[Qemu-devel] some questions about g_malloc in qemu

[Qemu-devel] some questions about g_malloc in qemu

[Zhi Hui Li]

I am not sure  whether it is  need to check the return of g_malloc  in 
qemu  ?

  Thank you very much !

Re: [Qemu-devel] some questions about g_malloc in qemu

[陳韋任]

> I am not sure  whether it is  need to check the return of g_malloc  in 
> qemu  ?

  I think not, that's why you use g_malloc as a wrapper.

Regards,
chenwj

-- 
Wei-Ren Chen (陳韋任)
Computer Systems Lab, Institute of Information Science,
Academia Sinica, Taiwan (R.O.C.)
Tel:886-2-2788-3799 #1667
Homepage: http://people.cs.nctu.edu.tw/~chenwj

Re: [Qemu-devel] some questions about g_malloc in qemu

[Andreas Färber]

Am 15.12.2011 09:10, schrieb Zhi Hui Li:
> 
> I am not sure  whether it is  need to check the return of g_malloc  in
> qemu  ?

Compare the glib manual:

http://developer.gnome.org/glib/2.30/glib-Memory-Allocation.html#g-malloc

http://developer.gnome.org/glib/2.30/glib-Memory-Allocation.html#g-try-malloc

So, g_malloc() may return NULL if the requested size is zero, and it may
abort but does not return NULL due to insufficient memory available.

HTH,
Andreas

-- 
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany
GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer; HRB 16746 AG Nürnberg

Re: [Qemu-devel] some questions about g_malloc in qemu

[陳韋任]

Hi,

 I found this in HACKING:

  Please note that NULL check for the g_malloc result is redundant and
  that g_malloc() call with zero size is not allowed.


Regards,
chenwj

-- 
Wei-Ren Chen (陳韋任)
Computer Systems Lab, Institute of Information Science,
Academia Sinica, Taiwan (R.O.C.)
Tel:886-2-2788-3799 #1667
Homepage: http://people.cs.nctu.edu.tw/~chenwj

Re: [Qemu-devel] some questions about g_malloc in qemu

[Stefan Hajnoczi]

On Thu, Dec 15, 2011 at 04:28:28PM +0800, 陳韋任 wrote:
>  I found this in HACKING:
> 
>   Please note that NULL check for the g_malloc result is redundant and
>   that g_malloc() call with zero size is not allowed.

So we have:

1. You should not request 0 bytes from g_malloc().
2. g_malloc() does not return NULL (if you follow rule #1).

There is no need to check for NULL return.

Stefan

Re: [Qemu-devel] some questions about g_malloc in qemu

[Zhi Hui Li]

On 2011年12月15日 17:36, Stefan Hajnoczi wrote:
> On Thu, Dec 15, 2011 at 04:28:28PM +0800, 陳韋任 wrote:
>>   I found this in HACKING:
>>
>>    Please note that NULL check for the g_malloc result is redundant and
>>    that g_malloc() call with zero size is not allowed.
>
> So we have:
>
> 1. You should not request 0 bytes from g_malloc().
> 2. g_malloc() does not return NULL (if you follow rule #1).
>
> There is no need to check for NULL return.
>
> Stefan
>
>

Maybe there is insufficient memory, the return is error,
Do we need to check the return ?


Thank you very much for your feedback!

Re: [Qemu-devel] some questions about g_malloc in qemu

[陳韋任]

> Maybe there is insufficient memory, the return is error,
> Do we need to check the return ?

  It'll abort if there is not enough memory.
  http://mail.gnome.org/archives/gtk-app-devel-list/2003-September/msg00260.html

Regards,
chenwj

-- 
Wei-Ren Chen (陳韋任)
Computer Systems Lab, Institute of Information Science,
Academia Sinica, Taiwan (R.O.C.)
Tel:886-2-2788-3799 #1667
Homepage: http://people.cs.nctu.edu.tw/~chenwj

Re: [Qemu-devel] some questions about g_malloc in qemu

[Kevin Wolf]

Am 15.12.2011 10:36, schrieb Stefan Hajnoczi:
> On Thu, Dec 15, 2011 at 04:28:28PM +0800, 陳韋任 wrote:
>>  I found this in HACKING:
>>
>>   Please note that NULL check for the g_malloc result is redundant and
>>   that g_malloc() call with zero size is not allowed.
> 
> So we have:
> 
> 1. You should not request 0 bytes from g_malloc().

I think this was related to qemu_malloc() and Anthony's sed run made it
refer to g_malloc(), even though it works just fine with 0 bytes. We
should probably remove this sentence.

Kevin

Re: [Qemu-devel] some questions about g_malloc in qemu

[Stefan Hajnoczi]

On Thu, Dec 15, 2011 at 11:02:39AM +0100, Kevin Wolf wrote:
> Am 15.12.2011 10:36, schrieb Stefan Hajnoczi:
> > On Thu, Dec 15, 2011 at 04:28:28PM +0800, 陳韋任 wrote:
> >>  I found this in HACKING:
> >>
> >>   Please note that NULL check for the g_malloc result is redundant and
> >>   that g_malloc() call with zero size is not allowed.
> > 
> > So we have:
> > 
> > 1. You should not request 0 bytes from g_malloc().
> 
> I think this was related to qemu_malloc() and Anthony's sed run made it
> refer to g_malloc(), even though it works just fine with 0 bytes. We
> should probably remove this sentence.

If you remove it then you can't interpret it the way I did.  It's not
longer possible to say that g_malloc() never returns NULL.  You always
have to qualify that with "unless you ask for 0 bytes". :)

Stefan

Re: [Qemu-devel] some questions about g_malloc in qemu

[Max Filippov]

>> >>  I found this in HACKING:
>> >>
>> >>   Please note that NULL check for the g_malloc result is redundant and
>> >>   that g_malloc() call with zero size is not allowed.
>> >
>> > So we have:
>> >
>> > 1. You should not request 0 bytes from g_malloc().
>>
>> I think this was related to qemu_malloc() and Anthony's sed run made it
>> refer to g_malloc(), even though it works just fine with 0 bytes. We
>> should probably remove this sentence.
>
> If you remove it then you can't interpret it the way I did.  It's not
> longer possible to say that g_malloc() never returns NULL.  You always
> have to qualify that with "unless you ask for 0 bytes". :)

I have a number of ELFs that result in ELF loader making zero-size malloc.
Previously I had to enable CONFIG_ZERO_MALLOC, now it just works.

-- 
Thanks.
-- Max

Re: [Qemu-devel] some questions about g_malloc in qemu

[Kevin Wolf]

Am 15.12.2011 11:19, schrieb Stefan Hajnoczi:
> On Thu, Dec 15, 2011 at 11:02:39AM +0100, Kevin Wolf wrote:
>> Am 15.12.2011 10:36, schrieb Stefan Hajnoczi:
>>> On Thu, Dec 15, 2011 at 04:28:28PM +0800, 陳韋任 wrote:
>>>>  I found this in HACKING:
>>>>
>>>>   Please note that NULL check for the g_malloc result is redundant and
>>>>   that g_malloc() call with zero size is not allowed.
>>>
>>> So we have:
>>>
>>> 1. You should not request 0 bytes from g_malloc().
>>
>> I think this was related to qemu_malloc() and Anthony's sed run made it
>> refer to g_malloc(), even though it works just fine with 0 bytes. We
>> should probably remove this sentence.
> 
> If you remove it then you can't interpret it the way I did.  It's not
> longer possible to say that g_malloc() never returns NULL.  You always
> have to qualify that with "unless you ask for 0 bytes". :)

Try this: "g_malloc() never returns an error". ;-)

Kevin

Re: [Qemu-devel] some questions about g_malloc in qemu

[Peter Maydell]

On 15 December 2011 10:19, Stefan Hajnoczi <stefanha@linux.vnet.ibm.com> wrote:
> On Thu, Dec 15, 2011 at 11:02:39AM +0100, Kevin Wolf wrote:
>> I think this was related to qemu_malloc() and Anthony's sed run made it
>> refer to g_malloc(), even though it works just fine with 0 bytes. We
>> should probably remove this sentence.
>
> If you remove it then you can't interpret it the way I did.  It's not
> longer possible to say that g_malloc() never returns NULL.  You always
> have to qualify that with "unless you ask for 0 bytes". :)

Well, we should reword it, then, because the thing it was warning
about is no longer true...

-- PMM