[Qemu-devel] [PATCH 0/1] Fix large memory chunks allocation with tcg_malloc

[Qemu-devel] [PATCH 0/1] Fix large memory chunks allocation with tcg_malloc

[Kirill Batuzov]

Currently large memory chunk allocation with tcg_malloc is broken.  An attempt
to allocate such chunk when pool_current field of TCGContext is not NULL will
result in circular links in list of memory pools:

p = new pool;
s->pool_current->next = p;
p->next = s->pool_current;
(in tcg_malloc_internal)

Later p became a current pool, and current pool became next pool.  Next
tcg_malloc will switch current pool to next pool ('previous' current pool)
and will start allocating memory from it's beginning.  But some memory in
the beginning of this pool was already allocated and will be used twice
for different arrays.

At the end of this cover letter there is a patch that demonstrates the
problem.  It breaks current trunk on the first translation block containing
labels.

Large memory pools can not be reused by memory allocator for big allocations
and an attempt to reuse them for small allocations may result in an infinite
increase of memory consumption during run time.  Memory consumption would
increase every time a new large chunk of memory is allocated.  If code
allocates such chunk on every translation block (like patch at the end of this
letter do) then memory consumption would increase with every new block
translated.

My fix for the problems mentioned above is in the second e-mail.  I moved large
memory pools to a separate list and free them on pool_reset.

By the way: is there any particular reason for labels array in TCGContex to be
allocated dynamically?  It has constant size and is allocated unconditionally
for each translation block.

Kirill Batuzov (1):
  Fix large memory chunks allocation with tcg_malloc.

 tcg/tcg.c |   14 +++++++++-----
 tcg/tcg.h |    2 +-
 2 files changed, 10 insertions(+), 6 deletions(-)

---
diff --git a/tcg/tcg.c b/tcg/tcg.c
index 351a0a3..6dd54e6 100644
--- a/tcg/tcg.c
+++ b/tcg/tcg.c
@@ -265,6 +265,8 @@ void tcg_set_frame(TCGContext *s, int reg,
     s->frame_reg = reg;
 }
 
+uint8_t *p;
+
 void tcg_func_start(TCGContext *s)
 {
     int i;
@@ -273,6 +275,7 @@ void tcg_func_start(TCGContext *s)
     for(i = 0; i < (TCG_TYPE_COUNT * 2); i++)
         s->first_free_temp[i] = -1;
     s->labels = tcg_malloc(sizeof(TCGLabel) * TCG_MAX_LABELS);
+    p = tcg_malloc(TCG_POOL_CHUNK_SIZE + 1);
     s->nb_labels = 0;
     s->current_frame_offset = s->frame_start;
-- 
1.7.5.4

[Qemu-devel] [PATCH 1/1] Fix large memory chunks allocation with tcg_malloc.

[Kirill Batuzov]

An attempt to allocate a large memory chunk after a small one resulted in
circular links in list of pools.  It caused the same memory being
allocated twice for different arrays.

Now pools for large memory chunks are kept in separate list and are
freed during pool reset because current allocator can not reuse them.

Signed-off-by: Kirill Batuzov <batuzovk@ispras.ru>
---
 tcg/tcg.c |   14 +++++++++-----
 tcg/tcg.h |    2 +-
 2 files changed, 10 insertions(+), 6 deletions(-)

diff --git a/tcg/tcg.c b/tcg/tcg.c
index 351a0a3..7db8340 100644
--- a/tcg/tcg.c
+++ b/tcg/tcg.c
@@ -173,11 +173,9 @@ void *tcg_malloc_internal(TCGContext *s, int size)
         /* big malloc: insert a new pool (XXX: could optimize) */
         p = g_malloc(sizeof(TCGPool) + size);
         p->size = size;
-        if (s->pool_current)
-            s->pool_current->next = p;
-        else
-            s->pool_first = p;
-        p->next = s->pool_current;
+        p->next = s->pool_first_large;
+        s->pool_first_large = p;
+        return p->data;
     } else {
         p = s->pool_current;
         if (!p) {
@@ -208,6 +206,12 @@ void *tcg_malloc_internal(TCGContext *s, int size)
 
 void tcg_pool_reset(TCGContext *s)
 {
+    TCGPool *p, *t;
+    for (p = s->pool_first_large; p; p = t) {
+        t = p->next;
+        g_free(p);
+    }
+    s->pool_first_large = NULL;
     s->pool_cur = s->pool_end = NULL;
     s->pool_current = NULL;
 }
diff --git a/tcg/tcg.h b/tcg/tcg.h
index 5c28239..48d3f17 100644
--- a/tcg/tcg.h
+++ b/tcg/tcg.h
@@ -337,7 +337,7 @@ typedef struct TCGContext TCGContext;
 
 struct TCGContext {
     uint8_t *pool_cur, *pool_end;
-    TCGPool *pool_first, *pool_current;
+    TCGPool *pool_first, *pool_current, *pool_first_large;
     TCGLabel *labels;
     int nb_labels;
     TCGTemp *temps; /* globals first, temps after */
-- 
1.7.5.4

Re: [Qemu-devel] [PATCH 0/1] Fix large memory chunks allocation with tcg_malloc

[Evgeny Voevodin]

On 02.03.2012 13:22, Kirill Batuzov wrote:
> Currently large memory chunk allocation with tcg_malloc is broken.  An attempt
> to allocate such chunk when pool_current field of TCGContext is not NULL will
> result in circular links in list of memory pools:
>
> p = new pool;
> s->pool_current->next = p;
> p->next = s->pool_current;
> (in tcg_malloc_internal)
>
> Later p became a current pool, and current pool became next pool.  Next
> tcg_malloc will switch current pool to next pool ('previous' current pool)
> and will start allocating memory from it's beginning.  But some memory in
> the beginning of this pool was already allocated and will be used twice
> for different arrays.
>
> At the end of this cover letter there is a patch that demonstrates the
> problem.  It breaks current trunk on the first translation block containing
> labels.
>
> Large memory pools can not be reused by memory allocator for big allocations
> and an attempt to reuse them for small allocations may result in an infinite
> increase of memory consumption during run time.  Memory consumption would
> increase every time a new large chunk of memory is allocated.  If code
> allocates such chunk on every translation block (like patch at the end of this
> letter do) then memory consumption would increase with every new block
> translated.
>
> My fix for the problems mentioned above is in the second e-mail.  I moved large
> memory pools to a separate list and free them on pool_reset.
>

As I understand, your approach removes linking back to the previous 
allocated chunk to avoid usage of already allocated and used memory 
again. Also you added g_free() to tcg_pool_reset(). Wouldn't it slow 
down emulation?
Maybe such linkage to previous chunk was made with assumption that big 
allocations will not happen twice one after another? For example, while 
loading kernel on realview and exynos, the code never reaches this 
string in tcg_malloc_internal():

p->next = s->pool_current;

If this assumption is correct, maybe it's better to just insert a 
verification that allocated chunk has enough space to hold requested block?

+   if (s->pool_cur - p->data + size > p->size) {
+       goto new_pool;
+   }

     s->pool_current = p;
     s->pool_cur = p->data + size;
     s->pool_end = p->data + p->size;
     return p->data;


-- 
Kind regards,
Evgeny Voevodin,
Leading Software Engineer,
ASWG, Moscow R&D center, Samsung Electronics
e-mail: e.voevodin@samsung.com

Re: [Qemu-devel] [PATCH 0/1] Fix large memory chunks allocation with tcg_malloc

[Kirill Batuzov]

On Mon, 5 Mar 2012, Evgeny Voevodin wrote:

> 
> As I understand, your approach removes linking back to the previous allocated
> chunk to avoid usage of already allocated and used memory again. Also you
> added g_free() to tcg_pool_reset(). Wouldn't it slow down emulation?

No, it would not.  I added g_free() for large memory chunks only.
Emulation runs which use only small allocations are unaffected.
If an emulation run uses a large memory allocation - well, it was broken
before.  So there is nothing to compare speed to in this case.

> Maybe such linkage to previous chunk was made with assumption that big
> allocations will not happen twice one after another? For example, while
> loading kernel on realview and exynos, the code never reaches this string in
> tcg_malloc_internal():
> 
> p->next = s->pool_current;
> 

The corresponding code had not changed since 2008.  And, probably, never
worked.  It would be hard to know now what the author wanted from this code
back then.

> If this assumption is correct, maybe it's better to just insert a verification
> that allocated chunk has enough space to hold requested block?
> 
> +   if (s->pool_cur - p->data + size > p->size) {
> +       goto new_pool;
> +   }
> 
>     s->pool_current = p;
>     s->pool_cur = p->data + size;
>     s->pool_end = p->data + p->size;
>     return p->data;
> 

First, this verification does not solve problem of circular links in
the list.
Second, it does not solve problem of unbound increase of memory
consumption.

I probably should clarify that I found this issue when I was writing
an unrelated modification to the register allocator which required large memory
allocations.  Current code in TCG probably never does such allocations
and does not trigger memory corruption.

-- 
Kirill Batuzov

Re: [Qemu-devel] [PATCH 0/1] Fix large memory chunks allocation with tcg_malloc

[Kirill Batuzov]

Ping?
Somebody please review this patch...

And whom should I Cc in case of changes to tcg/ ? File MAINTAINERS lists
only qemu-devel for this subsystem.

On Fri, 2 Mar 2012, Kirill Batuzov wrote:

> Currently large memory chunk allocation with tcg_malloc is broken.  An attempt
> to allocate such chunk when pool_current field of TCGContext is not NULL will
> result in circular links in list of memory pools:
> 
> p = new pool;
> s->pool_current->next = p;
> p->next = s->pool_current;
> (in tcg_malloc_internal)
> 
> Later p became a current pool, and current pool became next pool.  Next
> tcg_malloc will switch current pool to next pool ('previous' current pool)
> and will start allocating memory from it's beginning.  But some memory in
> the beginning of this pool was already allocated and will be used twice
> for different arrays.
> 
> At the end of this cover letter there is a patch that demonstrates the
> problem.  It breaks current trunk on the first translation block containing
> labels.
> 
> Large memory pools can not be reused by memory allocator for big allocations
> and an attempt to reuse them for small allocations may result in an infinite
> increase of memory consumption during run time.  Memory consumption would
> increase every time a new large chunk of memory is allocated.  If code
> allocates such chunk on every translation block (like patch at the end of this
> letter do) then memory consumption would increase with every new block
> translated.
> 
> My fix for the problems mentioned above is in the second e-mail.  I moved large
> memory pools to a separate list and free them on pool_reset.
> 
> By the way: is there any particular reason for labels array in TCGContex to be
> allocated dynamically?  It has constant size and is allocated unconditionally
> for each translation block.
> 
> Kirill Batuzov (1):
>   Fix large memory chunks allocation with tcg_malloc.
> 
>  tcg/tcg.c |   14 +++++++++-----
>  tcg/tcg.h |    2 +-
>  2 files changed, 10 insertions(+), 6 deletions(-)
> 
> ---
> diff --git a/tcg/tcg.c b/tcg/tcg.c
> index 351a0a3..6dd54e6 100644
> --- a/tcg/tcg.c
> +++ b/tcg/tcg.c
> @@ -265,6 +265,8 @@ void tcg_set_frame(TCGContext *s, int reg,
>      s->frame_reg = reg;
>  }
>  
> +uint8_t *p;
> +
>  void tcg_func_start(TCGContext *s)
>  {
>      int i;
> @@ -273,6 +275,7 @@ void tcg_func_start(TCGContext *s)
>      for(i = 0; i < (TCG_TYPE_COUNT * 2); i++)
>          s->first_free_temp[i] = -1;
>      s->labels = tcg_malloc(sizeof(TCGLabel) * TCG_MAX_LABELS);
> +    p = tcg_malloc(TCG_POOL_CHUNK_SIZE + 1);
>      s->nb_labels = 0;
>      s->current_frame_offset = s->frame_start;
> 

Re: [Qemu-devel] [PATCH 1/1] Fix large memory chunks allocation with tcg_malloc.

[Blue Swirl]

Thanks, applied.

On Fri, Mar 2, 2012 at 09:22, Kirill Batuzov <batuzovk@ispras.ru> wrote:
> An attempt to allocate a large memory chunk after a small one resulted in
> circular links in list of pools.  It caused the same memory being
> allocated twice for different arrays.
>
> Now pools for large memory chunks are kept in separate list and are
> freed during pool reset because current allocator can not reuse them.
>
> Signed-off-by: Kirill Batuzov <batuzovk@ispras.ru>
> ---
>  tcg/tcg.c |   14 +++++++++-----
>  tcg/tcg.h |    2 +-
>  2 files changed, 10 insertions(+), 6 deletions(-)
>
> diff --git a/tcg/tcg.c b/tcg/tcg.c
> index 351a0a3..7db8340 100644
> --- a/tcg/tcg.c
> +++ b/tcg/tcg.c
> @@ -173,11 +173,9 @@ void *tcg_malloc_internal(TCGContext *s, int size)
>         /* big malloc: insert a new pool (XXX: could optimize) */
>         p = g_malloc(sizeof(TCGPool) + size);
>         p->size = size;
> -        if (s->pool_current)
> -            s->pool_current->next = p;
> -        else
> -            s->pool_first = p;
> -        p->next = s->pool_current;
> +        p->next = s->pool_first_large;
> +        s->pool_first_large = p;
> +        return p->data;
>     } else {
>         p = s->pool_current;
>         if (!p) {
> @@ -208,6 +206,12 @@ void *tcg_malloc_internal(TCGContext *s, int size)
>
>  void tcg_pool_reset(TCGContext *s)
>  {
> +    TCGPool *p, *t;
> +    for (p = s->pool_first_large; p; p = t) {
> +        t = p->next;
> +        g_free(p);
> +    }
> +    s->pool_first_large = NULL;
>     s->pool_cur = s->pool_end = NULL;
>     s->pool_current = NULL;
>  }
> diff --git a/tcg/tcg.h b/tcg/tcg.h
> index 5c28239..48d3f17 100644
> --- a/tcg/tcg.h
> +++ b/tcg/tcg.h
> @@ -337,7 +337,7 @@ typedef struct TCGContext TCGContext;
>
>  struct TCGContext {
>     uint8_t *pool_cur, *pool_end;
> -    TCGPool *pool_first, *pool_current;
> +    TCGPool *pool_first, *pool_current, *pool_first_large;
>     TCGLabel *labels;
>     int nb_labels;
>     TCGTemp *temps; /* globals first, temps after */
> --
> 1.7.5.4
>
>