All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Alex Bennée" <alex.bennee@linaro.org>
To: "Daniel P. Berrangé" <berrange@redhat.com>
Cc: yamamoto@midokura.com, qemu-devel@nongnu.org,
	Laurent Vivier <laurent@vivier.eu>
Subject: Re: [RFC PATCH] linux-user: glib-ify is_proc_myself
Date: Mon, 24 May 2021 13:35:47 +0100	[thread overview]
Message-ID: <87o8d045j8.fsf@linaro.org> (raw)
In-Reply-To: <YKuPiul64Uu9vlHs@redhat.com>


Daniel P. Berrangé <berrange@redhat.com> writes:

> On Mon, May 24, 2021 at 12:23:23PM +0100, Alex Bennée wrote:
>> I'm not sure if this is neater than the original code but it does
>> remove a bunch of the !strcmp's in favour of glib's more natural bool
>> results. While we are at it make the function a bool return and fixup
>> the fake_open function prototypes.
>> 
>> Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
>> ---
>>  linux-user/syscall.c | 30 ++++++++++++------------------
>>  1 file changed, 12 insertions(+), 18 deletions(-)
>> 
>> diff --git a/linux-user/syscall.c b/linux-user/syscall.c
>> index e739921e86..18e953de9d 100644
>> --- a/linux-user/syscall.c
>> +++ b/linux-user/syscall.c
>> @@ -7987,33 +7987,27 @@ static int open_self_auxv(void *cpu_env, int fd)
>>      return 0;
>>  }
>>  
>> -static int is_proc_myself(const char *filename, const char *entry)
>> +static bool is_proc_myself(const char *filename, const char *entry)
>>  {
>> -    if (!strncmp(filename, "/proc/", strlen("/proc/"))) {
>> +    if (g_str_has_prefix(filename, "/proc/")) {
>>          filename += strlen("/proc/");
>> -        if (!strncmp(filename, "self/", strlen("self/"))) {
>> +        if (g_str_has_prefix(filename, "self/")) {
>>              filename += strlen("self/");
>> -        } else if (*filename >= '1' && *filename <= '9') {
>> -            char myself[80];
>> -            snprintf(myself, sizeof(myself), "%d/", getpid());
>> -            if (!strncmp(filename, myself, strlen(myself))) {
>> -                filename += strlen(myself);
>> -            } else {
>> -                return 0;
>> +        } else if (g_ascii_isdigit(*filename)) {
>> +            g_autofree char * myself = g_strdup_printf("%d/", getpid());
>> +            if (!g_str_has_prefix(filename, myself)) {
>> +                return false;
>>              }
>> -        } else {
>> -            return 0;
>> -        }
>> -        if (!strcmp(filename, entry)) {
>> -            return 1;
>> +            filename += strlen(myself);
>>          }
>> +        return g_str_has_prefix(filename, entry);
>>      }
>> -    return 0;
>> +    return false;
>>  }
>
> Diff is hard to compare, so this is what it looks like:
>
>  static bool is_proc_myself(const char *filename, const char *entry)
>  {
>      if (g_str_has_prefix(filename, "/proc/")) {
>          filename += strlen("/proc/");
>          if (g_str_has_prefix(filename, "self/")) {
>              filename += strlen("self/");
>          } else if (g_ascii_isdigit(*filename)) {
>              g_autofree char * myself = g_strdup_printf("%d/", getpid());
>              if (!g_str_has_prefix(filename, myself)) {
>                  return false;
>              }
>              filename += strlen(myself);
>          }
>          return g_str_has_prefix(filename, entry);
>      }
>      return false;
>  }
>
> I think if we don't mind two heap allocs it can be simplified to:
>
>   static int is_proc_myself(const char *filename, const char *entry)
>   {
>     g_autofree char *procself = g_strdup_printf("/proc/self/%s", entry);
>     g_autofree char *procpid = g_strdup_printf("/proc/%d/%s", getpid(), entry);
>     return g_str_equal(filename, procself) || g_str_equal(filename, procpid);
>   }

Ahh nice, even simpler and easy to follow. I don't think the double
alloc is too much of a concern because we are typically on a syscall
path anyway so have a bunch of stuff to check.

> This makes me wonder if the code needs to cope with non-canonicalized
> filenames though. eg /proc///self/maps  or /proc/./self/maps
>
> Is something further up ensuring canonicalization of 'filename' ?

It seems so from my cursory pokes but I'm not convinced all paths do. We
could throw in a g_canonicalize_filename to be sure.

>
>
>>  
>>  #if defined(HOST_WORDS_BIGENDIAN) != defined(TARGET_WORDS_BIGENDIAN) || \
>>      defined(TARGET_SPARC) || defined(TARGET_M68K) || defined(TARGET_HPPA)
>> -static int is_proc(const char *filename, const char *entry)
>> +static bool is_proc(const char *filename, const char *entry)
>>  {
>>      return strcmp(filename, entry) == 0;
>>  }
>> @@ -8097,7 +8091,7 @@ static int do_openat(void *cpu_env, int dirfd, const char *pathname, int flags,
>>      struct fake_open {
>>          const char *filename;
>>          int (*fill)(void *cpu_env, int fd);
>> -        int (*cmp)(const char *s1, const char *s2);
>> +        bool (*cmp)(const char *s1, const char *s2);
>>      };
>>      const struct fake_open *fake_open;
>>      static const struct fake_open fakes[] = {
>> -- 
>> 2.20.1
>> 
>> 
>
> Regards,
> Daniel


-- 
Alex Bennée


  reply	other threads:[~2021-05-24 13:38 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-05-24 11:23 [RFC PATCH] linux-user: glib-ify is_proc_myself Alex Bennée
2021-05-24 11:27 ` no-reply
2021-05-24 11:35 ` Daniel P. Berrangé
2021-05-24 12:35   ` Alex Bennée [this message]
2021-05-24 19:12 ` Richard Henderson

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=87o8d045j8.fsf@linaro.org \
    --to=alex.bennee@linaro.org \
    --cc=berrange@redhat.com \
    --cc=laurent@vivier.eu \
    --cc=qemu-devel@nongnu.org \
    --cc=yamamoto@midokura.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.