# [PATCH 1/2] code cleanup in sha1_file.c: xmalloc -> xcalloc 5 messages from 2009-02-10 to 2009-02-11. Participants: Guanqun Lu, Junio C Hamano, Brandon Casey. Thread: https://gitlist.dev/t/17701 ## Junio C Hamano, 2009-02-10 15:29 Subject: Re: [PATCH 2/2] lockfile: set lk->fd = -1 in 'rollback_lock_file()' Message-ID: <7v63jipax4.fsf@gitster.siamese.dyndns.org> URL: https://gitlist.dev/e/7v63jipax4.fsf%40gitster.siamese.dyndns.org In-Reply-To: <1234334612-24907-2-git-send-email-guanqun.lu@gmail.com> ``` Guanqun Lu writes: > Post-conditions when these functions return successfully: > lk->fd == -1? lk->filename[0] == '\0'? > close_lock_file() yes no > commit_lock_file() yes yes > rollback_lock_file() no* yes > > [*] This commit changes this 'no' in rollback_lock_file() to 'yes', > which achieves more robust and unified interface. Is there a broken caller this patch fixes? IOW, is there a codepath that can call rollback and then later notices lk->fd is not negative and uses it without checking lk->filename[0] first? > Signed-off-by: Guanqun Lu > --- > lockfile.c | 4 +++- > 1 files changed, 3 insertions(+), 1 deletions(-) > > diff --git a/lockfile.c b/lockfile.c > index 021c337..44e5253 100644 > --- a/lockfile.c > +++ b/lockfile.c > @@ -243,8 +243,10 @@ int commit_locked_index(struct lock_file *lk) > void rollback_lock_file(struct lock_file *lk) > { > if (lk->filename[0]) { > - if (lk->fd >= 0) > + if (lk->fd >= 0) { > close(lk->fd); > + lk->fd = -1; > + } > unlink(lk->filename); > } > lk->filename[0] = 0; > -- > 1.6.1.2.392.gb04d1 ``` ## Brandon Casey, 2009-02-10 16:29 Subject: Re: [PATCH 1/2] code cleanup in sha1_file.c: xmalloc -> xcalloc Message-ID: URL: https://gitlist.dev/e/EohcOnsg9EGsisIaM0B0iYUSHt-mz_fzqBCqg27mBU0yXXlkP9baOA%40cipher.nrlssc.navy.mil In-Reply-To: <1234334612-24907-1-git-send-email-guanqun.lu@gmail.com> ``` Guanqun Lu wrote: > 'xmalloc' followed immediately by 'memset' is replaced > with 'xcalloc', and a simple grep in this project seems > to show that it's the only place. > > Signed-off-by: Guanqun Lu > --- > sha1_file.c | 3 +-- > 1 files changed, 1 insertions(+), 2 deletions(-) > > diff --git a/sha1_file.c b/sha1_file.c > index 8868b80..93e5fc0 100644 > --- a/sha1_file.c > +++ b/sha1_file.c > @@ -858,8 +858,7 @@ unsigned char* use_pack(struct packed_git *p, > > static struct packed_git *alloc_packed_git(int extra) > { > - struct packed_git *p = xmalloc(sizeof(*p) + extra); > - memset(p, 0, sizeof(*p)); > + struct packed_git *p = xcalloc(1, sizeof(*p) + extra); > p->pack_fd = -1; > return p; > } I previously scanned through and did this. I left this one as is because the extra part is about as large as the sizeof(*p) part. 69 and 72 bytes respectively on 32-bit. The extra part is always filled in immediately by callers. It's only called once for each pack so it's not performance critical, so maybe your patch is worth it since it is simpler? -brandon ``` ## Junio C Hamano, 2009-02-10 16:43 Subject: Re: [PATCH 1/2] code cleanup in sha1_file.c: xmalloc -> xcalloc Message-ID: <7v3aemnsxp.fsf@gitster.siamese.dyndns.org> URL: https://gitlist.dev/e/7v3aemnsxp.fsf%40gitster.siamese.dyndns.org In-Reply-To: <1234334612-24907-1-git-send-email-guanqun.lu@gmail.com> ``` Guanqun Lu writes: > 'xmalloc' followed immediately by 'memset' is replaced > with 'xcalloc', and a simple grep in this project seems > to show that it's the only place. But isn't this memset() done only for the initial part of the allocated area, not the whole thing? You are not cleaning up but changing what it does, if I am reading this code correctly. > Signed-off-by: Guanqun Lu > --- > sha1_file.c | 3 +-- > 1 files changed, 1 insertions(+), 2 deletions(-) > > diff --git a/sha1_file.c b/sha1_file.c > index 8868b80..93e5fc0 100644 > --- a/sha1_file.c > +++ b/sha1_file.c > @@ -858,8 +858,7 @@ unsigned char* use_pack(struct packed_git *p, > > static struct packed_git *alloc_packed_git(int extra) > { > - struct packed_git *p = xmalloc(sizeof(*p) + extra); > - memset(p, 0, sizeof(*p)); > + struct packed_git *p = xcalloc(1, sizeof(*p) + extra); > p->pack_fd = -1; > return p; > } > -- > 1.6.1.2.392.gb04d1 ``` ## Guanqun Lu, 2009-02-11 06:43 Subject: [PATCH 1/2] code cleanup in sha1_file.c: xmalloc -> xcalloc Message-ID: <1234334612-24907-1-git-send-email-guanqun.lu@gmail.com> URL: https://gitlist.dev/e/1234334612-24907-1-git-send-email-guanqun.lu%40gmail.com ``` 'xmalloc' followed immediately by 'memset' is replaced with 'xcalloc', and a simple grep in this project seems to show that it's the only place. Signed-off-by: Guanqun Lu --- sha1_file.c | 3 +-- 1 files changed, 1 insertions(+), 2 deletions(-) diff --git a/sha1_file.c b/sha1_file.c index 8868b80..93e5fc0 100644 --- a/sha1_file.c +++ b/sha1_file.c @@ -858,8 +858,7 @@ unsigned char* use_pack(struct packed_git *p, static struct packed_git *alloc_packed_git(int extra) { - struct packed_git *p = xmalloc(sizeof(*p) + extra); - memset(p, 0, sizeof(*p)); + struct packed_git *p = xcalloc(1, sizeof(*p) + extra); p->pack_fd = -1; return p; } -- 1.6.1.2.392.gb04d1 ``` ## Guanqun Lu, 2009-02-11 06:43 Subject: [PATCH 2/2] lockfile: set lk->fd = -1 in 'rollback_lock_file()' Message-ID: <1234334612-24907-2-git-send-email-guanqun.lu@gmail.com> URL: https://gitlist.dev/e/1234334612-24907-2-git-send-email-guanqun.lu%40gmail.com In-Reply-To: <1234334612-24907-1-git-send-email-guanqun.lu@gmail.com> ``` Post-conditions when these functions return successfully: lk->fd == -1? lk->filename[0] == '\0'? close_lock_file() yes no commit_lock_file() yes yes rollback_lock_file() no* yes [*] This commit changes this 'no' in rollback_lock_file() to 'yes', which achieves more robust and unified interface. Signed-off-by: Guanqun Lu --- lockfile.c | 4 +++- 1 files changed, 3 insertions(+), 1 deletions(-) diff --git a/lockfile.c b/lockfile.c index 021c337..44e5253 100644 --- a/lockfile.c +++ b/lockfile.c @@ -243,8 +243,10 @@ int commit_locked_index(struct lock_file *lk) void rollback_lock_file(struct lock_file *lk) { if (lk->filename[0]) { - if (lk->fd >= 0) + if (lk->fd >= 0) { close(lk->fd); + lk->fd = -1; + } unlink(lk->filename); } lk->filename[0] = 0; -- 1.6.1.2.392.gb04d1 ```