tracing: Clean up persistent ring buffer code
Subsystems (4)
Steven Rostedt
maintainerv5 cover tracing: Clean up persistent ring buffer code
Now that I learned that the memory passed back from reserve_mem is part of the memory allocator and just "reserved" and the memory is already virtually mapped, it can simply use phys_to_virt() on the physical memory that is returned to get the virtual mapping for that memory!
(Thanks Mike!)
That makes things much easier, especially since it means that the memory returned by reserve_mem is no different than the memory retrieved by page_alloc(). This allows that memory to be memory mapped to user space no differently than it is mapped by the normal buffer.
This new series does the following:
- Enforce the memory mapping is page aligned (both the address and the size). If not, it errors out.
- Use phys_to_virt() to get to the virtual memory from the reserve_mem returned addresses. As the memory is already freed via reserve_mem_release_by_name() and it's not mapped by vmap() anymore, the free ring buffer code doesn't need to do anything special for this mapping.
- Treat the buffer allocated via memmap differently. It still needs to be virtually mapped (cannot use phys_to_virt) and it must not be freed nor memory mapped to user space. A new flag is added when a buffer is created this way to prevent it from ever being memory mapped to user space and the ref count is upped so that it can never be freed.
- Use vmap_page_range() instead of using kmalloc_array() to create an array of struct pages for vmap().
- Use flush_kernel_vmap_range() instead of flush_dcache_folio()
Changes since v4: https://lore.kernel.org/linux-trace-kernel/[email protected]/
- Fixed calling free_reserve_area() twice on the same memory
Steven Rostedt (4):
tracing: Enforce the persistent ring buffer to be page aligned
tracing: Have reserve_mem use phys_to_virt() and separate from memmap buffer
tracing: Use vmap_page_range() to map memmap ring buffer
ring-buffer: Use flush_kernel_vmap_range() over flush_dcache_folio()
----
Documentation/admin-guide/kernel-parameters.txt | 2 + Documentation/trace/debugging.rst | 2 + kernel/trace/ring_buffer.c | 5 +- kernel/trace/trace.c | 68 ++++++++++++++++--------- kernel/trace/trace.h | 1 + 5 files changed, 52 insertions(+), 26 deletions(-)
Steven Rostedt
maintainerPATCH v5 1/4 tracing: Enforce the persistent ring buffer to be page aligned
From: Steven Rostedt <[email protected]>
Enforce that the address and the size of the memory used by the persistent ring buffer is page aligned. Also update the documentation to reflect this requirement.
--- Documentation/admin-guide/kernel-parameters.txt | 2 ++ Documentation/trace/debugging.rst | 2 ++ kernel/trace/trace.c | 12 ++++++++++++ 3 files changed, 16 insertions(+)
Documentation/admin-guide/kernel-parameters.txt +2 -0
@@ -7266,6 +7266,8 @@ This is just one of many ways that can clear memory. Make sure your system keeps the content of memory across reboots before relying on this option. + NB: Both the mapped address and size must be page aligned for the architecture. + See also Documentation/trace/debugging.rst
Documentation/trace/debugging.rst +2 -0
@@ -136,6 +136,8 @@ kernel, so only the same kernel is guaranteed to work if the mapping is preserved. Switching to a different kernel version may find a different layout and mark the buffer as invalid. +NB: Both the mapped address and size must be page aligned for the architecture. + Using trace_printk() in the boot instance ----------------------------------------- By default, the content of trace_printk() goes into the top level tracing
kernel/trace/trace.c +12 -0
@@ -10788,6 +10788,18 @@ __init static void enable_instances(void) }if (start) { +/* Start and size must be page aligned */ +if (start & ~PAGE_MASK ) { +pr_warn ("Tracing: mapping start addr %lx is not page aligned\n" , + (unsigned long )start); +continue ; + } +if (size & ~PAGE_MASK ) { +pr_warn ("Tracing: mapping size %lx is not page aligned\n" , + (unsigned long )size); +continue ; + } + addr =map_pages (start, size);if (addr) { pr_info("Tracing: mapped boot instance %s at physical memory %pa of size 0x%lx\n",
Signature · 1 line
2.47.2
Re: [PATCH v5 1/4] tracing: Enforce the persistent ring buffer to be page aligned
Quoted 51 lines · Steven Rostedt
From: Steven Rostedt <[email protected]>
Enforce that the address and the size of the memory used by the persistent ring buffer is page aligned. Also update the documentation to reflect this requirement.
Suggested-by Linus Torvalds <[email protected]>Signed-off-by Steven Rostedt (Google) <[email protected]>--- Documentation/admin-guide/kernel-parameters.txt | 2 ++ Documentation/trace/debugging.rst | 2 ++ kernel/trace/trace.c | 12 ++++++++++++ 3 files changed, 16 insertions(+)Documentation/admin-guide/kernel-parameters.txt +2
-0@@ -7266,6 +7266,8 @@ This is just one of many ways that can clear memory. Make sure your system keeps the content of memory across reboots before relying on this option. + NB: Both the mapped address and size must be page aligned for the architecture. + See also Documentation/trace/debugging.rstDocumentation/trace/debugging.rst +2
-0@@ -136,6 +136,8 @@ kernel, so only the same kernel is guaranteed to work if the mapping is preserved. Switching to a different kernel version may find a different layout and mark the buffer as invalid. +NB: Both the mapped address and size must be page aligned for the architecture. + Using trace_printk() in the boot instance ----------------------------------------- By default, the content of trace_printk() goes into the top level tracingkernel/trace/trace.c +1
-0@@ -10788,6 +10788,18 @@ __init static void enable_instances(void) }if (start) { +/* Start and size must be page aligned */
kernel/trace/trace.c __init static void enable_instances(void) · line 10801
Quoted 10 lines
kernel/trace/trace.c +10
-0+if (start & ~PAGE_MASK ) { +pr_warn ("Tracing: mapping start addr %lx is not page aligned\n" , + (unsigned long )start); +continue ; + } +if (size & ~PAGE_MASK ) { +pr_warn ("Tracing: mapping size %lx is not page aligned\n" , + (unsigned long )size); +continue ; + }
Better use %pa for printing physical address as on 32-bit systems phys_addr_t may be unsigned long long:
pr_warn("Tracing: mapping size %pa is not page aligned\n", &size);Quoted 8 lines
kernel/trace/trace.c +1
-0+ addr =map_pages (start, size);if (addr) { pr_info("Tracing: mapped boot instance %s at physical memory %pa of size 0x%lx\n",Signature · 1 line
2.47.2
Signature · 2 lines
Sincerely yours, Mike.
Steven Rostedt
maintainerRe: [PATCH v5 1/4] tracing: Enforce the persistent ring buffer to be page aligned
Quoted 10 lines · On Wed, 2 Apr 2025 12:21:49 +0300 Mike Rapoport <[email protected]> wrote:
kernel/trace/trace.c __init static void enable_instances(void) · line 10801
Quoted 5 lines
kernel/trace/trace.c +5
-0+if (size & ~PAGE_MASK ) { +pr_warn ("Tracing: mapping size %lx is not page aligned\n" , + (unsigned long )size); +continue ; + }Better use %pa for printing physical address as on 32-bit systems
phys_addr_tmay be unsigned long long:pr_warn("Tracing: mapping size %pa is not page aligned\n", &size);
Thanks, will update.
-- Steve
Mathieu Desnoyers
reviewerRe: [PATCH v5 1/4] tracing: Enforce the persistent ring buffer to be page aligned
In reply to Mike RapoportQuoted 6 lines · Mike Rapoport
Quoted 5 lines · Steven Rostedt
From: Steven Rostedt <[email protected]>
Enforce that the address and the size of the memory used by the persistent ring buffer is page aligned. Also update the documentation to reflect this requirement.
I've been loosely following this thread, and I'm confused about one thing.
AFAIU the goal is to have the ftrace persistent ring buffer written to through a memory range mapped by vmap_page_range(), and userspace maps the buffer with its own virtual mappings.
With respect to architectures with aliasing dcache, is the plan:
A) To make sure all persistent ring buffer mappings are aligned on SHMLBA:
Quoting "Documentation/core-api/cachetlb.rst":
Is your port susceptible to virtual aliasing in its D-cache? Well, if your D-cache is virtually indexed, is larger in size than PAGE_SIZE, and does not prevent multiple cache lines for the same physical address from existing at once, you have this problem.
If your D-cache has this problem, first define asm/shmparam.h SHMLBA properly, it should essentially be the size of your virtually addressed D-cache (or if the size is variable, the largest possible size). This setting will force the SYSv IPC layer to only allow user processes to mmap shared memory at address which are a multiple of this value.
or
B) to flush both the kernel and userspace mappings when a ring buffer page is handed over from writer to reader ?
I've seen both approaches being discussed in the recent threads, with some participants recommending approach (A), but then the code revisions that follow take approach (B).
AFAIU, it we are aiming for approach (A), then I'm missing where vmap_page_range() guarantees that the _kernel_ virtual mapping is SHMLBA aligned. AFAIU, only user mappings are aligned on SHMLBA.
And if we aiming towards approach (A), then the explicit flushing is not needed when handing over pages from writer to reader.
Please let me know if I'm missing something,
Thanks,
Mathieu
Quoted 71 lines
Quoted 46 lines
Suggested-by Linus Torvalds <[email protected]>Signed-off-by Steven Rostedt (Google) <[email protected]>--- Documentation/admin-guide/kernel-parameters.txt | 2 ++ Documentation/trace/debugging.rst | 2 ++ kernel/trace/trace.c | 12 ++++++++++++ 3 files changed, 16 insertions(+)Documentation/admin-guide/kernel-parameters.txt +2
-0@@ -7266,6 +7266,8 @@ This is just one of many ways that can clear memory. Make sure your system keeps the content of memory across reboots before relying on this option. + NB: Both the mapped address and size must be page aligned for the architecture. + See also Documentation/trace/debugging.rstDocumentation/trace/debugging.rst +2
-0@@ -136,6 +136,8 @@ kernel, so only the same kernel is guaranteed to work if the mapping is preserved. Switching to a different kernel version may find a different layout and mark the buffer as invalid. +NB: Both the mapped address and size must be page aligned for the architecture. + Using trace_printk() in the boot instance ----------------------------------------- By default, the content of trace_printk() goes into the top level tracingkernel/trace/trace.c +1
-0@@ -10788,6 +10788,18 @@ __init static void enable_instances(void) }if (start) { +/* Start and size must be page aligned */ kernel/trace/trace.c __init static void enable_instances(void) · line 10801
Quoted 10 lines
kernel/trace/trace.c +10
-0+if (start & ~PAGE_MASK ) { +pr_warn ("Tracing: mapping start addr %lx is not page aligned\n" , + (unsigned long )start); +continue ; + } +if (size & ~PAGE_MASK ) { +pr_warn ("Tracing: mapping size %lx is not page aligned\n" , + (unsigned long )size); +continue ; + }Better use %pa for printing physical address as on 32-bit systems
phys_addr_tmay be unsigned long long:pr_warn("Tracing: mapping size %pa is not page aligned\n", &size);Quoted 8 lines
kernel/trace/trace.c +1
-0+ addr =map_pages (start, size);if (addr) { pr_info("Tracing: mapped boot instance %s at physical memory %pa of size 0x%lx\n",Signature · 1 line
2.47.2
Signature · 3 lines
Mathieu Desnoyers EfficiOS Inc. https://www.efficios.com
Mathieu Desnoyers
reviewerRe: [PATCH v5 1/4] tracing: Enforce the persistent ring buffer to be page aligned
Quoted 8 lines · Mathieu Desnoyers
Quoted 6 lines · Mike Rapoport
Quoted 5 lines · Steven Rostedt
From: Steven Rostedt <[email protected]>
Enforce that the address and the size of the memory used by the persistent ring buffer is page aligned. Also update the documentation to reflect this requirement.
[ Please disregard this duplicate message and consider
https://lore.kernel.org/lkml/[email protected]/T/#m461e6111397b33c037f6fb746ed74ffbd0a4340f
instead. ]
Quoted 121 lines
I've been loosely following this thread, and I'm confused about one thing.
AFAIU the goal is to have the ftrace persistent ring buffer written to through a memory range mapped by
vmap_page_range(), and userspace maps the buffer with its own virtual mappings.With respect to architectures with aliasing dcache, is the plan:
A) To make sure all persistent ring buffer mappings are aligned on SHMLBA:
Quoting "
Documentation/core-api/cachetlb.rst":Is your port susceptible to virtual aliasing in its D-cache? Well, if your D-cache is virtually indexed, is larger in size than
PAGE_SIZE, and does not prevent multiple cache lines for the same physical address from existing at once, you have this problem.If your D-cache has this problem, first define
asm/shmparam.hSHMLBA properly, it should essentially be the size of your virtually addressed D-cache (or if the size is variable, the largest possible size). This setting will force the SYSv IPC layer to only allow user processes to mmap shared memory at address which are a multiple of this value.or
B) to flush both the kernel and userspace mappings when a ring buffer page is handed over from writer to reader ?
I've seen both approaches being discussed in the recent threads, with some participants recommending approach (A), but then the code revisions that follow take approach (B).
AFAIU, it we are aiming for approach (A), then I'm missing where
vmap_page_range()guarantees that the _kernel_ virtual mapping is SHMLBA aligned. AFAIU, only user mappings are aligned on SHMLBA.And if we aiming towards approach (A), then the explicit flushing is not needed when handing over pages from writer to reader.
Please let me know if I'm missing something,
Thanks,
Mathieu
Quoted 71 lines
Quoted 46 lines
Suggested-by Linus Torvalds <[email protected]>Signed-off-by Steven Rostedt (Google) <[email protected]>---
Documentation/admin-guide/kernel-parameters.txt| 2 ++ Documentation/trace/debugging.rst | 2 ++ kernel/trace/trace.c | 12 ++++++++++++3 files changed, 16 insertions(+)Documentation/admin-guide/kernel-parameters.txt +0
-0@@ -7266,6 +7266,8 @@This is just one of many ways that can clear memory. Make sure your system keeps the content of memory across reboots before relying on this option. + NB: Both the mapped address and size must be page
aligned for the architecture.
+
See alsoDocumentation/trace/debugging.rstDocumentation/trace/debugging.rst +0
-0@@ -136,6 +136,8 @@ kernel, so only the same kernel is guaranteed towork if the mapping is
preserved. Switching to a different kernel version may find a different layout and mark the buffer as invalid.
+NB: Both the mapped address and size must be page aligned for the architecture.
+
Usingtrace_printk()in the boot instance
-----------------------------------------
By default, the content oftrace_printk()goes into the top level tracingkernel/trace/trace.c +3
-0@@ -10788,6 +10788,18 @@ __init static void enable_instances(void) } if (start) { + /* Start and size must be page aligned */ + if (start & ~PAGE_MASK) { + pr_warn("Tracing: mapping start addr %lx is not pagekernel/trace/trace.c __init static void enable_instances(void) · line 10802
Quoted 10 lines
kernel/trace/trace.c +8
-0aligned\n", + (unsigned long)start); + continue; + } + if (size & ~PAGE_MASK) { + pr_warn("Tracing: mapping size %lx is not page aligned\n", + (unsigned long)size); + continue; + }Better use %pa for printing physical address as on 32-bit systems
phys_addr_tmay be unsigned long long:pr_warn("Tracing: mapping size %pa is not page aligned\n", &size);
Quoted 8 lines
kernel/trace/trace.c +1
-0+ addr = map_pages(start, size); if (addr) {pr_info("Tracing: mapped boot instance %s at physical memory %pa of size 0x%lx\n",
Signature · 1 line
2.47.2
Signature · 3 lines
Mathieu Desnoyers EfficiOS Inc. https://www.efficios.com
Steven Rostedt
maintainerPATCH v5 2/4 tracing: Have reserve_mem use phys_to_virt() and separate from memmap buffer
In reply to Steven RostedtFrom: Steven Rostedt <[email protected]>
The reserve_mem kernel command line option may pass back a physical address, but the memory is still part of the normal memory just like using memblock_reserve() would be. This means that the physical memory returned by the reserve_mem command line option can be converted directly to virtual memory by simply using phys_to_virt().
When freeing the buffer there's no need to call vunmap() anymore as the memory allocated by reserve_mem is freed by the call to reserve_mem_release_by_name().
Because the persistent ring buffer can also be allocated via the memmap option, which *is* different than normal memory as it cannot be added back to the buddy system, it must be treated differently. It still needs to be virtually mapped to have access to it. It also can not be freed nor can it ever be memory mapped to user space.
Create a new trace_array flag called TRACE_ARRAY_FL_MEMMAP which gets set if the buffer is created by the memmap option, and this will prevent the buffer from being memory mapped by user space.
Also increment the ref count for memmap'ed buffers so that they can never be freed.
--- kernel/trace/trace.c | 23 ++++++++++++++++------- kernel/trace/trace.h | 1 + 2 files changed, 17 insertions(+), 7 deletions(-)
kernel/trace/trace.c +16 -7
@@ -8505,6 +8505,10 @@ static int tracing_buffers_mmap(struct file *filp, struct vm_area_struct *vma)struct trace_iterator *iter = &info->iter;int ret =0 ; +/* A memmap'ed buffer is not supported for user space mmap */ +if (iter->tr->flags &TRACE_ARRAY_FL_MEMMAP ) +return -ENODEV ; +/* Currently the boot mapped buffer is not supported for mmap */ if (iter->tr->flags &TRACE_ARRAY_FL_BOOT )return -ENODEV ;
@@ -9614,9 +9618,6 @@ static void free_trace_buffers(struct trace_array *tr)#ifdef CONFIG_TRACER_MAX_TRACE free_trace_buffer (&tr->max_buffer);#endif - -if (tr->range_addr_start) -vunmap ((void *)tr->range_addr_start); }static void init_trace_flags_index (struct trace_array *tr)
@@ -10710,6 +10711,7 @@ static inline void do_allocate_snapshot(const char *name) { }__init static void enable_instances (void ) {struct trace_array *tr; +bool memmap_area = false;char *curr_str;char *name;char *str;
@@ -10778,6 +10780,7 @@ __init static void enable_instances(void) name);continue ; } + memmap_area = true; }else if (tok) {if (!reserve_mem_find_by_name (tok, &start, &size)) { start =0 ;
@@ -10800,7 +10803,10 @@ __init static void enable_instances(void)continue ; } - addr =map_pages (start, size); +if (memmap_area) + addr =map_pages (start, size); +else + addr = (unsigned long )phys_to_virt (start);if (addr) {pr_info ("Tracing: mapped boot instance %s at physical memory %pa of size 0x%lx\n" , name, &start, (unsigned long )size);
@@ -10827,10 +10833,13 @@ __init static void enable_instances(void)update_printk_trace (tr);/* -* If start is set, then this is a mapped buffer, and -* cannot be deleted by user space, so keep the reference -* to it. +* memmap'd buffers can not be freed. */ +if (memmap_area) { + tr->flags |=TRACE_ARRAY_FL_MEMMAP ; + tr->ref++; + } +if (start) { tr->flags |=TRACE_ARRAY_FL_BOOT |TRACE_ARRAY_FL_LAST_BOOT ; tr->range_name =no_free_ptr (rname);
kernel/trace/trace.h +1 -0
@@ -447,6 +447,7 @@ enum {TRACE_ARRAY_FL_BOOT =BIT (1 ),TRACE_ARRAY_FL_LAST_BOOT =BIT (2 ),TRACE_ARRAY_FL_MOD_INIT =BIT (3 ), + TRACE_ARRAY_FL_MEMMAP = BIT(4), };#ifdef CONFIG_MODULES
Signature · 1 line
2.47.2
Re: [PATCH v5 2/4] tracing: Have reserve_mem use phys_to_virt() and separate from memmap buffer
Quoted 5 lines · Steven Rostedt
From: Steven Rostedt <[email protected]>
The
reserve_memkernel command line option may pass back a physical address, but the memory is still part of the normal memory just like usingmemblock_reserve()would be. This means that the physical memory
... using memblock_alloc() would be
Quoted 115 lines
returned by the
reserve_memcommand line option can be converted directly to virtual memory by simply usingphys_to_virt().When freeing the buffer there's no need to call
vunmap()anymore as the memory allocated byreserve_memis freed by the call toreserve_mem_release_by_name().Because the persistent ring buffer can also be allocated via the memmap option, which *is* different than normal memory as it cannot be added back to the buddy system, it must be treated differently. It still needs to be virtually mapped to have access to it. It also can not be freed nor can it ever be memory mapped to user space.
Create a new
trace_arrayflag calledTRACE_ARRAY_FL_MEMMAPwhich gets set if the buffer is created by the memmap option, and this will prevent the buffer from being memory mapped by user space.Also increment the ref count for memmap'ed buffers so that they can never be freed.
Suggested-by Mike Rapoport <[email protected]>Signed-off-by Steven Rostedt (Google) <[email protected]>--- kernel/trace/trace.c | 23 ++++++++++++++++------- kernel/trace/trace.h | 1 + 2 files changed, 17 insertions(+), 7 deletions(-)kernel/trace/trace.c +16
-7@@ -8505,6 +8505,10 @@ static int tracing_buffers_mmap(struct file *filp, struct vm_area_struct *vma)struct trace_iterator *iter = &info->iter;int ret =0 ; +/* A memmap'ed buffer is not supported for user space mmap */ +if (iter->tr->flags &TRACE_ARRAY_FL_MEMMAP ) +return -ENODEV ; +/* Currently the boot mapped buffer is not supported for mmap */ if (iter->tr->flags &TRACE_ARRAY_FL_BOOT )return -ENODEV ;@@ -9614,9 +9618,6 @@ static void free_trace_buffers(struct trace_array *tr)#ifdef CONFIG_TRACER_MAX_TRACE free_trace_buffer (&tr->max_buffer);#endif - -if (tr->range_addr_start) -vunmap ((void *)tr->range_addr_start); }static void init_trace_flags_index (struct trace_array *tr)@@ -10710,6 +10711,7 @@ static inline void do_allocate_snapshot(const char *name) { }__init static void enable_instances (void ) {struct trace_array *tr; +bool memmap_area = false;char *curr_str;char *name;char *str;@@ -10778,6 +10780,7 @@ __init static void enable_instances(void) name);continue ; } + memmap_area = true; }else if (tok) {if (!reserve_mem_find_by_name (tok, &start, &size)) { start =0 ;@@ -10800,7 +10803,10 @@ __init static void enable_instances(void)continue ; } - addr =map_pages (start, size); +if (memmap_area) + addr =map_pages (start, size); +else + addr = (unsigned long )phys_to_virt (start);if (addr) {pr_info ("Tracing: mapped boot instance %s at physical memory %pa of size 0x%lx\n" , name, &start, (unsigned long )size);@@ -10827,10 +10833,13 @@ __init static void enable_instances(void)update_printk_trace (tr);/* -* If start is set, then this is a mapped buffer, and -* cannot be deleted by user space, so keep the reference -* to it. +* memmap'd buffers can not be freed. */ +if (memmap_area) { + tr->flags |=TRACE_ARRAY_FL_MEMMAP ; + tr->ref++; + } +if (start) { tr->flags |=TRACE_ARRAY_FL_BOOT |TRACE_ARRAY_FL_LAST_BOOT ; tr->range_name =no_free_ptr (rname);kernel/trace/trace.h +1
-0@@ -447,6 +447,7 @@ enum {TRACE_ARRAY_FL_BOOT =BIT (1 ),TRACE_ARRAY_FL_LAST_BOOT =BIT (2 ),TRACE_ARRAY_FL_MOD_INIT =BIT (3 ), + TRACE_ARRAY_FL_MEMMAP = BIT(4), };#ifdef CONFIG_MODULES Signature · 1 line
2.47.2
Signature · 2 lines
Sincerely yours, Mike.
Steven Rostedt
maintainerRe: [PATCH v5 2/4] tracing: Have reserve_mem use phys_to_virt() and separate from memmap buffer
Quoted 5 lines · On Wed, 2 Apr 2025 12:24:12 +0300 Mike Rapoport <[email protected]> wrote:
Quoted 3 lines
The
reserve_memkernel command line option may pass back a physical address, but the memory is still part of the normal memory just like usingmemblock_reserve()would be. This means that the physical memory... using
memblock_alloc()would be
Heh, I looked for examples of code that used free_reserved_area() and found initramfs.c which uses memblock_reserve(), which is why I picked that.
I'll update.
Thanks for reviewing Mike!
-- Steve
Steven Rostedt
maintainerPATCH v5 3/4 tracing: Use vmap_page_range() to map memmap ring buffer
In reply to Steven RostedtFrom: Steven Rostedt <[email protected]>
The code to map the physical memory retrieved by memmap currently allocates an array of pages to cover the physical memory and then calls vmap() to map it to a virtual address. Instead of using this temporary array of struct page descriptors, simply use vmap_page_range() that can directly map the contiguous physical memory to a virtual address.
--- kernel/trace/trace.c | 33 ++++++++++++++++----------------- 1 file changed, 16 insertions(+), 17 deletions(-)
kernel/trace/trace.c +16 -17
@@ -50,6 +50,7 @@#include <linux/irq_work.h> #include <linux/workqueue.h> #include <linux/sort.h> +#include <linux/io.h> /* vmap_page_range() */ #include <asm/setup.h> /* COMMAND_LINE_SIZE */
@@ -9810,29 +9811,27 @@ static int instance_mkdir(const char *name)return ret; } -static u64 map_pages (u64 start,u64 size) +static u64 map_pages (unsigned long start,unsigned long size) { -struct page **pages; -phys_addr_t page_start; -unsigned int page_count; -unsigned int i; -void *vaddr; - - page_count =DIV_ROUND_UP (size,PAGE_SIZE ); +unsigned long vmap_start, vmap_end; +struct vm_struct *area; +int ret; - page_start = start; - pages =kmalloc_array (page_count,sizeof (struct page *),GFP_KERNEL ); -if (!pages) + area =get_vm_area (size,VM_IOREMAP ); +if (!area)return 0 ; -for (i =0 ; i < page_count; i++) { -phys_addr_t addr = page_start + i *PAGE_SIZE ; - pages[i] =pfn_to_page (addr >>PAGE_SHIFT ); + vmap_start = (unsigned long ) area->addr; + vmap_end = vmap_start + size; + + ret =vmap_page_range (vmap_start, vmap_end, + start,pgprot_nx (PAGE_KERNEL )); +if (ret <0 ) { +free_vm_area (area); +return 0 ; } - vaddr =vmap (pages, page_count,VM_MAP ,PAGE_KERNEL ); -kfree (pages); -return (u64 )(unsigned long )vaddr; +return (u64 )vmap_start; }/**
Signature · 1 line
2.47.2
Re: [PATCH v5 3/4] tracing: Use vmap_page_range() to map memmap ring buffer
Quoted 6 lines · Steven Rostedt
The code to map the physical memory retrieved by memmap currently allocates an array of pages to cover the physical memory and then calls
vmap()to map it to a virtual address. Instead of using this temporary array of struct page descriptors, simply usevmap_page_range()that can directly map the contiguous physical memory to a virtual address.
Wait, what?
Didn't you just say that you can just use page_address() for the kernel mapping?
So the whole vmap thing is entirely unnecessary in the first place.
Including the simpler vmap_page_range().
Linus
Steven Rostedt
maintainerRe: [PATCH v5 3/4] tracing: Use vmap_page_range() to map memmap ring buffer
Quoted 3 lines · On Wed, 2 Apr 2025 09:42:00 -0700 Linus Torvalds…
Wait, what?
Didn't you just say that you can just use
page_address()for the kernel mapping?
Correct. I misunderstood Mike when he created the reserve_mem and had it return physical addresses. I didn't realize he had it already mapped via the normal virtual mappings.
Quoted 4 lines
So the whole vmap thing is entirely unnecessary in the first place.
Including the simpler
vmap_page_range().
Not entirely. The original code only used memmap=, which does require the vmap_page_range(). The reserve_mem came later, and thinking it was just physical address space (not a mapping), I just had it use the memmap code.
That is, the persistent memory buffer was originally created on top of the memmap feature (and only for x86). I talked with Mike about making it more generic and he and I worked out the reserve_mem option. Having the memmap code already doing the vmap() I just had the reserve_mem version use the same, not knowing that it could just use phys_to_virt().
This patch series fixes that miscommunication and separates out a memmap'ed buffer from reserve_mem buffer and simplifies everything.
-- Steve
Steven Rostedt
maintainerRe: [PATCH v5 3/4] tracing: Use vmap_page_range() to map memmap ring buffer
Quoted 2 lines · On Wed, 2 Apr 2025 12:55:48 -0400 Steven Rostedt <[email protected]> wrote:
This patch series fixes that miscommunication and separates out a memmap'ed buffer from
reserve_membuffer and simplifies everything.
Since I only care about memory mapping a buffer from reserve_mem to user space, this series makes sure that a buffer created from memmap (or any other physical memory area) is never used for mapping to user space. It also prevents that buffer from being freed, as it is not part of the buddy allocator and can't be added later.
That also means all the tricks to determine where the page in the ring buffer came from can go away.
-- Steve
Steven Rostedt
maintainerRe: [PATCH v5 3/4] tracing: Use vmap_page_range() to map memmap ring buffer
Quoted 2 lines · On Wed, 2 Apr 2025 13:03:37 -0400 Steven Rostedt <[email protected]> wrote:
That also means all the tricks to determine where the page in the ring buffer came from can go away.
I wonder, just to be safe, if I should add in the ring buffer mapping code:
/* If the buffer is not part of the normal memory, do not allow mapping */ if (!virt_addr_valid(cpu_buffer->reader_page->page)) return -ENODEV;
?
As the buffer can come from anywhere, but we only allow a memory mapping to user space if the buffer is from the normal memory allocator.
Currently, this patch series has the caller (the tracing code) prevent allowing memmap to be user mapped. Perhaps the above should include a WARN_ON_ONCE()?
-- Steve
Re: [PATCH v5 3/4] tracing: Use vmap_page_range() to map memmap ring buffer
Quoted 4 lines · Steven Rostedt
/* If the buffer is not part of the normal memory, do not allow mapping */
if (!virt_addr_valid(cpu_buffer->reader_page->page))
return -ENODEV;Honestly, virt_addr_valid() is pretty much random.
It's not really a valid function outside of random debugging code. It has no real semantics. The comment saying that it validates some kind of 'virt_to_page()' thing is pure garbage.
Quoted 2 lines
As the buffer can come from anywhere, but we only allow a memory mapping to user space if the buffer is from the normal memory allocator.
You should damn well keep track of where the memory comes from.
You can't just say "I'll take random shit, and then I'll ask the VM what it is".
Stop it.
If the address came from something you consider trustworthy, then just trust it. If some admin person gave you garbage, it's better to just get a random oops than to have random bogus code.
Linus
Steven Rostedt
maintainerRe: [PATCH v5 3/4] tracing: Use vmap_page_range() to map memmap ring buffer
Quoted 1 line · On Wed, 2 Apr 2025 10:20:58 -0700 Linus Torvalds…
You should damn well keep track of where the memory comes from.
And it does.
Quoted 8 lines
You can't just say "I'll take random shit, and then I'll ask the VM what it is".
Stop it.
If the address came from something you consider trustworthy, then just trust it. If some admin person gave you garbage, it's better to just get a random oops than to have random bogus code.
This has nothing to do with admins. This would only occur if the kernel itself created a buffer from some random physical address and then tried to mmap it to user space (which would be a bug).
My early career came from the military industry (I worked on the C17 engine control systems in the early '90s). It was drilled in my head to have safety checks throughout the code, in case something buggy happened, it would be caught quickly later on.
The virt_addr_valid() would only be a debugging feature, hence the added "WARN_ON()" for it.
But I'm fine to not do that.
-- Steve
Re: [PATCH v5 3/4] tracing: Use vmap_page_range() to map memmap ring buffer
Quoted 4 lines · Steven Rostedt
This has nothing to do with admins. This would only occur if the kernel itself created a buffer from some random physical address and then tried to mmap it to user space (which would be a bug).
Do *not* try to check for bugs like that with virt_addr_valid().
It literally snakes debugging harder.
You're much better off getting an oops,. and then you have stack traces, distro bug trackers, and various other automated tools that give you information.
Trying to "validate" buggy data is crazy. It's absolutely the opposite of safety. It's going to cause more bugs, it's going to only work for the validation scenarios you thought about, and it's going to make it harder to debug the cases it actually catches.
And if you are trying to catch kernel bugs, *any* data could be that buggy data. So the whole concept is insane.
Yes, you could make every single line be a "WARN_ON()" with some random check for the particular data you are using.
Or you could just write good solid code that is actually readable and maintainable, and doesn't have random pointless checks in it.
Linus
Steven Rostedt
maintainerPATCH v5 4/4 ring-buffer: Use flush_kernel_vmap_range() over flush_dcache_folio()
In reply to Steven RostedtFrom: Steven Rostedt <[email protected]>
Some architectures do not have data cache coherency between user and kernel space. For these architectures, the cache needs to be flushed on both the kernel and user addresses so that user space can see the updates the kernel has made.
Instead of using flush_dcache_folio() and playing with virt_to_folio() within the call to that function, use flush_kernel_vmap_range() which takes the virtual address and does the work for those architectures that need it.
This also fixes a bug where the flush of the reader page only flushed one page. If the sub-buffer order is 1 or more, where the sub-buffer size would be greater than a page, it would miss the rest of the sub-buffer content, as the "reader page" is not just a page, but the size of a sub-buffer.
--- kernel/trace/ring_buffer.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)
kernel/trace/ring_buffer.c +3 -2
@@ -6016,7 +6016,7 @@ static void rb_update_meta_page(struct ring_buffer_per_cpu *cpu_buffer) meta->read = cpu_buffer->read;/* Some archs do not have data cache coherency between kernel and user-space */ -flush_dcache_folio (virt_to_folio (cpu_buffer->meta_page)); +flush_kernel_vmap_range (cpu_buffer->meta_page,PAGE_SIZE ); } static void
@@ -7319,7 +7319,8 @@ int ring_buffer_map_get_reader(struct trace_buffer *buffer, int cpu) out:/* Some archs do not have data cache coherency between kernel and user-space */ -flush_dcache_folio (virt_to_folio (cpu_buffer->reader_page->page)); +flush_kernel_vmap_range (cpu_buffer->reader_page->page, + buffer->subbuf_size +BUF_PAGE_HDR_SIZE );rb_update_meta_page (cpu_buffer);
Signature · 1 line
2.47.2