gfs2: better code for translating characters
PATCH 1/1 gfs2: better code for translating characters
the original code could work, but I think this code could work better.
--- fs/gfs2/ops_fstype.c | 3 ++- 1 files changed, 2 insertions(+), 1 deletions(-)
fs/gfs2/ops_fstype.c +2 -1
@@ -145,7 +145,8 @@ static int init_names(struct gfs2_sbd *sdp, int silent)snprintf (sdp->sd_proto_name,GFS2_FSNAME_LEN ,"%s" , proto);snprintf (sdp->sd_table_name,GFS2_FSNAME_LEN ,"%s" , table); -while ((table =strchr (sdp->sd_table_name,'/' ))) + table = sdp->sd_table_name; +while ((table =strchr (table,'/' ))) *table ='_' ; out:
Signature · 1 line
1.5.2.2
Quoted 12 lines · Denis Cheng
the original code could work, but I think this code could work better.
Signed-off-by Denis Cheng <[email protected]>--- fs/gfs2/ops_fstype.c | 3 ++- 1 files changed, 2 insertions(+), 1 deletions(-)fs/gfs2/ops_fstype.c +0
-0
fs/gfs2/ops_fstype.c static int init_names(struct gfs2_sbd *sdp, int silent) · line 150
Quoted 7 lines
fs/gfs2/ops_fstype.c +2
-1@@ -145,7 +145,8 @@ static int init_names(struct gfs2_sbd *sdp, int silent)snprintf (sdp->sd_proto_name,GFS2_FSNAME_LEN ,"%s" , proto);snprintf (sdp->sd_table_name,GFS2_FSNAME_LEN ,"%s" , table); -while ((table =strchr (sdp->sd_table_name,'/' ))) + table = sdp->sd_table_name; +while ((table =strchr (table,'/' ))) *table ='_' ;
Sorry, I don't know what the while loop really means, what's the common case that slash character exists? if the '/' appears multiple, the latter code would be better; however, if slash appears rarely, the original would still be better.
Quoted 6 lines
fs/gfs2/ops_fstype.c +0
-1out: --1.5.2.2
Signature · 5 lines
Denis Cheng Linux Application Developer "One of my most productive days was throwing away 1000 lines of code." - Ken Thompson.
Quoted 16 lines · rae l
Quoted 15 lines · Denis Cheng
the original code could work, but I think this code could work better.
Signed-off-by Denis Cheng <[email protected]>--- fs/gfs2/ops_fstype.c | 3 ++- 1 files changed, 2 insertions(+), 1 deletions(-)fs/gfs2/ops_fstype.c +0
-0@@ -145,7 +145,8 @@ static int init_names(struct gfs2_sbd *sdp, int silent)snprintf (sdp->sd_proto_name,GFS2_FSNAME_LEN ,"%s" , proto);snprintf (sdp->sd_table_name,GFS2_FSNAME_LEN ,"%s" , table);
Quoted 9 lines
fs/gfs2/ops_fstype.c static int init_names(struct gfs2_sbd *sdp, int silent) · line 150
Quoted 4 lines
fs/gfs2/ops_fstype.c +2
-1-while ((table =strchr (sdp->sd_table_name,'/' ))) + table = sdp->sd_table_name; +while ((table =strchr (table,'/' ))) *table ='_' ;Sorry, I don't know what the while loop really means, what's the common case that slash character exists? if the '/' appears multiple, the latter code would be better; however, if slash appears rarely, the original would still be better.
Only if the compiler is stupid.
-hpa
A message that is not in the archive (sent off-list or not yet received): [email protected]
Quoted 8 lines · rae l
Quoted 2 lines
Quoted 1 line
Only if the compiler is stupid.
What? Did you know I really say? Could you tell a little more clear?
if the string in
sdp->sd_table_namehas many '/' chars, the latter algorithm will be better.
but if there's no '/' char, one assignment will be wasted.
You seem to have confused modern compiled C with an old BASIC interpreter.
Consider the code in point:
- while ((table = strchr(sdp->sd_table_name, '/'))) + table = sdp->sd_table_name; + while ((table = strchr(table, '/')))
*table = '_';
sdp->sd_table_name refers to a memory location, and will have to be loaded from memory into a register before it can be transmitted to the strchr() function. In the latter case, we call this register "table"; since the value is immediately killed after the function call, there is no reason for the compiler to carry it across the function.
Consider x86-64 as an example:
# Assume sdp is held in %r15 at this point, and assume # the offset of sd_table_name is 0x30.
# First case
.L1:
movq 30(%r15), %rdi # First argument register movb '/', %sil # Second argument register call strchr testq %rax, %rax # Result register jz .L2 movb '_', (%rax) jmp .L1
.L2:
# Second case movq 30(%r15), %rdi .L1: movb '/', %sil call strchr testq %rax, %rax jz .L2 movq %rax, %rdi movb '_', (%rax) jmp .L1
As you can see, in the zero case, the instruction sequence is exactly the same, whereas in the nonzero case, we have replaced a memory load with a register-register copy. On most architectures (x86-64, Alpha and MIPS are the oddballs here) we wouldn't even need the copy.
-hpa
fs/gfs2/ops_fstype.c static int init_names(struct gfs2_sbd *sdp, int silent) · line 150
Quoted 8 lines · H. Peter Anvin
You seem to have confused modern compiled C with an old BASIC interpreter.
Consider the code in point:
fs/gfs2/ops_fstype.c +2
-1-while ((table =strchr (sdp->sd_table_name,'/' ))) + table = sdp->sd_table_name; +while ((table =strchr (table,'/' ))) *table ='_' ;
Sorry, I just mean for call to strchr, things are different, especially for multiple '/' chars appeared.
The while loop's purpose is to translate all '/' chars appeared in sdp->sd_table_name to '_' chars, consider the string:
'a////aa/a/a/...'
if strchr called with sdp->sd_table_name, every strchr would begun at index 0 of the string, but if called with table, every strchr begun at the last searched position.
So I wonder the common case is no existence of '/', or just one or multiple existence? Things are different for these cases.
Quoted 11 lines · rae l
fs/gfs2/ops_fstype.c static int init_names(struct gfs2_sbd *sdp, int silent) · line 150
Quoted 8 lines · H. Peter Anvin
You seem to have confused modern compiled C with an old BASIC interpreter.
Consider the code in point:
fs/gfs2/ops_fstype.c +2
-1-while ((table =strchr (sdp->sd_table_name,'/' ))) + table = sdp->sd_table_name; +while ((table =strchr (table,'/' ))) *table ='_' ;Sorry, I just mean for call to strchr, things are different, especially for multiple '/' chars appeared.
Things are different only for multiple '/' chars. In that case, the latter form is better, in the former case, there is no difference.
So it doesn't matter. One is unambiguously better.
-hpa
Hi,
Quoted 13 lines · H. Peter Anvin
Quoted 11 lines · rae l
Review
Quoted 8 lines · H. Peter Anvin
You seem to have confused modern compiled C with an old BASIC interpreter.
Consider the code in point:
(quoted diff) +2
-1- while ((table = strchr(sdp->sd_table_name, '/'))) + table = sdp->sd_table_name; + while ((table = strchr(table, '/'))) *table = '_';Sorry, I just mean for call to strchr, things are different, especially for multiple '/' chars appeared.
Quoted 6 lines
Things are different only for multiple '/' chars. In that case, the latter form is better, in the former case, there is no difference.
So it doesn't matter. One is unambiguously better.
-hpa
I agree, so I've applied the patch. Thanks Denis,
Steve.
Hi,
Quoted 19 lines · rae l
fs/gfs2/ops_fstype.c static int init_names(struct gfs2_sbd *sdp, int silent) · line 150
Quoted 8 lines · H. Peter Anvin
You seem to have confused modern compiled C with an old BASIC interpreter.
Consider the code in point:
fs/gfs2/ops_fstype.c +2
-1-while ((table =strchr (sdp->sd_table_name,'/' ))) + table = sdp->sd_table_name; +while ((table =strchr (table,'/' ))) *table ='_' ;Sorry, I just mean for call to strchr, things are different, especially for multiple '/' chars appeared.
The while loop's purpose is to translate all '/' chars appeared in
sdp->sd_table_nameto '_' chars, consider the string:
'a////aa/a/a/...'
if strchr called withsdp->sd_table_name, every strchr would begun at index 0 of the string, but if called with table, every strchr begun at the last searched position.
Quoted 2 lines
So I wonder the common case is no existence of '/', or just one or multiple existence? Things are different for these cases.
The common case is for there to be a few '/' characters since the table name is often the name of the device upon which the fs is mounted, e.g. /dev/sda so it general its not going to be a long string or have many '/' in it,
Steve.