* ggc_alloc_rtvec_sized allocates spaces more than necessary?
@ 2011-08-15 12:17 王亮
2011-08-15 12:20 ` Richard Guenther
2011-08-15 17:22 ` Andreas Schwab
0 siblings, 2 replies; 7+ messages in thread
From: 王亮 @ 2011-08-15 12:17 UTC (permalink / raw)
To: gcc
Hi,
Current implementation of ggc_alloc_rtvec_sized is
#define ggc_alloc_rtvec_sized(NELT) \
(ggc_alloc_zone_vec_rtvec_def (sizeof (rtx), \
sizeof (struct rtvec_def) + ((NELT) - 1), \
&rtl_zone))
The size it allocates is
(sizeof (struct rtvec_def) + ((NELT) - 1)) * sizeof (rtx)
// (1)
Originally, the allocated size is
sizeof (struct rtvec_def) + ((NELT) - 1) * sizeof (rtx)
// (2)
So current implementation allocates more spaces than before.
I replace the second parameter of ggc_alloc_zone_vec_rtvec_def with
(sizeof (struct rtvec_def) + sizeof (rtx) - 1) / sizeof (rtx) +
((NELT) - 1) // (3)
It bootstraps on x86 successfully. So I guess the extra spaces are
not used. Did I miss something?
Thanks,
Liang.
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: ggc_alloc_rtvec_sized allocates spaces more than necessary?
2011-08-15 12:17 ggc_alloc_rtvec_sized allocates spaces more than necessary? 王亮
@ 2011-08-15 12:20 ` Richard Guenther
2011-08-16 5:52 ` Laurynas Biveinis
2011-08-15 17:22 ` Andreas Schwab
1 sibling, 1 reply; 7+ messages in thread
From: Richard Guenther @ 2011-08-15 12:20 UTC (permalink / raw)
To: 王亮; +Cc: gcc, Laurynas Biveinis
On Mon, Aug 15, 2011 at 2:16 PM, 王亮 <netcasper@gmail.com> wrote:
> Hi,
>
> Current implementation of ggc_alloc_rtvec_sized is
>
> #define ggc_alloc_rtvec_sized(NELT) \
> (ggc_alloc_zone_vec_rtvec_def (sizeof (rtx), \
> sizeof (struct rtvec_def) + ((NELT) - 1), \
> &rtl_zone))
>
> The size it allocates is
>
> (sizeof (struct rtvec_def) + ((NELT) - 1)) * sizeof (rtx)
This looks indeed bogus.
> // (1)
>
> Originally, the allocated size is
>
> sizeof (struct rtvec_def) + ((NELT) - 1) * sizeof (rtx)
> // (2)
This one is correct.
Laurynas?
> So current implementation allocates more spaces than before.
>
> I replace the second parameter of ggc_alloc_zone_vec_rtvec_def with
>
> (sizeof (struct rtvec_def) + sizeof (rtx) - 1) / sizeof (rtx) +
> ((NELT) - 1) // (3)
>
> It bootstraps on x86 successfully. So I guess the extra spaces are
> not used. Did I miss something?
>
> Thanks,
> Liang.
>
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: ggc_alloc_rtvec_sized allocates spaces more than necessary?
2011-08-15 12:17 ggc_alloc_rtvec_sized allocates spaces more than necessary? 王亮
2011-08-15 12:20 ` Richard Guenther
@ 2011-08-15 17:22 ` Andreas Schwab
2011-08-16 5:52 ` Laurynas Biveinis
1 sibling, 1 reply; 7+ messages in thread
From: Andreas Schwab @ 2011-08-15 17:22 UTC (permalink / raw)
To: 王亮; +Cc: gcc
王亮 <netcasper@gmail.com> writes:
> Hi,
>
> Current implementation of ggc_alloc_rtvec_sized is
>
> #define ggc_alloc_rtvec_sized(NELT) \
> (ggc_alloc_zone_vec_rtvec_def (sizeof (rtx), \
> sizeof (struct rtvec_def) + ((NELT) - 1), \
> &rtl_zone))
>
> The size it allocates is
>
> (sizeof (struct rtvec_def) + ((NELT) - 1)) * sizeof (rtx)
> // (1)
>
> Originally, the allocated size is
>
> sizeof (struct rtvec_def) + ((NELT) - 1) * sizeof (rtx)
> // (2)
>
> So current implementation allocates more spaces than before.
>
> I replace the second parameter of ggc_alloc_zone_vec_rtvec_def with
>
> (sizeof (struct rtvec_def) + sizeof (rtx) - 1) / sizeof (rtx) +
> ((NELT) - 1) // (3)
I think it was meant to be this:
#define ggc_alloc_rtvec_sized(NELT) \
ggc_alloc_zone_rtvec_def (sizeof (struct rtvec_def) \
+ ((NELT) - 1) * sizeof (rtx), \
&rtl_zone)
Andreas.
--
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 58CA 54C7 6D53 942B 1756 01D3 44D5 214B 8276 4ED5
"And now for something completely different."
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: ggc_alloc_rtvec_sized allocates spaces more than necessary?
2011-08-15 12:20 ` Richard Guenther
@ 2011-08-16 5:52 ` Laurynas Biveinis
2011-08-16 9:37 ` Liang Wang
0 siblings, 1 reply; 7+ messages in thread
From: Laurynas Biveinis @ 2011-08-16 5:52 UTC (permalink / raw)
To: Richard Guenther; +Cc: 王亮, gcc
> On Mon, Aug 15, 2011 at 2:16 PM, 王亮 <netcasper@gmail.com> wrote:
>> The size it allocates is
>>
>> (sizeof (struct rtvec_def) + ((NELT) - 1)) * sizeof (rtx)
>> Originally, the allocated size is
>>
>> sizeof (struct rtvec_def) + ((NELT) - 1) * sizeof (rtx)
Yes, this is correct, good catch.
>> (sizeof (struct rtvec_def) + sizeof (rtx) - 1) / sizeof (rtx) +
>> ((NELT) - 1) // (3)
Due to the way those macros expand, right now replacing the first arg
with "1" and the second one with straightforward "sizeof (struct
rtvec_def) + ((NELT) - 1) * sizeof (rtx)" will work and will be easier
to read than division. Liang, would you submit such patch?
Thanks again,
--
Laurynas
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: ggc_alloc_rtvec_sized allocates spaces more than necessary?
2011-08-15 17:22 ` Andreas Schwab
@ 2011-08-16 5:52 ` Laurynas Biveinis
2011-08-16 7:24 ` Andreas Schwab
0 siblings, 1 reply; 7+ messages in thread
From: Laurynas Biveinis @ 2011-08-16 5:52 UTC (permalink / raw)
To: Andreas Schwab; +Cc: 王亮, gcc
2011/8/15 Andreas Schwab <schwab@linux-m68k.org>:
> I think it was meant to be this:
>
> #define ggc_alloc_rtvec_sized(NELT) \
> ggc_alloc_zone_rtvec_def (sizeof (struct rtvec_def) \
> + ((NELT) - 1) * sizeof (rtx), \
> &rtl_zone)
Note that ggc_alloc_zone_rtvec_def takes three args, not two.
--
Laurynas
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: ggc_alloc_rtvec_sized allocates spaces more than necessary?
2011-08-16 5:52 ` Laurynas Biveinis
@ 2011-08-16 7:24 ` Andreas Schwab
0 siblings, 0 replies; 7+ messages in thread
From: Andreas Schwab @ 2011-08-16 7:24 UTC (permalink / raw)
To: Laurynas Biveinis; +Cc: 王亮, gcc
Laurynas Biveinis <laurynas.biveinis@gmail.com> writes:
> 2011/8/15 Andreas Schwab <schwab@linux-m68k.org>:
>> I think it was meant to be this:
>>
>> #define ggc_alloc_rtvec_sized(NELT) \
>> ggc_alloc_zone_rtvec_def (sizeof (struct rtvec_def) \
>> + ((NELT) - 1) * sizeof (rtx), \
>> &rtl_zone)
>
> Note that ggc_alloc_zone_rtvec_def takes three args, not two.
No, it does not of course.
Andreas.
--
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 58CA 54C7 6D53 942B 1756 01D3 44D5 214B 8276 4ED5
"And now for something completely different."
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: ggc_alloc_rtvec_sized allocates spaces more than necessary?
2011-08-16 5:52 ` Laurynas Biveinis
@ 2011-08-16 9:37 ` Liang Wang
0 siblings, 0 replies; 7+ messages in thread
From: Liang Wang @ 2011-08-16 9:37 UTC (permalink / raw)
To: Laurynas Biveinis; +Cc: Richard Guenther, gcc
On Tue, Aug 16, 2011 at 1:51 PM, Laurynas Biveinis
<laurynas.biveinis@gmail.com> wrote:
>> On Mon, Aug 15, 2011 at 2:16 PM, 王亮 <netcasper@gmail.com> wrote:
>>> The size it allocates is
>>>
>>> (sizeof (struct rtvec_def) + ((NELT) - 1)) * sizeof (rtx)
>
>>> Originally, the allocated size is
>>>
>>> sizeof (struct rtvec_def) + ((NELT) - 1) * sizeof (rtx)
>
> Yes, this is correct, good catch.
>
>>> (sizeof (struct rtvec_def) + sizeof (rtx) - 1) / sizeof (rtx) +
>>> ((NELT) - 1) // (3)
>
> Due to the way those macros expand, right now replacing the first arg
> with "1" and the second one with straightforward "sizeof (struct
> rtvec_def) + ((NELT) - 1) * sizeof (rtx)" will work and will be easier
> to read than division. Liang, would you submit such patch?
Yes, I have sent patch to gcc-patches.
>
> Thanks again,
> --
> Laurynas
>
^ permalink raw reply [flat|nested] 7+ messages in thread
end of thread, other threads:[~2011-08-16 9:37 UTC | newest]
Thread overview: 7+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2011-08-15 12:17 ggc_alloc_rtvec_sized allocates spaces more than necessary? 王亮
2011-08-15 12:20 ` Richard Guenther
2011-08-16 5:52 ` Laurynas Biveinis
2011-08-16 9:37 ` Liang Wang
2011-08-15 17:22 ` Andreas Schwab
2011-08-16 5:52 ` Laurynas Biveinis
2011-08-16 7:24 ` Andreas Schwab
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).