Skip to contentSkip to conversation

linux-user/syscall.c: malloc to g_try_malloc

PATCH Applied Ahmed Abouzied 3 participants 4 messages Jan 4, 2022 – Jan 10, 2022 1 message not in the archive

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 linux-user/syscall.c: malloc to g_try_malloc

Use g_try_malloc instead of malloc to alocate the target ifconfig. Also replace the corresponding free with g_free.

---

Hello,

I noticed that there was a malloc call in this file. It seems that it was added by the commit 22e4a267 (3 years ago) which was after the commit 0e173b24 (6 years ago) that replaced malloc calls with glib alternative calls.

There is no issue for this on Gitlab. Should I have created an issue first?

Best regards,

 linux-user/syscall.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
linux-user/syscall.c +2 -2
index 56a3e17183..715f9430e1 100644
@@ -4867,7 +4867,7 @@ static abi_long do_ioctl_ifconf(const IOCTLEntry *ie, uint8_t *buf_temp,
              * We can't fit all the extents into the fixed size buffer.
              * Allocate one that is large enough and use it instead.
              */
-            host_ifconf = malloc(outbufsz);
+            host_ifconf = g_try_malloc(outbufsz);
             if (!host_ifconf) {
                 return -TARGET_ENOMEM;
             }
@@ -4915,7 +4915,7 @@ static abi_long do_ioctl_ifconf(const IOCTLEntry *ie, uint8_t *buf_temp,
     }
 
     if (free_buf) {
-        free(host_ifconf);
+        g_free(host_ifconf);
     }
 
     return ret;
Signature · 1 line
2.25.1
#1 lore raw

Le 04/01/2022 à 15:38, Ahmed Abouzied a écrit :

Quoted 34 lines

Use g_try_malloc instead of malloc to alocate the target ifconfig. Also replace the corresponding free with g_free.

---

Hello,

I noticed that there was a malloc call in this file. It seems that it was added by the commit 22e4a267 (3 years ago) which was after the commit 0e173b24 (6 years ago) that replaced malloc calls with glib alternative calls.

There is no issue for this on Gitlab. Should I have created an issue first?

Best regards,

  linux-user/syscall.c | 4 ++--
  1 file changed, 2 insertions(+), 2 deletions(-)
linux-user/syscall.c +1 -1
index 56a3e17183..715f9430e1 100644
@@ -4867,7 +4867,7 @@ static abi_long do_ioctl_ifconf(const IOCTLEntry *ie, uint8_t *buf_temp,
               * We can't fit all the extents into the fixed size buffer.
               * Allocate one that is large enough and use it instead.
               */
-            host_ifconf = malloc(outbufsz);
+            host_ifconf = g_try_malloc(outbufsz);
              if (!host_ifconf) {
                  return -TARGET_ENOMEM;
              }

linux-user/syscall.c static abi_long do_ioctl_ifconf(const IOCTLEntry *ie, uint8_t *buf_temp, · line 4921

Quoted 8 lines
linux-user/syscall.c +1 -1
@@ -4915,7 +4915,7 @@ static abi_long do_ioctl_ifconf(const IOCTLEntry *ie, uint8_t *buf_temp,
      }
 
      if (free_buf) {
-        free(host_ifconf);
+        g_free(host_ifconf);
      }
 
      return ret;

Applied to my linux-user-for-7.0 branch.

Thanks,
Laurent

A message that is not in the archive (sent off-list or not yet received): CANL98e9RO38rmP8oSmEcoop3Pr_qC741Om91Bw9FQxAZcqf2QQ@mail.gmail.com

#3 lore raw

Byte Sized task (was Re: [PATCH] linux-user/syscall.c: malloc to g_try_malloc)

Hi Ahmed,

the email address you use for the trivial ML is broken. I've fixed it for my answer and added qemu-devel.

Le 05/01/2022 à 20:09, Ahmed Abouzied a écrit :

Quoted 3 lines

Hello,

Thanks for the review. I also found some occurrences of "cpu_physical_memor_*" calls that should be replaced with "address_space_*" around the codebase as mentioned [here](https://wiki.qemu.org/Contribute/BiteSizedTasks <https://wiki.qemu.org/Contribute/BiteSizedTasks>) in the API conversion section. The "/hw" directory for instance. Should I go about replacing those as well? Should I create a Gitlab issue first to track the changes?

According to the wiki, the gitlab issue are already created with the "Byte Sized" label.

https://gitlab.com/qemu-project/qemu/-/issues?scope=all&state=opened&label_name[]=Bite%20Sized

But this one is not, I think John can help you for that (added in cc:)

Thanks,
Laurent

Quoted 54 lines

Best regards,
Ahmed

On Wed, 5 Jan 2022 at 12:26, Laurent Vivier <[email protected] <mailto:[email protected]>> wrote:

    Le 04/01/2022 à 15:38, Ahmed Abouzied a écrit :
     > Use g_try_malloc instead of malloc to alocate the target ifconfig.
     > Also replace the corresponding free with g_free.
     >
     > Signed-off-by: Ahmed Abouzied <[email protected] <mailto:[email protected]>>
     > ---
     >
     > Hello,
     >
     > I noticed that there was a `malloc` call in this file. It seems that it
     > was added by the commit 22e4a267 (3 years ago) which was after the commit
     > 0e173b24 (6 years ago) that replaced malloc calls with glib alternative calls.
     >
     > There is no issue for this on Gitlab. Should I have created an issue
     > first?
     >
     > Best regards,
     >
     >   linux-user/syscall.c | 4 ++--
     >   1 file changed, 2 insertions(+), 2 deletions(-)
     >
     > diff --git a/linux-user/syscall.c b/linux-user/syscall.c
     > index 56a3e17183..715f9430e1 100644
     > --- a/linux-user/syscall.c
     > +++ b/linux-user/syscall.c
     > @@ -4867,7 +4867,7 @@ static abi_long do_ioctl_ifconf(const IOCTLEntry *ie, uint8_t *buf_temp,
     >                * We can't fit all the extents into the fixed size buffer.
     >                * Allocate one that is large enough and use it instead.
     >                */
     > -            host_ifconf = malloc(outbufsz);
     > +            host_ifconf = g_try_malloc(outbufsz);
     >               if (!host_ifconf) {
     >                   return -TARGET_ENOMEM;
     >               }
     > @@ -4915,7 +4915,7 @@ static abi_long do_ioctl_ifconf(const IOCTLEntry *ie, uint8_t *buf_temp,
     >       }
     >
     >       if (free_buf) {
     > -        free(host_ifconf);
     > +        g_free(host_ifconf);
     >       }
     >
     >       return ret;

    Applied to my linux-user-for-7.0 branch.

    Thanks,
    Laurent
#4 lore raw

Re: Byte Sized task (was Re: [PATCH] linux-user/syscall.c: malloc to g_try_malloc)

Quoted 73 lines · Laurent Vivier

Hi Ahmed,

the email address you use for the trivial ML is broken. I've fixed it for my answer and added
qemu-devel.

Quoted 3 lines · Ahmed Abouzied

Hello,

Thanks for the review. I also found some occurrences of

"cpu_physical_memor_*" calls that should be

Quoted 3 lines

replaced with "address_space_*" around the codebase as mentioned [here](https://wiki.qemu.org/Contribute/BiteSizedTasks
<https://wiki.qemu.org/Contribute/BiteSizedTasks>) in the API

conversion section. The "/hw"

Quoted 1 line

directory for instance. Should I go about replacing those as well?

Should I create a Gitlab issue

Quoted 1 line

first to track the changes?

According to the wiki, the gitlab issue are already created with the "Byte Sized" label.

https://gitlab.com/qemu-project/qemu/-/issues?scope=all&state=opened&label_name[]=Bite%20Sized

But this one is not, I think John can help you for that (added in cc:)

Thanks,
Laurent

Quoted 4 lines

Best regards,
Ahmed

On Wed, 5 Jan 2022 at 12:26, Laurent Vivier <[email protected] <mailto:

Quoted 6 lines · [email protected]>> wrote:

Le 04/01/2022 à 15:38, Ahmed Abouzied a écrit :
> Use g_try_malloc instead of malloc to alocate the target ifconfig. > Also replace the corresponding free with g_free.
>
> Signed-off-by: Ahmed Abouzied <[email protected] <mailto:

[email protected]>>

Quoted 5 lines
     > ---
     >
     > Hello,
     >
     > I noticed that there was a `malloc` call in this file. It seems

that it

Quoted 1 line

> was added by the commit 22e4a267 (3 years ago) which was after

the commit

Quoted 1 line

> 0e173b24 (6 years ago) that replaced malloc calls with glib

alternative calls.

Quoted 2 lines

>
> There is no issue for this on Gitlab. Should I have created an

issue

Quoted 4 lines
     > first?
     >
     > Best regards,
     >
Quoted 8 lines
     >   linux-user/syscall.c | 4 ++--
     >   1 file changed, 2 insertions(+), 2 deletions(-)
     >
     > diff --git a/linux-user/syscall.c b/linux-user/syscall.c
     > index 56a3e17183..715f9430e1 100644
     > --- a/linux-user/syscall.c
     > +++ b/linux-user/syscall.c
     > @@ -4867,7 +4867,7 @@ static abi_long do_ioctl_ifconf(const

IOCTLEntry *ie, uint8_t *buf_temp,

Quoted 1 line

> * We can't fit all the extents into the fixed size

buffer.

Quoted 1 line

> * Allocate one that is large enough and use it

instead.

Quoted 24 lines
Quoted 7 lines
     >                */
     > -            host_ifconf = malloc(outbufsz);
     > +            host_ifconf = g_try_malloc(outbufsz);
     >               if (!host_ifconf) {
     >                   return -TARGET_ENOMEM;
     >               }
     > @@ -4915,7 +4915,7 @@ static abi_long do_ioctl_ifconf(const

IOCTLEntry *ie, uint8_t *buf_temp,

Quoted 14 lines
     >       }
     >
     >       if (free_buf) {
     > -        free(host_ifconf);
     > +        g_free(host_ifconf);
     >       }
     >
     >       return ret;

    Applied to my linux-user-for-7.0 branch.

    Thanks,
    Laurent

IIRC, consensus was that this task was not "bite sized", as it might possibly be a large task without an easily attainable "end". You don't need to worry about creating a GitLab issue to track it (Unless you WANT to for your own tracking, then go ahead and make one), but if you believe that the task is now *done* in its entirety, the task should be removed from the wiki.

Feel free to send me questions if you've got any.

--js