kp" Received: by smtp.kernel.org (Postfix) with ESMTPSA id 91B64C2BCB8; Tue, 10 Mar 2026 14:35:54 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=k20201202; t=1773153354; bh=pxqba9wmZxQ1nmSRgyrt1Z9u0BBfiTIH8Z8HAWwBMIY=; h=Date:From:To:Cc:Subject:References:From; b=e9Ys+0kpPkBegDc91VSk9tM9lKBL2yRWbNPkEqGuKwYuHAgL4AwA5b6w9Dmyh8az+ 3qiUwsMsmK3sSZRsU+/LLC6RJhtPrQCU7dUVCGz6HhAR7yoylrjA0itIeLZPJmBxBG nj6iFq6YCmpNGjugvHAIWa9Jm25V5Mcr5Rtj2a2XuQ2slA5oNAmRcutUTN5DuMxnwi 2dIBRB9ClBZrehPYPa3rM62lA1p3g3r6O6uooIAJJa4fILnafkxOMz1nY4y+MyIvVy YwVn0uPWaUV37DaOHnbpm6roPntzzEIStTUpTeF7Qf3e74GYg26k5Y6YewOoPGhAiY oy0kyEk/TFU6g== Received: from rostedt by gandalf with local (Exim 4.99.1) (envelope-from ) id 1vzyBo-00000002Zfl-2ykJ; Tue, 10 Mar 2026 10:36:04 -0400 Message-ID: <20260310143604.573386927@kernel.org> User-Agent: quilt/0.69 Date: Tue, 10 Mar 2026 10:35:28 -0400 From: Steven Rostedt To: linux-kernel@vger.kernel.org Cc: Masami Hiramatsu , Mark Rutland , Mathieu Desnoyers , Andrew Morton , Vincent Donnefort Subject: [for-next][PATCH 13/18] tracing: Introduce simple_ring_buffer References: <20260310143515.132579088@kernel.org> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 From: Vincent Donnefort Add a simple implementation of the kernel ring-buffer. This intends to be used later by ring-buffer remotes such as the pKVM hypervisor, hence the need for a cut down version (write only) without any dependency. Link: https://patch.msgid.link/20260309162516.2623589-14-vdonnefort@google.com Reviewed-by: Steven Rostedt (Google) Signed-off-by: Vincent Donnefort Signed-off-by: Steven Rostedt (Google) --- include/linux/simple_ring_buffer.h | 57 ++++ kernel/trace/Kconfig | 3 + kernel/trace/Makefile | 1 + kernel/trace/simple_ring_buffer.c | 464 +++++++++++++++++++++++++++++ 4 files changed, 525 insertions(+) create mode 100644 include/linux/simple_ring_buffer.h create mode 100644 kernel/trace/simple_ring_buffer.c diff --git a/include/linux/simple_ring_buffer.h b/include/linux/simple_ring_buffer.h new file mode 100644 index 000000000000..2c4c0ae336bc --- /dev/null +++ b/include/linux/simple_ring_buffer.h @@ -0,0 +1,57 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef _LINUX_SIMPLE_RING_BUFFER_H +#define _LINUX_SIMPLE_RING_BUFFER_H + +#include +#include +#include +#include + +/* + * Ideally those struct would stay private but the caller needs to know + * the allocation size for simple_ring_buffer_init(). + */ +struct simple_buffer_page { + struct list_head link; + struct buffer_data_page *page; + u64 entries; + u32 write; + u32 id; +}; + +struct simple_rb_per_cpu { + struct simple_buffer_page *tail_page; + struct simple_buffer_page *reader_page; + struct simple_buffer_page *head_page; + struct simple_buffer_page *bpages; + struct trace_buffer_meta *meta; + u32 nr_pages; + +#define SIMPLE_RB_UNAVAILABLE 0 +#define SIMPLE_RB_READY 1 +#define SIMPLE_RB_WRITING 2 + u32 status; + + u64 last_overrun; + u64 write_stamp; + + struct simple_rb_cbs *cbs; +}; + +int simple_ring_buffer_init(struct simple_rb_per_cpu *cpu_buffer, struct simple_buffer_page *bpages, + const struct ring_buffer_desc *desc); + +void simple_ring_buffer_unload(struct simple_rb_per_cpu *cpu_buffer); + +void *simple_ring_buffer_reserve(struct simple_rb_per_cpu *cpu_buffer, unsigned long length, + u64 timestamp); + +void simple_ring_buffer_commit(struct simple_rb_per_cpu *cpu_buffer); + +int simple_ring_buffer_enable_tracing(struct simple_rb_per_cpu *cpu_buffer, bool enable); + +int simple_ring_buffer_reset(struct simple_rb_per_cpu *cpu_buffer); + +int simple_ring_buffer_swap_reader_page(struct simple_rb_per_cpu *cpu_buffer); + +#endif diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index 384dd36c8e29..edbdd7b38f61 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -1284,4 +1284,7 @@ source "kernel/trace/rv/Kconfig" config TRACE_REMOTE bool +config SIMPLE_RING_BUFFER + bool + endif # FTRACE diff --git a/kernel/trace/Makefile b/kernel/trace/Makefile index 318923ce39f5..2e39b09398b3 100644 --- a/kernel/trace/Makefile +++ b/kernel/trace/Makefile @@ -129,4 +129,5 @@ obj-$(CONFIG_TRACEPOINT_BENCHMARK) += trace_benchmark.o obj-$(CONFIG_RV) += rv/ obj-$(CONFIG_TRACE_REMOTE) += trace_remote.o +obj-$(CONFIG_SIMPLE_RING_BUFFER) += simple_ring_buffer.o libftrace-y := ftrace.o diff --git a/kernel/trace/simple_ring_buffer.c b/kernel/trace/simple_ring_buffer.c new file mode 100644 index 000000000000..15df9781411b --- /dev/null +++ b/kernel/trace/simple_ring_buffer.c @@ -0,0 +1,464 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2025 - Google LLC + * Author: Vincent Donnefort + */ + +#include +#include + +#include +#include + +enum simple_rb_link_type { + SIMPLE_RB_LINK_NORMAL = 0, + SIMPLE_RB_LINK_HEAD = 1, + SIMPLE_RB_LINK_HEAD_MOVING +}; + +#define SIMPLE_RB_LINK_MASK ~(SIMPLE_RB_LINK_HEAD | SIMPLE_RB_LINK_HEAD_MOVING) + +static void simple_bpage_set_head_link(struct simple_buffer_page *bpage) +{ + unsigned long link = (unsigned long)bpage->link.next; + + link &= SIMPLE_RB_LINK_MASK; + link |= SIMPLE_RB_LINK_HEAD; + + /* + * Paired with simple_rb_find_head() to order access between the head + * link and overrun. It ensures we always report an up-to-date value + * after swapping the reader page. + */ + smp_store_release(&bpage->link.next, (struct list_head *)link); +} + +static bool simple_bpage_unset_head_link(struct simple_buffer_page *bpage, + struct simple_buffer_page *dst, + enum simple_rb_link_type new_type) +{ + unsigned long *link = (unsigned long *)(&bpage->link.next); + unsigned long old = (*link & SIMPLE_RB_LINK_MASK) | SIMPLE_RB_LINK_HEAD; + unsigned long new = (unsigned long)(&dst->link) | new_type; + + return try_cmpxchg(link, &old, new); +} + +static void simple_bpage_set_normal_link(struct simple_buffer_page *bpage) +{ + unsigned long link = (unsigned long)bpage->link.next; + + WRITE_ONCE(bpage->link.next, (struct list_head *)(link & SIMPLE_RB_LINK_MASK)); +} + +static struct simple_buffer_page *simple_bpage_from_link(struct list_head *link) +{ + unsigned long ptr = (unsigned long)link & SIMPLE_RB_LINK_MASK; + + return container_of((struct list_head *)ptr, struct simple_buffer_page, link); +} + +static struct simple_buffer_page *simple_bpage_next_page(struct simple_buffer_page *bpage) +{ + return simple_bpage_from_link(bpage->link.next); +} + +static void simple_bpage_reset(struct simple_buffer_page *bpage) +{ + bpage->write = 0; + bpage->entries = 0; + + local_set(&bpage->page->commit, 0); +} + +static void simple_bpage_init(struct simple_buffer_page *bpage, unsigned long page) +{ + INIT_LIST_HEAD(&bpage->link); + bpage->page = (struct buffer_data_page *)page; + + simple_bpage_reset(bpage); +} + +#define simple_rb_meta_inc(__meta, __inc) \ + WRITE_ONCE((__meta), (__meta + __inc)) + +static bool simple_rb_loaded(struct simple_rb_per_cpu *cpu_buffer) +{ + return !!cpu_buffer->bpages; +} + +static int simple_rb_find_head(struct simple_rb_per_cpu *cpu_buffer) +{ + int retry = cpu_buffer->nr_pages * 2; + struct simple_buffer_page *head; + + head = cpu_buffer->head_page; + + while (retry--) { + unsigned long link; + +spin: + /* See smp_store_release in simple_bpage_set_head_link() */ + link = (unsigned long)smp_load_acquire(&head->link.prev->next); + + switch (link & ~SIMPLE_RB_LINK_MASK) { + /* Found the head */ + case SIMPLE_RB_LINK_HEAD: + cpu_buffer->head_page = head; + return 0; + /* The writer caught the head, we can spin, that won't be long */ + case SIMPLE_RB_LINK_HEAD_MOVING: + goto spin; + } + + head = simple_bpage_next_page(head); + } + + return -EBUSY; +} + +/** + * simple_ring_buffer_swap_reader_page - Swap ring-buffer head with the reader + * @cpu_buffer: A simple_rb_per_cpu + * + * This function enables consuming reading. It ensures the current head page will not be overwritten + * and can be safely read. + * + * Returns 0 on success, -ENODEV if @cpu_buffer was unloaded or -EBUSY if we failed to catch the + * head page. + */ +int simple_ring_buffer_swap_reader_page(struct simple_rb_per_cpu *cpu_buffer) +{ + struct simple_buffer_page *last, *head, *reader; + unsigned long overrun; + int retry = 8; + int ret; + + if (!simple_rb_loaded(cpu_buffer)) + return -ENODEV; + + reader = cpu_buffer->reader_page; + + do { + /* Run after the writer to find the head */ + ret = simple_rb_find_head(cpu_buffer); + if (ret) + return ret; + + head = cpu_buffer->head_page; + + /* Connect the reader page around the header page */ + reader->link.next = head->link.next; + reader->link.prev = head->link.prev; + + /* The last page before the head */ + last = simple_bpage_from_link(head->link.prev); + + /* The reader page points to the new header page */ + simple_bpage_set_head_link(reader); + + overrun = cpu_buffer->meta->overrun; + } while (!simple_bpage_unset_head_link(last, reader, SIMPLE_RB_LINK_NORMAL) && retry--); + + if (!retry) + return -EINVAL; + + cpu_buffer->head_page = simple_bpage_from_link(reader->link.next); + cpu_buffer->head_page->link.prev = &reader->link; + cpu_buffer->reader_page = head; + cpu_buffer->meta->reader.lost_events = overrun - cpu_buffer->last_overrun; + cpu_buffer->meta->reader.id = cpu_buffer->reader_page->id; + cpu_buffer->last_overrun = overrun; + + return 0; +} +EXPORT_SYMBOL_GPL(simple_ring_buffer_swap_reader_page); + +static struct simple_buffer_page *simple_rb_move_tail(struct simple_rb_per_cpu *cpu_buffer) +{ + struct simple_buffer_page *tail, *new_tail; + + tail = cpu_buffer->tail_page; + new_tail = simple_bpage_next_page(tail); + + if (simple_bpage_unset_head_link(tail, new_tail, SIMPLE_RB_LINK_HEAD_MOVING)) { + /* + * Oh no! we've caught the head. There is none anymore and + * swap_reader will spin until we set the new one. Overrun must + * be written first, to make sure we report the correct number + * of lost events. + */ + simple_rb_meta_inc(cpu_buffer->meta->overrun, new_tail->entries); + simple_rb_meta_inc(cpu_buffer->meta->pages_lost, 1); + + simple_bpage_set_head_link(new_tail); + simple_bpage_set_normal_link(tail); + } + + simple_bpage_reset(new_tail); + cpu_buffer->tail_page = new_tail; + + simple_rb_meta_inc(cpu_buffer->meta->pages_touched, 1); + + return new_tail; +} + +static unsigned long rb_event_size(unsigned long length) +{ + struct ring_buffer_event *event; + + return length + RB_EVNT_HDR_SIZE + sizeof(event->array[0]); +} + +static struct ring_buffer_event * +rb_event_add_ts_extend(struct ring_buffer_event *event, u64 delta) +{ + event->type_len = RINGBUF_TYPE_TIME_EXTEND; + event->time_delta = delta & TS_MASK; + event->array[0] = delta >> TS_SHIFT; + + return (struct ring_buffer_event *)((unsigned long)event + 8); +} + +static struct ring_buffer_event * +simple_rb_reserve_next(struct simple_rb_per_cpu *cpu_buffer, unsigned long length, u64 timestamp) +{ + unsigned long ts_ext_size = 0, event_size = rb_event_size(length); + struct simple_buffer_page *tail = cpu_buffer->tail_page; + struct ring_buffer_event *event; + u32 write, prev_write; + u64 time_delta; + + time_delta = timestamp - cpu_buffer->write_stamp; + + if (test_time_stamp(time_delta)) + ts_ext_size = 8; + + prev_write = tail->write; + write = prev_write + event_size + ts_ext_size; + + if (unlikely(write > (PAGE_SIZE - BUF_PAGE_HDR_SIZE))) + tail = simple_rb_move_tail(cpu_buffer); + + if (!tail->entries) { + tail->page->time_stamp = timestamp; + time_delta = 0; + ts_ext_size = 0; + write = event_size; + prev_write = 0; + } + + tail->write = write; + tail->entries++; + + cpu_buffer->write_stamp = timestamp; + + event = (struct ring_buffer_event *)(tail->page->data + prev_write); + if (ts_ext_size) { + event = rb_event_add_ts_extend(event, time_delta); + time_delta = 0; + } + + event->type_len = 0; + event->time_delta = time_delta; + event->array[0] = event_size - RB_EVNT_HDR_SIZE; + + return event; +} + +/** + * simple_ring_buffer_reserve - Reserve an entry in @cpu_buffer + * @cpu_buffer: A simple_rb_per_cpu + * @length: Size of the entry in bytes + * @timestamp: Timestamp of the entry + * + * Returns the address of the entry where to write data or NULL + */ +void *simple_ring_buffer_reserve(struct simple_rb_per_cpu *cpu_buffer, unsigned long length, + u64 timestamp) +{ + struct ring_buffer_event *rb_event; + + if (cmpxchg(&cpu_buffer->status, SIMPLE_RB_READY, SIMPLE_RB_WRITING) != SIMPLE_RB_READY) + return NULL; + + rb_event = simple_rb_reserve_next(cpu_buffer, length, timestamp); + + return &rb_event->array[1]; +} +EXPORT_SYMBOL_GPL(simple_ring_buffer_reserve); + +/** + * simple_ring_buffer_commit - Commit the entry reserved with simple_ring_buffer_reserve() + * @cpu_buffer: The simple_rb_per_cpu where the entry has been reserved + */ +void simple_ring_buffer_commit(struct simple_rb_per_cpu *cpu_buffer) +{ + local_set(&cpu_buffer->tail_page->page->commit, + cpu_buffer->tail_page->write); + simple_rb_meta_inc(cpu_buffer->meta->entries, 1); + + /* + * Paired with simple_rb_enable_tracing() to ensure data is + * written to the ring-buffer before teardown. + */ + smp_store_release(&cpu_buffer->status, SIMPLE_RB_READY); +} +EXPORT_SYMBOL_GPL(simple_ring_buffer_commit); + +static u32 simple_rb_enable_tracing(struct simple_rb_per_cpu *cpu_buffer, bool enable) +{ + u32 prev_status; + + if (enable) + return cmpxchg(&cpu_buffer->status, SIMPLE_RB_UNAVAILABLE, SIMPLE_RB_READY); + + /* Wait for the buffer to be released */ + do { + prev_status = cmpxchg_acquire(&cpu_buffer->status, + SIMPLE_RB_READY, + SIMPLE_RB_UNAVAILABLE); + } while (prev_status == SIMPLE_RB_WRITING); + + return prev_status; +} + +/** + * simple_ring_buffer_reset - Reset @cpu_buffer + * @cpu_buffer: A simple_rb_per_cpu + * + * This will not clear the content of the data, only reset counters and pointers + * + * Returns 0 on success or -ENODEV if @cpu_buffer was unloaded. + */ +int simple_ring_buffer_reset(struct simple_rb_per_cpu *cpu_buffer) +{ + struct simple_buffer_page *bpage; + u32 prev_status; + int ret; + + if (!simple_rb_loaded(cpu_buffer)) + return -ENODEV; + + prev_status = simple_rb_enable_tracing(cpu_buffer, false); + + ret = simple_rb_find_head(cpu_buffer); + if (ret) + return ret; + + bpage = cpu_buffer->tail_page = cpu_buffer->head_page; + do { + simple_bpage_reset(bpage); + bpage = simple_bpage_next_page(bpage); + } while (bpage != cpu_buffer->head_page); + + simple_bpage_reset(cpu_buffer->reader_page); + + cpu_buffer->last_overrun = 0; + cpu_buffer->write_stamp = 0; + + cpu_buffer->meta->reader.read = 0; + cpu_buffer->meta->reader.lost_events = 0; + cpu_buffer->meta->entries = 0; + cpu_buffer->meta->overrun = 0; + cpu_buffer->meta->read = 0; + cpu_buffer->meta->pages_lost = 0; + cpu_buffer->meta->pages_touched = 0; + + if (prev_status == SIMPLE_RB_READY) + simple_rb_enable_tracing(cpu_buffer, true); + + return 0; +} +EXPORT_SYMBOL_GPL(simple_ring_buffer_reset); + +/** + * simple_ring_buffer_init - Init @cpu_buffer based on @desc + * @cpu_buffer: A simple_rb_per_cpu buffer to init, allocated by the caller. + * @bpages: Array of simple_buffer_pages, with as many elements as @desc->nr_page_va + * @desc: A ring_buffer_desc + * + * Returns 0 on success or -EINVAL if the content of @desc is invalid + */ +int simple_ring_buffer_init(struct simple_rb_per_cpu *cpu_buffer, struct simple_buffer_page *bpages, + const struct ring_buffer_desc *desc) +{ + struct simple_buffer_page *bpage = bpages; + int i; + + /* At least 1 reader page and two pages in the ring-buffer */ + if (desc->nr_page_va < 3) + return -EINVAL; + + memset(cpu_buffer, 0, sizeof(*cpu_buffer)); + + cpu_buffer->bpages = bpages; + + cpu_buffer->meta = (void *)desc->meta_va; + memset(cpu_buffer->meta, 0, sizeof(*cpu_buffer->meta)); + cpu_buffer->meta->meta_page_size = PAGE_SIZE; + cpu_buffer->meta->nr_subbufs = cpu_buffer->nr_pages; + + /* The reader page is not part of the ring initially */ + simple_bpage_init(bpage, desc->page_va[0]); + bpage->id = 0; + + cpu_buffer->nr_pages = 1; + + cpu_buffer->reader_page = bpage; + cpu_buffer->tail_page = bpage + 1; + cpu_buffer->head_page = bpage + 1; + + for (i = 1; i < desc->nr_page_va; i++) { + simple_bpage_init(++bpage, desc->page_va[i]); + + bpage->link.next = &(bpage + 1)->link; + bpage->link.prev = &(bpage - 1)->link; + bpage->id = i; + + cpu_buffer->nr_pages = i + 1; + } + + /* Close the ring */ + bpage->link.next = &cpu_buffer->tail_page->link; + cpu_buffer->tail_page->link.prev = &bpage->link; + + /* The last init'ed page points to the head page */ + simple_bpage_set_head_link(bpage); + + return 0; +} +EXPORT_SYMBOL_GPL(simple_ring_buffer_init); + +/** + * simple_ring_buffer_unload - Prepare @cpu_buffer for deletion + * @cpu_buffer: A simple_rb_per_cpu that will be deleted. + */ +void simple_ring_buffer_unload(struct simple_rb_per_cpu *cpu_buffer) +{ + if (!simple_rb_loaded(cpu_buffer)) + return; + + simple_rb_enable_tracing(cpu_buffer, false); + + cpu_buffer->bpages = NULL; +} +EXPORT_SYMBOL_GPL(simple_ring_buffer_unload); + +/** + * simple_ring_buffer_enable_tracing - Enable or disable writing to @cpu_buffer + * @cpu_buffer: A simple_rb_per_cpu + * @enable: True to enable tracing, False to disable it + * + * Returns 0 on success or -ENODEV if @cpu_buffer was unloaded + */ +int simple_ring_buffer_enable_tracing(struct simple_rb_per_cpu *cpu_buffer, bool enable) +{ + if (!simple_rb_loaded(cpu_buffer)) + return -ENODEV; + + simple_rb_enable_tracing(cpu_buffer, enable); + + return 0; +} +EXPORT_SYMBOL_GPL(simple_ring_buffer_enable_tracing); -- 2.51.0