[PATCH 2/3] Convert SystemTap probe interface to C++ (and perform some cleanups)
Sergio Durigan Junior
sergiodj@redhat.com
Wed Nov 15 22:49:00 GMT 2017
On Tuesday, November 14 2017, Simon Marchi wrote:
> On 2017-11-13 12:59 PM, Sergio Durigan Junior wrote:
>> This patch converts the SystemTap probe
>> interface (gdb/stap-probe.[ch]) to C++, and also performs some
>> cleanups that were on my TODO list for a while.
>>
>> The main changes were the conversion of 'struct stap_probe' to 'class
>> stap_probe', and a new 'class stap_static_probe_ops' to replace the
>> use of 'stap_probe_ops'. Both classes implement the virtual methods
>> exported by their parents, 'class probe' and 'class static_probe_ops',
>> respectively. I believe it's now a bit simpler to understand the
>> logic behind the stap-probe interface.
>>
>> There are several helper functions used to parse parts of a stap
>> probe, and since they are generic and don't need to know about the
>> probe they're working on, I decided to leave them as simple static
>> functions (instead of e.g. converting them to class methods).
>>
>> I've also converted a few uses of "VEC" to "std::vector", which makes
>> the code simpler and easier to maintain. And, as usual, some cleanups
>> here and there.
>>
>> Even though I'm sending a series of patches, they need to be tested
>> and committed as a single unit, because of inter-dependencies. But it
>> should be easier to review in separate logical units.
>
> Hi Sergio,
>
> This look very good to me. In general, try to use const-references when
> iterating on vector of objects (in all 3 patches). A few comments below.
Hey Simon,
Thanks for the review. I also saw your comment on IRC; good catch,
thanks. I'll update the code.
>> I've regtested this patch on BuildBot, no regressions found.
>>
>> gdb/ChangeLog:
>> 2017-11-13 Sergio Durigan Junior <sergiodj@redhat.com>
>>
>> * stap-probe.c (struct probe_ops stap_probe_ops): Delete
>> variable.
>> (stap_probe_arg_s): Delete type and VEC.
>> (struct stap_probe): Delete. Replace by...
>> (class stap_static_probe_ops): ...this and...
>> (class stap_probe): ...this. Rename variables to add 'm_'
>> prefix. Do not use 'union' for arguments anymore.
>> (stap_get_expected_argument_type): Receive probe name instead
>> of 'struct stap_probe'. Adjust code.
>> (stap_parse_probe_arguments): Rename to...
>> (stap_probe::parse_arguments): ...this. Adjust code to
>> reflect change.
>> (stap_get_probe_address): Rename to...
>> (stap_probe::get_relocated_address): ...this. Adjust code
>> to reflect change.
>> (stap_get_probe_argument_count): Rename to...
>> (stap_probe::get_argument_count): ...this. Adjust code
>> to reflect change.
>> (stap_get_arg): Rename to...
>> (stap_probe::get_arg_by_number'): ...this. Adjust code to
>> reflect change.
>> (can_evaluate_probe_arguments): Rename to...
>> (stap_probe::can_evaluate_arguments): ...this. Adjust code
>> to reflect change.
>> (stap_evaluate_probe_argument): Rename to...
>> (stap_probe::evaluate_argument): ...this. Adjust code
>> to reflect change.
>> (stap_compile_to_ax): Rename to...
>> (stap_probe::compile_to_ax): ...this. Adjust code to
>> reflect change.
>> (stap_probe_destroy): Delete.
>> (stap_modify_semaphore): Adjust comment.
>> (stap_set_semaphore): Rename to...
>> (stap_probe::set_semaphore): ...this. Adjust code to reflect
>> change.
>> (stap_clear_semaphore): Rename to...
>> (stap_probe::clear_semaphore): ...this. Adjust code to
>> reflect change.
>> (stap_probe::get_static_ops): New method.
>> (handle_stap_probe): Adjust code to create instance of
>> 'stap_probe'.
>> (stap_get_probes): Rename to...
>> (stap_static_probe_ops::get_probes): ...this. Adjust code to
>> reflect change.
>> (stap_probe_is_linespec): Rename to...
>> (stap_static_probe_ops::is_linespec): ...this. Adjust code to
>> reflect change.
>> (stap_type_name): Rename to...
>> (stap_static_probe_ops::type_name): ...this. Adjust code to
>> reflect change.
>> (stap_static_probe_ops::emit_info_probes_extra_fields): New
>> method.
>> (stap_gen_info_probes_table_header): Rename to...
>> (stap_static_probe_ops::gen_info_probes_table_header):
>> ...this. Adjust code to reflect change.
>> (stap_gen_info_probes_table_values): Rename to...
>> (stap_probe::gen_info_probes_table_values): ...this. Adjust
>> code to reflect change.
>> (struct probe_ops stap_probe_ops): Delete.
>> (info_probes_stap_command): Use 'info_probes_for_spops'
>> instead of 'info_probes_for_ops'.
>> (_initialize_stap_probe): Use 'all_static_probe_ops' instead
>> of 'all_probe_ops'.
>> ---
>> gdb/stap-probe.c | 478 ++++++++++++++++++++++++++++---------------------------
>> 1 file changed, 244 insertions(+), 234 deletions(-)
>>
>> diff --git a/gdb/stap-probe.c b/gdb/stap-probe.c
>> index 6fa0d20280..c169ffc488 100644
>> --- a/gdb/stap-probe.c
>> +++ b/gdb/stap-probe.c
>> @@ -45,10 +45,6 @@
>>
>> #define STAP_BASE_SECTION_NAME ".stapsdt.base"
>>
>> -/* Forward declaration. */
>> -
>> -extern const struct probe_ops stap_probe_ops;
>> -
>> /* Should we display debug information for the probe's argument expression
>> parsing? */
>>
>> @@ -95,32 +91,135 @@ struct stap_probe_arg
>> struct expression *aexpr;
>> };
>>
>> -typedef struct stap_probe_arg stap_probe_arg_s;
>> -DEF_VEC_O (stap_probe_arg_s);
>> +/* Class that implements the static probe methods for "stap" probes. */
>>
>> -struct stap_probe
>> +class stap_static_probe_ops : public static_probe_ops
>> {
>> - /* Generic information about the probe. This shall be the first element
>> - of this struct, in order to maintain binary compatibility with the
>> - `struct probe' and be able to fully abstract it. */
>> - struct probe p;
>> +public:
>> + /* See probe.h. */
>> + bool is_linespec (const char **linespecp) const override;
>> +
>> + /* See probe.h. */
>> + void get_probes (std::vector<probe *> *probesp,
>> + struct objfile *objfile) const override;
>> +
>> + /* See probe.h. */
>> + const char *type_name () const override;
>> +
>> + /* See probe.h. */
>> + bool emit_info_probes_extra_fields () const override;
>> +
>> + /* See probe.h. */
>> + void gen_info_probes_table_header
>> + (std::vector<struct info_probe_column> *heads) const override;
>> +};
>> +
>> +/* SystemTap static_probe_ops. */
>> +
>> +const stap_static_probe_ops stap_static_probe_ops;
>>
>> +class stap_probe : public probe
>> +{
>> +public:
>> + /* Constructor for stap_probe. */
>> + stap_probe (std::string &&name_, std::string &&provider_, CORE_ADDR address_,
>> + struct gdbarch *arch_, CORE_ADDR sem_addr, const char *args_text)
>> + : probe (std::move (name_), std::move (provider_), address_, arch_),
>> + m_sem_addr (sem_addr),
>> + m_have_parsed_args (false), m_unparsed_args_text (args_text)
>> + {}
>> +
>> + /* Destructor for stap_probe. */
>> + ~stap_probe ()
>> + {
>> + if (m_have_parsed_args)
>> + for (struct stap_probe_arg arg : m_parsed_args)
>> + xfree (arg.aexpr);
>> + }
>
> I would suggest adding a destructor to stap_probe_arg instead, since it's
> the object that "owns" the expression. You would need to disable copy
> construction and assignment (using DISABLE_COPY_AND_ASSIGN) to avoid double
> free. You can then add a simple constructor and use emplace_back when
> inserting in the vector.
Good point. As we discussed on IRC, I implemented things the way you
proposed, with expression_up. I'll send the updated patch.
>> +
>> + /* See probe.h. */
>> + CORE_ADDR get_relocated_address (struct objfile *objfile) override;
>> +
>> + /* See probe.h. */
>> + unsigned get_argument_count (struct frame_info *frame) override;
>> +
>> + /* See probe.h. */
>> + int can_evaluate_arguments () const override;
>> +
>> + /* See probe.h. */
>> + struct value *evaluate_argument (unsigned n,
>> + struct frame_info *frame) override;
>> +
>> + /* See probe.h. */
>> + void compile_to_ax (struct agent_expr *aexpr,
>> + struct axs_value *axs_value,
>> + unsigned n) override;
>> +
>> + /* See probe.h. */
>> + void set_semaphore (struct objfile *objfile,
>> + struct gdbarch *gdbarch) override;
>> +
>> + /* See probe.h. */
>> + void clear_semaphore (struct objfile *objfile,
>> + struct gdbarch *gdbarch) override;
>> +
>> + /* See probe.h. */
>> + const static_probe_ops *get_static_ops () const override;
>> +
>> + /* See probe.h. */
>> + void gen_info_probes_table_values
>> + (std::vector<const char *> *values) const override;
>> +
>> + /* Return argument N of probe.
>> +
>> + If the probe's arguments have not been parsed yet, parse them. If
>> + there are no arguments, throw an exception (error). Otherwise,
>> + return the requested argument. */
>> + struct stap_probe_arg *get_arg_by_number (unsigned n,
>> + struct gdbarch *gdbarch)
>> + {
>> + if (!m_have_parsed_args)
>> + this->parse_arguments (gdbarch);
>> +
>> + gdb_assert (m_have_parsed_args);
>> + if (m_parsed_args.empty ())
>> + internal_error (__FILE__, __LINE__,
>> + _("Probe '%s' apparently does not have arguments, but \n"
>> + "GDB is requesting its argument number %u anyway. "
>> + "This should not happen. Please report this bug."),
>> + this->get_name (), n);
>> +
>> + return &m_parsed_args[n];
>
> There wasn't one before, but it sounds to me like there should be a bound check here...
Will do.
>> @@ -1081,26 +1179,16 @@ stap_parse_argument (const char **arg, struct type *atype,
>> return p.pstate.expout;
>> }
>>
>> -/* Function which parses an argument string from PROBE, correctly splitting
>> - the arguments and storing their information in properly ways.
>> +/* Implementation of 'parse_probe_arguments' method. */
>
> of 'parse_arguments' method ?
Fixed.
>> +void
>> +stap_probe::set_semaphore (struct objfile *objfile, struct gdbarch *gdbarch)
>> {
>> - struct stap_probe *probe = (struct stap_probe *) probe_generic;
>> CORE_ADDR addr;
>>
>> - gdb_assert (probe_generic->pops == &stap_probe_ops);
>> -
>> - addr = (probe->sem_addr
>> + addr = (m_sem_addr
>> + ANOFFSET (objfile->section_offsets, SECT_OFF_DATA (objfile)));
>
> While at it, you could replace this with get_relocated_address() so that this
> computation is not duplicated. Same with clear_semaphore.
Sure thing, done.
Thanks for the review.
--
Sergio
GPG key ID: 237A 54B1 0287 28BF 00EF 31F4 D0EB 7628 65FC 5E36
Please send encrypted e-mail if possible
http://sergiodj.net/
More information about the Gdb-patches
mailing list