* [PATCH 0/6] another batch of fixes
@ 2022-11-16 9:26 Eric Wong
2022-11-16 9:26 ` [PATCH 1/6] avoid leaking array returned by backtrace_symbols() Eric Wong
` (5 more replies)
0 siblings, 6 replies; 7+ messages in thread
From: Eric Wong @ 2022-11-16 9:26 UTC (permalink / raw)
To: mwrap-perl
Every bug I fix makes me feel dumber :<
Eric Wong (6):
avoid leaking array returned by backtrace_symbols()
mwrap_core: document the reason for padding the mutex
wrap jhash calls to hash_loc
fix under-allocated kbuf size
mwrap_get: golf out redundant assignment across branches
location_name: handle malloc failure on backtrace_symbols(3)
Mwrap.xs | 18 +++++++++++-------
mwrap_core.h | 13 +++++++++----
2 files changed, 20 insertions(+), 11 deletions(-)
^ permalink raw reply [flat|nested] 7+ messages in thread
* [PATCH 1/6] avoid leaking array returned by backtrace_symbols()
2022-11-16 9:26 [PATCH 0/6] another batch of fixes Eric Wong
@ 2022-11-16 9:26 ` Eric Wong
2022-11-16 9:26 ` [PATCH 2/6] mwrap_core: document the reason for padding the mutex Eric Wong
` (4 subsequent siblings)
5 siblings, 0 replies; 7+ messages in thread
From: Eric Wong @ 2022-11-16 9:26 UTC (permalink / raw)
To: mwrap-perl
Oops :x Fortunately, most allocations are tracked to a Perl
location and the non-Perl code path is infrequently hit.
---
Mwrap.xs | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/Mwrap.xs b/Mwrap.xs
index d5336fa..8aeb425 100644
--- a/Mwrap.xs
+++ b/Mwrap.xs
@@ -20,8 +20,8 @@ static SV *location_string(struct src_loc *l)
char **s = backtrace_symbols((void *)l->k, 1);
ret = newSVpvn(s[0], strlen(s[0]));
- }
- else {
+ free(s);
+ } else {
ret = newSVpvn(l->k, l->capa - 1);
}
^ permalink raw reply related [flat|nested] 7+ messages in thread
* [PATCH 2/6] mwrap_core: document the reason for padding the mutex
2022-11-16 9:26 [PATCH 0/6] another batch of fixes Eric Wong
2022-11-16 9:26 ` [PATCH 1/6] avoid leaking array returned by backtrace_symbols() Eric Wong
@ 2022-11-16 9:26 ` Eric Wong
2022-11-16 9:26 ` [PATCH 3/6] wrap jhash calls to hash_loc Eric Wong
` (3 subsequent siblings)
5 siblings, 0 replies; 7+ messages in thread
From: Eric Wong @ 2022-11-16 9:26 UTC (permalink / raw)
To: mwrap-perl
While 64 has an obvious meaning to me, I reject Moore's law
and have no idea what cache line sizes modern CPUs use.
---
mwrap_core.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/mwrap_core.h b/mwrap_core.h
index 09b579d..94a6af3 100644
--- a/mwrap_core.h
+++ b/mwrap_core.h
@@ -68,7 +68,7 @@ static size_t *root_locating; /* determines if PL_curcop is our thread */
static struct cds_lfht *totals;
union padded_mutex {
pthread_mutex_t mtx;
- char pad[64];
+ char pad[64]; /* cache alignment for common CPUs */
};
/* a round-robin pool of mutexes */
^ permalink raw reply related [flat|nested] 7+ messages in thread
* [PATCH 3/6] wrap jhash calls to hash_loc
2022-11-16 9:26 [PATCH 0/6] another batch of fixes Eric Wong
2022-11-16 9:26 ` [PATCH 1/6] avoid leaking array returned by backtrace_symbols() Eric Wong
2022-11-16 9:26 ` [PATCH 2/6] mwrap_core: document the reason for padding the mutex Eric Wong
@ 2022-11-16 9:26 ` Eric Wong
2022-11-16 9:26 ` [PATCH 4/6] fix under-allocated kbuf size Eric Wong
` (2 subsequent siblings)
5 siblings, 0 replies; 7+ messages in thread
From: Eric Wong @ 2022-11-16 9:26 UTC (permalink / raw)
To: mwrap-perl
This will make it easier to keep a consistent hash seed
and/or switch to different hashes in the future.
---
Mwrap.xs | 4 ++--
mwrap_core.h | 9 +++++++--
2 files changed, 9 insertions(+), 4 deletions(-)
diff --git a/Mwrap.xs b/Mwrap.xs
index 8aeb425..a5ffeed 100644
--- a/Mwrap.xs
+++ b/Mwrap.xs
@@ -164,12 +164,12 @@ CODE:
k = (void *)kbuf;
memcpy(k->k, &p, sizeof(p));
k->capa = 0;
- k->hval = jhash(k->k, sizeof(p), 0xdeadbeef);
+ hash_loc(k, sizeof(p));
} else {
k = (void *)kbuf;
memcpy(k->k, str, len + 1);
k->capa = len + 1;
- k->hval = jhash(k->k, k->capa, 0xdeadbeef);
+ hash_loc(k, k->capa);
}
if (!k)
diff --git a/mwrap_core.h b/mwrap_core.h
index 94a6af3..d9396e6 100644
--- a/mwrap_core.h
+++ b/mwrap_core.h
@@ -272,6 +272,11 @@ static const COP *mwp_curcop(void)
return NULL;
}
+static void hash_loc(struct src_loc *k, size_t len)
+{
+ k->hval = jhash(k->k, len, 0xdeadbeef);
+}
+
static struct src_loc *assign_line(size_t size, const char *file, unsigned line)
{
/* avoid vsnprintf or anything which could call malloc here: */
@@ -298,7 +303,7 @@ static struct src_loc *assign_line(size_t size, const char *file, unsigned line)
assert(dst && "bad math");
*dst = 0; /* terminate string */
k->capa = (uint32_t)(dst - k->k + 1);
- k->hval = jhash(k->k, k->capa, 0xdeadbeef);
+ hash_loc(k, k->capa);
return totals_add_rcu(k);
}
@@ -325,7 +330,7 @@ update_stats_rcu_lock(size_t *generation, size_t size, uintptr_t caller)
k->total = size;
memcpy(k->k, &caller, xlen);
k->capa = 0;
- k->hval = jhash(k->k, xlen, 0xdeadbeef);
+ hash_loc(k, xlen);
ret = totals_add_rcu(k);
}
out:
^ permalink raw reply related [flat|nested] 7+ messages in thread
* [PATCH 4/6] fix under-allocated kbuf size
2022-11-16 9:26 [PATCH 0/6] another batch of fixes Eric Wong
` (2 preceding siblings ...)
2022-11-16 9:26 ` [PATCH 3/6] wrap jhash calls to hash_loc Eric Wong
@ 2022-11-16 9:26 ` Eric Wong
2022-11-16 9:26 ` [PATCH 5/6] mwrap_get: golf out redundant assignment across branches Eric Wong
2022-11-16 9:26 ` [PATCH 6/6] location_name: handle malloc failure on backtrace_symbols(3) Eric Wong
5 siblings, 0 replies; 7+ messages in thread
From: Eric Wong @ 2022-11-16 9:26 UTC (permalink / raw)
To: mwrap-perl
kbuf is always `struct src_loc', not `struct alloc_hdr'.
Fortunately, nobody really needs PATH_MAX size for pathnames, so
this wasn't a problem in practice.
---
mwrap_core.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/mwrap_core.h b/mwrap_core.h
index d9396e6..3abc3dd 100644
--- a/mwrap_core.h
+++ b/mwrap_core.h
@@ -179,7 +179,7 @@ struct alloc_hdr {
/* $PATHNAME:$LINENO */
static MWRAP_TSD char kbuf[
- PATH_MAX + sizeof(":") + UINT2STR_MAX + sizeof(struct alloc_hdr)
+ sizeof(struct src_loc) + PATH_MAX + sizeof(":") + UINT2STR_MAX
];
static struct alloc_hdr *ptr2hdr(void *p)
^ permalink raw reply related [flat|nested] 7+ messages in thread
* [PATCH 5/6] mwrap_get: golf out redundant assignment across branches
2022-11-16 9:26 [PATCH 0/6] another batch of fixes Eric Wong
` (3 preceding siblings ...)
2022-11-16 9:26 ` [PATCH 4/6] fix under-allocated kbuf size Eric Wong
@ 2022-11-16 9:26 ` Eric Wong
2022-11-16 9:26 ` [PATCH 6/6] location_name: handle malloc failure on backtrace_symbols(3) Eric Wong
5 siblings, 0 replies; 7+ messages in thread
From: Eric Wong @ 2022-11-16 9:26 UTC (permalink / raw)
To: mwrap-perl
Terminal space is precious with my font sizes.
---
Mwrap.xs | 4 +---
1 file changed, 1 insertion(+), 3 deletions(-)
diff --git a/Mwrap.xs b/Mwrap.xs
index a5ffeed..c8a4568 100644
--- a/Mwrap.xs
+++ b/Mwrap.xs
@@ -147,7 +147,7 @@ mwrap_get(loc)
PREINIT:
STRLEN len;
const char *str;
- struct src_loc *k = 0;
+ struct src_loc *k = (void *)kbuf;
uintptr_t p;
struct cds_lfht_iter iter;
struct cds_lfht_node *cur;
@@ -161,12 +161,10 @@ CODE:
if (len > PATH_MAX)
XSRETURN_UNDEF;
if (extract_addr(str, len, (void **)&p)) {
- k = (void *)kbuf;
memcpy(k->k, &p, sizeof(p));
k->capa = 0;
hash_loc(k, sizeof(p));
} else {
- k = (void *)kbuf;
memcpy(k->k, str, len + 1);
k->capa = len + 1;
hash_loc(k, k->capa);
^ permalink raw reply related [flat|nested] 7+ messages in thread
* [PATCH 6/6] location_name: handle malloc failure on backtrace_symbols(3)
2022-11-16 9:26 [PATCH 0/6] another batch of fixes Eric Wong
` (4 preceding siblings ...)
2022-11-16 9:26 ` [PATCH 5/6] mwrap_get: golf out redundant assignment across branches Eric Wong
@ 2022-11-16 9:26 ` Eric Wong
5 siblings, 0 replies; 7+ messages in thread
From: Eric Wong @ 2022-11-16 9:26 UTC (permalink / raw)
To: mwrap-perl
backtrace_symbols(3) may fail due to malloc failures, so we
must account for it and return `undef' in that case.
---
Mwrap.xs | 6 ++++++
1 file changed, 6 insertions(+)
diff --git a/Mwrap.xs b/Mwrap.xs
index c8a4568..b515894 100644
--- a/Mwrap.xs
+++ b/Mwrap.xs
@@ -19,6 +19,12 @@ static SV *location_string(struct src_loc *l)
if (loc_is_addr(l)) {
char **s = backtrace_symbols((void *)l->k, 1);
+ if (!s) {
+ fprintf(stderr, "backtrace_symbols => NULL: %s\n",
+ strerror(errno));
+ return &PL_sv_undef;
+ }
+
ret = newSVpvn(s[0], strlen(s[0]));
free(s);
} else {
^ permalink raw reply related [flat|nested] 7+ messages in thread
end of thread, other threads:[~2022-11-16 9:26 UTC | newest]
Thread overview: 7+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-11-16 9:26 [PATCH 0/6] another batch of fixes Eric Wong
2022-11-16 9:26 ` [PATCH 1/6] avoid leaking array returned by backtrace_symbols() Eric Wong
2022-11-16 9:26 ` [PATCH 2/6] mwrap_core: document the reason for padding the mutex Eric Wong
2022-11-16 9:26 ` [PATCH 3/6] wrap jhash calls to hash_loc Eric Wong
2022-11-16 9:26 ` [PATCH 4/6] fix under-allocated kbuf size Eric Wong
2022-11-16 9:26 ` [PATCH 5/6] mwrap_get: golf out redundant assignment across branches Eric Wong
2022-11-16 9:26 ` [PATCH 6/6] location_name: handle malloc failure on backtrace_symbols(3) Eric Wong
Code repositories for project(s) associated with this public inbox
https://80x24.org/mwrap-perl.git
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).