On 1/29/25 09:52, Sebastian Andrzej Siewior wrote:
> On 2025-01-27 13:22:17 [+0100], Petr Pavlu wrote:
>> On 1/24/25 18:49, Sebastian Andrzej Siewior wrote:
>>> On 2025-01-13 12:09:27 [+0100], Petr Pavlu wrote:
>>>> Thanks for this cleanup. I've queued the fi
On 1/24/25 18:49, Sebastian Andrzej Siewior wrote:
> On 2025-01-13 12:09:27 [+0100], Petr Pavlu wrote:
>> Thanks for this cleanup. I've queued the fix in patch #1 on
>> modules-fixes. For the rest, I plan to give folks more time to look at
>> the changes as this affects
The following changes since commit ffd294d346d185b70e28b1a28abe367bbfe53c04:
Linux 6.13 (2025-01-19 15:51:45 -0800)
are available in the Git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/modules/linux.git/
tags/modules-6.14-rc1
for you to fetch changes up to f3b93547b91ad849
On 1/9/25 11:54, Andreas Hindborg wrote:
> This patch includes changes required for Rust kernel modules to utilize
> module parameters. This code implements read only support for integer
> types without `sysfs` support.
>
> Signed-off-by: Andreas Hindborg
> ---
> rust/kernel/lib.rs |
On 1/10/25 20:16, Luis Chamberlain wrote:
> On Thu, Jan 09, 2025 at 11:52:27AM +0100, Arnout Engelen wrote:
>> On Fri, 3 Jan 2025 17:37:52 -0800, Luis Chamberlain wrote:
>>> What distro which is using module signatures would switch
>>> to this as an alternative instead?
>>
>> In NixOS, we disable M
On 1/8/25 10:04, Sebastian Andrzej Siewior wrote:
> This is an updated version of the initial post after PeterZ made me
> aware that there are users outside of the module directory.
> The goal is replace the mix auf rcu_read_lock(), rcu_read_lock_sched()
> and preempt_disable() with just rcu_read_l
On 1/4/25 10:43, Thorsten Leemhuis wrote:
> On 20.10.24 00:57, Luis Chamberlain wrote:
>> On Wed, Oct 16, 2024 at 04:18:41PM +0200, Thorsten Leemhuis wrote:
>>> Switch away from using sha1 for module signing by default and use the
>>> more modern sha512 instead, which is what among others Arch, Fed
On 1/4/25 08:39, Christophe Leroy wrote:
> Le 03/01/2025 à 17:13, Petr Pavlu a écrit :
>> On 12/5/24 20:46, Christophe Leroy wrote:
>>> This series reworks module loading to avoid leaving the module in a
>>> stale state when protecting ro_after_init section fails.
&
On 1/3/25 15:06, Christophe Leroy wrote:
> Le 23/12/2024 à 10:37, Petr Pavlu a écrit :
>> Section .static_call_sites holds data structures that need to be sorted and
>> processed only at module load time. The section is never modified afterwards.
>> Make it therefore r
On 12/5/24 20:46, Christophe Leroy wrote:
> This series reworks module loading to avoid leaving the module in a
> stale state when protecting ro_after_init section fails.
>
> Once module init has succeded it is too late to cancel loading.
> If setting ro_after_init data section to read-only fails,
On 12/10/24 11:49, Daniel Gomez wrote:
> On 12/4/2024 4:14 PM, Petr Pavlu wrote:
>> On 11/28/24 21:23, Daniel Gomez wrote:
>>> On 11/12/2024 3:35 PM, Petr Pavlu wrote:
>>>> On 11/12/24 10:43, Daniel Gomez wrote:
>>>>> On Mon Nov 11
On 12/23/24 21:14, Linus Torvalds wrote:
> On Mon, 23 Dec 2024 at 02:55, Petr Pavlu wrote:
>>
>> Luis asked me to look after the modules maintenance for a while, with the
>> plan to rotate it with other recently added modules maintainers/reviewers.
>> This is my first
On 12/20/24 18:41, Sebastian Andrzej Siewior wrote:
> __module_address() can be invoked within a RCU section, there is no
> requirement to have preemption disabled.
>
> I'm not sure if using rcu_read_lock() will introduce the regression that
> has been fixed in commit 14c4c8e41511a ("cfi: Use
> rc
n.c:
>
> crcval = *crc;
>
> Signed-off-by: Masahiro Yamada
Reviewed-by: Petr Pavlu
I understand the plan is for this to go through the kbuild tree with the
rest of the extended modversions + Rust support.
--
Thanks,
Petr
The following changes since commit 4bbf9020becbfd8fc2c3da790855b7042fad455b:
Linux 6.13-rc4 (2024-12-22 13:22:21 -0800)
are available in the Git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/modules/linux.git/
tags/modules-6.13-rc5
for you to fetch changes up to 0b7a66a2c864
Section .static_call_sites holds data structures that need to be sorted and
processed only at module load time. The section is never modified afterwards.
Make it therefore read-only after module initialization to avoid any
(non-)accidental modifications.
Petr Pavlu (3):
module: Constify
Move the logic to mark special sections as read-only after module
initialization into a separate function, along other related code in
strict_rwx.c. Use a table with names of such sections to make it easier to
add more.
Signed-off-by: Petr Pavlu
---
kernel/module/internal.h | 2 ++
kernel
modified afterwards. Make it therefore read-only after
module initialization to avoid any (non-)accidental modifications.
Signed-off-by: Petr Pavlu
---
kernel/module/strict_rwx.c | 9 +
1 file changed, 9 insertions(+)
diff --git a/kernel/module/strict_rwx.c b/kernel/module/strict_rwx.c
Minor cleanup, this is a non-functional change.
Signed-off-by: Petr Pavlu
---
kernel/module/internal.h | 5 +++--
kernel/module/strict_rwx.c | 5 +++--
2 files changed, 6 insertions(+), 4 deletions(-)
diff --git a/kernel/module/internal.h b/kernel/module/internal.h
index daef2be83902
b517bcbfe7ec0d0186f9b0b8
> change-id: 20241204-sysfs-const-attr-module-927afe76eda4
Reviewed-by: Petr Pavlu
I'm going to wait for a few days if others want to comment and then plan
to queue this on the modules tree for the 6.14 merge window.
--
Thanks,
Petr
On 12/2/24 15:59, Peter Zijlstra wrote:
> Hi!
>
> Implement a means for exports to be available only to an explicit list of
> named
> modules. By explicitly limiting the usage of certain exports, the abuse
> potential/risk is greatly reduced.
>
> The first 'patch' is an awk scripts that cleans u
e string accurate instead of
> removing it.
>
> Move the PREEMPT_RT check before the PREEMPT so that it takes precedence
> if both symbols are enabled.
>
> Fixes: 35772d627b55c ("sched: Enable PREEMPT_DYNAMIC for PREEMPT_RT")
> Signed-off-by: Sebastian Andrzej Siewior
On 11/28/24 21:23, Daniel Gomez wrote:
> On 11/12/2024 3:35 PM, Petr Pavlu wrote:
>> On 11/12/24 10:43, Daniel Gomez wrote:
>>> On Mon Nov 11, 2024 at 7:53 PM CET, Christophe Leroy wrote:
>>>>
>>>>
>>>> Le 09/11/2024 à 23:17, Daniel Go
he preprocessed C-style format
> genksyms produces.
>
> Signed-off-by: Sami Tolvanen
Reviewed-by: Petr Pavlu
--
Thanks,
Petr
On 11/21/24 21:42, Sami Tolvanen wrote:
> Calculate symbol versions from the fully expanded type strings in
> type_map, and output the versions in a genksyms-compatible format.
>
> Signed-off-by: Sami Tolvanen
Reviewed-by: Petr Pavlu
--
Thanks,
Petr
&sym_mem, &xndx);
> +
> + if (!sym ||
> + GELF_ST_BIND(sym->st_info) == STB_LOCAL)
> + continue;
Doesn't NULL returned by gelf_getsymshndx() indicate an error? If yes
then I think it should be reported instead of silently skipping the
symbol.
With this addressed, feel free to add:
Reviewed-by: Petr Pavlu
--
Thanks,
Petr
e that genksyms has.
I was also recently thinking that it would be great if genksyms could
skip definitions that are in internal header files, for example,
kernel/events/internal.h. Perhaps something that could be added in the
future..
I've noted one nit below, but the patch looks sensible to me
On 11/20/24 03:17, Song Chen wrote:
> Hi Petr,
>
> 在 2024/11/18 20:54, Petr Pavlu 写道:
>> On 11/13/24 03:15, Song Chen wrote:
>>> 在 2024/11/12 20:56, Petr Pavlu 写道:
>>>> On 11/10/24 12:42, Song Chen wrote:
>>>>> Sometimes when kernel calls requ
On 11/13/24 03:15, Song Chen wrote:
> 在 2024/11/12 20:56, Petr Pavlu 写道:
>> On 11/10/24 12:42, Song Chen wrote:
>>> Sometimes when kernel calls request_module to load a module
>>> into kernel space, it doesn't pass the module name appropriately,
>>> and
On 11/12/24 10:43, Daniel Gomez wrote:
> On Mon Nov 11, 2024 at 7:53 PM CET, Christophe Leroy wrote:
>>
>>
>> Le 09/11/2024 à 23:17, Daniel Gomez a écrit :
>>> On Sat Nov 9, 2024 at 11:35 AM CET, Christophe Leroy wrote:
Once module init has succeded it is too late to cancel loading.
If se
On 11/10/24 12:42, Song Chen wrote:
> Sometimes when kernel calls request_module to load a module
> into kernel space, it doesn't pass the module name appropriately,
> and request_module doesn't verify it as well.
>
> As a result, modprobe is invoked anyway and spend a lot of time
> searching a no
On 11/9/24 11:35, Christophe Leroy wrote:
> Once module init has succeded it is too late to cancel loading.
> If setting ro_after_init data section to read-only fails, all we
> can do is to inform the user through a warning.
Makes sense to me. If I'm looking correctly, set_memory_ro() could
mostly
On 10/8/24 20:38, Sami Tolvanen wrote:
> When MODVERSIONS is enabled, allow selecting gendwarfksyms as the
> implementation, but default to genksyms.
>
> Signed-off-by: Sami Tolvanen
> Acked-by: Neal Gompa
> ---
> kernel/module/Kconfig | 25 -
> scripts/Makefile |
iscard.gendwarfksyms") = &sym;
>
> extern int external_symbol(void);
> GENDWARFKSYMS_PTR(external_symbol);
>
> Signed-off-by: Sami Tolvanen
> Acked-by: Neal Gompa
Looks ok to me, feel free to add:
Reviewed-by: Petr Pavlu
-- Petr
mber_location(24)
> + * STABLE-NEXT: } byte_size(32)
> + *
> + * VERSION-DAG: #SYMVER ex2b 0x[[#%.08x,EX2]]
> + */
> +
> +struct {
> + int a;
> + KABI_IGNORE(0, unsigned int n);
> + unsigned long b;
> + int c;
> + KABI_IGNORE(1, unsigned int m);
> + unsigned long d;
> +} ex2c;
> +
> +_Static_assert(sizeof(ex2a) == sizeof(ex2c), "ex2a size doesn't match ex2c");
> +
> +/*
> + * STABLE: variable structure_type {
> + * STABLE-NEXT: member base_type int byte_size(4) encoding(5) a
> data_member_location(0) ,
> + * STABLE-NEXT: member base_type [[ULONG]] byte_size(8) encoding(7) b
> data_member_location(8)
> + * STABLE-NEXT: member base_type int byte_size(4) encoding(5) c
> data_member_location(16) ,
> + * STABLE-NEXT: member base_type [[ULONG]] byte_size(8) encoding(7) d
> data_member_location(24)
> + * STABLE-NEXT: } byte_size(32)
> + *
> + * VERSION-DAG: #SYMVER ex2c 0x[[#%.08x,EX2]]
> + */
> diff --git a/scripts/gendwarfksyms/gendwarfksyms.h
> b/scripts/gendwarfksyms/gendwarfksyms.h
> index f32ad4389b58..1cff868bacdb 100644
> --- a/scripts/gendwarfksyms/gendwarfksyms.h
> +++ b/scripts/gendwarfksyms/gendwarfksyms.h
> @@ -222,6 +222,20 @@ void cache_clear_expanded(struct expansion_cache *ec);
> /*
> * dwarf.c
> */
> +
> +/* See dwarf.c:get_union_kabi_status */
> +#define KABI_PREFIX "__kabi_"
> +#define KABI_PREFIX_LEN (sizeof(KABI_PREFIX) - 1)
> +#define KABI_RESERVED_PREFIX "reserved"
> +#define KABI_RESERVED_PREFIX_LEN (sizeof(KABI_RESERVED_PREFIX) - 1)
> +#define KABI_IGNORED_PREFIX "ignored"
> +#define KABI_IGNORED_PREFIX_LEN (sizeof(KABI_IGNORED_PREFIX) - 1)
> +
> +static inline bool is_kabi_prefix(const char *name)
> +{
> + return name && !strncmp(name, KABI_PREFIX, KABI_PREFIX_LEN);
> +}
> +
Nit: The new KABI_* macros and the is_kabi_prefix() function are used
only in dwarf.c so could be all moved there.
> struct expansion_state {
> bool expand;
> unsigned int ptr_depth;
> @@ -229,6 +243,18 @@ struct expansion_state {
> const char *current_fqn;
> };
>
> +enum kabi_status {
> + /* >0 to stop DIE processing */
> + KABI_NORMAL = 1,
> + KABI_RESERVED,
> + KABI_IGNORED,
> +};
> +
> +struct kabi_state {
> + int members;
> + Dwarf_Die placeholder;
> +};
> +
> struct state {
> struct symbol *sym;
> Dwarf_Die die;
> @@ -239,6 +265,9 @@ struct state {
> /* Structure expansion */
> struct expansion_state expand;
> struct expansion_cache expansion_cache;
> +
> + /* Reserved or ignored members */
> + struct kabi_state kabi;
> };
>
> typedef int (*die_callback_t)(struct state *state, struct die *cache,
I've noted some nits above which you might want to trivially address.
Otherwise this looks ok to me, feel free to add:
Reviewed-by: Petr Pavlu
--
Thanks,
Petr
On 10/8/24 20:38, Sami Tolvanen wrote:
> Distributions that want to maintain a stable kABI need the ability
> to make ABI compatible changes to kernel without affecting symbol
> versions, either because of LTS updates or backports.
>
> With genksyms, developers would typically hide these changes f
+
> + hash_for_each_safe(symbol_names, sym, tmp, name_hash) {
> + func(sym, arg);
> + }
> +}
> +
> typedef void (*elf_symbol_callback_t)(const char *name, GElf_Sym *sym,
> Elf32_Word xndx, void *arg);
>
> @@ -231,3 +271,22 @@ void symbol_read_symtab(int fd)
> {
> elf_for_each_global(fd, elf_set_symbol_addr, NULL);
> }
> +
> +void symbol_print_versions(void)
> +{
> + struct hlist_node *tmp;
> + struct symbol *sym;
> +
> + hash_for_each_safe(symbol_names, sym, tmp, name_hash) {
> + if (sym->state != SYMBOL_PROCESSED)
> + warn("no information for symbol %s", sym->name);
> +
> + printf("#SYMVER %s 0x%08lx\n", sym->name, sym->crc);
> +
> + free((void *)sym->name);
> + free(sym);
> + }
> +
> + hash_init(symbol_addrs);
> + hash_init(symbol_names);
> +}
I had some minor comment about adjusting the name of function
symbol_print_versions() and possibly changing sym->name to 'char *' on
the v2 of the patch:
https://lore.kernel.org/all/286b1cc5-1757-4f0a-bb66-0875f4608...@suse.com/
Please have a look, it seems it felt through the cracks.
In any case, the patch looks ok to me, feel free to add:
Reviewed-by: Petr Pavlu
--
Thanks,
Petr
) alignment(8) msg
> )
> -> base_type void
>
> Signed-off-by: Sami Tolvanen
> Acked-by: Neal Gompa
Looks ok to me, feel free to add:
Reviewed-by: Petr Pavlu
-- Petr
, free free to add:
Reviewed-by: Petr Pavlu
-- Petr
On 10/8/24 20:38, Sami Tolvanen wrote:
> Add support for producing genksyms-style symtypes files. Process
> die_map to find the longest expansions for each type, and use symtypes
> references in type definitions. The basic file format is similar to
> genksyms, with two notable exceptions:
>
> 1.
-by: Sami Tolvanen
> Acked-by: Neal Gompa
Looks ok to me, feel free to use:
Reviewed-by: Petr Pavlu
-- Petr
On 10/8/24 20:38, Sami Tolvanen wrote:
> The compiler may choose not to emit type information in DWARF for all
> aliases, but it's possible for each alias to be exported separately.
> To ensure we find type information for the aliases as well, read
> {section, address} tuples from the symbol table
On 10/8/24 20:38, Sami Tolvanen wrote:
> Expand each structure type only once per exported symbol. This
> is necessary to support self-referential structures, which would
> otherwise result in infinite recursion, but is still sufficient for
> catching ABI changes.
>
> For pointers, limit structure
if (!name || !*name)
> + return 0;
> +
> + hash_for_each_possible_safe(symbol_names, match, tmp, name_hash,
> + hash_str(name)) {
> + if (strcmp(match->name, name))
> + continue;
> +
> + if (func)
> + func(match, data);
> +
> + return 1;
> + }
> +
> + return 0;
> +}
> +
> +static bool is_exported(const char *name)
> +{
> + return checkp(for_each(name, NULL, NULL)) > 0;
> +}
> +
> +void symbol_read_exports(FILE *file)
> +{
> + struct symbol *sym;
> + char *line = NULL;
> + char *name = NULL;
> + size_t size = 0;
> + int nsym = 0;
> +
> + while (getline(&line, &size, file) > 0) {
> + if (sscanf(line, "%ms\n", &name) != 1)
> + error("malformed input line: %s", line);
> +
> + if (is_exported(name)) {
> + /* Ignore duplicates */
> + free(name);
> + continue;
> + }
> +
> + sym = xcalloc(1, sizeof(struct symbol));
> + sym->name = name;
> +
> + hash_add(symbol_names, &sym->name_hash, hash_str(sym->name));
> + ++nsym;
> +
> + debug("%s", sym->name);
> + }
> +
> + free(line);
> + debug("%d exported symbols", nsym);
> +}
> +
> +static void get_symbol(struct symbol *sym, void *arg)
> +{
> + struct symbol **res = arg;
> +
> + *res = sym;
> +}
> +
> +struct symbol *symbol_get(const char *name)
> +{
> + struct symbol *sym = NULL;
> +
> + for_each(name, get_symbol, &sym);
> + return sym;
> +}
Nit: The code inconsistently checks for a potential error from the
function for_each(). Looking at the whole series, the value is checked
using checkp() in functions symbol_set_crc(), symbol_set_ptr(),
symbol_set_die(), is_exported(), but not in symbol_get() and
elf_set_symbol_addr(). It would be good to unify this, or perhaps even
make for_each() return an unsigned int to indicate it never fails?
Looks otherwise ok to me, feel free to add:
Reviewed-by: Petr Pavlu
--
Thanks,
Petr
On 9/23/24 20:18, Sami Tolvanen wrote:
> Basic types in DWARF repeat frequently and traversing the DIEs using
> libdw is relatively slow. Add a simple hashtable based cache for the
> processed DIEs.
>
> Signed-off-by: Sami Tolvanen
> ---
> scripts/gendwarfksyms/Makefile| 1 +
> scripts
On 9/23/24 20:18, Sami Tolvanen wrote:
> The compiler may choose not to emit type information in DWARF for all
> aliases, but it's possible for each alias to be exported separately.
> To ensure we find type information for the aliases as well, read
> {section, address} tuples from the symbol table
pointer_type {
> const_type {
> base_type char byte_size(1) encoding(6)
> }
> } byte_size(8)
> }
>
> Signed-off-by: Sami Tolvanen
> [...]
Looks ok to me, feel free to add:
Reviewed-by: Petr Pavlu
--
Thanks,
Petr
On 9/23/24 20:18, Sami Tolvanen wrote:
> Recursively expand DWARF structure types, i.e. structs, unions, and
> enums. Also include relevant DWARF attributes in type strings to
> encode structure layout, for example.
>
> Example output with --dump-dies:
>
> subprogram (
> formal_parameter st
and indentation to debugging output to make it
> more readable.
>
> Signed-off-by: Sami Tolvanen
> [...]
Looks ok to me, feel free to add:
Reviewed-by: Petr Pavlu
--
Thanks,
Petr
{
> const_type {
> base_type char byte_size(1) encoding(6)
> }
> }
> )
> -> base_type unsigned long byte_size(8) encoding(7)
>
> Signed-off-by: Sami Tolvanen
> [...]
Looks ok to me, feel free to add:
Reviewed-by: Petr Pavlu
--
Thanks,
Petr
ug --dump-dies vmlinux.o
> ...
> gendwarfksyms: process_symbol: loops_per_jiffy
> variable base_type unsigned long byte_size(8) encoding(7)
> ...
>
> Signed-off-by: Sami Tolvanen
> [...]
Looks ok to me, feel free to add:
Reviewed-by: Petr Pavlu
--
Thanks,
Petr
On 9/11/24 18:03, Sami Tolvanen wrote:
> On Wed, Sep 11, 2024 at 3:08 AM Petr Pavlu wrote:
>> On 8/15/24 19:39, Sami Tolvanen wrote:
>> More importantly, it made me think which names are included in the CRC
>> calculation and which ones are omitted.
>>
>> I
On 8/31/24 02:05, Sami Tolvanen wrote:
> On Fri, Aug 30, 2024 at 9:34 AM Miroslav Benes wrote:
>>
>> yes, this is one of the approaches we use in SLES. We add kabi paddings
>> to some structures in advance (see [1] as a random example) and then use
>> it later if needed.
>>
>> It is not the only a
On 8/15/24 19:39, Sami Tolvanen wrote:
> Add a basic CRC32 implementation adapted from genksyms, and produce
> matching output from symtypes strings in type_map.
>
> Signed-off-by: Sami Tolvanen
> ---
> scripts/gendwarfksyms/Makefile| 1 +
> scripts/gendwarfksyms/crc32.c | 69
On 8/15/24 19:39, Sami Tolvanen wrote:
> Add support for producing genksyms-style symtypes files. Process
> die_map to find the longest expansions for each type, and use symtypes
> references in type definitions. The basic file format is similar to
> genksyms, with two notable exceptions:
>
> 1.
On 8/15/24 19:39, Sami Tolvanen wrote:
> Expand each structure type only once per exported symbol. This
> is necessary to support self-referential structures, which would
> otherwise result in infinite recursion, but is still sufficient for
> catching ABI changes.
>
> For pointers to structure typ
On 8/15/24 19:39, Sami Tolvanen wrote:
> Add support for expanding DW_TAG_subroutine_type and the parameters
> in DW_TAG_formal_parameter. Use this to also expand subprograms.
>
> Example output with --debug:
>
> subprogram(
> formal_parameter base_type usize byte_size(8),
> formal_para
On 8/30/24 07:43, Chunhui Li wrote:
> When insmod a kernel module, if fails in add_notes_attrs or
> add_sysfs_attrs such as memory allocation fail, mod_sysfs_setup
> will still return success, but we can't access user interface
> on android device.
>
> Patch for make mod_sysfs_setup can check the
On 8/15/24 19:39, Sami Tolvanen wrote:
> Basic types in DWARF repeat frequently and traversing the DIEs using
> libdw is relatively slow. Add a simple hashtable based cache for the
> processed DIEs.
>
> Signed-off-by: Sami Tolvanen
> ---
> scripts/gendwarfksyms/Makefile| 1 +
> scripts
On 8/29/24 00:53, Sami Tolvanen wrote:
> On Wed, Aug 28, 2024 at 04:04:09PM +0900, Masahiro Yamada wrote:
>> On Fri, Aug 16, 2024 at 2:39 AM Sami Tolvanen
>> wrote:
>>> - Added a --symtypes flag for generating a genksyms-style
>>> symtypes output based on Petr's feedback, and refactored
>>> s
On 8/29/24 01:09, Sami Tolvanen wrote:
> On Wed, Aug 28, 2024 at 02:35:29PM +0200, Petr Pavlu wrote:
>> On 8/15/24 19:39, Sami Tolvanen wrote:
>>> diff --git a/scripts/gendwarfksyms/dwarf.c b/scripts/gendwarfksyms/dwarf.c
>>> index 65a29d0bd8f4..71cfab0553da 100644
>
On 8/15/24 19:39, Sami Tolvanen wrote:
> Start making gendwarfksyms more useful by adding support for
> expanding DW_TAG_base_type types and basic DWARF attributes.
>
> Example:
>
> $ echo loops_per_jiffy | \
> scripts/gendwarfksyms/gendwarfksyms --debug vmlinux.o
> ...
> gendwarfksym
On 8/15/24 19:39, Sami Tolvanen wrote:
> diff --git a/scripts/gendwarfksyms/dwarf.c b/scripts/gendwarfksyms/dwarf.c
> index 65a29d0bd8f4..71cfab0553da 100644
> --- a/scripts/gendwarfksyms/dwarf.c
> +++ b/scripts/gendwarfksyms/dwarf.c
> @@ -5,6 +5,48 @@
> [...]
> +
> +static bool is_export_symbol(st
On 8/26/24 20:47, Sami Tolvanen wrote:
> On Mon, Aug 26, 2024 at 10:42 AM Petr Pavlu wrote:
>> On 8/15/24 19:39, Sami Tolvanen wrote:
>>> [...]
>>> +int main(int argc, const char **argv)
>>> +{
>>> + unsigned int n;
>>> +
>>> +
On 8/15/24 19:39, Sami Tolvanen wrote:
> The compiler may choose not to emit type information in DWARF for all
> aliases, but it's possible for each alias to be exported separately.
> To ensure we find type information for the aliases as well, read
> {section, address} tuples from the symbol table
On 8/15/24 19:39, Sami Tolvanen wrote:
> Add support for passing a list of exported symbols to gendwarfksyms
> via stdin and filter out non-exported symbols from the output.
>
> Signed-off-by: Sami Tolvanen
> ---
> scripts/gendwarfksyms/Makefile| 1 +
> scripts/gendwarfksyms/dwarf.c
On 8/15/24 19:39, Sami Tolvanen wrote:
> Add a basic DWARF parser, which uses libdw to traverse the debugging
> information in an object file and looks for functions and variables.
> In follow-up patches, this will be expanded to produce symbol versions
> for CONFIG_MODVERSIONS from DWARF.
>
> Sig
les so we can ensure we get proper
> reviews for both parts of the code and so that we can scale.
>
> Add those who have stepped up to help.
>
> Signed-off-by: Luis Chamberlain
Acked-by: Petr Pavlu
Cheers,
Petr
ted-by: Miroslav Benes
Fixes: 7aa1eaef9f42 ("function_graph: Allow multiple users to attach to
function graph")
Signed-off-by: Petr Pavlu
---
kernel/trace/fgraph.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/kernel/trace/fgraph.c b/kernel/trace/fgraph.c
index
On 7/26/24 23:05, Sami Tolvanen wrote:
> On Mon, Jul 22, 2024 at 8:20 AM Petr Pavlu wrote:
>>
>> From my perspective, I'm okay if gendwarfksyms doesn't provide
>> functionality to compare a new object file with its reference symtypes
>> file.
>>
>&
On 7/22/24 12:23, Masahiro Yamada wrote:
> On Mon, Jul 22, 2024 at 6:07 PM Petr Pavlu wrote:
>>
>> The kernel configuration allows specifying a module compression mode. If
>> one is selected then each module gets compressed during
>> 'make modules_install'
The MODULE_SIG_ config choice has an inconsistent prompt styled as
a question and lengthy option names.
Simplify the prompt and option names to be consistent with other module
options.
Signed-off-by: Petr Pavlu
---
kernel/module/Kconfig | 16
1 file changed, 8 insertions(+), 8
t
defaults to Y.
* "Support in-kernel module decompression" (MODULE_DECOMPRESS) enables
in-kernel decompression.
Signed-off-by: Petr Pavlu
---
kernel/module/Kconfig| 61
scripts/Makefile.modinst | 2 ++
2 files changed, 33 insertions(+), 30 d
Allow enabling the in-kernel module decompression support separately,
without requiring to enable also the automatic compression during
'make modules_install'.
Petr Pavlu (2):
module: Split modules_install compression and in-kernel decompression
module: Clean up the description of M
On 7/15/24 22:39, Sami Tolvanen wrote:
> On Wed, Jul 10, 2024 at 7:30 AM Petr Pavlu wrote:
>> On 6/17/24 19:58, Sami Tolvanen wrote:
>>> The first 12 patches of this series add a small tool for computing
>>> symbol versions from DWARF, called gendwarfksyms. When pa
for longterm or distribution kernels. Additionally, I would like to
propose a small improvement to refcount_t and this gives me an actual
problem to point to about its motivation.
Cheers,
Petr
---
>From b0dde62cc5268a7d728cfdb360cb5170266a5e11 Mon Sep 17 00:00:00 2001
From: Petr Pavlu
Date: Thu
ividual pages. Introduce
for this purpose a new variable ring_buffer_per_cpu.cnt which is bumped
any time the list is modified. The value is used by rb_check_pages() to
detect such a change and restart the check.
Signed-off-by: Petr Pavlu
---
This is a follow-up to the discussion about improving
On 6/17/24 19:58, Sami Tolvanen wrote:
> Hi folks,
>
> This series implements CONFIG_MODVERSIONS for Rust, an important
> feature for distributions like Android that want to ship Rust
> kernel modules, and depend on modversions to help ensure module ABI
> compatibility.
Thanks for working on this
On 7/3/24 23:44, Steven Rostedt wrote:
> On Wed, 3 Jul 2024 09:53:14 +0200
> Petr Pavlu wrote:
>
>> The function rb_check_pages() validates the integrity of a specified
>> per-CPU tracing ring buffer. It does so by traversing the underlying
>> linked list and check
ividual pages. Introduce
for this purpose a new variable ring_buffer_per_cpu.pages_era which is
bumped any time the list is modified. The value is used by
rb_check_pages() to detect such a change and restart the check.
Signed-off-by: Petr Pavlu
---
This is a follow-up to the discussion about improving
On 6/21/24 17:09, Petr Pavlu wrote:
> Function rb_check_pages() validates the integrity of a specified per-CPU
> tracing ring buffer. It does so by walking the underlying linked
> list and checking its next and prev links.
>
> To guarantee that the list doesn't get modified d
ze. This particularly
affects PREEMPT_RT because the reader_lock is a raw spinlock which
doesn't become sleepable on PREEMPT_RT kernels.
Modify the check so it still tries to walk the whole list but gives up
the reader_lock between checking individual pages.
Signed-off-by: Petr Pavlu
-
On 5/28/24 01:43, Steven Rostedt wrote:
> On Mon, 27 May 2024 11:36:55 +0200
> Petr Pavlu wrote:
>
>>>> static void rb_check_pages(struct ring_buffer_per_cpu *cpu_buffer)
>>>> {
>>>> @@ -2200,8 +2205,13 @@ int ring_buffer_resize(struct t
On 5/20/24 15:50, Steven Rostedt wrote:
> On Fri, 17 May 2024 15:40:08 +0200
> Petr Pavlu wrote:
>
>> The reader code in rb_get_reader_page() swaps a new reader page into the
>> ring buffer by doing cmpxchg on old->list.prev->next to point it to the
>> new page.
g on the same
ring_buffer_per_cpu by taking its cpu_buffer->reader_lock.
Fixes: 659f451ff213 ("ring-buffer: Add integrity check at end of iter read")
Signed-off-by: Petr Pavlu
---
kernel/trace/ring_buffer.c | 10 ++
1 file changed, 10 insertions(+)
diff --git a/kernel/trace/ri
. Nonetheless, the lock is still
needed because a reader swapping a page into the ring buffer can make
the underlying doubly-linked list temporarily inconsistent.
This is a non-functional change.
Signed-off-by: Petr Pavlu
---
kernel/trace/ring_buffer.c | 16 +++-
1 file chang
to exist for ring_buffer_subbuf_order_set()
too. I plan to tackle it in a separate series.
Petr Pavlu (2):
ring-buffer: Correct stale comments related to non-consuming readers
ring-buffer: Fix a race between readers and resize checks
kernel/trace/ring_buffer.c | 26
he missing "CONFIG_" prefix for the logic to work as intended.
Fixes: a8b9cf62ade1 ("ftrace: Fix DIRECT_CALLS to use SAVE_REGS by default")
Signed-off-by: Petr Pavlu
---
kernel/trace/ftrace.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/kernel/trace/ft
y
expected potential races which result in duplicates but merged them
later when they occurred.
Fixes: c193707dde77 ("tracing: Remove code which merges duplicates")
Signed-off-by: Petr Pavlu
---
kernel/trace/tracing_map.c | 7 ++-
1 file changed, 6 insertions(+), 1 deletion(-)
;tracing: Use temp buffer when filtering events")
Signed-off-by: Petr Pavlu
---
kernel/trace/trace.c | 12
1 file changed, 8 insertions(+), 4 deletions(-)
diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
index dd45020fcdde..dc234b5dde47 100644
--- a/kernel/trace/trace.c
+++
the first failure and left any
cleanup later to when trace_buffered_event_disable() is called.
Fixes: 0fc1b09ff1ff ("tracing: Use temp buffer when filtering events")
Signed-off-by: Petr Pavlu
---
kernel/trace/trace.c | 11 +--
1 file changed, 5 insertions(+), 6 deletions(-)
diff
h invokes a given callback on all CPUs.
Fixes: 0fc1b09ff1ff ("tracing: Use temp buffer when filtering events")
Fixes: dea499781a11 ("tracing: Fix warning in trace_buffered_event_disable()")
Signed-off-by: Petr Pavlu
---
kernel/trace/trace.c | 12
1 file changed, 4 inserti
/20231127151248.7232-1-petr.pa...@suse.com/
Petr Pavlu (3):
tracing: Fix incomplete locking when disabling buffered events
tracing: Fix a warning when allocating buffered events fails
tracing: Fix a possible race when disabling buffered events
kernel/trace/trace.c | 35 +--
1
On 12/1/23 15:46, Steven Rostedt wrote:
> On Fri, 1 Dec 2023 15:17:35 +0100
> Petr Pavlu wrote:
>
>> Ok, keeping the current approach, my plan for v2 is to prepare the
>> following patches:
>>
>> [...]
>> * Fix the potential race be
On 11/29/23 15:58, Steven Rostedt wrote:
> On Wed, 29 Nov 2023 10:22:23 +0100
> Petr Pavlu wrote:
>
>> Yes, I believe this should address this potential race condition.
>>
>> An alternative would be instead to update
>> trace_event_buffer_lock_re
On 11/28/23 16:27, Steven Rostedt wrote:
> On Tue, 28 Nov 2023 16:05:00 +0100
> Petr Pavlu wrote:
>
>> On 11/27/23 18:41, Steven Rostedt wrote:
>>> On Mon, 27 Nov 2023 16:12:47 +0100
>>> Petr Pavlu wrote:
>>>
>>>> The following warning
On 11/27/23 18:41, Steven Rostedt wrote:
> On Mon, 27 Nov 2023 16:12:47 +0100
> Petr Pavlu wrote:
>
>> The following warning appears when using buffered events:
>> [ 203.556451] WARNING: CPU: 53 PID: 10220 at
>> kernel/trace/ring_buffer.c:3912 ring_buffer_discard_com
event would be still registered and could be hit, but would
unnecessarily go directly to a ring buffer.
* The SOFT_DISABLED flag is now consistently set only when SOFT_MODE is
also set.
Signed-off-by: Petr Pavlu
---
kernel/trace/trace_events.c | 17 ++---
1 file changed, 10
Rework the maintenance synchronization for buffered events to simplify
it and to address several found problems.
Petr Pavlu (2):
tracing: Simplify and fix "buffered event" synchronization
tracing: Disable events in reverse order of their enable operation
kernel/trace/trace.c
cing: Use temp buffer when filtering events")
Fixes: dea499781a11 ("tracing: Fix warning in trace_buffered_event_disable()")
Signed-off-by: Petr Pavlu
---
kernel/trace/trace.c | 66 ++--
1 file changed, 27 insertions(+), 39 deletions(-)
diff --gi
100 matches
Mail list logo
