Re: a strlen update

From: Henrik Holst via curl-library <curl-library_at_lists.haxx.se>
Date: Mon, 14 Feb 2022 20:23:40 +0100

Those single byte memcpy() comes from alloc_addbyter() which in turn is
called by Curl_dyn_vprintf() and curl_mvaprintf() and I can see from traces
that they are e.g called like this:

   0.180 us [ 66160] | strlen(".curlrc") = 7;
            [ 66160] | alloc_addbyter() {
            [ 66160] | Curl_dyn_addn() {
            [ 66160] | dyn_nappend() {
   0.110 us [ 66160] | memcpy(0x5616b0c976c5,
0x7fff3533def7, 1);
   0.290 us [ 66160] | } /* dyn_nappend */
   0.401 us [ 66160] | } /* Curl_dyn_addn */
   0.501 us [ 66160] | } /* alloc_addbyter */
            [ 66160] | alloc_addbyter() {
            [ 66160] | Curl_dyn_addn() {
            [ 66160] | dyn_nappend() {
   0.100 us [ 66160] | memcpy(0x5616b0c976c6,
0x7fff3533def7, 1);
   0.280 us [ 66160] | } /* dyn_nappend */
   0.380 us [ 66160] | } /* Curl_dyn_addn */
   0.481 us [ 66160] | } /* alloc_addbyter */
            [ 66160] | alloc_addbyter() {
            [ 66160] | Curl_dyn_addn() {
            [ 66160] | dyn_nappend() {
   0.100 us [ 66160] | memcpy(0x5616b0c976c7,
0x7fff3533def7, 1);
   0.280 us [ 66160] | } /* dyn_nappend */
   0.380 us [ 66160] | } /* Curl_dyn_addn */
   0.480 us [ 66160] | } /* alloc_addbyter */
            [ 66160] | alloc_addbyter() {
            [ 66160] | Curl_dyn_addn() {
            [ 66160] | dyn_nappend() {
   0.100 us [ 66160] | memcpy(0x5616b0c976c8,
0x7fff3533def7, 1);
   0.280 us [ 66160] | } /* dyn_nappend */
   0.380 us [ 66160] | } /* Curl_dyn_addn */
   0.491 us [ 66160] | } /* alloc_addbyter */
            [ 66160] | alloc_addbyter() {
            [ 66160] | Curl_dyn_addn() {
            [ 66160] | dyn_nappend() {
   0.100 us [ 66160] | memcpy(0x5616b0c976c9,
0x7fff3533def7, 1);
   0.280 us [ 66160] | } /* dyn_nappend */
   0.380 us [ 66160] | } /* Curl_dyn_addn */
   0.481 us [ 66160] | } /* alloc_addbyter */
            [ 66160] | alloc_addbyter() {
            [ 66160] | Curl_dyn_addn() {
            [ 66160] | dyn_nappend() {
   0.100 us [ 66160] | memcpy(0x5616b0c976ca,
0x7fff3533def7, 1);
   0.280 us [ 66160] | } /* dyn_nappend */
   0.381 us [ 66160] | } /* Curl_dyn_addn */
   0.481 us [ 66160] | } /* alloc_addbyter */
            [ 66160] | alloc_addbyter() {
            [ 66160] | Curl_dyn_addn() {
            [ 66160] | dyn_nappend() {
   0.100 us [ 66160] | memcpy(0x5616b0c976cb,
0x7fff3533def7, 1);
   0.270 us [ 66160] | } /* dyn_nappend */
   0.381 us [ 66160] | } /* Curl_dyn_addn */
   0.481 us [ 66160] | } /* alloc_addbyter */
  17.002 us [ 66160] | } /* dprintf_formatf */

So first ".curlrc" is strlen:ed and then 7 calls to alloc_addbyter() which
is the exact length of ".curlrc" so it looks like some function goes over
strings one byte at a time and adds them to a dynamic buffer one by one.
Perhaps this is done to exclude invalid characters or something?

Many times a file path is first strlen:ed when I see these one byte memcpy
(but far from every time). alloc_addbyter() could most likely be optimized
to do a single write to the buffer instead of calling memcpy() by way of
Curl_dyn_addn(buf, out, 1) but I think some one should take a look at why
those file paths are gone over one by one (there could be a good reason)
and if there might be a better approach.

/HH

Den mån 14 feb. 2022 kl 09:33 skrev Daniel Stenberg via curl-library <
curl-library_at_lists.haxx.se>:

> Other fun observations running this same command line:
>
> The top-25 lib calls done:
>
> 2239 free
> 635 malloc
> 545 memcpy
> 300 memchr
> 245 strcmp
> 225 strlen
> 162 inet_pton
> 134 clock_gettime
> 110 OPENSSL_sk_value
> 110 ASN1_STRING_length
> 110 ASN1_STRING_get0_data
> 75 strdup
> 63 sigaction
> 56 calloc
> 50 nghttp2_session_get_stream_user_data
> 38 realloc
> 37 ERR_clear_error
> 36 strchr
> 31 poll
> 31 nghttp2_session_want_read
> 28 nghttp2_session_send
> 26 nghttp2_session_mem_recv
> 26 nghttp2_session_check_request_allowed
> 25 getenv
> 24 SSL_set_ex_data
>
> Those 545 memcpy() calls are interesting because 384 of them copy only a
> single byte!
>
> --
>
> / daniel.haxx.se
> | Commercial curl support up to 24x7 is available!
> | Private help, bug fixes, support, ports, new features
> | https://curl.se/support.html
> --
> Unsubscribe: https://lists.haxx.se/listinfo/curl-library
> Etiquette: https://curl.haxx.se/mail/etiquette.html
>

-- 
Unsubscribe: https://lists.haxx.se/listinfo/curl-library
Etiquette:   https://curl.haxx.se/mail/etiquette.html

Received on 2022-02-14