From: "Alex Bennée" <alex.bennee@linaro.org>
To: Sergey Fedorov <sergey.fedorov@linaro.org>
Cc: qemu-devel@nongnu.org, Sergey Fedorov <serge.fdrv@gmail.com>,
Paolo Bonzini <pbonzini@redhat.com>,
Peter Crosthwaite <crosthwaite.peter@gmail.com>,
Richard Henderson <rth@twiddle.net>
Subject: Re: [Qemu-devel] [PATCH v3 02/10] tcg: Use uintptr_t type for jmp_list_{next|first} fields of TB
Date: Tue, 19 Apr 2016 11:34:13 +0100 [thread overview]
Message-ID: <87d1plx4yi.fsf@linaro.org> (raw)
In-Reply-To: <1460324732-30330-3-git-send-email-sergey.fedorov@linaro.org>
Sergey Fedorov <sergey.fedorov@linaro.org> writes:
> From: Sergey Fedorov <serge.fdrv@gmail.com>
>
> These fields do not contain pure pointers to a TranslationBlock
> structure. So uintptr_t is the most appropriate type for them.
> Also put an assert to assure that the two least significant bits of the
> pointer are zero before assigning it to jmp_list_first.
>
> Signed-off-by: Sergey Fedorov <serge.fdrv@gmail.com>
> Signed-off-by: Sergey Fedorov <sergey.fedorov@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
> ---
>
> Changes in v2:
> * Eliminated duplicate dereference of 'ptb' in tb_jmp_remove()
>
> include/exec/exec-all.h | 13 ++++++++-----
> translate-all.c | 38 ++++++++++++++++++++------------------
> 2 files changed, 28 insertions(+), 23 deletions(-)
>
> diff --git a/include/exec/exec-all.h b/include/exec/exec-all.h
> index 38a149cc5e0c..b055716ed690 100644
> --- a/include/exec/exec-all.h
> +++ b/include/exec/exec-all.h
> @@ -275,14 +275,16 @@ struct TranslationBlock {
> * jmp_list_first points to the first TB jumping to this one.
> * jmp_list_next is used to point to the next TB in a list.
> * Since each TB can have two jumps, it can participate in two lists.
> - * The two least significant bits of a pointer are used to choose which
> - * data field holds a pointer to the next TB:
> + * jmp_list_first and jmp_list_next are 4-byte aligned pointers to a
> + * TranslationBlock structure, but the two least significant bits of
> + * them are used to encode which data field of the pointed TB should
> + * be used to traverse the list further from that TB:
> * 0 => jmp_list_next[0], 1 => jmp_list_next[1], 2 => jmp_list_first.
> * In other words, 0/1 tells which jump is used in the pointed TB,
> * and 2 means that this is a pointer back to the target TB of this list.
> */
> - struct TranslationBlock *jmp_list_next[2];
> - struct TranslationBlock *jmp_list_first;
> + uintptr_t jmp_list_next[2];
> + uintptr_t jmp_list_first;
> };
>
> #include "qemu/thread.h"
> @@ -401,7 +403,8 @@ static inline void tb_add_jump(TranslationBlock *tb, int n,
>
> /* add in TB jmp circular list */
> tb->jmp_list_next[n] = tb_next->jmp_list_first;
> - tb_next->jmp_list_first = (TranslationBlock *)((uintptr_t)tb | n);
> + assert(((uintptr_t)tb & 3) == 0);
> + tb_next->jmp_list_first = (uintptr_t)tb | n;
> }
> }
>
> diff --git a/translate-all.c b/translate-all.c
> index 33ca06c663d4..ba71ff73f55f 100644
> --- a/translate-all.c
> +++ b/translate-all.c
> @@ -927,17 +927,17 @@ static inline void tb_page_remove(TranslationBlock **ptb, TranslationBlock *tb)
>
> static inline void tb_jmp_remove(TranslationBlock *tb, int n)
> {
> - TranslationBlock *tb1, **ptb;
> + TranslationBlock *tb1;
> + uintptr_t *ptb, ntb;
> unsigned int n1;
>
> ptb = &tb->jmp_list_next[n];
> - tb1 = *ptb;
> - if (tb1) {
> + if (*ptb) {
> /* find tb(n) in circular list */
> for (;;) {
> - tb1 = *ptb;
> - n1 = (uintptr_t)tb1 & 3;
> - tb1 = (TranslationBlock *)((uintptr_t)tb1 & ~3);
> + ntb = *ptb;
> + n1 = ntb & 3;
> + tb1 = (TranslationBlock *)(ntb & ~3);
> if (n1 == n && tb1 == tb) {
> break;
> }
> @@ -950,7 +950,7 @@ static inline void tb_jmp_remove(TranslationBlock *tb, int n)
> /* now we can suppress tb(n) from the list */
> *ptb = tb->jmp_list_next[n];
>
> - tb->jmp_list_next[n] = NULL;
> + tb->jmp_list_next[n] = (uintptr_t)NULL;
> }
> }
>
> @@ -969,7 +969,7 @@ void tb_phys_invalidate(TranslationBlock *tb, tb_page_addr_t page_addr)
> PageDesc *p;
> unsigned int h, n1;
> tb_page_addr_t phys_pc;
> - TranslationBlock *tb1, *tb2;
> + uintptr_t tb1, tb2;
>
> /* remove the TB from the hash list */
> phys_pc = tb->page_addr[0] + (tb->pc & ~TARGET_PAGE_MASK);
> @@ -1005,19 +1005,20 @@ void tb_phys_invalidate(TranslationBlock *tb, tb_page_addr_t page_addr)
> /* suppress any remaining jumps to this TB */
> tb1 = tb->jmp_list_first;
> for (;;) {
> - n1 = (uintptr_t)tb1 & 3;
> + TranslationBlock *tmp_tb;
> + n1 = tb1 & 3;
> if (n1 == 2) {
> break;
> }
> - tb1 = (TranslationBlock *)((uintptr_t)tb1 & ~3);
> - tb2 = tb1->jmp_list_next[n1];
> - tb_reset_jump(tb1, n1);
> - tb1->jmp_list_next[n1] = NULL;
> + tmp_tb = (TranslationBlock *)(tb1 & ~3);
> + tb2 = tmp_tb->jmp_list_next[n1];
> + tb_reset_jump(tmp_tb, n1);
> + tmp_tb->jmp_list_next[n1] = (uintptr_t)NULL;
> tb1 = tb2;
> }
>
> - /* fail safe */
> - tb->jmp_list_first = (TranslationBlock *)((uintptr_t)tb | 2);
> + assert(((uintptr_t)tb & 3) == 0);
> + tb->jmp_list_first = (uintptr_t)tb | 2; /* fail safe */
>
> tcg_ctx.tb_ctx.tb_phys_invalidate_count++;
> }
> @@ -1491,9 +1492,10 @@ static void tb_link_page(TranslationBlock *tb, tb_page_addr_t phys_pc,
> tb->page_addr[1] = -1;
> }
>
> - tb->jmp_list_first = (TranslationBlock *)((uintptr_t)tb | 2);
> - tb->jmp_list_next[0] = NULL;
> - tb->jmp_list_next[1] = NULL;
> + assert(((uintptr_t)tb & 3) == 0);
> + tb->jmp_list_first = (uintptr_t)tb | 2;
> + tb->jmp_list_next[0] = (uintptr_t)NULL;
> + tb->jmp_list_next[1] = (uintptr_t)NULL;
>
> /* init original jump addresses */
> if (tb->jmp_reset_offset[0] != TB_JMP_RESET_OFFSET_INVALID) {
--
Alex Bennée
next prev parent reply other threads:[~2016-04-19 10:34 UTC|newest]
Thread overview: 35+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-04-10 21:45 [Qemu-devel] [PATCH v3 00/10] tcg: Direct block chaining clean-up Sergey Fedorov
2016-04-10 21:45 ` [PATCH v3 01/10] tcg: Clean up direct block chaining data fields Sergey Fedorov
2016-04-10 21:45 ` [Qemu-devel] " Sergey Fedorov
2016-04-19 10:02 ` Alex Bennée
2016-04-19 10:02 ` [Qemu-devel] " Alex Bennée
2016-04-10 21:45 ` [Qemu-devel] [PATCH v3 02/10] tcg: Use uintptr_t type for jmp_list_{next|first} fields of TB Sergey Fedorov
2016-04-19 10:34 ` Alex Bennée [this message]
2016-04-10 21:45 ` [Qemu-devel] [PATCH v3 03/10] tcg: Rearrange tb_link_page() to avoid forward declaration Sergey Fedorov
2016-04-18 17:20 ` Alex Bennée
2016-04-18 17:59 ` Sergey Fedorov
2016-04-10 21:45 ` [Qemu-devel] [PATCH v3 04/10] tcg: Init TB's direct jumps before making it visible Sergey Fedorov
2016-04-19 10:55 ` Alex Bennée
2016-04-19 12:42 ` Sergey Fedorov
2016-04-19 13:07 ` Alex Bennée
2016-04-10 21:45 ` [Qemu-devel] [PATCH v3 05/10] tcg: Clarify thread safety check in tb_add_jump() Sergey Fedorov
2016-04-19 11:01 ` Alex Bennée
2016-04-19 12:49 ` Sergey Fedorov
2016-04-19 15:27 ` Alex Bennée
2016-04-10 21:45 ` [Qemu-devel] [PATCH v3 06/10] tcg: Rename tb_jmp_remove() to tb_remove_from_jmp_list() Sergey Fedorov
2016-04-10 21:45 ` [Qemu-devel] [PATCH v3 07/10] tcg: Extract removing of jumps to TB from tb_phys_invalidate() Sergey Fedorov
2016-04-10 21:45 ` [Qemu-devel] [PATCH v3 08/10] tcg: Clean up tb_jmp_unlink() Sergey Fedorov
2016-04-10 21:45 ` [PATCH v3 09/10] tcg: Clean up direct block chaining safety checks Sergey Fedorov
2016-04-10 21:45 ` [Qemu-devel] " Sergey Fedorov
2016-04-19 11:37 ` Alex Bennée
2016-04-19 11:37 ` [Qemu-devel] " Alex Bennée
2016-04-19 13:02 ` Sergey Fedorov
2016-04-19 13:02 ` [Qemu-devel] " Sergey Fedorov
2016-04-19 14:53 ` Alex Bennée
2016-04-19 14:53 ` [Qemu-devel] " Alex Bennée
2016-04-10 21:45 ` [PATCH v3 10/10] tcg: Moderate direct block chaining safety checks in user mode Sergey Fedorov
2016-04-10 21:45 ` [Qemu-devel] " Sergey Fedorov
2016-04-19 13:10 ` Alex Bennée
2016-04-19 13:10 ` [Qemu-devel] " Alex Bennée
2016-04-19 13:17 ` Sergey Fedorov
2016-04-19 13:17 ` [Qemu-devel] " Sergey Fedorov
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=87d1plx4yi.fsf@linaro.org \
--to=alex.bennee@linaro.org \
--cc=crosthwaite.peter@gmail.com \
--cc=pbonzini@redhat.com \
--cc=qemu-devel@nongnu.org \
--cc=rth@twiddle.net \
--cc=serge.fdrv@gmail.com \
--cc=sergey.fedorov@linaro.org \
/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.