[U-Boot] [PATCH v2 0/8] arm: a few steps to reduce the boot time

[U-Boot] [PATCH v2 0/8] arm: a few steps to reduce the boot time

[Przemyslaw Marczak]

This patchset reduces the boot time for ARM architecture,
Exynos boards, and boards with DFU enabled.

For tested Trats2 and Odroid X2 devices, this was done in four steps.

1. Enable the arch memcpy and memset

2. Enable arch memset for .bss clear

3. Reduce the .bss section small as possible by removing the static dfu buffer
   (32MiB in .bss - Trats2), and use the malloc.

4. Skip zeroing the memory reserved for malloc at malloc init.
   For Trats2 it was 80MiB of memory.

The .bss section will grow if we have a lot of static variables.
This section is cleared before jump to the relocated U-Boot,
and it was done word by word. To reduce the time for this step,
we can enable arch memset, which uses multiple ARM registers.

For configs with DFU enabled, we can find the dfu buffer in this section,
which has at least 8MB (32MB for Trats2). This is a lot of useless data,
which is not required for standard boot. So this buffer should be dynamic
allocated.

The next issue was about the malloc reserved memory, which was zeroed
at malloc init in one of early init call. Some boards has more than one MiB
of reserved malloc memory. Zeroing this, adds another boot delay.
Now the memory is zeroed only on calloc call.

So, actually the all was about unnecessary operations on 'big' data.

Przemyslaw Marczak (8):
  exynos: config: enable arch memcpy and arch memset
  arm: relocation: clear .bss section with arch memset if defined
  dfu: mmc: file buffer: remove static allocation
  dlmalloc: add option for skip memset in malloc init
  README: add info about skip memset at malloc init
  kconfig: malloc: add option for skip memset at malloc init
  trats2: defconfig: enable expert and skip memset at malloc init
  odroid: defconfig: enable expert and skip malloc memset

 Kconfig                         | 26 +++++++++++++++++++-------
 README                          |  7 +++++++
 arch/arm/lib/crt0.S             | 10 +++++++++-
 common/dlmalloc.c               | 10 +++++++---
 configs/odroid_defconfig        |  2 ++
 configs/trats2_defconfig        |  2 ++
 drivers/dfu/dfu_mmc.c           | 25 ++++++++++++++++++++++---
 include/configs/exynos-common.h |  3 +++
 8 files changed, 71 insertions(+), 14 deletions(-)

-- 
1.9.1

[U-Boot] [PATCH v2 1/8] exynos: config: enable arch memcpy and arch memset

[Przemyslaw Marczak]

This commit enables the following configs:
- CONFIG_USE_ARCH_MEMCPY
- CONFIG_USE_ARCH_MEMSET
This increases the performance of memcpy/memset
and also reduces the boot time.

This was tested on Trats2.
A quick test with trace. Boot time from start to main_loop() entry:
- ~1527ms - before this change (arch memset enabled for .bss clear)
- ~1384ms - after this change

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Cc: Minkyu Kang <mk7.kang@samsung.com>
Cc: Akshay Saraswat <akshay.s@samsung.com>
Cc: Simon Glass <sjg@chromium.org>
Cc: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
---
 include/configs/exynos-common.h | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/include/configs/exynos-common.h b/include/configs/exynos-common.h
index 1f3ee55..5c14c40 100644
--- a/include/configs/exynos-common.h
+++ b/include/configs/exynos-common.h
@@ -30,6 +30,9 @@
 #define CONFIG_SKIP_LOWLEVEL_INIT
 #define CONFIG_BOARD_EARLY_INIT_F
 
+#define CONFIG_USE_ARCH_MEMCPY
+#define CONFIG_USE_ARCH_MEMSET
+
 /* Keep L2 Cache Disabled */
 #define CONFIG_CMD_CACHE
 
-- 
1.9.1

[U-Boot] [PATCH v2 2/8] arm: relocation: clear .bss section with arch memset if defined

[Przemyslaw Marczak]

For ARM architecture, enable the CONFIG_USE_ARCH_MEMSET/MEMCPY,
will highly increase the memset/memcpy performance. This is able
thanks to the ARM multiple register instructions.

Unfortunatelly the relocation is done without the cache enabled,
so it takes some time, but zeroing the BSS memory takes much more
longer, especially for the configs with big static buffers.

A quick test confirms, that the boot time improvement after using
the arch memcpy for relocation has no significant meaning.
The same test confirms that enable the memset for zeroing BSS,
reduces the boot time.

So this patch enables the arch memset for zeroing the BSS after
the relocation process. For ARM boards, this can be enabled
in board configs by defining: 'CONFIG_USE_ARCH_MEMSET'.

This was tested on Trats2.
A quick test with trace. Boot time from start to main_loop() entry:
- ~1384ms - before this change
-  ~888ms - after this change

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Cc: Albert Aribaud <albert.u.boot@aribaud.net>
Cc: Tom Rini <trini@ti.com>
---
 arch/arm/lib/crt0.S | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

diff --git a/arch/arm/lib/crt0.S b/arch/arm/lib/crt0.S
index 22df3e5..fab3d2c 100644
--- a/arch/arm/lib/crt0.S
+++ b/arch/arm/lib/crt0.S
@@ -115,14 +115,22 @@ here:
 	bl	c_runtime_cpu_setup	/* we still call old routine here */
 
 	ldr	r0, =__bss_start	/* this is auto-relocated! */
-	ldr	r1, =__bss_end		/* this is auto-relocated! */
 
+#ifdef CONFIG_USE_ARCH_MEMSET
+	ldr	r3, =__bss_end		/* this is auto-relocated! */
+	mov	r1, #0x00000000		/* prepare zero to clear BSS */
+
+	subs	r2, r3, r0		/* r2 = memset len */
+	bl	memset
+#else
+	ldr	r1, =__bss_end		/* this is auto-relocated! */
 	mov	r2, #0x00000000		/* prepare zero to clear BSS */
 
 clbss_l:cmp	r0, r1			/* while not at end of BSS */
 	strlo	r2, [r0]		/* clear 32-bit BSS word */
 	addlo	r0, r0, #4		/* move to next */
 	blo	clbss_l
+#endif
 
 	bl coloured_LED_init
 	bl red_led_on
-- 
1.9.1

[U-Boot] [PATCH v2 3/8] dfu: mmc: file buffer: remove static allocation

[Przemyslaw Marczak]

For writing files, DFU implementation requires the file buffer
with the len at least of file size. For big files it requires
the same big buffer.

Previously the file buffer was allocated as a static variable,
so it was a part of U-Boot .bss section. For 32MiB len of buffer
we have 32MiB of additional space, required for this section.

The .bss needs to be cleared after the relocation.
This introduces an additional boot delay at every start, but usually
the dfu feature is not required at the standard boot, so the buffer
should be allocated only if required.

This patch removes the static allocation of this buffer,
and alloc it with memalign after first call of function:
- dfu_fill_entity_mmc()
and the buffer is freed on dfu_free_entity() call.

This was tested on Trats2.
A quick test with trace. Boot time from start to main_loop() entry:
- ~888ms - before this change (arch memset enabled for .bss clear)
- ~464ms - after this change

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Cc: Lukasz Majewski <l.majewski@samsung.com>
Cc: Stephen Warren <swarren@nvidia.com>
Cc: Pantelis Antoniou <panto@antoniou-consulting.com>
Cc: Tom Rini <trini@ti.com>
Cc: Marek Vasut <marek.vasut@gmail.com>
---
 drivers/dfu/dfu_mmc.c | 25 ++++++++++++++++++++++---
 1 file changed, 22 insertions(+), 3 deletions(-)

diff --git a/drivers/dfu/dfu_mmc.c b/drivers/dfu/dfu_mmc.c
index 62d72fe..fd865e1 100644
--- a/drivers/dfu/dfu_mmc.c
+++ b/drivers/dfu/dfu_mmc.c
@@ -16,8 +16,7 @@
 #include <fat.h>
 #include <mmc.h>
 
-static unsigned char __aligned(CONFIG_SYS_CACHELINE_SIZE)
-				dfu_file_buf[CONFIG_SYS_DFU_MAX_FILE_SIZE];
+static unsigned char *dfu_file_buf;
 static long dfu_file_buf_len;
 
 static int mmc_access_part(struct dfu_entity *dfu, struct mmc *mmc, int part)
@@ -211,7 +210,7 @@ int dfu_flush_medium_mmc(struct dfu_entity *dfu)
 
 	if (dfu->layout != DFU_RAW_ADDR) {
 		/* Do stuff here. */
-		ret = mmc_file_op(DFU_OP_WRITE, dfu, &dfu_file_buf,
+		ret = mmc_file_op(DFU_OP_WRITE, dfu, dfu_file_buf,
 				&dfu_file_buf_len);
 
 		/* Now that we're done */
@@ -263,6 +262,14 @@ int dfu_read_medium_mmc(struct dfu_entity *dfu, u64 offset, void *buf,
 	return ret;
 }
 
+void dfu_free_entity_mmc(struct dfu_entity *dfu)
+{
+	if (dfu_file_buf) {
+		free(dfu_file_buf);
+		dfu_file_buf = NULL;
+	}
+}
+
 /*
  * @param s Parameter string containing space-separated arguments:
  *	1st:
@@ -370,6 +377,18 @@ int dfu_fill_entity_mmc(struct dfu_entity *dfu, char *devstr, char *s)
 	dfu->write_medium = dfu_write_medium_mmc;
 	dfu->flush_medium = dfu_flush_medium_mmc;
 	dfu->inited = 0;
+	dfu->free_entity = dfu_free_entity_mmc;
+
+	/* Check if file buffer is ready */
+	if (!dfu_file_buf) {
+		dfu_file_buf = memalign(CONFIG_SYS_CACHELINE_SIZE,
+					CONFIG_SYS_DFU_MAX_FILE_SIZE);
+		if (!dfu_file_buf) {
+			error("Could not memalign 0x%x bytes",
+			      CONFIG_SYS_DFU_MAX_FILE_SIZE);
+			return -ENOMEM;
+		}
+	}
 
 	return 0;
 }
-- 
1.9.1

[U-Boot] [PATCH v2 4/8] dlmalloc: add option for skip memset in malloc init

[Przemyslaw Marczak]

This commit introduces new config: CONFIG_SYS_MALLOC_INIT_SKIP_ZEROING.

Before this change, the all amount of memory reserved for the malloc,
was set to zero in mem_malloc_init(). When the malloc reserved memory
exceeds few MiB, then the boot process can slow down.

So enabling this config, is an option to reduce the boot time.

Note:
After enable this option, only calloc() will return the pointer to zeroed
memory area. Previously, without this option, the memory pointed to untouched
malloc memory region, was filled with zeros. So it means, that code with
malloc() calls should be reexamined.

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
---
 common/dlmalloc.c | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)

diff --git a/common/dlmalloc.c b/common/dlmalloc.c
index 6453ee9..63f68ed 100644
--- a/common/dlmalloc.c
+++ b/common/dlmalloc.c
@@ -1535,9 +1535,9 @@ void mem_malloc_init(ulong start, ulong size)
 
 	debug("using memory %#lx-%#lx for malloc()\n", mem_malloc_start,
 	      mem_malloc_end);
-
-	memset((void *)mem_malloc_start, 0, size);
-
+#ifndef CONFIG_SYS_MALLOC_INIT_SKIP_ZEROING
+	memset((void *)mem_malloc_start, 0x0, size);
+#endif
 	malloc_bin_reloc();
 }
 
@@ -2948,10 +2948,12 @@ Void_t* cALLOc(n, elem_size) size_t n; size_t elem_size;
 
 
   /* check if expand_top called, in which case don't need to clear */
+#ifndef CONFIG_SYS_MALLOC_INIT_SKIP_ZEROING
 #if MORECORE_CLEARS
   mchunkptr oldtop = top;
   INTERNAL_SIZE_T oldtopsize = chunksize(top);
 #endif
+#endif
   Void_t* mem = mALLOc (sz);
 
   if ((long)n < 0) return NULL;
@@ -2977,6 +2979,7 @@ Void_t* cALLOc(n, elem_size) size_t n; size_t elem_size;
 
     csz = chunksize(p);
 
+#ifndef CONFIG_SYS_MALLOC_INIT_SKIP_ZEROING
 #if MORECORE_CLEARS
     if (p == oldtop && csz > oldtopsize)
     {
@@ -2984,6 +2987,7 @@ Void_t* cALLOc(n, elem_size) size_t n; size_t elem_size;
       csz = oldtopsize;
     }
 #endif
+#endif
 
     MALLOC_ZERO(mem, csz - SIZE_SZ);
     return mem;
-- 
1.9.1

[U-Boot] [PATCH v2 5/8] README: add info about skip memset at malloc init

[Przemyslaw Marczak]

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
---
 README | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/README b/README
index fefa71c..8673640 100644
--- a/README
+++ b/README
@@ -3989,6 +3989,13 @@ Configuration Settings:
 - CONFIG_SYS_MALLOC_LEN:
 		Size of DRAM reserved for malloc() use.
 
+- CONFIG_SYS_MALLOC_INIT_SKIP_ZEROING:
+		Do not set to zero the reserved DRAM area when init malloc.
+		For very big CONFIG_SYS_MALLOC_LEN(more than one MB), this will
+		reduce the boot time.
+		Before enabling this, please check if malloc calls, maybe
+		should be replaced by calloc - if expects zeroed memory.
+
 - CONFIG_SYS_MALLOC_F_LEN
 		Size of the malloc() pool for use before relocation. If
 		this is defined, then a very simple malloc() implementation
-- 
1.9.1

[U-Boot] [PATCH v2 6/8] kconfig: malloc: add option for skip memset at malloc init

[Przemyslaw Marczak]

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
---
 Kconfig | 26 +++++++++++++++++++-------
 1 file changed, 19 insertions(+), 7 deletions(-)

diff --git a/Kconfig b/Kconfig
index 4157da3..e08e44a 100644
--- a/Kconfig
+++ b/Kconfig
@@ -57,13 +57,25 @@ config CC_OPTIMIZE_FOR_SIZE
 	  This option is enabled by default for U-Boot.
 
 menuconfig EXPERT
-        bool "Configure standard U-Boot features (expert users)"
-        help
-          This option allows certain base U-Boot options and settings
-          to be disabled or tweaked. This is for specialized
-          environments which can tolerate a "non-standard" U-Boot.
-          Only use this if you really know what you are doing.
-
+	bool "Configure standard U-Boot features (expert users)"
+	help
+	  This option allows certain base U-Boot options and settings
+	  to be disabled or tweaked. This is for specialized
+	  environments which can tolerate a "non-standard" U-Boot.
+	  Only use this if you really know what you are doing.
+
+if EXPERT
+	config SYS_MALLOC_INIT_SKIP_ZEROING
+	bool "Skip memset at malloc init (reduce boot time)"
+	help
+	 This avoids zeroing memory reserved for malloc at malloc init.
+	 Significant boot time reduction is visible for configs in which
+	 CONFIG_SYS_MALLOC_LEN value, has more than few MiB.
+	 Useful for bzip2, bmp logo.
+	 Warning:
+	 When enable, make sure that calloc() is used when zeroed
+	 memory is needed.
+endif
 endmenu		# General setup
 
 menu "Boot images"
-- 
1.9.1

[U-Boot] [PATCH v2 7/8] trats2: defconfig: enable expert and skip memset at malloc init

[Przemyslaw Marczak]

Reduce the boot time of Trats2 by disabling the memset
at malloc init.

This was tested on Trats2.
A quick test with trace. Boot time from start to main_loop() entry:
- ~464ms - before this change (arch memset enabled for .bss clear)
- ~341ms - after this change

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
---
 configs/trats2_defconfig | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/configs/trats2_defconfig b/configs/trats2_defconfig
index 1b98b73..21cc37a 100644
--- a/configs/trats2_defconfig
+++ b/configs/trats2_defconfig
@@ -3,3 +3,5 @@ CONFIG_ARCH_EXYNOS=y
 CONFIG_TARGET_TRATS2=y
 CONFIG_OF_CONTROL=y
 CONFIG_DEFAULT_DEVICE_TREE="exynos4412-trats2"
+CONFIG_EXPERT=y
+CONFIG_SYS_MALLOC_INIT_SKIP_ZEROING=y
-- 
1.9.1

[U-Boot] [PATCH v2 8/8] odroid: defconfig: enable expert and skip malloc memset

[Przemyslaw Marczak]

Reduce the boot time of Odroid X2/U3 by disabling the memset
at malloc init.

This was tested on Odroid X2.
A quick test with checking gpio pin state using the oscilloscope.
Boot time from start to bootcmd (change gpio state by memory write command):
- ~228ms - before this change (arch memset enabled for .bss clear)
- ~100ms - after this change

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
---
 configs/odroid_defconfig | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/configs/odroid_defconfig b/configs/odroid_defconfig
index a842837..24af164 100644
--- a/configs/odroid_defconfig
+++ b/configs/odroid_defconfig
@@ -3,3 +3,5 @@ CONFIG_ARCH_EXYNOS=y
 CONFIG_TARGET_ODROID=y
 CONFIG_OF_CONTROL=y
 CONFIG_DEFAULT_DEVICE_TREE="exynos4412-odroid"
+CONFIG_EXPERT=y
+CONFIG_SYS_MALLOC_INIT_SKIP_ZEROING=y
-- 
1.9.1

[U-Boot] [PATCH v2 2/8] arm: relocation: clear .bss section with arch memset if defined

[Przemyslaw Marczak]

Hello,

On 02/16/2015 04:13 PM, Przemyslaw Marczak wrote:
> For ARM architecture, enable the CONFIG_USE_ARCH_MEMSET/MEMCPY,
> will highly increase the memset/memcpy performance. This is able
> thanks to the ARM multiple register instructions.
>
> Unfortunatelly the relocation is done without the cache enabled,
> so it takes some time, but zeroing the BSS memory takes much more
> longer, especially for the configs with big static buffers.
>
> A quick test confirms, that the boot time improvement after using
> the arch memcpy for relocation has no significant meaning.
> The same test confirms that enable the memset for zeroing BSS,
> reduces the boot time.
>
> So this patch enables the arch memset for zeroing the BSS after
> the relocation process. For ARM boards, this can be enabled
> in board configs by defining: 'CONFIG_USE_ARCH_MEMSET'.
>
> This was tested on Trats2.
> A quick test with trace. Boot time from start to main_loop() entry:
> - ~1384ms - before this change
> -  ~888ms - after this change
>
> Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
> Cc: Albert Aribaud <albert.u.boot@aribaud.net>
> Cc: Tom Rini <trini@ti.com>
> ---
>   arch/arm/lib/crt0.S | 10 +++++++++-
>   1 file changed, 9 insertions(+), 1 deletion(-)
>
> diff --git a/arch/arm/lib/crt0.S b/arch/arm/lib/crt0.S
> index 22df3e5..fab3d2c 100644
> --- a/arch/arm/lib/crt0.S
> +++ b/arch/arm/lib/crt0.S
> @@ -115,14 +115,22 @@ here:
>   	bl	c_runtime_cpu_setup	/* we still call old routine here */
>
>   	ldr	r0, =__bss_start	/* this is auto-relocated! */
> -	ldr	r1, =__bss_end		/* this is auto-relocated! */
>
> +#ifdef CONFIG_USE_ARCH_MEMSET
> +	ldr	r3, =__bss_end		/* this is auto-relocated! */
> +	mov	r1, #0x00000000		/* prepare zero to clear BSS */
> +
> +	subs	r2, r3, r0		/* r2 = memset len */
> +	bl	memset
> +#else
> +	ldr	r1, =__bss_end		/* this is auto-relocated! */
>   	mov	r2, #0x00000000		/* prepare zero to clear BSS */
>
>   clbss_l:cmp	r0, r1			/* while not at end of BSS */
>   	strlo	r2, [r0]		/* clear 32-bit BSS word */
>   	addlo	r0, r0, #4		/* move to next */
>   	blo	clbss_l
> +#endif
>
>   	bl coloured_LED_init
>   	bl red_led_on
>

This commit left unchanged. After boot time test using oscilloscope and 
the clock cycle counter I didn't noticed a time difference in more then 
one ms. In this case I think that insert a duplicated code here, has no 
sense.

Best regards,
-- 
Przemyslaw Marczak
Samsung R&D Institute Poland
Samsung Electronics
p.marczak at samsung.com

[U-Boot] [PATCH v2 0/8] arm: a few steps to reduce the boot time

[Stephen Warren]

On 02/16/2015 08:13 AM, Przemyslaw Marczak wrote:
> This patchset reduces the boot time for ARM architecture,
> Exynos boards, and boards with DFU enabled.

I tested this series on NVIDIA's Jetson TK1 board. It doesn't seem to 
introduce any new issues, but I did find a couple have crept in recently:

I'm running the following in U-Boot:

setenv dfu_alt_info "/dfu_test.bin ext4 0 1;/dfudummy.bin ext4 0 1"
dfu 0 mmc 0

1)

Whenever any file is uploaded through DFU, I see:

#File System is consistent
file found deleting
update journal finished
File System is consistent
update journal finished
18425346722729591336 bytes written in 4070 ms (3.9 EiB/s)

Notice that the byte count is way off (that's from a 4KB file). The byte 
count is always the same invalid number. I'm not sure if this message 
comes from the ext4 or DFU code.

2)

Both the 4096 and 1048576 DFU tests fail in the read-back stage, with 
the following appearing on the U-Boot console:

#File System is consistent
file found deleting
update journal finished
File System is consistent
update journal finished
18425346722734241092 bytes written in 4070 ms (3.9 EiB/s)
DOWNLOAD ... OK
Ctrl+C to exit ...
#File System is consistent
file found deleting
update journal finished
File System is consistent
update journal finished
18425346722734241092 bytes written in 4091 ms (3.9 EiB/s)
DOWNLOAD ... OK
Ctrl+C to exit ...
4096 bytes read in 127 ms (31.3 KiB/s)
#
UPLOAD ... done
Ctrl+C to exit ...
4096 bytes read in 126 ms (31.3 KiB/s)
#dfu_read: Wrong sequence number! [1] [3]

The host-side DFU test log ends with:

Opening DFU USB device... ID 0955:701a
Warning: Assuming DFU version 1.0
Run-time device DFU version 0100
Found DFU: [0955:701a] devnum=0, cfg=1, intf=0, alt=0, name="/dfu_test.bin"
Claiming USB DFU Interface...
Setting Alternate Setting #0 ...
Determining device status: state = dfuIDLE, status = 0
dfuIDLE, continuing
DFU mode device DFU version 0110
Device returned transfer size 4096
bytes_per_hash=4096
Copying data from DFU device to PC
Starting upload: [###dfu_upload: libusb_control_msg returned -9

[U-Boot] [PATCH v2 0/8] arm: a few steps to reduce the boot time

[Stephen Warren]

On 02/17/2015 02:43 PM, Stephen Warren wrote:
> On 02/16/2015 08:13 AM, Przemyslaw Marczak wrote:
>> This patchset reduces the boot time for ARM architecture,
>> Exynos boards, and boards with DFU enabled.
>
> I tested this series on NVIDIA's Jetson TK1 board. It doesn't seem to
> introduce any new issues, but I did find a couple have crept in recently:
>
> I'm running the following in U-Boot:
>
> setenv dfu_alt_info "/dfu_test.bin ext4 0 1;/dfudummy.bin ext4 0 1"
> dfu 0 mmc 0
>
> 1)
>
> Whenever any file is uploaded through DFU, I see:
>
> #File System is consistent
> file found deleting
> update journal finished
> File System is consistent
> update journal finished
> 18425346722729591336 bytes written in 4070 ms (3.9 EiB/s)
>
> Notice that the byte count is way off (that's from a 4KB file). The byte
> count is always the same invalid number. I'm not sure if this message
> comes from the ext4 or DFU code.

For the record in this thread, this is fixed by:
[PATCH] fs: ext4 write: return file len on success

[U-Boot] [PATCH v2 1/8] exynos: config: enable arch memcpy and arch memset

[Simon Glass]

On 16 February 2015 at 08:13, Przemyslaw Marczak <p.marczak@samsung.com> wrote:
> This commit enables the following configs:
> - CONFIG_USE_ARCH_MEMCPY
> - CONFIG_USE_ARCH_MEMSET
> This increases the performance of memcpy/memset
> and also reduces the boot time.
>
> This was tested on Trats2.
> A quick test with trace. Boot time from start to main_loop() entry:
> - ~1527ms - before this change (arch memset enabled for .bss clear)
> - ~1384ms - after this change
>
> Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
> Cc: Minkyu Kang <mk7.kang@samsung.com>
> Cc: Akshay Saraswat <akshay.s@samsung.com>
> Cc: Simon Glass <sjg@chromium.org>
> Cc: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
> ---
>  include/configs/exynos-common.h | 3 +++
>  1 file changed, 3 insertions(+)

Reviewed-by: Simon Glass <sjg@chromium.org>

BTW in case you are interested, in the Chromium U-Boot tree
(chromeos-v2013.06 branch) we have exynos support for turning on the
cache in SPL and leaving it on through to the end of U-Boot. It runs
two SPLs and two U-Boots (with verified boot and kernel verification)
in a total of about 750ms. This shipped last year with Pit and Pi
(Samsung Chromebook 2).

Might be some interesting patches there...

Regards,
Simon

[U-Boot] [PATCH v2 2/8] arm: relocation: clear .bss section with arch memset if defined

[Simon Glass]

Hi Przemyslaw,

On 16 February 2015 at 08:21, Przemyslaw Marczak <p.marczak@samsung.com> wrote:
> Hello,
>
>
> On 02/16/2015 04:13 PM, Przemyslaw Marczak wrote:
>>
>> For ARM architecture, enable the CONFIG_USE_ARCH_MEMSET/MEMCPY,
>> will highly increase the memset/memcpy performance. This is able
>> thanks to the ARM multiple register instructions.
>>
>> Unfortunatelly the relocation is done without the cache enabled,
>> so it takes some time, but zeroing the BSS memory takes much more
>> longer, especially for the configs with big static buffers.
>>
>> A quick test confirms, that the boot time improvement after using
>> the arch memcpy for relocation has no significant meaning.
>> The same test confirms that enable the memset for zeroing BSS,
>> reduces the boot time.
>>
>> So this patch enables the arch memset for zeroing the BSS after
>> the relocation process. For ARM boards, this can be enabled
>> in board configs by defining: 'CONFIG_USE_ARCH_MEMSET'.
>>
>> This was tested on Trats2.
>> A quick test with trace. Boot time from start to main_loop() entry:
>> - ~1384ms - before this change
>> -  ~888ms - after this change
>>
>> Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
>> Cc: Albert Aribaud <albert.u.boot@aribaud.net>
>> Cc: Tom Rini <trini@ti.com>
>> ---
>>   arch/arm/lib/crt0.S | 10 +++++++++-
>>   1 file changed, 9 insertions(+), 1 deletion(-)
>>
>> diff --git a/arch/arm/lib/crt0.S b/arch/arm/lib/crt0.S
>> index 22df3e5..fab3d2c 100644
>> --- a/arch/arm/lib/crt0.S
>> +++ b/arch/arm/lib/crt0.S
>> @@ -115,14 +115,22 @@ here:
>>         bl      c_runtime_cpu_setup     /* we still call old routine here
>> */
>>
>>         ldr     r0, =__bss_start        /* this is auto-relocated! */
>> -       ldr     r1, =__bss_end          /* this is auto-relocated! */
>>
>> +#ifdef CONFIG_USE_ARCH_MEMSET
>> +       ldr     r3, =__bss_end          /* this is auto-relocated! */
>> +       mov     r1, #0x00000000         /* prepare zero to clear BSS */
>> +
>> +       subs    r2, r3, r0              /* r2 = memset len */
>> +       bl      memset
>> +#else
>> +       ldr     r1, =__bss_end          /* this is auto-relocated! */
>>         mov     r2, #0x00000000         /* prepare zero to clear BSS */
>>
>>   clbss_l:cmp   r0, r1                  /* while not at end of BSS */
>>         strlo   r2, [r0]                /* clear 32-bit BSS word */
>>         addlo   r0, r0, #4              /* move to next */
>>         blo     clbss_l
>> +#endif
>>
>>         bl coloured_LED_init
>>         bl red_led_on
>>
>
> This commit left unchanged. After boot time test using oscilloscope and the
> clock cycle counter I didn't noticed a time difference in more then one ms.
> In this case I think that insert a duplicated code here, has no sense.

I don't understand this comment, sorry.

Regards,
Simon

[U-Boot] [PATCH v2 3/8] dfu: mmc: file buffer: remove static allocation

[Simon Glass]

On 16 February 2015 at 08:13, Przemyslaw Marczak <p.marczak@samsung.com> wrote:
> For writing files, DFU implementation requires the file buffer
> with the len at least of file size. For big files it requires
> the same big buffer.
>
> Previously the file buffer was allocated as a static variable,
> so it was a part of U-Boot .bss section. For 32MiB len of buffer
> we have 32MiB of additional space, required for this section.
>
> The .bss needs to be cleared after the relocation.
> This introduces an additional boot delay at every start, but usually
> the dfu feature is not required at the standard boot, so the buffer
> should be allocated only if required.
>
> This patch removes the static allocation of this buffer,
> and alloc it with memalign after first call of function:
> - dfu_fill_entity_mmc()
> and the buffer is freed on dfu_free_entity() call.
>
> This was tested on Trats2.
> A quick test with trace. Boot time from start to main_loop() entry:
> - ~888ms - before this change (arch memset enabled for .bss clear)
> - ~464ms - after this change
>

Wow.

Reviewed-by: Simon Glass <sjg@chromium.org>

[U-Boot] [PATCH v2 4/8] dlmalloc: add option for skip memset in malloc init

[Simon Glass]

Hi Przemyslaw,

On 16 February 2015 at 08:13, Przemyslaw Marczak <p.marczak@samsung.com> wrote:
> This commit introduces new config: CONFIG_SYS_MALLOC_INIT_SKIP_ZEROING.
>
> Before this change, the all amount of memory reserved for the malloc,
> was set to zero in mem_malloc_init(). When the malloc reserved memory
> exceeds few MiB, then the boot process can slow down.
>
> So enabling this config, is an option to reduce the boot time.
>
> Note:
> After enable this option, only calloc() will return the pointer to zeroed
> memory area. Previously, without this option, the memory pointed to untouched
> malloc memory region, was filled with zeros. So it means, that code with
> malloc() calls should be reexamined.

Can this go in Kconfig somewhere?

>
> Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
> ---
>  common/dlmalloc.c | 10 +++++++---
>  1 file changed, 7 insertions(+), 3 deletions(-)
>
> diff --git a/common/dlmalloc.c b/common/dlmalloc.c
> index 6453ee9..63f68ed 100644
> --- a/common/dlmalloc.c
> +++ b/common/dlmalloc.c
> @@ -1535,9 +1535,9 @@ void mem_malloc_init(ulong start, ulong size)
>
>         debug("using memory %#lx-%#lx for malloc()\n", mem_malloc_start,
>               mem_malloc_end);
> -
> -       memset((void *)mem_malloc_start, 0, size);
> -
> +#ifndef CONFIG_SYS_MALLOC_INIT_SKIP_ZEROING
> +       memset((void *)mem_malloc_start, 0x0, size);
> +#endif
>         malloc_bin_reloc();
>  }
>
> @@ -2948,10 +2948,12 @@ Void_t* cALLOc(n, elem_size) size_t n; size_t elem_size;
>
>
>    /* check if expand_top called, in which case don't need to clear */
> +#ifndef CONFIG_SYS_MALLOC_INIT_SKIP_ZEROING
>  #if MORECORE_CLEARS
>    mchunkptr oldtop = top;
>    INTERNAL_SIZE_T oldtopsize = chunksize(top);
>  #endif
> +#endif
>    Void_t* mem = mALLOc (sz);
>
>    if ((long)n < 0) return NULL;
> @@ -2977,6 +2979,7 @@ Void_t* cALLOc(n, elem_size) size_t n; size_t elem_size;
>
>      csz = chunksize(p);
>
> +#ifndef CONFIG_SYS_MALLOC_INIT_SKIP_ZEROING
>  #if MORECORE_CLEARS
>      if (p == oldtop && csz > oldtopsize)
>      {
> @@ -2984,6 +2987,7 @@ Void_t* cALLOc(n, elem_size) size_t n; size_t elem_size;
>        csz = oldtopsize;
>      }
>  #endif
> +#endif
>
>      MALLOC_ZERO(mem, csz - SIZE_SZ);
>      return mem;
> --
> 1.9.1
>

Regards,
Simon

[U-Boot] [PATCH v2 5/8] README: add info about skip memset at malloc init

[Simon Glass]

Hi Przemyslaw,

On 16 February 2015 at 08:13, Przemyslaw Marczak <p.marczak@samsung.com> wrote:
> Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
> ---
>  README | 7 +++++++
>  1 file changed, 7 insertions(+)
>
> diff --git a/README b/README
> index fefa71c..8673640 100644
> --- a/README
> +++ b/README
> @@ -3989,6 +3989,13 @@ Configuration Settings:
>  - CONFIG_SYS_MALLOC_LEN:
>                 Size of DRAM reserved for malloc() use.
>
> +- CONFIG_SYS_MALLOC_INIT_SKIP_ZEROING:
> +               Do not set to zero the reserved DRAM area when init malloc.
> +               For very big CONFIG_SYS_MALLOC_LEN(more than one MB), this will
> +               reduce the boot time.
> +               Before enabling this, please check if malloc calls, maybe
> +               should be replaced by calloc - if expects zeroed memory.
> +

I think if you put this in Kconfig you can put this help there.

>  - CONFIG_SYS_MALLOC_F_LEN
>                 Size of the malloc() pool for use before relocation. If
>                 this is defined, then a very simple malloc() implementation

(this one is in Kconfig now too)

Regards,
Simon

[U-Boot] [PATCH v2 6/8] kconfig: malloc: add option for skip memset at malloc init

[Simon Glass]

Hi Przemyslaw,

On 16 February 2015 at 08:13, Przemyslaw Marczak <p.marczak@samsung.com> wrote:
> Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
> ---
>  Kconfig | 26 +++++++++++++++++++-------
>  1 file changed, 19 insertions(+), 7 deletions(-)
>
> diff --git a/Kconfig b/Kconfig
> index 4157da3..e08e44a 100644
> --- a/Kconfig
> +++ b/Kconfig
> @@ -57,13 +57,25 @@ config CC_OPTIMIZE_FOR_SIZE
>           This option is enabled by default for U-Boot.

Ah, you have done this. Then I think you can merge this patch with the
dlmalloc patch and drop the README one.

>
>  menuconfig EXPERT
> -        bool "Configure standard U-Boot features (expert users)"
> -        help
> -          This option allows certain base U-Boot options and settings
> -          to be disabled or tweaked. This is for specialized
> -          environments which can tolerate a "non-standard" U-Boot.
> -          Only use this if you really know what you are doing.
> -
> +       bool "Configure standard U-Boot features (expert users)"
> +       help
> +         This option allows certain base U-Boot options and settings
> +         to be disabled or tweaked. This is for specialized
> +         environments which can tolerate a "non-standard" U-Boot.
> +         Only use this if you really know what you are doing.
> +
> +if EXPERT
> +       config SYS_MALLOC_INIT_SKIP_ZEROING
> +       bool "Skip memset at malloc init (reduce boot time)"
> +       help
> +        This avoids zeroing memory reserved for malloc at malloc init.
> +        Significant boot time reduction is visible for configs in which
> +        CONFIG_SYS_MALLOC_LEN value, has more than few MiB.
> +        Useful for bzip2, bmp logo.
> +        Warning:
> +        When enable, make sure that calloc() is used when zeroed
> +        memory is needed.
> +endif
>  endmenu                # General setup
>
>  menu "Boot images"
> --
> 1.9.1
>

Regards,
Simon

[U-Boot] [PATCH v2 7/8] trats2: defconfig: enable expert and skip memset at malloc init

[Simon Glass]

On 16 February 2015 at 08:13, Przemyslaw Marczak <p.marczak@samsung.com> wrote:
> Reduce the boot time of Trats2 by disabling the memset
> at malloc init.
>
> This was tested on Trats2.
> A quick test with trace. Boot time from start to main_loop() entry:
> - ~464ms - before this change (arch memset enabled for .bss clear)
> - ~341ms - after this change
>
> Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
> ---
>  configs/trats2_defconfig | 2 ++
>  1 file changed, 2 insertions(+)

Reviewed-by: Simon Glass <sjg@chromium.org>

[U-Boot] [PATCH v2 8/8] odroid: defconfig: enable expert and skip malloc memset

[Simon Glass]

On 16 February 2015 at 08:13, Przemyslaw Marczak <p.marczak@samsung.com> wrote:
> Reduce the boot time of Odroid X2/U3 by disabling the memset
> at malloc init.
>
> This was tested on Odroid X2.
> A quick test with checking gpio pin state using the oscilloscope.
> Boot time from start to bootcmd (change gpio state by memory write command):
> - ~228ms - before this change (arch memset enabled for .bss clear)
> - ~100ms - after this change
>
> Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>

Reviewed-by: Simon Glass <sjg@chromium.org>

[U-Boot] [PATCH v2 1/8] exynos: config: enable arch memcpy and arch memset

[Przemyslaw Marczak]

Hello,

On 02/18/2015 05:23 AM, Simon Glass wrote:
> On 16 February 2015 at 08:13, Przemyslaw Marczak <p.marczak@samsung.com> wrote:
>> This commit enables the following configs:
>> - CONFIG_USE_ARCH_MEMCPY
>> - CONFIG_USE_ARCH_MEMSET
>> This increases the performance of memcpy/memset
>> and also reduces the boot time.
>>
>> This was tested on Trats2.
>> A quick test with trace. Boot time from start to main_loop() entry:
>> - ~1527ms - before this change (arch memset enabled for .bss clear)
>> - ~1384ms - after this change
>>
>> Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
>> Cc: Minkyu Kang <mk7.kang@samsung.com>
>> Cc: Akshay Saraswat <akshay.s@samsung.com>
>> Cc: Simon Glass <sjg@chromium.org>
>> Cc: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
>> ---
>>   include/configs/exynos-common.h | 3 +++
>>   1 file changed, 3 insertions(+)
>
> Reviewed-by: Simon Glass <sjg@chromium.org>
>
> BTW in case you are interested, in the Chromium U-Boot tree
> (chromeos-v2013.06 branch) we have exynos support for turning on the
> cache in SPL and leaving it on through to the end of U-Boot. It runs
> two SPLs and two U-Boots (with verified boot and kernel verification)
> in a total of about 750ms. This shipped last year with Pit and Pi
> (Samsung Chromebook 2).
>
> Might be some interesting patches there...
>
> Regards,
> Simon
>

This is very interesting. Some time ago I made some tests witch the 
cache on/off cases for s-boot(bl1/Bl2 for trats2). Enabling the cache 
incredible improve the performance.
Since it is easy to break the Trats2, such changes in the s-boot has no 
sense. But it could be easy in the future to modify the bl2 for Odroid.

Best regards,
-- 
Przemyslaw Marczak
Samsung R&D Institute Poland
Samsung Electronics
p.marczak at samsung.com

[U-Boot] [PATCH v2 2/8] arm: relocation: clear .bss section with arch memset if defined

[Przemyslaw Marczak]

Hello,

On 02/18/2015 05:32 AM, Simon Glass wrote:
> Hi Przemyslaw,
>
> On 16 February 2015 at 08:21, Przemyslaw Marczak <p.marczak@samsung.com> wrote:
>> Hello,
>>
>>
>> On 02/16/2015 04:13 PM, Przemyslaw Marczak wrote:
>>>
>>> For ARM architecture, enable the CONFIG_USE_ARCH_MEMSET/MEMCPY,
>>> will highly increase the memset/memcpy performance. This is able
>>> thanks to the ARM multiple register instructions.
>>>
>>> Unfortunatelly the relocation is done without the cache enabled,
>>> so it takes some time, but zeroing the BSS memory takes much more
>>> longer, especially for the configs with big static buffers.
>>>
>>> A quick test confirms, that the boot time improvement after using
>>> the arch memcpy for relocation has no significant meaning.
>>> The same test confirms that enable the memset for zeroing BSS,
>>> reduces the boot time.
>>>
>>> So this patch enables the arch memset for zeroing the BSS after
>>> the relocation process. For ARM boards, this can be enabled
>>> in board configs by defining: 'CONFIG_USE_ARCH_MEMSET'.
>>>
>>> This was tested on Trats2.
>>> A quick test with trace. Boot time from start to main_loop() entry:
>>> - ~1384ms - before this change
>>> -  ~888ms - after this change
>>>
>>> Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
>>> Cc: Albert Aribaud <albert.u.boot@aribaud.net>
>>> Cc: Tom Rini <trini@ti.com>
>>> ---
>>>    arch/arm/lib/crt0.S | 10 +++++++++-
>>>    1 file changed, 9 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/arch/arm/lib/crt0.S b/arch/arm/lib/crt0.S
>>> index 22df3e5..fab3d2c 100644
>>> --- a/arch/arm/lib/crt0.S
>>> +++ b/arch/arm/lib/crt0.S
>>> @@ -115,14 +115,22 @@ here:
>>>          bl      c_runtime_cpu_setup     /* we still call old routine here
>>> */
>>>
>>>          ldr     r0, =__bss_start        /* this is auto-relocated! */
>>> -       ldr     r1, =__bss_end          /* this is auto-relocated! */
>>>
>>> +#ifdef CONFIG_USE_ARCH_MEMSET
>>> +       ldr     r3, =__bss_end          /* this is auto-relocated! */
>>> +       mov     r1, #0x00000000         /* prepare zero to clear BSS */
>>> +
>>> +       subs    r2, r3, r0              /* r2 = memset len */
>>> +       bl      memset
>>> +#else
>>> +       ldr     r1, =__bss_end          /* this is auto-relocated! */
>>>          mov     r2, #0x00000000         /* prepare zero to clear BSS */
>>>
>>>    clbss_l:cmp   r0, r1                  /* while not at end of BSS */
>>>          strlo   r2, [r0]                /* clear 32-bit BSS word */
>>>          addlo   r0, r0, #4              /* move to next */
>>>          blo     clbss_l
>>> +#endif
>>>
>>>          bl coloured_LED_init
>>>          bl red_led_on
>>>
>>
>> This commit left unchanged. After boot time test using oscilloscope and the
>> clock cycle counter I didn't noticed a time difference in more then one ms.
>> In this case I think that insert a duplicated code here, has no sense.
>
> I don't understand this comment, sorry.
>
> Regards,
> Simon
>

Sorry for the misleading message.
When I send this patch set, I forgot about adding the message-id of the 
previous thread as "in-reply-to".

There was a discussion about insert the memory zeroing routines as an 
asm here, instead of using the 'memset' call. But I tested that there is 
no difference in the performance. So in this case, it's better to use 
the common lib and this commit is the same as it was in the first version.

(I missed the changelog)

Best regards,
-- 
Przemyslaw Marczak
Samsung R&D Institute Poland
Samsung Electronics
p.marczak at samsung.com

[U-Boot] [PATCH v2 6/8] kconfig: malloc: add option for skip memset at malloc init

[Przemyslaw Marczak]

Hi Simon,

On 02/18/2015 05:32 AM, Simon Glass wrote:
> Hi Przemyslaw,
>
> On 16 February 2015 at 08:13, Przemyslaw Marczak <p.marczak@samsung.com> wrote:
>> Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
>> ---
>>   Kconfig | 26 +++++++++++++++++++-------
>>   1 file changed, 19 insertions(+), 7 deletions(-)
>>
>> diff --git a/Kconfig b/Kconfig
>> index 4157da3..e08e44a 100644
>> --- a/Kconfig
>> +++ b/Kconfig
>> @@ -57,13 +57,25 @@ config CC_OPTIMIZE_FOR_SIZE
>>            This option is enabled by default for U-Boot.
>
> Ah, you have done this. Then I think you can merge this patch with the
> dlmalloc patch and drop the README one.
>

Shouldn't we keep both, README and Kconfig help?
Kconfig is just a configuration tool, README is a documentation.
Sometimes it could be faster to find something in the text instead of 
config.

>>
>>   menuconfig EXPERT
>> -        bool "Configure standard U-Boot features (expert users)"
>> -        help
>> -          This option allows certain base U-Boot options and settings
>> -          to be disabled or tweaked. This is for specialized
>> -          environments which can tolerate a "non-standard" U-Boot.
>> -          Only use this if you really know what you are doing.
>> -
>> +       bool "Configure standard U-Boot features (expert users)"
>> +       help
>> +         This option allows certain base U-Boot options and settings
>> +         to be disabled or tweaked. This is for specialized
>> +         environments which can tolerate a "non-standard" U-Boot.
>> +         Only use this if you really know what you are doing.
>> +
>> +if EXPERT
>> +       config SYS_MALLOC_INIT_SKIP_ZEROING
>> +       bool "Skip memset at malloc init (reduce boot time)"
>> +       help
>> +        This avoids zeroing memory reserved for malloc at malloc init.
>> +        Significant boot time reduction is visible for configs in which
>> +        CONFIG_SYS_MALLOC_LEN value, has more than few MiB.
>> +        Useful for bzip2, bmp logo.
>> +        Warning:
>> +        When enable, make sure that calloc() is used when zeroed
>> +        memory is needed.
>> +endif
>>   endmenu                # General setup
>>
>>   menu "Boot images"
>> --
>> 1.9.1
>>
>
> Regards,
> Simon
>

Best regards,
-- 
Przemyslaw Marczak
Samsung R&D Institute Poland
Samsung Electronics
p.marczak at samsung.com

[U-Boot] [PATCH v2 8/8] odroid: defconfig: enable expert and skip malloc memset

[Przemyslaw Marczak]

Hello Simon,

On 02/18/2015 05:32 AM, Simon Glass wrote:
> On 16 February 2015 at 08:13, Przemyslaw Marczak <p.marczak@samsung.com> wrote:
>> Reduce the boot time of Odroid X2/U3 by disabling the memset
>> at malloc init.
>>
>> This was tested on Odroid X2.
>> A quick test with checking gpio pin state using the oscilloscope.
>> Boot time from start to bootcmd (change gpio state by memory write command):
>> - ~228ms - before this change (arch memset enabled for .bss clear)
>> - ~100ms - after this change
>>
>> Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
>
> Reviewed-by: Simon Glass <sjg@chromium.org>
>

Thanks for the review.

Best regards,
-- 
Przemyslaw Marczak
Samsung R&D Institute Poland
Samsung Electronics
p.marczak at samsung.com

[U-Boot] [PATCH v2 0/8] arm: a few steps to reduce the boot time

[Przemyslaw Marczak]

Hello Stephen,

On 02/17/2015 11:39 PM, Stephen Warren wrote:
> On 02/17/2015 02:43 PM, Stephen Warren wrote:
>> On 02/16/2015 08:13 AM, Przemyslaw Marczak wrote:
>>> This patchset reduces the boot time for ARM architecture,
>>> Exynos boards, and boards with DFU enabled.
>>
>> I tested this series on NVIDIA's Jetson TK1 board. It doesn't seem to
>> introduce any new issues, but I did find a couple have crept in recently:
>>
>> I'm running the following in U-Boot:
>>
>> setenv dfu_alt_info "/dfu_test.bin ext4 0 1;/dfudummy.bin ext4 0 1"
>> dfu 0 mmc 0
>>
>> 1)
>>
>> Whenever any file is uploaded through DFU, I see:
>>
>> #File System is consistent
>> file found deleting
>> update journal finished
>> File System is consistent
>> update journal finished
>> 18425346722729591336 bytes written in 4070 ms (3.9 EiB/s)
>>
>> Notice that the byte count is way off (that's from a 4KB file). The byte
>> count is always the same invalid number. I'm not sure if this message
>> comes from the ext4 or DFU code.
>
> For the record in this thread, this is fixed by:
> [PATCH] fs: ext4 write: return file len on success
>

Thank you for testing. I should add the ext4 fix to this patchset, so it 
will be linked in the next version.

Best regards,
-- 
Przemyslaw Marczak
Samsung R&D Institute Poland
Samsung Electronics
p.marczak at samsung.com

[U-Boot] [PATCH v2 6/8] kconfig: malloc: add option for skip memset at malloc init

[Simon Glass]

+Masahiro

Hi,

On 18 February 2015 at 05:40, Przemyslaw Marczak <p.marczak@samsung.com> wrote:
> Hi Simon,
>
> On 02/18/2015 05:32 AM, Simon Glass wrote:
>>
>> Hi Przemyslaw,
>>
>> On 16 February 2015 at 08:13, Przemyslaw Marczak <p.marczak@samsung.com>
>> wrote:
>>>
>>> Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
>>> ---
>>>   Kconfig | 26 +++++++++++++++++++-------
>>>   1 file changed, 19 insertions(+), 7 deletions(-)
>>>
>>> diff --git a/Kconfig b/Kconfig
>>> index 4157da3..e08e44a 100644
>>> --- a/Kconfig
>>> +++ b/Kconfig
>>> @@ -57,13 +57,25 @@ config CC_OPTIMIZE_FOR_SIZE
>>>            This option is enabled by default for U-Boot.
>>
>>
>> Ah, you have done this. Then I think you can merge this patch with the
>> dlmalloc patch and drop the README one.
>>
>
> Shouldn't we keep both, README and Kconfig help?
> Kconfig is just a configuration tool, README is a documentation.
> Sometimes it could be faster to find something in the text instead of
> config.

Agreed, but isn't it going to be a pain to add it in both places and
keep it in sync? Maybe we could create a script which creates a
README.kconfig containing all the options and help.

>
>
>>>
>>>   menuconfig EXPERT
>>> -        bool "Configure standard U-Boot features (expert users)"
>>> -        help
>>> -          This option allows certain base U-Boot options and settings
>>> -          to be disabled or tweaked. This is for specialized
>>> -          environments which can tolerate a "non-standard" U-Boot.
>>> -          Only use this if you really know what you are doing.
>>> -
>>> +       bool "Configure standard U-Boot features (expert users)"
>>> +       help
>>> +         This option allows certain base U-Boot options and settings
>>> +         to be disabled or tweaked. This is for specialized
>>> +         environments which can tolerate a "non-standard" U-Boot.
>>> +         Only use this if you really know what you are doing.
>>> +
>>> +if EXPERT
>>> +       config SYS_MALLOC_INIT_SKIP_ZEROING
>>> +       bool "Skip memset at malloc init (reduce boot time)"
>>> +       help
>>> +        This avoids zeroing memory reserved for malloc at malloc init.
>>> +        Significant boot time reduction is visible for configs in which
>>> +        CONFIG_SYS_MALLOC_LEN value, has more than few MiB.
>>> +        Useful for bzip2, bmp logo.
>>> +        Warning:
>>> +        When enable, make sure that calloc() is used when zeroed
>>> +        memory is needed.
>>> +endif
>>>   endmenu                # General setup
>>>
>>>   menu "Boot images"

Regards,
Simon

[U-Boot] [PATCH v2 2/8] arm: relocation: clear .bss section with arch memset if defined

[Simon Glass]

On 18 February 2015 at 05:31, Przemyslaw Marczak <p.marczak@samsung.com> wrote:
> Hello,
>
>
> On 02/18/2015 05:32 AM, Simon Glass wrote:
>>
>> Hi Przemyslaw,
>>
>> On 16 February 2015 at 08:21, Przemyslaw Marczak <p.marczak@samsung.com>
>> wrote:
>>>
>>> Hello,
>>>
>>>
>>> On 02/16/2015 04:13 PM, Przemyslaw Marczak wrote:
>>>>
>>>>
>>>> For ARM architecture, enable the CONFIG_USE_ARCH_MEMSET/MEMCPY,
>>>> will highly increase the memset/memcpy performance. This is able
>>>> thanks to the ARM multiple register instructions.
>>>>
>>>> Unfortunatelly the relocation is done without the cache enabled,
>>>> so it takes some time, but zeroing the BSS memory takes much more
>>>> longer, especially for the configs with big static buffers.
>>>>
>>>> A quick test confirms, that the boot time improvement after using
>>>> the arch memcpy for relocation has no significant meaning.
>>>> The same test confirms that enable the memset for zeroing BSS,
>>>> reduces the boot time.
>>>>
>>>> So this patch enables the arch memset for zeroing the BSS after
>>>> the relocation process. For ARM boards, this can be enabled
>>>> in board configs by defining: 'CONFIG_USE_ARCH_MEMSET'.
>>>>
>>>> This was tested on Trats2.
>>>> A quick test with trace. Boot time from start to main_loop() entry:
>>>> - ~1384ms - before this change
>>>> -  ~888ms - after this change
>>>>
>>>> Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
>>>> Cc: Albert Aribaud <albert.u.boot@aribaud.net>
>>>> Cc: Tom Rini <trini@ti.com>
>>>> ---
>>>>    arch/arm/lib/crt0.S | 10 +++++++++-
>>>>    1 file changed, 9 insertions(+), 1 deletion(-)
>>>>
>>>> diff --git a/arch/arm/lib/crt0.S b/arch/arm/lib/crt0.S
>>>> index 22df3e5..fab3d2c 100644
>>>> --- a/arch/arm/lib/crt0.S
>>>> +++ b/arch/arm/lib/crt0.S
>>>> @@ -115,14 +115,22 @@ here:
>>>>          bl      c_runtime_cpu_setup     /* we still call old routine
>>>> here
>>>> */
>>>>
>>>>          ldr     r0, =__bss_start        /* this is auto-relocated! */
>>>> -       ldr     r1, =__bss_end          /* this is auto-relocated! */
>>>>
>>>> +#ifdef CONFIG_USE_ARCH_MEMSET
>>>> +       ldr     r3, =__bss_end          /* this is auto-relocated! */
>>>> +       mov     r1, #0x00000000         /* prepare zero to clear BSS */
>>>> +
>>>> +       subs    r2, r3, r0              /* r2 = memset len */
>>>> +       bl      memset
>>>> +#else
>>>> +       ldr     r1, =__bss_end          /* this is auto-relocated! */
>>>>          mov     r2, #0x00000000         /* prepare zero to clear BSS */
>>>>
>>>>    clbss_l:cmp   r0, r1                  /* while not at end of BSS */
>>>>          strlo   r2, [r0]                /* clear 32-bit BSS word */
>>>>          addlo   r0, r0, #4              /* move to next */
>>>>          blo     clbss_l
>>>> +#endif
>>>>
>>>>          bl coloured_LED_init
>>>>          bl red_led_on
>>>>
>>>
>>> This commit left unchanged. After boot time test using oscilloscope and
>>> the
>>> clock cycle counter I didn't noticed a time difference in more then one
>>> ms.
>>> In this case I think that insert a duplicated code here, has no sense.
>>
>>
>> I don't understand this comment, sorry.
>>
>> Regards,
>> Simon
>>
>
> Sorry for the misleading message.
> When I send this patch set, I forgot about adding the message-id of the
> previous thread as "in-reply-to".
>
> There was a discussion about insert the memory zeroing routines as an asm
> here, instead of using the 'memset' call. But I tested that there is no
> difference in the performance. So in this case, it's better to use the
> common lib and this commit is the same as it was in the first version.
>
> (I missed the changelog)

I see, thanks.

Reviewed-by: Simon Glass <sjg@chromium.org>

[U-Boot] [PATCH v2 6/8] kconfig: malloc: add option for skip memset at malloc init

[Masahiro Yamada]

On Thu, 19 Feb 2015 11:59:07 -0700
Simon Glass <sjg@chromium.org> wrote:

> +Masahiro
> 
> Hi,
> 
> On 18 February 2015 at 05:40, Przemyslaw Marczak <p.marczak@samsung.com> wrote:
> > Hi Simon,
> >
> > On 02/18/2015 05:32 AM, Simon Glass wrote:
> >>
> >> Hi Przemyslaw,
> >>
> >> On 16 February 2015 at 08:13, Przemyslaw Marczak <p.marczak@samsung.com>
> >> wrote:
> >>>
> >>> Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
> >>> ---
> >>>   Kconfig | 26 +++++++++++++++++++-------
> >>>   1 file changed, 19 insertions(+), 7 deletions(-)
> >>>
> >>> diff --git a/Kconfig b/Kconfig
> >>> index 4157da3..e08e44a 100644
> >>> --- a/Kconfig
> >>> +++ b/Kconfig
> >>> @@ -57,13 +57,25 @@ config CC_OPTIMIZE_FOR_SIZE
> >>>            This option is enabled by default for U-Boot.
> >>
> >>
> >> Ah, you have done this. Then I think you can merge this patch with the
> >> dlmalloc patch and drop the README one.
> >>
> >
> > Shouldn't we keep both, README and Kconfig help?
> > Kconfig is just a configuration tool, README is a documentation.
> > Sometimes it could be faster to find something in the text instead of
> > config.
> 
> Agreed, but isn't it going to be a pain to add it in both places and
> keep it in sync? Maybe we could create a script which creates a
> README.kconfig containing all the options and help.
> 

I agree with Simon.

README is useful when we need a wider and more general explanation about a feature,
for ex. doc/driver-model/README.txt

Per-config explanation should be documented in Kconfig only.

Another reasone I prefer Kconfig help is:
When we remove a CONFIG from Kconfig, the documentation in a separete README
might be left over.


BTW, I did not know that U-Boot filled all the malloc space with zero.

I guess our consensus is that malloc() returns uninitialized memory.

So, I am happy with this patch.  (Perhaps, could it be enabled by default? )



Best Regards
Masahiro Yamada

[U-Boot] [PATCH v2 6/8] kconfig: malloc: add option for skip memset at malloc init

[Przemyslaw Marczak]

Hello,

On 02/20/2015 08:32 AM, Masahiro Yamada wrote:
>
> On Thu, 19 Feb 2015 11:59:07 -0700
> Simon Glass <sjg@chromium.org> wrote:
>
>> +Masahiro
>>
>> Hi,
>>
>> On 18 February 2015 at 05:40, Przemyslaw Marczak <p.marczak@samsung.com> wrote:
>>> Hi Simon,
>>>
>>> On 02/18/2015 05:32 AM, Simon Glass wrote:
>>>>
>>>> Hi Przemyslaw,
>>>>
>>>> On 16 February 2015 at 08:13, Przemyslaw Marczak <p.marczak@samsung.com>
>>>> wrote:
>>>>>
>>>>> Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
>>>>> ---
>>>>>    Kconfig | 26 +++++++++++++++++++-------
>>>>>    1 file changed, 19 insertions(+), 7 deletions(-)
>>>>>
>>>>> diff --git a/Kconfig b/Kconfig
>>>>> index 4157da3..e08e44a 100644
>>>>> --- a/Kconfig
>>>>> +++ b/Kconfig
>>>>> @@ -57,13 +57,25 @@ config CC_OPTIMIZE_FOR_SIZE
>>>>>             This option is enabled by default for U-Boot.
>>>>
>>>>
>>>> Ah, you have done this. Then I think you can merge this patch with the
>>>> dlmalloc patch and drop the README one.
>>>>
>>>
>>> Shouldn't we keep both, README and Kconfig help?
>>> Kconfig is just a configuration tool, README is a documentation.
>>> Sometimes it could be faster to find something in the text instead of
>>> config.
>>
>> Agreed, but isn't it going to be a pain to add it in both places and
>> keep it in sync? Maybe we could create a script which creates a
>> README.kconfig containing all the options and help.
>>
>
> I agree with Simon.
>
> README is useful when we need a wider and more general explanation about a feature,
> for ex. doc/driver-model/README.txt
>
> Per-config explanation should be documented in Kconfig only.
>
> Another reasone I prefer Kconfig help is:
> When we remove a CONFIG from Kconfig, the documentation in a separete README
> might be left over.
>

Yes, but this is the maintainer role, to keep it in sync. It's no 
problem for me. I can remove the README entry.
I just think that it may be useful for people who just starting the fun 
with U-Boot.

>
> BTW, I did not know that U-Boot filled all the malloc space with zero.

The same as I, but checking the execution time of some functions, shows 
that something takes too long time.

>
> I guess our consensus is that malloc() returns uninitialized memory.

Yes, as should it do.

>
> So, I am happy with this patch.  (Perhaps, could it be enabled by default? )
>

This could potentially break something. Let's give a free hand to 
maintainers to enable this after test. And this is why I add such config 
option.

>
>
> Best Regards
> Masahiro Yamada
>
>

So as Simon and Masahiro wish, I will remove the README entry for this.

Best regards,
-- 
Przemyslaw Marczak
Samsung R&D Institute Poland
Samsung Electronics
p.marczak at samsung.com

[U-Boot] [PATCH v3 0/6] arm: a few steps to improve boot time

[Przemyslaw Marczak]

This patchset reduces the boot time for ARM architecture,
Exynos boards, and (ARM) boards with DFU enabled.

For tested Trats2 and Odroid X2 devices, this was done in four steps.

1. Enable the arch memcpy and memset - ARCH specific

2. Enable arch memset for .bss clear - ARCH specific

3. Reduce the .bss section small as possible by: (board specific)
  - remove static dfu buffer (32MiB in .bss - Trats2), and use malloc

4. Skip zeroing the memory reserved for malloc at malloc init.
   For Trats2 it was 80MiB of memory.

The .bss section will grow if we have a lot of static variables.
This section is cleared before jump to the relocated U-Boot,
and it was done word by word. To reduce the time for this step,
we can enable arch memset, which uses multiple ARM registers.

For configs with DFU enabled, we can find the dfu buffer in this section,
which has at least 8MB (32MB for Trats2). This is a lot of useless data,
which is not required for standard boot. So this buffer should be dynamic
allocated.

So, actually the all was about unnecessary operations on 'big' data.

Przemyslaw Marczak (6):
  exynos: config: enable arch memcpy and arch memset
  arm: relocation: clear .bss section with arch memset if defined
  dfu: mmc: file buffer: remove static allocation
  dlmalloc: add option for skip memset in malloc init
  trats2: defconfig: enable expert and skip memset at malloc init
  odroid: defconfig: enable expert and skip memset at malloc init

 Kconfig                         | 26 +++++++++++++++++++-------
 arch/arm/lib/crt0.S             | 10 +++++++++-
 common/dlmalloc.c               | 10 +++++++---
 configs/odroid_defconfig        |  2 ++
 configs/trats2_defconfig        |  2 ++
 drivers/dfu/dfu_mmc.c           | 25 ++++++++++++++++++++++---
 include/configs/exynos-common.h |  3 +++
 7 files changed, 64 insertions(+), 14 deletions(-)

-- 
1.9.1

[U-Boot] [PATCH v3 1/6] exynos: config: enable arch memcpy and arch memset

[Przemyslaw Marczak]

This commit enables the following configs:
- CONFIG_USE_ARCH_MEMCPY
- CONFIG_USE_ARCH_MEMSET
This increases the performance of memcpy/memset
and also reduces the boot time.

This was tested on Trats2.
A quick test with trace. Boot time from start to main_loop() entry:
- ~1527ms - before this change (arch memset enabled for .bss clear)
- ~1384ms - after this change

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Cc: Minkyu Kang <mk7.kang@samsung.com>
Cc: Akshay Saraswat <akshay.s@samsung.com>
Cc: Simon Glass <sjg@chromium.org>
Cc: Sjoerd Simons <sjoerd.simons@collabora.co.uk>

---
Changes V3
- none
---
 include/configs/exynos-common.h | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/include/configs/exynos-common.h b/include/configs/exynos-common.h
index 59676ae..87f8db0 100644
--- a/include/configs/exynos-common.h
+++ b/include/configs/exynos-common.h
@@ -24,6 +24,9 @@
 #define CONFIG_SKIP_LOWLEVEL_INIT
 #define CONFIG_BOARD_EARLY_INIT_F
 
+#define CONFIG_USE_ARCH_MEMCPY
+#define CONFIG_USE_ARCH_MEMSET
+
 /* Keep L2 Cache Disabled */
 #define CONFIG_CMD_CACHE
 
-- 
1.9.1

[U-Boot] [PATCH v3 2/6] arm: relocation: clear .bss section with arch memset if defined

[Przemyslaw Marczak]

For ARM architecture, enable the CONFIG_USE_ARCH_MEMSET/MEMCPY,
will highly increase the memset/memcpy performance. This is able
thanks to the ARM multiple register instructions.

Unfortunatelly the relocation is done without the cache enabled,
so it takes some time, but zeroing the BSS memory takes much more
longer, especially for the configs with big static buffers.

A quick test confirms, that the boot time improvement after using
the arch memcpy for relocation has no significant meaning.
The same test confirms that enable the memset for zeroing BSS,
reduces the boot time.

So this patch enables the arch memset for zeroing the BSS after
the relocation process. For ARM boards, this can be enabled
in board configs by defining: 'CONFIG_USE_ARCH_MEMSET'.

This was tested on Trats2.
A quick test with trace. Boot time from start to main_loop() entry:
- ~1384ms - before this change
-  ~888ms - after this change

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Cc: Albert Aribaud <albert.u.boot@aribaud.net>
Cc: Tom Rini <trini@ti.com>
---
Changes V3
- none
---
 arch/arm/lib/crt0.S | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

diff --git a/arch/arm/lib/crt0.S b/arch/arm/lib/crt0.S
index 22df3e5..fab3d2c 100644
--- a/arch/arm/lib/crt0.S
+++ b/arch/arm/lib/crt0.S
@@ -115,14 +115,22 @@ here:
 	bl	c_runtime_cpu_setup	/* we still call old routine here */
 
 	ldr	r0, =__bss_start	/* this is auto-relocated! */
-	ldr	r1, =__bss_end		/* this is auto-relocated! */
 
+#ifdef CONFIG_USE_ARCH_MEMSET
+	ldr	r3, =__bss_end		/* this is auto-relocated! */
+	mov	r1, #0x00000000		/* prepare zero to clear BSS */
+
+	subs	r2, r3, r0		/* r2 = memset len */
+	bl	memset
+#else
+	ldr	r1, =__bss_end		/* this is auto-relocated! */
 	mov	r2, #0x00000000		/* prepare zero to clear BSS */
 
 clbss_l:cmp	r0, r1			/* while not at end of BSS */
 	strlo	r2, [r0]		/* clear 32-bit BSS word */
 	addlo	r0, r0, #4		/* move to next */
 	blo	clbss_l
+#endif
 
 	bl coloured_LED_init
 	bl red_led_on
-- 
1.9.1

[U-Boot] [PATCH v3 3/6] dfu: mmc: file buffer: remove static allocation

[Przemyslaw Marczak]

For writing files, DFU implementation requires the file buffer
with the len at least of file size. For big files it requires
the same big buffer.

Previously the file buffer was allocated as a static variable,
so it was a part of U-Boot .bss section. For 32MiB len of buffer
we have 32MiB of additional space, required for this section.

The .bss needs to be cleared after the relocation.
This introduces an additional boot delay at every start, but usually
the dfu feature is not required at the standard boot, so the buffer
should be allocated only if required.

This patch removes the static allocation of this buffer,
and alloc it with memalign after first call of function:
- dfu_fill_entity_mmc()
and the buffer is freed on dfu_free_entity() call.

This was tested on Trats2.
A quick test with trace. Boot time from start to main_loop() entry:
- ~888ms - before this change (arch memset enabled for .bss clear)
- ~464ms - after this change

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Cc: Lukasz Majewski <l.majewski@samsung.com>
Cc: Stephen Warren <swarren@nvidia.com>
Cc: Pantelis Antoniou <panto@antoniou-consulting.com>
Cc: Tom Rini <trini@ti.com>
Cc: Marek Vasut <marek.vasut@gmail.com>

---
Changes V3
- none
---
 drivers/dfu/dfu_mmc.c | 25 ++++++++++++++++++++++---
 1 file changed, 22 insertions(+), 3 deletions(-)

diff --git a/drivers/dfu/dfu_mmc.c b/drivers/dfu/dfu_mmc.c
index 62d72fe..fd865e1 100644
--- a/drivers/dfu/dfu_mmc.c
+++ b/drivers/dfu/dfu_mmc.c
@@ -16,8 +16,7 @@
 #include <fat.h>
 #include <mmc.h>
 
-static unsigned char __aligned(CONFIG_SYS_CACHELINE_SIZE)
-				dfu_file_buf[CONFIG_SYS_DFU_MAX_FILE_SIZE];
+static unsigned char *dfu_file_buf;
 static long dfu_file_buf_len;
 
 static int mmc_access_part(struct dfu_entity *dfu, struct mmc *mmc, int part)
@@ -211,7 +210,7 @@ int dfu_flush_medium_mmc(struct dfu_entity *dfu)
 
 	if (dfu->layout != DFU_RAW_ADDR) {
 		/* Do stuff here. */
-		ret = mmc_file_op(DFU_OP_WRITE, dfu, &dfu_file_buf,
+		ret = mmc_file_op(DFU_OP_WRITE, dfu, dfu_file_buf,
 				&dfu_file_buf_len);
 
 		/* Now that we're done */
@@ -263,6 +262,14 @@ int dfu_read_medium_mmc(struct dfu_entity *dfu, u64 offset, void *buf,
 	return ret;
 }
 
+void dfu_free_entity_mmc(struct dfu_entity *dfu)
+{
+	if (dfu_file_buf) {
+		free(dfu_file_buf);
+		dfu_file_buf = NULL;
+	}
+}
+
 /*
  * @param s Parameter string containing space-separated arguments:
  *	1st:
@@ -370,6 +377,18 @@ int dfu_fill_entity_mmc(struct dfu_entity *dfu, char *devstr, char *s)
 	dfu->write_medium = dfu_write_medium_mmc;
 	dfu->flush_medium = dfu_flush_medium_mmc;
 	dfu->inited = 0;
+	dfu->free_entity = dfu_free_entity_mmc;
+
+	/* Check if file buffer is ready */
+	if (!dfu_file_buf) {
+		dfu_file_buf = memalign(CONFIG_SYS_CACHELINE_SIZE,
+					CONFIG_SYS_DFU_MAX_FILE_SIZE);
+		if (!dfu_file_buf) {
+			error("Could not memalign 0x%x bytes",
+			      CONFIG_SYS_DFU_MAX_FILE_SIZE);
+			return -ENOMEM;
+		}
+	}
 
 	return 0;
 }
-- 
1.9.1

[U-Boot] [PATCH v3 4/6] dlmalloc: add option for skip memset in malloc init

[Przemyslaw Marczak]

This commit introduces new config: CONFIG_SYS_MALLOC_INIT_SKIP_ZEROING.

Before this change, the all amount of memory reserved for the malloc,
was set to zero in mem_malloc_init(). When the malloc reserved memory
exceeds few MiB, then the boot process can slow down.

So enabling this config, is an option to reduce the boot time.

This option can be enabled by Kconfig.

Note:
After enable this option, only calloc() will return the pointer to zeroed
memory area. Previously, without this option, the memory pointed to untouched
malloc memory region, was filled with zeros. So it means, that code with
malloc() calls should be reexamined.

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Reviewed-by: Simon Glass <sjg@chromium.org>

---
Changes v3:
- squash the commit with the Kconfig option
---
 Kconfig           | 26 +++++++++++++++++++-------
 common/dlmalloc.c | 10 +++++++---
 2 files changed, 26 insertions(+), 10 deletions(-)

diff --git a/Kconfig b/Kconfig
index 75bab7f..87d4daf 100644
--- a/Kconfig
+++ b/Kconfig
@@ -76,13 +76,25 @@ config SYS_MALLOC_F_LEN
 	  initial serial device and any others that are needed.
 
 menuconfig EXPERT
-        bool "Configure standard U-Boot features (expert users)"
-        help
-          This option allows certain base U-Boot options and settings
-          to be disabled or tweaked. This is for specialized
-          environments which can tolerate a "non-standard" U-Boot.
-          Only use this if you really know what you are doing.
-
+	bool "Configure standard U-Boot features (expert users)"
+	help
+	  This option allows certain base U-Boot options and settings
+	  to be disabled or tweaked. This is for specialized
+	  environments which can tolerate a "non-standard" U-Boot.
+	  Only use this if you really know what you are doing.
+
+if EXPERT
+	config SYS_MALLOC_INIT_SKIP_ZEROING
+	bool "Skip memset at malloc init (reduce boot time)"
+	help
+	  This avoids zeroing memory reserved for malloc at malloc init.
+	  Significant boot time reduction is visible for configs in which
+	  CONFIG_SYS_MALLOC_LEN value, has more than few MiB.
+	  Useful for bzip2, bmp logo.
+	  Warning:
+	  When enabling this, please check if malloc calls, maybe
+	  should be replaced by calloc - if expects zeroed memory.
+endif
 endmenu		# General setup
 
 menu "Boot images"
diff --git a/common/dlmalloc.c b/common/dlmalloc.c
index 6453ee9..63f68ed 100644
--- a/common/dlmalloc.c
+++ b/common/dlmalloc.c
@@ -1535,9 +1535,9 @@ void mem_malloc_init(ulong start, ulong size)
 
 	debug("using memory %#lx-%#lx for malloc()\n", mem_malloc_start,
 	      mem_malloc_end);
-
-	memset((void *)mem_malloc_start, 0, size);
-
+#ifndef CONFIG_SYS_MALLOC_INIT_SKIP_ZEROING
+	memset((void *)mem_malloc_start, 0x0, size);
+#endif
 	malloc_bin_reloc();
 }
 
@@ -2948,10 +2948,12 @@ Void_t* cALLOc(n, elem_size) size_t n; size_t elem_size;
 
 
   /* check if expand_top called, in which case don't need to clear */
+#ifndef CONFIG_SYS_MALLOC_INIT_SKIP_ZEROING
 #if MORECORE_CLEARS
   mchunkptr oldtop = top;
   INTERNAL_SIZE_T oldtopsize = chunksize(top);
 #endif
+#endif
   Void_t* mem = mALLOc (sz);
 
   if ((long)n < 0) return NULL;
@@ -2977,6 +2979,7 @@ Void_t* cALLOc(n, elem_size) size_t n; size_t elem_size;
 
     csz = chunksize(p);
 
+#ifndef CONFIG_SYS_MALLOC_INIT_SKIP_ZEROING
 #if MORECORE_CLEARS
     if (p == oldtop && csz > oldtopsize)
     {
@@ -2984,6 +2987,7 @@ Void_t* cALLOc(n, elem_size) size_t n; size_t elem_size;
       csz = oldtopsize;
     }
 #endif
+#endif
 
     MALLOC_ZERO(mem, csz - SIZE_SZ);
     return mem;
-- 
1.9.1

[U-Boot] [PATCH v3 5/6] trats2: defconfig: enable expert and skip memset at malloc init

[Przemyslaw Marczak]

Reduce the boot time of Trats2 by disabling the memset
at malloc init.

This was tested on Trats2.
A quick test with trace. Boot time from start to main_loop() entry:
- ~464ms - before this change (arch memset enabled for .bss clear)
- ~341ms - after this change

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
---
Changes V3
- none
---
 configs/trats2_defconfig | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/configs/trats2_defconfig b/configs/trats2_defconfig
index 1b98b73..21cc37a 100644
--- a/configs/trats2_defconfig
+++ b/configs/trats2_defconfig
@@ -3,3 +3,5 @@ CONFIG_ARCH_EXYNOS=y
 CONFIG_TARGET_TRATS2=y
 CONFIG_OF_CONTROL=y
 CONFIG_DEFAULT_DEVICE_TREE="exynos4412-trats2"
+CONFIG_EXPERT=y
+CONFIG_SYS_MALLOC_INIT_SKIP_ZEROING=y
-- 
1.9.1

[U-Boot] [PATCH v3 6/6] odroid: defconfig: enable expert and skip memset at malloc init

[Przemyslaw Marczak]

Reduce the boot time of Odroid X2/U3 by disabling the memset
at malloc init.

This was tested on Odroid X2.
A quick test with checking gpio pin state using the oscilloscope.
Boot time from start to bootcmd (change gpio state by memory write command):
- ~228ms - before this change (arch memset enabled for .bss clear)
- ~100ms - after this change

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Reviewed-by: Simon Glass <sjg@chromium.org>

---
Changes V3
- update commit head
---
 configs/odroid_defconfig | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/configs/odroid_defconfig b/configs/odroid_defconfig
index a842837..24af164 100644
--- a/configs/odroid_defconfig
+++ b/configs/odroid_defconfig
@@ -3,3 +3,5 @@ CONFIG_ARCH_EXYNOS=y
 CONFIG_TARGET_ODROID=y
 CONFIG_OF_CONTROL=y
 CONFIG_DEFAULT_DEVICE_TREE="exynos4412-odroid"
+CONFIG_EXPERT=y
+CONFIG_SYS_MALLOC_INIT_SKIP_ZEROING=y
-- 
1.9.1

[U-Boot] [PATCH v3 4/6] dlmalloc: add option for skip memset in malloc init

[Masahiro Yamada]

Hi Przemyslaw,


On Fri, 20 Feb 2015 12:06:17 +0100
Przemyslaw Marczak <p.marczak@samsung.com> wrote:

> This commit introduces new config: CONFIG_SYS_MALLOC_INIT_SKIP_ZEROING.
> 
> Before this change, the all amount of memory reserved for the malloc,
> was set to zero in mem_malloc_init(). When the malloc reserved memory
> exceeds few MiB, then the boot process can slow down.
> 
> So enabling this config, is an option to reduce the boot time.
> 
> This option can be enabled by Kconfig.
> 
> Note:
> After enable this option, only calloc() will return the pointer to zeroed
> memory area. Previously, without this option, the memory pointed to untouched
> malloc memory region, was filled with zeros. So it means, that code with
> malloc() calls should be reexamined.
> 
> Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
> Reviewed-by: Simon Glass <sjg@chromium.org>
> 
> ---
> Changes v3:
> - squash the commit with the Kconfig option
> ---
>  Kconfig           | 26 +++++++++++++++++++-------
>  common/dlmalloc.c | 10 +++++++---
>  2 files changed, 26 insertions(+), 10 deletions(-)
> 
> diff --git a/Kconfig b/Kconfig
> index 75bab7f..87d4daf 100644
> --- a/Kconfig
> +++ b/Kconfig
> @@ -76,13 +76,25 @@ config SYS_MALLOC_F_LEN
>  	  initial serial device and any others that are needed.
>  
>  menuconfig EXPERT
> -        bool "Configure standard U-Boot features (expert users)"
> -        help
> -          This option allows certain base U-Boot options and settings
> -          to be disabled or tweaked. This is for specialized
> -          environments which can tolerate a "non-standard" U-Boot.
> -          Only use this if you really know what you are doing.
> -
> +	bool "Configure standard U-Boot features (expert users)"
> +	help
> +	  This option allows certain base U-Boot options and settings
> +	  to be disabled or tweaked. This is for specialized
> +	  environments which can tolerate a "non-standard" U-Boot.
> +	  Only use this if you really know what you are doing.
> +
> +if EXPERT
> +	config SYS_MALLOC_INIT_SKIP_ZEROING
> +	bool "Skip memset at malloc init (reduce boot time)"
> +	help
> +	  This avoids zeroing memory reserved for malloc at malloc init.
> +	  Significant boot time reduction is visible for configs in which
> +	  CONFIG_SYS_MALLOC_LEN value, has more than few MiB.
> +	  Useful for bzip2, bmp logo.
> +	  Warning:
> +	  When enabling this, please check if malloc calls, maybe
> +	  should be replaced by calloc - if expects zeroed memory.
> +endif
>  endmenu		# General setup
>  
>  menu "Boot images"
> diff --git a/common/dlmalloc.c b/common/dlmalloc.c
> index 6453ee9..63f68ed 100644
> --- a/common/dlmalloc.c
> +++ b/common/dlmalloc.c
> @@ -1535,9 +1535,9 @@ void mem_malloc_init(ulong start, ulong size)
>  
>  	debug("using memory %#lx-%#lx for malloc()\n", mem_malloc_start,
>  	      mem_malloc_end);
> -
> -	memset((void *)mem_malloc_start, 0, size);
> -
> +#ifndef CONFIG_SYS_MALLOC_INIT_SKIP_ZEROING
> +	memset((void *)mem_malloc_start, 0x0, size);
> +#endif
>  	malloc_bin_reloc();
>  }
>  
> @@ -2948,10 +2948,12 @@ Void_t* cALLOc(n, elem_size) size_t n; size_t elem_size;
>  
>  
>    /* check if expand_top called, in which case don't need to clear */
> +#ifndef CONFIG_SYS_MALLOC_INIT_SKIP_ZEROING
>  #if MORECORE_CLEARS
>    mchunkptr oldtop = top;
>    INTERNAL_SIZE_T oldtopsize = chunksize(top);
>  #endif
> +#endif
>    Void_t* mem = mALLOc (sz);
>  
>    if ((long)n < 0) return NULL;
> @@ -2977,6 +2979,7 @@ Void_t* cALLOc(n, elem_size) size_t n; size_t elem_size;
>  
>      csz = chunksize(p);
>  
> +#ifndef CONFIG_SYS_MALLOC_INIT_SKIP_ZEROING
>  #if MORECORE_CLEARS
>      if (p == oldtop && csz > oldtopsize)
>      {
> @@ -2984,6 +2987,7 @@ Void_t* cALLOc(n, elem_size) size_t n; size_t elem_size;
>        csz = oldtopsize;
>      }
>  #endif
> +#endif
>  
>      MALLOC_ZERO(mem, csz - SIZE_SZ);
>      return mem;


You are adding only "ifndef" conditionals.

IMHO,
Isn't "#ifdef CONFIG_SYS_MALLOC_INIT_ZEROING" better?
(Generally speaking, CONFIG options that disable a feature are not preferable.)

You also need to add "default y" to the Kconfig
and add "# CONIFG_SYS_MALLOC_INIT_ZEROING is not set" to your defconfig.


Best Regards
Masahiro Yamada

[U-Boot] [PATCH v3 4/6] dlmalloc: add option for skip memset in malloc init

[Przemyslaw Marczak]

Hello,

On 02/20/2015 01:52 PM, Masahiro Yamada wrote:
> Hi Przemyslaw,
>
>
> On Fri, 20 Feb 2015 12:06:17 +0100
> Przemyslaw Marczak <p.marczak@samsung.com> wrote:
>
>> This commit introduces new config: CONFIG_SYS_MALLOC_INIT_SKIP_ZEROING.
>>
>> Before this change, the all amount of memory reserved for the malloc,
>> was set to zero in mem_malloc_init(). When the malloc reserved memory
>> exceeds few MiB, then the boot process can slow down.
>>
>> So enabling this config, is an option to reduce the boot time.
>>
>> This option can be enabled by Kconfig.
>>
>> Note:
>> After enable this option, only calloc() will return the pointer to zeroed
>> memory area. Previously, without this option, the memory pointed to untouched
>> malloc memory region, was filled with zeros. So it means, that code with
>> malloc() calls should be reexamined.
>>
>> Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
>> Reviewed-by: Simon Glass <sjg@chromium.org>
>>
>> ---
>> Changes v3:
>> - squash the commit with the Kconfig option
>> ---
>>   Kconfig           | 26 +++++++++++++++++++-------
>>   common/dlmalloc.c | 10 +++++++---
>>   2 files changed, 26 insertions(+), 10 deletions(-)
>>
>> diff --git a/Kconfig b/Kconfig
>> index 75bab7f..87d4daf 100644
>> --- a/Kconfig
>> +++ b/Kconfig
>> @@ -76,13 +76,25 @@ config SYS_MALLOC_F_LEN
>>   	  initial serial device and any others that are needed.
>>
>>   menuconfig EXPERT
>> -        bool "Configure standard U-Boot features (expert users)"
>> -        help
>> -          This option allows certain base U-Boot options and settings
>> -          to be disabled or tweaked. This is for specialized
>> -          environments which can tolerate a "non-standard" U-Boot.
>> -          Only use this if you really know what you are doing.
>> -
>> +	bool "Configure standard U-Boot features (expert users)"
>> +	help
>> +	  This option allows certain base U-Boot options and settings
>> +	  to be disabled or tweaked. This is for specialized
>> +	  environments which can tolerate a "non-standard" U-Boot.
>> +	  Only use this if you really know what you are doing.
>> +
>> +if EXPERT
>> +	config SYS_MALLOC_INIT_SKIP_ZEROING
>> +	bool "Skip memset at malloc init (reduce boot time)"
>> +	help
>> +	  This avoids zeroing memory reserved for malloc at malloc init.
>> +	  Significant boot time reduction is visible for configs in which
>> +	  CONFIG_SYS_MALLOC_LEN value, has more than few MiB.
>> +	  Useful for bzip2, bmp logo.
>> +	  Warning:
>> +	  When enabling this, please check if malloc calls, maybe
>> +	  should be replaced by calloc - if expects zeroed memory.
>> +endif
>>   endmenu		# General setup
>>
>>   menu "Boot images"
>> diff --git a/common/dlmalloc.c b/common/dlmalloc.c
>> index 6453ee9..63f68ed 100644
>> --- a/common/dlmalloc.c
>> +++ b/common/dlmalloc.c
>> @@ -1535,9 +1535,9 @@ void mem_malloc_init(ulong start, ulong size)
>>
>>   	debug("using memory %#lx-%#lx for malloc()\n", mem_malloc_start,
>>   	      mem_malloc_end);
>> -
>> -	memset((void *)mem_malloc_start, 0, size);
>> -
>> +#ifndef CONFIG_SYS_MALLOC_INIT_SKIP_ZEROING
>> +	memset((void *)mem_malloc_start, 0x0, size);
>> +#endif
>>   	malloc_bin_reloc();
>>   }
>>
>> @@ -2948,10 +2948,12 @@ Void_t* cALLOc(n, elem_size) size_t n; size_t elem_size;
>>
>>
>>     /* check if expand_top called, in which case don't need to clear */
>> +#ifndef CONFIG_SYS_MALLOC_INIT_SKIP_ZEROING
>>   #if MORECORE_CLEARS
>>     mchunkptr oldtop = top;
>>     INTERNAL_SIZE_T oldtopsize = chunksize(top);
>>   #endif
>> +#endif
>>     Void_t* mem = mALLOc (sz);
>>
>>     if ((long)n < 0) return NULL;
>> @@ -2977,6 +2979,7 @@ Void_t* cALLOc(n, elem_size) size_t n; size_t elem_size;
>>
>>       csz = chunksize(p);
>>
>> +#ifndef CONFIG_SYS_MALLOC_INIT_SKIP_ZEROING
>>   #if MORECORE_CLEARS
>>       if (p == oldtop && csz > oldtopsize)
>>       {
>> @@ -2984,6 +2987,7 @@ Void_t* cALLOc(n, elem_size) size_t n; size_t elem_size;
>>         csz = oldtopsize;
>>       }
>>   #endif
>> +#endif
>>
>>       MALLOC_ZERO(mem, csz - SIZE_SZ);
>>       return mem;
>
>
> You are adding only "ifndef" conditionals.
>
> IMHO,
> Isn't "#ifdef CONFIG_SYS_MALLOC_INIT_ZEROING" better?
> (Generally speaking, CONFIG options that disable a feature are not preferable.)
>
> You also need to add "default y" to the Kconfig
> and add "# CONIFG_SYS_MALLOC_INIT_ZEROING is not set" to your defconfig.
>
>
> Best Regards
> Masahiro Yamada
>

Ok, will change this to positive.

Best regards,
-- 
Przemyslaw Marczak
Samsung R&D Institute Poland
Samsung Electronics
p.marczak at samsung.com

[U-Boot] [PATCH v4 0/6] arm: a few steps to improve boot time

[Przemyslaw Marczak]

This patchset reduces the boot time for ARM architecture,
Exynos boards, and (ARM) boards with DFU enabled.

For tested Trats2 and Odroid X2 devices, this was done in four steps.

1. Enable the arch memcpy and memset - ARCH specific

2. Enable arch memset for .bss clear - ARCH specific

3. Reduce the .bss section small as possible by: (board specific)
  - remove static dfu buffer (32MiB in .bss - Trats2), and use malloc

4. Skip zeroing the memory reserved for malloc at malloc init.
   For Trats2 it was 80MiB of memory.

The .bss section will grow if we have a lot of static variables.
This section is cleared before jump to the relocated U-Boot,
and it was done word by word. To reduce the time for this step,
we can enable arch memset, which uses multiple ARM registers.

For configs with DFU enabled, we can find the dfu buffer in this section,
which has at least 8MB (32MB for Trats2). This is a lot of useless data,
which is not required for standard boot. So this buffer should be dynamic
allocated.

So, actually the all was about unnecessary operations on 'big' data.

Przemyslaw Marczak (6):
  exynos: config: enable arch memcpy and arch memset
  arm: relocation: clear .bss section with arch memset if defined
  dfu: mmc: file buffer: remove static allocation
  dlmalloc: do memset in malloc init as new default config
  trats2: defconfig: disable memset at malloc init
  odroid: defconfig: disable memset at malloc init

 Kconfig                         | 32 +++++++++++++++++++++++++-------
 arch/arm/lib/crt0.S             | 10 +++++++++-
 common/dlmalloc.c               | 10 +++++++---
 configs/odroid_defconfig        |  1 +
 configs/trats2_defconfig        |  1 +
 drivers/dfu/dfu_mmc.c           | 25 ++++++++++++++++++++++---
 include/configs/exynos-common.h |  3 +++
 7 files changed, 68 insertions(+), 14 deletions(-)

-- 
1.9.1

[U-Boot] [PATCH v4 1/6] exynos: config: enable arch memcpy and arch memset

[Przemyslaw Marczak]

This commit enables the following configs:
- CONFIG_USE_ARCH_MEMCPY
- CONFIG_USE_ARCH_MEMSET
This increases the performance of memcpy/memset
and also reduces the boot time.

This was tested on Trats2.
A quick test with trace. Boot time from start to main_loop() entry:
- ~1527ms - before this change (arch memset enabled for .bss clear)
- ~1384ms - after this change

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Cc: Minkyu Kang <mk7.kang@samsung.com>
Cc: Akshay Saraswat <akshay.s@samsung.com>
Cc: Simon Glass <sjg@chromium.org>
Cc: Sjoerd Simons <sjoerd.simons@collabora.co.uk>

---
Changes V3
- none

Changes V4
- none
---
 include/configs/exynos-common.h | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/include/configs/exynos-common.h b/include/configs/exynos-common.h
index 59676ae..87f8db0 100644
--- a/include/configs/exynos-common.h
+++ b/include/configs/exynos-common.h
@@ -24,6 +24,9 @@
 #define CONFIG_SKIP_LOWLEVEL_INIT
 #define CONFIG_BOARD_EARLY_INIT_F
 
+#define CONFIG_USE_ARCH_MEMCPY
+#define CONFIG_USE_ARCH_MEMSET
+
 /* Keep L2 Cache Disabled */
 #define CONFIG_CMD_CACHE
 
-- 
1.9.1

[U-Boot] [PATCH v4 2/6] arm: relocation: clear .bss section with arch memset if defined

[Przemyslaw Marczak]

For ARM architecture, enable the CONFIG_USE_ARCH_MEMSET/MEMCPY,
will highly increase the memset/memcpy performance. This is able
thanks to the ARM multiple register instructions.

Unfortunatelly the relocation is done without the cache enabled,
so it takes some time, but zeroing the BSS memory takes much more
longer, especially for the configs with big static buffers.

A quick test confirms, that the boot time improvement after using
the arch memcpy for relocation has no significant meaning.
The same test confirms that enable the memset for zeroing BSS,
reduces the boot time.

So this patch enables the arch memset for zeroing the BSS after
the relocation process. For ARM boards, this can be enabled
in board configs by defining: 'CONFIG_USE_ARCH_MEMSET'.

This was tested on Trats2.
A quick test with trace. Boot time from start to main_loop() entry:
- ~1384ms - before this change
-  ~888ms - after this change

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Cc: Albert Aribaud <albert.u.boot@aribaud.net>
Cc: Tom Rini <trini@ti.com>
---
Changes V3
- none

Changes V4
- none
---
 arch/arm/lib/crt0.S | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

diff --git a/arch/arm/lib/crt0.S b/arch/arm/lib/crt0.S
index 22df3e5..fab3d2c 100644
--- a/arch/arm/lib/crt0.S
+++ b/arch/arm/lib/crt0.S
@@ -115,14 +115,22 @@ here:
 	bl	c_runtime_cpu_setup	/* we still call old routine here */
 
 	ldr	r0, =__bss_start	/* this is auto-relocated! */
-	ldr	r1, =__bss_end		/* this is auto-relocated! */
 
+#ifdef CONFIG_USE_ARCH_MEMSET
+	ldr	r3, =__bss_end		/* this is auto-relocated! */
+	mov	r1, #0x00000000		/* prepare zero to clear BSS */
+
+	subs	r2, r3, r0		/* r2 = memset len */
+	bl	memset
+#else
+	ldr	r1, =__bss_end		/* this is auto-relocated! */
 	mov	r2, #0x00000000		/* prepare zero to clear BSS */
 
 clbss_l:cmp	r0, r1			/* while not at end of BSS */
 	strlo	r2, [r0]		/* clear 32-bit BSS word */
 	addlo	r0, r0, #4		/* move to next */
 	blo	clbss_l
+#endif
 
 	bl coloured_LED_init
 	bl red_led_on
-- 
1.9.1

[U-Boot] [PATCH v4 3/6] dfu: mmc: file buffer: remove static allocation

[Przemyslaw Marczak]

For writing files, DFU implementation requires the file buffer
with the len at least of file size. For big files it requires
the same big buffer.

Previously the file buffer was allocated as a static variable,
so it was a part of U-Boot .bss section. For 32MiB len of buffer
we have 32MiB of additional space, required for this section.

The .bss needs to be cleared after the relocation.
This introduces an additional boot delay at every start, but usually
the dfu feature is not required at the standard boot, so the buffer
should be allocated only if required.

This patch removes the static allocation of this buffer,
and alloc it with memalign after first call of function:
- dfu_fill_entity_mmc()
and the buffer is freed on dfu_free_entity() call.

This was tested on Trats2.
A quick test with trace. Boot time from start to main_loop() entry:
- ~888ms - before this change (arch memset enabled for .bss clear)
- ~464ms - after this change

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Cc: Lukasz Majewski <l.majewski@samsung.com>
Cc: Stephen Warren <swarren@nvidia.com>
Cc: Pantelis Antoniou <panto@antoniou-consulting.com>
Cc: Tom Rini <trini@ti.com>
Cc: Marek Vasut <marek.vasut@gmail.com>

---
Changes V3
- none

Changes V4
- none
---
 drivers/dfu/dfu_mmc.c | 25 ++++++++++++++++++++++---
 1 file changed, 22 insertions(+), 3 deletions(-)

diff --git a/drivers/dfu/dfu_mmc.c b/drivers/dfu/dfu_mmc.c
index 62d72fe..fd865e1 100644
--- a/drivers/dfu/dfu_mmc.c
+++ b/drivers/dfu/dfu_mmc.c
@@ -16,8 +16,7 @@
 #include <fat.h>
 #include <mmc.h>
 
-static unsigned char __aligned(CONFIG_SYS_CACHELINE_SIZE)
-				dfu_file_buf[CONFIG_SYS_DFU_MAX_FILE_SIZE];
+static unsigned char *dfu_file_buf;
 static long dfu_file_buf_len;
 
 static int mmc_access_part(struct dfu_entity *dfu, struct mmc *mmc, int part)
@@ -211,7 +210,7 @@ int dfu_flush_medium_mmc(struct dfu_entity *dfu)
 
 	if (dfu->layout != DFU_RAW_ADDR) {
 		/* Do stuff here. */
-		ret = mmc_file_op(DFU_OP_WRITE, dfu, &dfu_file_buf,
+		ret = mmc_file_op(DFU_OP_WRITE, dfu, dfu_file_buf,
 				&dfu_file_buf_len);
 
 		/* Now that we're done */
@@ -263,6 +262,14 @@ int dfu_read_medium_mmc(struct dfu_entity *dfu, u64 offset, void *buf,
 	return ret;
 }
 
+void dfu_free_entity_mmc(struct dfu_entity *dfu)
+{
+	if (dfu_file_buf) {
+		free(dfu_file_buf);
+		dfu_file_buf = NULL;
+	}
+}
+
 /*
  * @param s Parameter string containing space-separated arguments:
  *	1st:
@@ -370,6 +377,18 @@ int dfu_fill_entity_mmc(struct dfu_entity *dfu, char *devstr, char *s)
 	dfu->write_medium = dfu_write_medium_mmc;
 	dfu->flush_medium = dfu_flush_medium_mmc;
 	dfu->inited = 0;
+	dfu->free_entity = dfu_free_entity_mmc;
+
+	/* Check if file buffer is ready */
+	if (!dfu_file_buf) {
+		dfu_file_buf = memalign(CONFIG_SYS_CACHELINE_SIZE,
+					CONFIG_SYS_DFU_MAX_FILE_SIZE);
+		if (!dfu_file_buf) {
+			error("Could not memalign 0x%x bytes",
+			      CONFIG_SYS_DFU_MAX_FILE_SIZE);
+			return -ENOMEM;
+		}
+	}
 
 	return 0;
 }
-- 
1.9.1

[U-Boot] [PATCH v4 4/6] dlmalloc: do memset in malloc init as new default config

[Przemyslaw Marczak]

This commit introduces new config: CONFIG_SYS_MALLOC_INIT_DO_ZEROING.

This config is an expert option and is enabled by default.

The all amount of memory reserved for the malloc, is by default set
to zero in mem_malloc_init(). When the malloc reserved memory exceeds
few MiB, then the boot process can slow down.

So disabling this config, is an expert option to reduce the boot time,
and can be disabled by Kconfig.

Note:
After disable this option, only calloc() will return the pointer
to the zeroed memory area. Previously, without this option,
the memory pointed to untouched malloc memory region, was filled
with zeros. So it means, that code with malloc() calls should
be reexamined.

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>

---
Changes v3:
- squash the commit with the Kconfig option

Changes v4:
- adjust commit message for the new config
---
 Kconfig           | 32 +++++++++++++++++++++++++-------
 common/dlmalloc.c | 10 +++++++---
 2 files changed, 32 insertions(+), 10 deletions(-)

diff --git a/Kconfig b/Kconfig
index 75bab7f..9ea99b5 100644
--- a/Kconfig
+++ b/Kconfig
@@ -76,13 +76,31 @@ config SYS_MALLOC_F_LEN
 	  initial serial device and any others that are needed.
 
 menuconfig EXPERT
-        bool "Configure standard U-Boot features (expert users)"
-        help
-          This option allows certain base U-Boot options and settings
-          to be disabled or tweaked. This is for specialized
-          environments which can tolerate a "non-standard" U-Boot.
-          Only use this if you really know what you are doing.
-
+	bool "Configure standard U-Boot features (expert users)"
+	default y
+	help
+	  This option allows certain base U-Boot options and settings
+	  to be disabled or tweaked. This is for specialized
+	  environments which can tolerate a "non-standard" U-Boot.
+	  Only use this if you really know what you are doing.
+
+if EXPERT
+	config SYS_MALLOC_INIT_DO_ZEROING
+	bool "Init with zeros the memory reserved for malloc (slow)"
+	default y
+	help
+	  This setting is enabled by default. The reserved malloc
+	  memory is initialized with zeros, so first malloc calls
+	  will return the pointer to the zeroed memory. But this
+	  slows the boot time.
+
+	  It is recommended to disable it, when CONFIG_SYS_MALLOC_LEN
+	  value, has more than few MiB, e.g. when uses bzip2 or bmp logo.
+	  Then the boot time can be significantly reduced.
+	  Warning:
+	  When disabling this, please check if malloc calls, maybe
+	  should be replaced by calloc - if expects zeroed memory.
+endif
 endmenu		# General setup
 
 menu "Boot images"
diff --git a/common/dlmalloc.c b/common/dlmalloc.c
index 6453ee9..21e103b 100644
--- a/common/dlmalloc.c
+++ b/common/dlmalloc.c
@@ -1535,9 +1535,9 @@ void mem_malloc_init(ulong start, ulong size)
 
 	debug("using memory %#lx-%#lx for malloc()\n", mem_malloc_start,
 	      mem_malloc_end);
-
-	memset((void *)mem_malloc_start, 0, size);
-
+#ifdef CONFIG_SYS_MALLOC_INIT_DO_ZEROING
+	memset((void *)mem_malloc_start, 0x0, size);
+#endif
 	malloc_bin_reloc();
 }
 
@@ -2948,10 +2948,12 @@ Void_t* cALLOc(n, elem_size) size_t n; size_t elem_size;
 
 
   /* check if expand_top called, in which case don't need to clear */
+#ifdef CONFIG_SYS_MALLOC_INIT_DO_ZEROING
 #if MORECORE_CLEARS
   mchunkptr oldtop = top;
   INTERNAL_SIZE_T oldtopsize = chunksize(top);
 #endif
+#endif
   Void_t* mem = mALLOc (sz);
 
   if ((long)n < 0) return NULL;
@@ -2977,6 +2979,7 @@ Void_t* cALLOc(n, elem_size) size_t n; size_t elem_size;
 
     csz = chunksize(p);
 
+#ifdef CONFIG_SYS_MALLOC_INIT_DO_ZEROING
 #if MORECORE_CLEARS
     if (p == oldtop && csz > oldtopsize)
     {
@@ -2984,6 +2987,7 @@ Void_t* cALLOc(n, elem_size) size_t n; size_t elem_size;
       csz = oldtopsize;
     }
 #endif
+#endif
 
     MALLOC_ZERO(mem, csz - SIZE_SZ);
     return mem;
-- 
1.9.1

[U-Boot] [PATCH v4 5/6] trats2: defconfig: disable memset at malloc init

[Przemyslaw Marczak]

Reduce the boot time of Trats2 by disabling the memset
at malloc init.

This was tested on Trats2.
A quick test with trace. Boot time from start to main_loop() entry:
- ~464ms - before this change (arch memset enabled for .bss clear)
- ~341ms - after this change

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>

---
Changes V3
- none

Changes V4
- trats2_defconfig: remove CONFIG_EXPERT
- trats2_defconfig: disable CONFIG_SYS_MALLOC_INIT_DO_ZEROING
---
 configs/trats2_defconfig | 1 +
 1 file changed, 1 insertion(+)

diff --git a/configs/trats2_defconfig b/configs/trats2_defconfig
index 1b98b73..e6d2062 100644
--- a/configs/trats2_defconfig
+++ b/configs/trats2_defconfig
@@ -3,3 +3,4 @@ CONFIG_ARCH_EXYNOS=y
 CONFIG_TARGET_TRATS2=y
 CONFIG_OF_CONTROL=y
 CONFIG_DEFAULT_DEVICE_TREE="exynos4412-trats2"
+# CONFIG_SYS_MALLOC_INIT_DO_ZEROING is not set
-- 
1.9.1

[U-Boot] [PATCH v4 6/6] odroid: defconfig: disable memset at malloc init

[Przemyslaw Marczak]

Reduce the boot time of Odroid X2/U3 by disabling the memset
at malloc init.

This was tested on Odroid X2.
A quick test with checking gpio pin state using the oscilloscope.
Boot time from start to bootcmd (change gpio state by memory write command):
- ~228ms - before this change (arch memset enabled for .bss clear)
- ~100ms - after this change

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>

---
Changes V3
- update commit head

Changes V4
- odroid_defconfig: remove CONFIG_EXPERT
- odroid_defconfig: disable CONFIG_SYS_MALLOC_INIT_DO_ZEROING
---
 configs/odroid_defconfig | 1 +
 1 file changed, 1 insertion(+)

diff --git a/configs/odroid_defconfig b/configs/odroid_defconfig
index a842837..592232f 100644
--- a/configs/odroid_defconfig
+++ b/configs/odroid_defconfig
@@ -3,3 +3,4 @@ CONFIG_ARCH_EXYNOS=y
 CONFIG_TARGET_ODROID=y
 CONFIG_OF_CONTROL=y
 CONFIG_DEFAULT_DEVICE_TREE="exynos4412-odroid"
+# CONFIG_SYS_MALLOC_INIT_DO_ZEROING is not set
-- 
1.9.1

[U-Boot] [PATCH v4 4/6] dlmalloc: do memset in malloc init as new default config

[Simon Glass]

Hi Przemyslaw,

On 23 February 2015 at 10:16, Przemyslaw Marczak <p.marczak@samsung.com> wrote:
> This commit introduces new config: CONFIG_SYS_MALLOC_INIT_DO_ZEROING.
>

Minor nit: CONFIG_SYS_MALLOC_CLEAR_ON_INIT might be better.

> This config is an expert option and is enabled by default.
>
> The all amount of memory reserved for the malloc, is by default set
> to zero in mem_malloc_init(). When the malloc reserved memory exceeds
> few MiB, then the boot process can slow down.
>
> So disabling this config, is an expert option to reduce the boot time,
> and can be disabled by Kconfig.
>
> Note:
> After disable this option, only calloc() will return the pointer
> to the zeroed memory area. Previously, without this option,
> the memory pointed to untouched malloc memory region, was filled
> with zeros. So it means, that code with malloc() calls should
> be reexamined.
>
> Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
>

Reviewed-by: Simon Glass <sjg@chromium.org>

> ---
> Changes v3:
> - squash the commit with the Kconfig option
>
> Changes v4:
> - adjust commit message for the new config
> ---
>  Kconfig           | 32 +++++++++++++++++++++++++-------
>  common/dlmalloc.c | 10 +++++++---
>  2 files changed, 32 insertions(+), 10 deletions(-)
>
> diff --git a/Kconfig b/Kconfig
> index 75bab7f..9ea99b5 100644
> --- a/Kconfig
> +++ b/Kconfig
> @@ -76,13 +76,31 @@ config SYS_MALLOC_F_LEN
>           initial serial device and any others that are needed.
>
>  menuconfig EXPERT
> -        bool "Configure standard U-Boot features (expert users)"
> -        help
> -          This option allows certain base U-Boot options and settings
> -          to be disabled or tweaked. This is for specialized
> -          environments which can tolerate a "non-standard" U-Boot.
> -          Only use this if you really know what you are doing.
> -
> +       bool "Configure standard U-Boot features (expert users)"
> +       default y
> +       help
> +         This option allows certain base U-Boot options and settings
> +         to be disabled or tweaked. This is for specialized
> +         environments which can tolerate a "non-standard" U-Boot.
> +         Only use this if you really know what you are doing.
> +
> +if EXPERT
> +       config SYS_MALLOC_INIT_DO_ZEROING
> +       bool "Init with zeros the memory reserved for malloc (slow)"
> +       default y
> +       help
> +         This setting is enabled by default. The reserved malloc
> +         memory is initialized with zeros, so first malloc calls
> +         will return the pointer to the zeroed memory. But this
> +         slows the boot time.
> +
> +         It is recommended to disable it, when CONFIG_SYS_MALLOC_LEN
> +         value, has more than few MiB, e.g. when uses bzip2 or bmp logo.
> +         Then the boot time can be significantly reduced.
> +         Warning:
> +         When disabling this, please check if malloc calls, maybe
> +         should be replaced by calloc - if expects zeroed memory.
> +endif
>  endmenu                # General setup
>
>  menu "Boot images"
> diff --git a/common/dlmalloc.c b/common/dlmalloc.c
> index 6453ee9..21e103b 100644
> --- a/common/dlmalloc.c
> +++ b/common/dlmalloc.c
> @@ -1535,9 +1535,9 @@ void mem_malloc_init(ulong start, ulong size)
>
>         debug("using memory %#lx-%#lx for malloc()\n", mem_malloc_start,
>               mem_malloc_end);
> -
> -       memset((void *)mem_malloc_start, 0, size);
> -
> +#ifdef CONFIG_SYS_MALLOC_INIT_DO_ZEROING
> +       memset((void *)mem_malloc_start, 0x0, size);
> +#endif
>         malloc_bin_reloc();
>  }
>
> @@ -2948,10 +2948,12 @@ Void_t* cALLOc(n, elem_size) size_t n; size_t elem_size;
>
>
>    /* check if expand_top called, in which case don't need to clear */
> +#ifdef CONFIG_SYS_MALLOC_INIT_DO_ZEROING
>  #if MORECORE_CLEARS
>    mchunkptr oldtop = top;
>    INTERNAL_SIZE_T oldtopsize = chunksize(top);
>  #endif
> +#endif
>    Void_t* mem = mALLOc (sz);
>
>    if ((long)n < 0) return NULL;
> @@ -2977,6 +2979,7 @@ Void_t* cALLOc(n, elem_size) size_t n; size_t elem_size;
>
>      csz = chunksize(p);
>
> +#ifdef CONFIG_SYS_MALLOC_INIT_DO_ZEROING
>  #if MORECORE_CLEARS
>      if (p == oldtop && csz > oldtopsize)
>      {
> @@ -2984,6 +2987,7 @@ Void_t* cALLOc(n, elem_size) size_t n; size_t elem_size;
>        csz = oldtopsize;
>      }
>  #endif
> +#endif
>
>      MALLOC_ZERO(mem, csz - SIZE_SZ);
>      return mem;
> --
> 1.9.1
>

[U-Boot] [PATCH v5 0/7] arm: a few steps to improve boot time

[Przemyslaw Marczak]

This patchset reduces the boot time for ARM architecture,
Exynos boards, and (ARM) boards with DFU enabled.

For tested Trats2 and Odroid X2/XU3 devices, this was done in four steps.

1. Enable the arch memcpy and memset - ARCH specific

2. Enable arch memset for .bss clear - ARCH specific

3. Reduce the .bss section small as possible by: (board specific)
  - remove static dfu buffer (32MiB in .bss - Trats2), and use malloc

4. Skip zeroing the memory reserved for malloc at malloc init.
   For Trats2 it was 80MiB of memory.

The .bss section will grow if we have a lot of static variables.
This section is cleared before jump to the relocated U-Boot,
and it was done word by word. To reduce the time for this step,
we can enable arch memset, which uses multiple ARM registers.

For configs with DFU enabled, we can find the dfu buffer in this section,
which has at least 8MB (32MB for Trats2). This is a lot of useless data,
which is not required for standard boot. So this buffer should be dynamic
allocated.

So, actually the all was about unnecessary operations on 'big' data.

Przemyslaw Marczak (7):
  exynos: config: enable arch memcpy and arch memset
  arm: relocation: clear .bss section with arch memset if defined
  dfu: mmc: file buffer: remove static allocation
  dlmalloc: do memset in malloc init as new default config
  trats2: defconfig: disable memset at malloc init
  odroid: defconfig: disable memset at malloc init
  odroid-xu3: defconfig: disable memset at malloc init

 Kconfig                         | 32 +++++++++++++++++++++++++-------
 arch/arm/lib/crt0.S             | 10 +++++++++-
 common/dlmalloc.c               | 10 +++++++---
 configs/odroid-xu3_defconfig    |  2 ++
 configs/odroid_defconfig        |  1 +
 configs/trats2_defconfig        |  1 +
 drivers/dfu/dfu_mmc.c           | 25 ++++++++++++++++++++++---
 include/configs/exynos-common.h |  3 +++
 8 files changed, 70 insertions(+), 14 deletions(-)

-- 
1.9.1

[U-Boot] [PATCH v5 1/7] exynos: config: enable arch memcpy and arch memset

[Przemyslaw Marczak]

This commit enables the following configs:
- CONFIG_USE_ARCH_MEMCPY
- CONFIG_USE_ARCH_MEMSET
This increases the performance of memcpy/memset
and also reduces the boot time.

This was tested on Trats2.
A quick test with trace. Boot time from start to main_loop() entry:
- ~1527ms - before this change (arch memset enabled for .bss clear)
- ~1384ms - after this change

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Cc: Minkyu Kang <mk7.kang@samsung.com>
Cc: Akshay Saraswat <akshay.s@samsung.com>
Cc: Simon Glass <sjg@chromium.org>
Cc: Sjoerd Simons <sjoerd.simons@collabora.co.uk>

---
Changes V3, V4, V5
- none
---
 include/configs/exynos-common.h | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/include/configs/exynos-common.h b/include/configs/exynos-common.h
index 59676ae..87f8db0 100644
--- a/include/configs/exynos-common.h
+++ b/include/configs/exynos-common.h
@@ -24,6 +24,9 @@
 #define CONFIG_SKIP_LOWLEVEL_INIT
 #define CONFIG_BOARD_EARLY_INIT_F
 
+#define CONFIG_USE_ARCH_MEMCPY
+#define CONFIG_USE_ARCH_MEMSET
+
 /* Keep L2 Cache Disabled */
 #define CONFIG_CMD_CACHE
 
-- 
1.9.1

[U-Boot] [PATCH v5 2/7] arm: relocation: clear .bss section with arch memset if defined

[Przemyslaw Marczak]

For ARM architecture, enable the CONFIG_USE_ARCH_MEMSET/MEMCPY,
will highly increase the memset/memcpy performance. This is able
thanks to the ARM multiple register instructions.

Unfortunatelly the relocation is done without the cache enabled,
so it takes some time, but zeroing the BSS memory takes much more
longer, especially for the configs with big static buffers.

A quick test confirms, that the boot time improvement after using
the arch memcpy for relocation has no significant meaning.
The same test confirms that enable the memset for zeroing BSS,
reduces the boot time.

So this patch enables the arch memset for zeroing the BSS after
the relocation process. For ARM boards, this can be enabled
in board configs by defining: 'CONFIG_USE_ARCH_MEMSET'.

This was tested on Trats2.
A quick test with trace. Boot time from start to main_loop() entry:
- ~1384ms - before this change
-  ~888ms - after this change

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Cc: Albert Aribaud <albert.u.boot@aribaud.net>
Cc: Tom Rini <trini@ti.com>
---
Changes V3, V4, V5
- none
---
 arch/arm/lib/crt0.S | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

diff --git a/arch/arm/lib/crt0.S b/arch/arm/lib/crt0.S
index 22df3e5..fab3d2c 100644
--- a/arch/arm/lib/crt0.S
+++ b/arch/arm/lib/crt0.S
@@ -115,14 +115,22 @@ here:
 	bl	c_runtime_cpu_setup	/* we still call old routine here */
 
 	ldr	r0, =__bss_start	/* this is auto-relocated! */
-	ldr	r1, =__bss_end		/* this is auto-relocated! */
 
+#ifdef CONFIG_USE_ARCH_MEMSET
+	ldr	r3, =__bss_end		/* this is auto-relocated! */
+	mov	r1, #0x00000000		/* prepare zero to clear BSS */
+
+	subs	r2, r3, r0		/* r2 = memset len */
+	bl	memset
+#else
+	ldr	r1, =__bss_end		/* this is auto-relocated! */
 	mov	r2, #0x00000000		/* prepare zero to clear BSS */
 
 clbss_l:cmp	r0, r1			/* while not at end of BSS */
 	strlo	r2, [r0]		/* clear 32-bit BSS word */
 	addlo	r0, r0, #4		/* move to next */
 	blo	clbss_l
+#endif
 
 	bl coloured_LED_init
 	bl red_led_on
-- 
1.9.1

[U-Boot] [PATCH v5 3/7] dfu: mmc: file buffer: remove static allocation

[Przemyslaw Marczak]

For writing files, DFU implementation requires the file buffer
with the len at least of file size. For big files it requires
the same big buffer.

Previously the file buffer was allocated as a static variable,
so it was a part of U-Boot .bss section. For 32MiB len of buffer
we have 32MiB of additional space, required for this section.

The .bss needs to be cleared after the relocation.
This introduces an additional boot delay at every start, but usually
the dfu feature is not required at the standard boot, so the buffer
should be allocated only if required.

This patch removes the static allocation of this buffer,
and alloc it with memalign after first call of function:
- dfu_fill_entity_mmc()
and the buffer is freed on dfu_free_entity() call.

This was tested on Trats2.
A quick test with trace. Boot time from start to main_loop() entry:
- ~888ms - before this change (arch memset enabled for .bss clear)
- ~464ms - after this change

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Cc: Lukasz Majewski <l.majewski@samsung.com>
Cc: Stephen Warren <swarren@nvidia.com>
Cc: Pantelis Antoniou <panto@antoniou-consulting.com>
Cc: Tom Rini <trini@ti.com>
Cc: Marek Vasut <marek.vasut@gmail.com>

---
Changes V3, V4, V5
- none
---
 drivers/dfu/dfu_mmc.c | 25 ++++++++++++++++++++++---
 1 file changed, 22 insertions(+), 3 deletions(-)

diff --git a/drivers/dfu/dfu_mmc.c b/drivers/dfu/dfu_mmc.c
index 62d72fe..fd865e1 100644
--- a/drivers/dfu/dfu_mmc.c
+++ b/drivers/dfu/dfu_mmc.c
@@ -16,8 +16,7 @@
 #include <fat.h>
 #include <mmc.h>
 
-static unsigned char __aligned(CONFIG_SYS_CACHELINE_SIZE)
-				dfu_file_buf[CONFIG_SYS_DFU_MAX_FILE_SIZE];
+static unsigned char *dfu_file_buf;
 static long dfu_file_buf_len;
 
 static int mmc_access_part(struct dfu_entity *dfu, struct mmc *mmc, int part)
@@ -211,7 +210,7 @@ int dfu_flush_medium_mmc(struct dfu_entity *dfu)
 
 	if (dfu->layout != DFU_RAW_ADDR) {
 		/* Do stuff here. */
-		ret = mmc_file_op(DFU_OP_WRITE, dfu, &dfu_file_buf,
+		ret = mmc_file_op(DFU_OP_WRITE, dfu, dfu_file_buf,
 				&dfu_file_buf_len);
 
 		/* Now that we're done */
@@ -263,6 +262,14 @@ int dfu_read_medium_mmc(struct dfu_entity *dfu, u64 offset, void *buf,
 	return ret;
 }
 
+void dfu_free_entity_mmc(struct dfu_entity *dfu)
+{
+	if (dfu_file_buf) {
+		free(dfu_file_buf);
+		dfu_file_buf = NULL;
+	}
+}
+
 /*
  * @param s Parameter string containing space-separated arguments:
  *	1st:
@@ -370,6 +377,18 @@ int dfu_fill_entity_mmc(struct dfu_entity *dfu, char *devstr, char *s)
 	dfu->write_medium = dfu_write_medium_mmc;
 	dfu->flush_medium = dfu_flush_medium_mmc;
 	dfu->inited = 0;
+	dfu->free_entity = dfu_free_entity_mmc;
+
+	/* Check if file buffer is ready */
+	if (!dfu_file_buf) {
+		dfu_file_buf = memalign(CONFIG_SYS_CACHELINE_SIZE,
+					CONFIG_SYS_DFU_MAX_FILE_SIZE);
+		if (!dfu_file_buf) {
+			error("Could not memalign 0x%x bytes",
+			      CONFIG_SYS_DFU_MAX_FILE_SIZE);
+			return -ENOMEM;
+		}
+	}
 
 	return 0;
 }
-- 
1.9.1

[U-Boot] [PATCH v5 4/7] dlmalloc: do memset in malloc init as new default config

[Przemyslaw Marczak]

This commit introduces new config: CONFIG_SYS_MALLOC_CLEAR_ON_INIT.

This config is an expert option and is enabled by default.

The all amount of memory reserved for the malloc, is by default set
to zero in mem_malloc_init(). When the malloc reserved memory exceeds
few MiB, then the boot process can slow down.

So disabling this config, is an expert option to reduce the boot time,
and can be disabled by Kconfig.

Note:
After disable this option, only calloc() will return the pointer
to the zeroed memory area. Previously, without this option,
the memory pointed to untouched malloc memory region, was filled
with zeros. So it means, that code with malloc() calls should
be reexamined.

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Reviewed-by: Simon Glass <sjg@chromium.org>

---
Changes v3:
- squash the commit with the Kconfig option

Changes v4:
- adjust commit message for the new config

Changes v5:
- change config name to CONFIG_SYS_MALLOC_CLEAR_ON_INIT
---
 Kconfig           | 32 +++++++++++++++++++++++++-------
 common/dlmalloc.c | 10 +++++++---
 2 files changed, 32 insertions(+), 10 deletions(-)

diff --git a/Kconfig b/Kconfig
index 75bab7f..d6c75d5 100644
--- a/Kconfig
+++ b/Kconfig
@@ -76,13 +76,31 @@ config SYS_MALLOC_F_LEN
 	  initial serial device and any others that are needed.
 
 menuconfig EXPERT
-        bool "Configure standard U-Boot features (expert users)"
-        help
-          This option allows certain base U-Boot options and settings
-          to be disabled or tweaked. This is for specialized
-          environments which can tolerate a "non-standard" U-Boot.
-          Only use this if you really know what you are doing.
-
+	bool "Configure standard U-Boot features (expert users)"
+	default y
+	help
+	  This option allows certain base U-Boot options and settings
+	  to be disabled or tweaked. This is for specialized
+	  environments which can tolerate a "non-standard" U-Boot.
+	  Only use this if you really know what you are doing.
+
+if EXPERT
+	config SYS_MALLOC_CLEAR_ON_INIT
+	bool "Init with zeros the memory reserved for malloc (slow)"
+	default y
+	help
+	  This setting is enabled by default. The reserved malloc
+	  memory is initialized with zeros, so first malloc calls
+	  will return the pointer to the zeroed memory. But this
+	  slows the boot time.
+
+	  It is recommended to disable it, when CONFIG_SYS_MALLOC_LEN
+	  value, has more than few MiB, e.g. when uses bzip2 or bmp logo.
+	  Then the boot time can be significantly reduced.
+	  Warning:
+	  When disabling this, please check if malloc calls, maybe
+	  should be replaced by calloc - if expects zeroed memory.
+endif
 endmenu		# General setup
 
 menu "Boot images"
diff --git a/common/dlmalloc.c b/common/dlmalloc.c
index 6453ee9..b2ce063 100644
--- a/common/dlmalloc.c
+++ b/common/dlmalloc.c
@@ -1535,9 +1535,9 @@ void mem_malloc_init(ulong start, ulong size)
 
 	debug("using memory %#lx-%#lx for malloc()\n", mem_malloc_start,
 	      mem_malloc_end);
-
-	memset((void *)mem_malloc_start, 0, size);
-
+#ifdef CONFIG_SYS_MALLOC_CLEAR_ON_INIT
+	memset((void *)mem_malloc_start, 0x0, size);
+#endif
 	malloc_bin_reloc();
 }
 
@@ -2948,10 +2948,12 @@ Void_t* cALLOc(n, elem_size) size_t n; size_t elem_size;
 
 
   /* check if expand_top called, in which case don't need to clear */
+#ifdef CONFIG_SYS_MALLOC_CLEAR_ON_INIT
 #if MORECORE_CLEARS
   mchunkptr oldtop = top;
   INTERNAL_SIZE_T oldtopsize = chunksize(top);
 #endif
+#endif
   Void_t* mem = mALLOc (sz);
 
   if ((long)n < 0) return NULL;
@@ -2977,6 +2979,7 @@ Void_t* cALLOc(n, elem_size) size_t n; size_t elem_size;
 
     csz = chunksize(p);
 
+#ifdef CONFIG_SYS_MALLOC_CLEAR_ON_INIT
 #if MORECORE_CLEARS
     if (p == oldtop && csz > oldtopsize)
     {
@@ -2984,6 +2987,7 @@ Void_t* cALLOc(n, elem_size) size_t n; size_t elem_size;
       csz = oldtopsize;
     }
 #endif
+#endif
 
     MALLOC_ZERO(mem, csz - SIZE_SZ);
     return mem;
-- 
1.9.1

[U-Boot] [PATCH v5 5/7] trats2: defconfig: disable memset at malloc init

[Przemyslaw Marczak]

Reduce the boot time of Trats2 by disabling the memset
at malloc init.

This was tested on Trats2.
A quick test with trace. Boot time from start to main_loop() entry:
- ~464ms - before this change (arch memset enabled for .bss clear)
- ~341ms - after this change

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Reviewed-by: Simon Glass <sjg@chromium.org>

---
Changes V3
- none

Changes V4
- trats2_defconfig: remove CONFIG_EXPERT
- trats2_defconfig: disable CONFIG_SYS_MALLOC_INIT_DO_ZEROING

Changes v5:
- update disabled config name to CONFIG_SYS_MALLOC_CLEAR_ON_INIT
---
 configs/trats2_defconfig | 1 +
 1 file changed, 1 insertion(+)

diff --git a/configs/trats2_defconfig b/configs/trats2_defconfig
index 1b98b73..9359706 100644
--- a/configs/trats2_defconfig
+++ b/configs/trats2_defconfig
@@ -3,3 +3,4 @@ CONFIG_ARCH_EXYNOS=y
 CONFIG_TARGET_TRATS2=y
 CONFIG_OF_CONTROL=y
 CONFIG_DEFAULT_DEVICE_TREE="exynos4412-trats2"
+# CONFIG_SYS_MALLOC_CLEAR_ON_INIT is not set
-- 
1.9.1

[U-Boot] [PATCH v5 6/7] odroid: defconfig: disable memset at malloc init

[Przemyslaw Marczak]

Reduce the boot time of Odroid X2/U3 by disabling the memset
at malloc init.

This was tested on Odroid X2.
A quick test with checking gpio pin state using the oscilloscope.
Boot time from start to bootcmd (change gpio state by memory write command):
- ~228ms - before this change (arch memset enabled for .bss clear)
- ~100ms - after this change

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Reviewed-by: Simon Glass <sjg@chromium.org>

---
Changes V3
- update commit head

Changes V4
- odroid_defconfig: remove CONFIG_EXPERT
- odroid_defconfig: disable CONFIG_SYS_MALLOC_INIT_DO_ZEROING

Changes v5
- update disabled config name to CONFIG_SYS_MALLOC_CLEAR_ON_INIT
---
 configs/odroid_defconfig | 1 +
 1 file changed, 1 insertion(+)

diff --git a/configs/odroid_defconfig b/configs/odroid_defconfig
index a842837..aac9f5a 100644
--- a/configs/odroid_defconfig
+++ b/configs/odroid_defconfig
@@ -3,3 +3,4 @@ CONFIG_ARCH_EXYNOS=y
 CONFIG_TARGET_ODROID=y
 CONFIG_OF_CONTROL=y
 CONFIG_DEFAULT_DEVICE_TREE="exynos4412-odroid"
+# CONFIG_SYS_MALLOC_CLEAR_ON_INIT is not set
-- 
1.9.1

[U-Boot] [PATCH v5 7/7] odroid-xu3: defconfig: disable memset at malloc init

[Przemyslaw Marczak]

Reduce the boot time of Odroid XU3 by disabling the memset
at malloc init.

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
---
Changes v5
- new commit

---
 configs/odroid-xu3_defconfig | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/configs/odroid-xu3_defconfig b/configs/odroid-xu3_defconfig
index 74aa0cf..0fb4623 100644
--- a/configs/odroid-xu3_defconfig
+++ b/configs/odroid-xu3_defconfig
@@ -2,3 +2,5 @@ CONFIG_ARM=y
 CONFIG_ARCH_EXYNOS=y
 CONFIG_TARGET_ODROID_XU3=y
 CONFIG_DEFAULT_DEVICE_TREE="exynos5422-odroidxu3"
+# CONFIG_SYS_MALLOC_CLEAR_ON_INIT is not set
+
-- 
1.9.1

[U-Boot] [PATCH v4 4/6] dlmalloc: do memset in malloc init as new default config

[Przemyslaw Marczak]

Hello Simon,

On 02/23/2015 06:38 PM, Simon Glass wrote:
> Hi Przemyslaw,
>
> On 23 February 2015 at 10:16, Przemyslaw Marczak <p.marczak@samsung.com> wrote:
>> This commit introduces new config: CONFIG_SYS_MALLOC_INIT_DO_ZEROING.
>>
>
> Minor nit: CONFIG_SYS_MALLOC_CLEAR_ON_INIT might be better.
>
...snip...

The config name is updated and I get XU3 for test so I disabled the 
memset also for this board in V5.

Best regards,
-- 
Przemyslaw Marczak
Samsung R&D Institute Poland
Samsung Electronics
p.marczak at samsung.com

[U-Boot] [PATCH v5 1/7] exynos: config: enable arch memcpy and arch memset

[Lukasz Majewski]

Hi Przemyslaw,

> This commit enables the following configs:
> - CONFIG_USE_ARCH_MEMCPY
> - CONFIG_USE_ARCH_MEMSET
> This increases the performance of memcpy/memset
> and also reduces the boot time.
> 
> This was tested on Trats2.
> A quick test with trace. Boot time from start to main_loop() entry:
> - ~1527ms - before this change (arch memset enabled for .bss clear)
> - ~1384ms - after this change
> 
> Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
> Reviewed-by: Simon Glass <sjg@chromium.org>
> Cc: Minkyu Kang <mk7.kang@samsung.com>
> Cc: Akshay Saraswat <akshay.s@samsung.com>
> Cc: Simon Glass <sjg@chromium.org>
> Cc: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
> 
> ---
> Changes V3, V4, V5
> - none
> ---
>  include/configs/exynos-common.h | 3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/include/configs/exynos-common.h
> b/include/configs/exynos-common.h index 59676ae..87f8db0 100644
> --- a/include/configs/exynos-common.h
> +++ b/include/configs/exynos-common.h
> @@ -24,6 +24,9 @@
>  #define CONFIG_SKIP_LOWLEVEL_INIT
>  #define CONFIG_BOARD_EARLY_INIT_F
>  
> +#define CONFIG_USE_ARCH_MEMCPY
> +#define CONFIG_USE_ARCH_MEMSET
> +
>  /* Keep L2 Cache Disabled */
>  #define CONFIG_CMD_CACHE
>  

Acked-by: Lukasz Majewski <l.majewski@samsung.com>

-- 
Best regards,

Lukasz Majewski

Samsung R&D Institute Poland (SRPOL) | Linux Platform Group

[U-Boot] [PATCH v5 2/7] arm: relocation: clear .bss section with arch memset if defined

[Lukasz Majewski]

Hi Przemyslaw,

> For ARM architecture, enable the CONFIG_USE_ARCH_MEMSET/MEMCPY,
> will highly increase the memset/memcpy performance. This is able
> thanks to the ARM multiple register instructions.
> 
> Unfortunatelly the relocation is done without the cache enabled,
> so it takes some time, but zeroing the BSS memory takes much more
> longer, especially for the configs with big static buffers.
> 
> A quick test confirms, that the boot time improvement after using
> the arch memcpy for relocation has no significant meaning.
> The same test confirms that enable the memset for zeroing BSS,
> reduces the boot time.
> 
> So this patch enables the arch memset for zeroing the BSS after
> the relocation process. For ARM boards, this can be enabled
> in board configs by defining: 'CONFIG_USE_ARCH_MEMSET'.
> 
> This was tested on Trats2.
> A quick test with trace. Boot time from start to main_loop() entry:
> - ~1384ms - before this change
> -  ~888ms - after this change
> 
> Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
> Reviewed-by: Simon Glass <sjg@chromium.org>
> Cc: Albert Aribaud <albert.u.boot@aribaud.net>
> Cc: Tom Rini <trini@ti.com>
> ---
> Changes V3, V4, V5
> - none
> ---
>  arch/arm/lib/crt0.S | 10 +++++++++-
>  1 file changed, 9 insertions(+), 1 deletion(-)
> 
> diff --git a/arch/arm/lib/crt0.S b/arch/arm/lib/crt0.S
> index 22df3e5..fab3d2c 100644
> --- a/arch/arm/lib/crt0.S
> +++ b/arch/arm/lib/crt0.S
> @@ -115,14 +115,22 @@ here:
>  	bl	c_runtime_cpu_setup	/* we still call old
> routine here */ 
>  	ldr	r0, =__bss_start	/* this is
> auto-relocated! */
> -	ldr	r1, =__bss_end		/* this is
> auto-relocated! */ 
> +#ifdef CONFIG_USE_ARCH_MEMSET
> +	ldr	r3, =__bss_end		/* this is
> auto-relocated! */
> +	mov	r1, #0x00000000		/* prepare zero to
> clear BSS */ +
> +	subs	r2, r3, r0		/* r2 = memset len */
> +	bl	memset
> +#else
> +	ldr	r1, =__bss_end		/* this is
> auto-relocated! */ mov	r2, #0x00000000		/*
> prepare zero to clear BSS */ 
>  clbss_l:cmp	r0, r1			/* while not at end
> of BSS */ strlo	r2, [r0]		/* clear 32-bit BSS
> word */ addlo	r0, r0, #4		/* move to next */
>  	blo	clbss_l
> +#endif
>  
>  	bl coloured_LED_init
>  	bl red_led_on

Acked-by: Lukasz Majewski <l.majewski@samsung.com>

-- 
Best regards,

Lukasz Majewski

Samsung R&D Institute Poland (SRPOL) | Linux Platform Group

[U-Boot] [PATCH v5 3/7] dfu: mmc: file buffer: remove static allocation

[Lukasz Majewski]

Hi Przemyslaw,

> For writing files, DFU implementation requires the file buffer
> with the len at least of file size. For big files it requires
> the same big buffer.
> 
> Previously the file buffer was allocated as a static variable,
> so it was a part of U-Boot .bss section. For 32MiB len of buffer
> we have 32MiB of additional space, required for this section.
> 
> The .bss needs to be cleared after the relocation.
> This introduces an additional boot delay at every start, but usually
> the dfu feature is not required at the standard boot, so the buffer
> should be allocated only if required.
> 
> This patch removes the static allocation of this buffer,
> and alloc it with memalign after first call of function:
> - dfu_fill_entity_mmc()
> and the buffer is freed on dfu_free_entity() call.
> 
> This was tested on Trats2.
> A quick test with trace. Boot time from start to main_loop() entry:
> - ~888ms - before this change (arch memset enabled for .bss clear)
> - ~464ms - after this change
> 
> Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
> Reviewed-by: Simon Glass <sjg@chromium.org>
> Cc: Lukasz Majewski <l.majewski@samsung.com>
> Cc: Stephen Warren <swarren@nvidia.com>
> Cc: Pantelis Antoniou <panto@antoniou-consulting.com>
> Cc: Tom Rini <trini@ti.com>
> Cc: Marek Vasut <marek.vasut@gmail.com>
> 
> ---
> Changes V3, V4, V5
> - none
> ---
>  drivers/dfu/dfu_mmc.c | 25 ++++++++++++++++++++++---
>  1 file changed, 22 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/dfu/dfu_mmc.c b/drivers/dfu/dfu_mmc.c
> index 62d72fe..fd865e1 100644
> --- a/drivers/dfu/dfu_mmc.c
> +++ b/drivers/dfu/dfu_mmc.c
> @@ -16,8 +16,7 @@
>  #include <fat.h>
>  #include <mmc.h>
>  
> -static unsigned char __aligned(CONFIG_SYS_CACHELINE_SIZE)
> -
> dfu_file_buf[CONFIG_SYS_DFU_MAX_FILE_SIZE]; +static unsigned char
> *dfu_file_buf; static long dfu_file_buf_len;
>  
>  static int mmc_access_part(struct dfu_entity *dfu, struct mmc *mmc,
> int part) @@ -211,7 +210,7 @@ int dfu_flush_medium_mmc(struct
> dfu_entity *dfu) 
>  	if (dfu->layout != DFU_RAW_ADDR) {
>  		/* Do stuff here. */
> -		ret = mmc_file_op(DFU_OP_WRITE, dfu, &dfu_file_buf,
> +		ret = mmc_file_op(DFU_OP_WRITE, dfu, dfu_file_buf,
>  				&dfu_file_buf_len);
>  
>  		/* Now that we're done */
> @@ -263,6 +262,14 @@ int dfu_read_medium_mmc(struct dfu_entity *dfu,
> u64 offset, void *buf, return ret;
>  }
>  
> +void dfu_free_entity_mmc(struct dfu_entity *dfu)
> +{
> +	if (dfu_file_buf) {
> +		free(dfu_file_buf);
> +		dfu_file_buf = NULL;
> +	}
> +}
> +
>  /*
>   * @param s Parameter string containing space-separated arguments:
>   *	1st:
> @@ -370,6 +377,18 @@ int dfu_fill_entity_mmc(struct dfu_entity *dfu,
> char *devstr, char *s) dfu->write_medium = dfu_write_medium_mmc;
>  	dfu->flush_medium = dfu_flush_medium_mmc;
>  	dfu->inited = 0;
> +	dfu->free_entity = dfu_free_entity_mmc;
> +
> +	/* Check if file buffer is ready */
> +	if (!dfu_file_buf) {
> +		dfu_file_buf = memalign(CONFIG_SYS_CACHELINE_SIZE,
> +
> CONFIG_SYS_DFU_MAX_FILE_SIZE);
> +		if (!dfu_file_buf) {
> +			error("Could not memalign 0x%x bytes",
> +			      CONFIG_SYS_DFU_MAX_FILE_SIZE);
> +			return -ENOMEM;
> +		}
> +	}
>  
>  	return 0;
>  }

Acked-by: Lukasz Majewski <l.majewski@samsung.com>

-- 
Best regards,

Lukasz Majewski

Samsung R&D Institute Poland (SRPOL) | Linux Platform Group

[U-Boot] [PATCH v5 4/7] dlmalloc: do memset in malloc init as new default config

[Lukasz Majewski]

Hi Przemyslaw,

> This commit introduces new config: CONFIG_SYS_MALLOC_CLEAR_ON_INIT.
> 
> This config is an expert option and is enabled by default.
> 
> The all amount of memory reserved for the malloc, is by default set
> to zero in mem_malloc_init(). When the malloc reserved memory exceeds
> few MiB, then the boot process can slow down.
> 
> So disabling this config, is an expert option to reduce the boot time,
> and can be disabled by Kconfig.
> 
> Note:
> After disable this option, only calloc() will return the pointer
> to the zeroed memory area. Previously, without this option,
> the memory pointed to untouched malloc memory region, was filled
> with zeros. So it means, that code with malloc() calls should
> be reexamined.
> 
> Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
> Reviewed-by: Simon Glass <sjg@chromium.org>
> 
> ---
> Changes v3:
> - squash the commit with the Kconfig option
> 
> Changes v4:
> - adjust commit message for the new config
> 
> Changes v5:
> - change config name to CONFIG_SYS_MALLOC_CLEAR_ON_INIT
> ---
>  Kconfig           | 32 +++++++++++++++++++++++++-------
>  common/dlmalloc.c | 10 +++++++---
>  2 files changed, 32 insertions(+), 10 deletions(-)
> 
> diff --git a/Kconfig b/Kconfig
> index 75bab7f..d6c75d5 100644
> --- a/Kconfig
> +++ b/Kconfig
> @@ -76,13 +76,31 @@ config SYS_MALLOC_F_LEN
>  	  initial serial device and any others that are needed.
>  
>  menuconfig EXPERT
> -        bool "Configure standard U-Boot features (expert users)"
> -        help
> -          This option allows certain base U-Boot options and settings
> -          to be disabled or tweaked. This is for specialized
> -          environments which can tolerate a "non-standard" U-Boot.
> -          Only use this if you really know what you are doing.
> -
> +	bool "Configure standard U-Boot features (expert users)"
> +	default y
> +	help
> +	  This option allows certain base U-Boot options and settings
> +	  to be disabled or tweaked. This is for specialized
> +	  environments which can tolerate a "non-standard" U-Boot.
> +	  Only use this if you really know what you are doing.
> +
> +if EXPERT
> +	config SYS_MALLOC_CLEAR_ON_INIT
> +	bool "Init with zeros the memory reserved for malloc (slow)"
> +	default y
> +	help
> +	  This setting is enabled by default. The reserved malloc
> +	  memory is initialized with zeros, so first malloc calls
> +	  will return the pointer to the zeroed memory. But this
> +	  slows the boot time.
> +
> +	  It is recommended to disable it, when CONFIG_SYS_MALLOC_LEN
> +	  value, has more than few MiB, e.g. when uses bzip2 or bmp
> logo.
> +	  Then the boot time can be significantly reduced.
> +	  Warning:
> +	  When disabling this, please check if malloc calls, maybe
> +	  should be replaced by calloc - if expects zeroed memory.
> +endif
>  endmenu		# General setup
>  
>  menu "Boot images"
> diff --git a/common/dlmalloc.c b/common/dlmalloc.c
> index 6453ee9..b2ce063 100644
> --- a/common/dlmalloc.c
> +++ b/common/dlmalloc.c
> @@ -1535,9 +1535,9 @@ void mem_malloc_init(ulong start, ulong size)
>  
>  	debug("using memory %#lx-%#lx for malloc()\n",
> mem_malloc_start, mem_malloc_end);
> -
> -	memset((void *)mem_malloc_start, 0, size);
> -
> +#ifdef CONFIG_SYS_MALLOC_CLEAR_ON_INIT
> +	memset((void *)mem_malloc_start, 0x0, size);
> +#endif
>  	malloc_bin_reloc();
>  }
>  
> @@ -2948,10 +2948,12 @@ Void_t* cALLOc(n, elem_size) size_t n; size_t
> elem_size; 
>  
>    /* check if expand_top called, in which case don't need to clear */
> +#ifdef CONFIG_SYS_MALLOC_CLEAR_ON_INIT
>  #if MORECORE_CLEARS
>    mchunkptr oldtop = top;
>    INTERNAL_SIZE_T oldtopsize = chunksize(top);
>  #endif
> +#endif
>    Void_t* mem = mALLOc (sz);
>  
>    if ((long)n < 0) return NULL;
> @@ -2977,6 +2979,7 @@ Void_t* cALLOc(n, elem_size) size_t n; size_t
> elem_size; 
>      csz = chunksize(p);
>  
> +#ifdef CONFIG_SYS_MALLOC_CLEAR_ON_INIT
>  #if MORECORE_CLEARS
>      if (p == oldtop && csz > oldtopsize)
>      {
> @@ -2984,6 +2987,7 @@ Void_t* cALLOc(n, elem_size) size_t n; size_t
> elem_size; csz = oldtopsize;
>      }
>  #endif
> +#endif
>  
>      MALLOC_ZERO(mem, csz - SIZE_SZ);
>      return mem;

Acked-by: Lukasz Majewski <l.majewski@samsung.com>

-- 
Best regards,

Lukasz Majewski

Samsung R&D Institute Poland (SRPOL) | Linux Platform Group

[U-Boot] [PATCH v5 5/7] trats2: defconfig: disable memset at malloc init

[Lukasz Majewski]

Hi Przemyslaw,

> Reduce the boot time of Trats2 by disabling the memset
> at malloc init.
> 
> This was tested on Trats2.
> A quick test with trace. Boot time from start to main_loop() entry:
> - ~464ms - before this change (arch memset enabled for .bss clear)
> - ~341ms - after this change
> 
> Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
> Reviewed-by: Simon Glass <sjg@chromium.org>
> 
> ---
> Changes V3
> - none
> 
> Changes V4
> - trats2_defconfig: remove CONFIG_EXPERT
> - trats2_defconfig: disable CONFIG_SYS_MALLOC_INIT_DO_ZEROING
> 
> Changes v5:
> - update disabled config name to CONFIG_SYS_MALLOC_CLEAR_ON_INIT
> ---
>  configs/trats2_defconfig | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/configs/trats2_defconfig b/configs/trats2_defconfig
> index 1b98b73..9359706 100644
> --- a/configs/trats2_defconfig
> +++ b/configs/trats2_defconfig
> @@ -3,3 +3,4 @@ CONFIG_ARCH_EXYNOS=y
>  CONFIG_TARGET_TRATS2=y
>  CONFIG_OF_CONTROL=y
>  CONFIG_DEFAULT_DEVICE_TREE="exynos4412-trats2"
> +# CONFIG_SYS_MALLOC_CLEAR_ON_INIT is not set

Acked-by: Lukasz Majewski <l.majewski@samsung.com>

-- 
Best regards,

Lukasz Majewski

Samsung R&D Institute Poland (SRPOL) | Linux Platform Group

[U-Boot] [PATCH v5 6/7] odroid: defconfig: disable memset at malloc init

[Lukasz Majewski]

Hi Przemyslaw,

> Reduce the boot time of Odroid X2/U3 by disabling the memset
> at malloc init.
> 
> This was tested on Odroid X2.
> A quick test with checking gpio pin state using the oscilloscope.
> Boot time from start to bootcmd (change gpio state by memory write
> command):
> - ~228ms - before this change (arch memset enabled for .bss clear)
> - ~100ms - after this change
> 
> Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
> Reviewed-by: Simon Glass <sjg@chromium.org>
> 
> ---
> Changes V3
> - update commit head
> 
> Changes V4
> - odroid_defconfig: remove CONFIG_EXPERT
> - odroid_defconfig: disable CONFIG_SYS_MALLOC_INIT_DO_ZEROING
> 
> Changes v5
> - update disabled config name to CONFIG_SYS_MALLOC_CLEAR_ON_INIT
> ---
>  configs/odroid_defconfig | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/configs/odroid_defconfig b/configs/odroid_defconfig
> index a842837..aac9f5a 100644
> --- a/configs/odroid_defconfig
> +++ b/configs/odroid_defconfig
> @@ -3,3 +3,4 @@ CONFIG_ARCH_EXYNOS=y
>  CONFIG_TARGET_ODROID=y
>  CONFIG_OF_CONTROL=y
>  CONFIG_DEFAULT_DEVICE_TREE="exynos4412-odroid"
> +# CONFIG_SYS_MALLOC_CLEAR_ON_INIT is not set

Acked-by: Lukasz Majewski <l.majewski@samsung.com>

-- 
Best regards,

Lukasz Majewski

Samsung R&D Institute Poland (SRPOL) | Linux Platform Group

[U-Boot] [PATCH v5 7/7] odroid-xu3: defconfig: disable memset at malloc init

[Lukasz Majewski]

Hi Przemyslaw,

> Reduce the boot time of Odroid XU3 by disabling the memset
> at malloc init.
> 
> Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
> ---
> Changes v5
> - new commit
> 
> ---
>  configs/odroid-xu3_defconfig | 2 ++
>  1 file changed, 2 insertions(+)
> 
> diff --git a/configs/odroid-xu3_defconfig
> b/configs/odroid-xu3_defconfig index 74aa0cf..0fb4623 100644
> --- a/configs/odroid-xu3_defconfig
> +++ b/configs/odroid-xu3_defconfig
> @@ -2,3 +2,5 @@ CONFIG_ARM=y
>  CONFIG_ARCH_EXYNOS=y
>  CONFIG_TARGET_ODROID_XU3=y
>  CONFIG_DEFAULT_DEVICE_TREE="exynos5422-odroidxu3"
> +# CONFIG_SYS_MALLOC_CLEAR_ON_INIT is not set
> +

Acked-by: Lukasz Majewski <l.majewski@samsung.com>

-- 
Best regards,

Lukasz Majewski

Samsung R&D Institute Poland (SRPOL) | Linux Platform Group

[U-Boot] [PATCH v5 0/7] arm: a few steps to improve boot time

[Przemyslaw Marczak]

Hello Tom,

On 02/24/2015 11:38 AM, Przemyslaw Marczak wrote:
> This patchset reduces the boot time for ARM architecture,
> Exynos boards, and (ARM) boards with DFU enabled.
>
> For tested Trats2 and Odroid X2/XU3 devices, this was done in four steps.
>
> 1. Enable the arch memcpy and memset - ARCH specific
>
> 2. Enable arch memset for .bss clear - ARCH specific
>
> 3. Reduce the .bss section small as possible by: (board specific)
>    - remove static dfu buffer (32MiB in .bss - Trats2), and use malloc
>
> 4. Skip zeroing the memory reserved for malloc at malloc init.
>     For Trats2 it was 80MiB of memory.
>
> The .bss section will grow if we have a lot of static variables.
> This section is cleared before jump to the relocated U-Boot,
> and it was done word by word. To reduce the time for this step,
> we can enable arch memset, which uses multiple ARM registers.
>
> For configs with DFU enabled, we can find the dfu buffer in this section,
> which has at least 8MB (32MB for Trats2). This is a lot of useless data,
> which is not required for standard boot. So this buffer should be dynamic
> allocated.
>
> So, actually the all was about unnecessary operations on 'big' data.
>
> Przemyslaw Marczak (7):
>    exynos: config: enable arch memcpy and arch memset
>    arm: relocation: clear .bss section with arch memset if defined
>    dfu: mmc: file buffer: remove static allocation
>    dlmalloc: do memset in malloc init as new default config
>    trats2: defconfig: disable memset at malloc init
>    odroid: defconfig: disable memset at malloc init
>    odroid-xu3: defconfig: disable memset at malloc init
>
>   Kconfig                         | 32 +++++++++++++++++++++++++-------
>   arch/arm/lib/crt0.S             | 10 +++++++++-
>   common/dlmalloc.c               | 10 +++++++---
>   configs/odroid-xu3_defconfig    |  2 ++
>   configs/odroid_defconfig        |  1 +
>   configs/trats2_defconfig        |  1 +
>   drivers/dfu/dfu_mmc.c           | 25 ++++++++++++++++++++++---
>   include/configs/exynos-common.h |  3 +++
>   8 files changed, 70 insertions(+), 14 deletions(-)
>

If you will merge this patchset, please take also this one patch:

http://patchwork.ozlabs.org/patch/440623/

Best regards,
-- 
Przemyslaw Marczak
Samsung R&D Institute Poland
Samsung Electronics
p.marczak at samsung.com

[U-Boot] [PATCH v6 00/10] arm: a few steps to improve boot time

[Przemyslaw Marczak]

This patchset reduces the boot time for ARM architecture,
Exynos boards, and (ARM) boards with DFU enabled.

For tested Trats2 and Odroid X2/XU3 devices, this was done in four steps.

1. Enable the arch memcpy and memset - ARCH specific

2. Enable arch memset for .bss clear - ARCH specific

3. Reduce the .bss section small as possible by: (board specific)
  - remove static dfu buffer (32MiB in .bss - Trats2), and use malloc

4. Skip zeroing the memory reserved for malloc at malloc init.
   For Trats2 it was 80MiB of memory.

The .bss section will grow if we have a lot of static variables.
This section is cleared before jump to the relocated U-Boot,
and it was done word by word. To reduce the time for this step,
we can enable arch memset, which uses multiple ARM registers.

For configs with DFU enabled, we can find the dfu buffer in this section,
which has at least 8MB (32MB for Trats2). This is a lot of useless data,
which is not required for standard boot. So this buffer should be dynamic
allocated.

Changes v6:
This version fixes merge conflict on latest master, which is:
"Prepare v2015.04-rc3"
It also extends malloc pool, for some board configs.

Przemyslaw Marczak (10):
  exynos: config: enable arch memcpy and arch memset
  arm: relocation: clear .bss section with arch memset if defined
  dfu: mmc: file buffer: remove static allocation
  dlmalloc: do memset in malloc init as new default config
  trats2: defconfig: disable memset at malloc init
  odroid: defconfig: disable memset at malloc init
  odroid-xu3: defconfig: disable memset at malloc init
  zynq-common: increase malloc pool len by dfu mmc file buffer size
  ti-armv7-common: increase malloc pool len by dfu mmc file buffer size
  tegra-common: increase malloc pool len by dfu mmc file buffer size

 Kconfig                           | 32 +++++++++++++++++++++++++-------
 arch/arm/lib/crt0.S               | 10 +++++++++-
 common/dlmalloc.c                 | 10 +++++++---
 configs/odroid-xu3_defconfig      |  2 ++
 configs/odroid_defconfig          |  2 +-
 configs/trats2_defconfig          |  1 +
 drivers/dfu/dfu_mmc.c             | 25 ++++++++++++++++++++++---
 include/configs/exynos-common.h   |  3 +++
 include/configs/tegra-common.h    |  5 +++++
 include/configs/ti_armv7_common.h |  6 +++++-
 include/configs/zynq-common.h     |  2 +-
 11 files changed, 81 insertions(+), 17 deletions(-)

-- 
1.9.1

[U-Boot] [PATCH v6 01/10] exynos: config: enable arch memcpy and arch memset

[Przemyslaw Marczak]

This commit enables the following configs:
- CONFIG_USE_ARCH_MEMCPY
- CONFIG_USE_ARCH_MEMSET
This increases the performance of memcpy/memset
and also reduces the boot time.

This was tested on Trats2.
A quick test with trace. Boot time from start to main_loop() entry:
- ~1527ms - before this change (arch memset enabled for .bss clear)
- ~1384ms - after this change

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Cc: Minkyu Kang <mk7.kang@samsung.com>
Cc: Akshay Saraswat <akshay.s@samsung.com>
Cc: Simon Glass <sjg@chromium.org>
Cc: Sjoerd Simons <sjoerd.simons@collabora.co.uk>

---
Changes V3, V4, V5, V6
- none
---
 include/configs/exynos-common.h | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/include/configs/exynos-common.h b/include/configs/exynos-common.h
index 59676ae..87f8db0 100644
--- a/include/configs/exynos-common.h
+++ b/include/configs/exynos-common.h
@@ -24,6 +24,9 @@
 #define CONFIG_SKIP_LOWLEVEL_INIT
 #define CONFIG_BOARD_EARLY_INIT_F
 
+#define CONFIG_USE_ARCH_MEMCPY
+#define CONFIG_USE_ARCH_MEMSET
+
 /* Keep L2 Cache Disabled */
 #define CONFIG_CMD_CACHE
 
-- 
1.9.1

[U-Boot] [PATCH v6 02/10] arm: relocation: clear .bss section with arch memset if defined

[Przemyslaw Marczak]

For ARM architecture, enable the CONFIG_USE_ARCH_MEMSET/MEMCPY,
will highly increase the memset/memcpy performance. This is able
thanks to the ARM multiple register instructions.

Unfortunatelly the relocation is done without the cache enabled,
so it takes some time, but zeroing the BSS memory takes much more
longer, especially for the configs with big static buffers.

A quick test confirms, that the boot time improvement after using
the arch memcpy for relocation has no significant meaning.
The same test confirms that enable the memset for zeroing BSS,
reduces the boot time.

So this patch enables the arch memset for zeroing the BSS after
the relocation process. For ARM boards, this can be enabled
in board configs by defining: 'CONFIG_USE_ARCH_MEMSET'.

This was tested on Trats2.
A quick test with trace. Boot time from start to main_loop() entry:
- ~1384ms - before this change
-  ~888ms - after this change

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Cc: Albert Aribaud <albert.u.boot@aribaud.net>
Cc: Tom Rini <trini@konsulko.com>
---
Changes V3, V4, V5, V6
- none
---
 arch/arm/lib/crt0.S | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

diff --git a/arch/arm/lib/crt0.S b/arch/arm/lib/crt0.S
index 22df3e5..fab3d2c 100644
--- a/arch/arm/lib/crt0.S
+++ b/arch/arm/lib/crt0.S
@@ -115,14 +115,22 @@ here:
 	bl	c_runtime_cpu_setup	/* we still call old routine here */
 
 	ldr	r0, =__bss_start	/* this is auto-relocated! */
-	ldr	r1, =__bss_end		/* this is auto-relocated! */
 
+#ifdef CONFIG_USE_ARCH_MEMSET
+	ldr	r3, =__bss_end		/* this is auto-relocated! */
+	mov	r1, #0x00000000		/* prepare zero to clear BSS */
+
+	subs	r2, r3, r0		/* r2 = memset len */
+	bl	memset
+#else
+	ldr	r1, =__bss_end		/* this is auto-relocated! */
 	mov	r2, #0x00000000		/* prepare zero to clear BSS */
 
 clbss_l:cmp	r0, r1			/* while not at end of BSS */
 	strlo	r2, [r0]		/* clear 32-bit BSS word */
 	addlo	r0, r0, #4		/* move to next */
 	blo	clbss_l
+#endif
 
 	bl coloured_LED_init
 	bl red_led_on
-- 
1.9.1

[U-Boot] [PATCH v6 03/10] dfu: mmc: file buffer: remove static allocation

[Przemyslaw Marczak]

For writing files, DFU implementation requires the file buffer
with the len at least of file size. For big files it requires
the same big buffer.

Previously the file buffer was allocated as a static variable,
so it was a part of U-Boot .bss section. For 32MiB len of buffer
we have 32MiB of additional space, required for this section.

The .bss needs to be cleared after the relocation.
This introduces an additional boot delay at every start, but usually
the dfu feature is not required at the standard boot, so the buffer
should be allocated only if required.

This patch removes the static allocation of this buffer,
and alloc it with memalign after first call of function:
- dfu_fill_entity_mmc()
and the buffer is freed on dfu_free_entity() call.

This was tested on Trats2.
A quick test with trace. Boot time from start to main_loop() entry:
- ~888ms - before this change (arch memset enabled for .bss clear)
- ~464ms - after this change

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Cc: Lukasz Majewski <l.majewski@samsung.com>
Cc: Stephen Warren <swarren@nvidia.com>
Cc: Pantelis Antoniou <panto@antoniou-consulting.com>
Cc: Tom Rini <trini@konsulko.com>
Cc: Marek Vasut <marek.vasut@gmail.com>

---
Changes V3, V4, V5, V6
- none
---
 drivers/dfu/dfu_mmc.c | 25 ++++++++++++++++++++++---
 1 file changed, 22 insertions(+), 3 deletions(-)

diff --git a/drivers/dfu/dfu_mmc.c b/drivers/dfu/dfu_mmc.c
index 62d72fe..fd865e1 100644
--- a/drivers/dfu/dfu_mmc.c
+++ b/drivers/dfu/dfu_mmc.c
@@ -16,8 +16,7 @@
 #include <fat.h>
 #include <mmc.h>
 
-static unsigned char __aligned(CONFIG_SYS_CACHELINE_SIZE)
-				dfu_file_buf[CONFIG_SYS_DFU_MAX_FILE_SIZE];
+static unsigned char *dfu_file_buf;
 static long dfu_file_buf_len;
 
 static int mmc_access_part(struct dfu_entity *dfu, struct mmc *mmc, int part)
@@ -211,7 +210,7 @@ int dfu_flush_medium_mmc(struct dfu_entity *dfu)
 
 	if (dfu->layout != DFU_RAW_ADDR) {
 		/* Do stuff here. */
-		ret = mmc_file_op(DFU_OP_WRITE, dfu, &dfu_file_buf,
+		ret = mmc_file_op(DFU_OP_WRITE, dfu, dfu_file_buf,
 				&dfu_file_buf_len);
 
 		/* Now that we're done */
@@ -263,6 +262,14 @@ int dfu_read_medium_mmc(struct dfu_entity *dfu, u64 offset, void *buf,
 	return ret;
 }
 
+void dfu_free_entity_mmc(struct dfu_entity *dfu)
+{
+	if (dfu_file_buf) {
+		free(dfu_file_buf);
+		dfu_file_buf = NULL;
+	}
+}
+
 /*
  * @param s Parameter string containing space-separated arguments:
  *	1st:
@@ -370,6 +377,18 @@ int dfu_fill_entity_mmc(struct dfu_entity *dfu, char *devstr, char *s)
 	dfu->write_medium = dfu_write_medium_mmc;
 	dfu->flush_medium = dfu_flush_medium_mmc;
 	dfu->inited = 0;
+	dfu->free_entity = dfu_free_entity_mmc;
+
+	/* Check if file buffer is ready */
+	if (!dfu_file_buf) {
+		dfu_file_buf = memalign(CONFIG_SYS_CACHELINE_SIZE,
+					CONFIG_SYS_DFU_MAX_FILE_SIZE);
+		if (!dfu_file_buf) {
+			error("Could not memalign 0x%x bytes",
+			      CONFIG_SYS_DFU_MAX_FILE_SIZE);
+			return -ENOMEM;
+		}
+	}
 
 	return 0;
 }
-- 
1.9.1

[U-Boot] [PATCH v6 04/10] dlmalloc: do memset in malloc init as new default config

[Przemyslaw Marczak]

This commit introduces new config: CONFIG_SYS_MALLOC_CLEAR_ON_INIT.

This config is an expert option and is enabled by default.

The all amount of memory reserved for the malloc, is by default set
to zero in mem_malloc_init(). When the malloc reserved memory exceeds
few MiB, then the boot process can slow down.

So disabling this config, is an expert option to reduce the boot time,
and can be disabled by Kconfig.

Note:
After disable this option, only calloc() will return the pointer
to the zeroed memory area. Previously, without this option,
the memory pointed to untouched malloc memory region, was filled
with zeros. So it means, that code with malloc() calls should
be reexamined.

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Reviewed-by: Simon Glass <sjg@chromium.org>

---
Changes v3:
- squash the commit with the Kconfig option

Changes v4:
- adjust commit message for the new config

Changes v5:
- change config name to CONFIG_SYS_MALLOC_CLEAR_ON_INIT

Changes v6:
- none
---
 Kconfig           | 32 +++++++++++++++++++++++++-------
 common/dlmalloc.c | 10 +++++++---
 2 files changed, 32 insertions(+), 10 deletions(-)

diff --git a/Kconfig b/Kconfig
index 91a0618..4cbbfc2 100644
--- a/Kconfig
+++ b/Kconfig
@@ -72,13 +72,31 @@ config SYS_MALLOC_F_LEN
 	  initial serial device and any others that are needed.
 
 menuconfig EXPERT
-        bool "Configure standard U-Boot features (expert users)"
-        help
-          This option allows certain base U-Boot options and settings
-          to be disabled or tweaked. This is for specialized
-          environments which can tolerate a "non-standard" U-Boot.
-          Only use this if you really know what you are doing.
-
+	bool "Configure standard U-Boot features (expert users)"
+	default y
+	help
+	  This option allows certain base U-Boot options and settings
+	  to be disabled or tweaked. This is for specialized
+	  environments which can tolerate a "non-standard" U-Boot.
+	  Only use this if you really know what you are doing.
+
+if EXPERT
+	config SYS_MALLOC_CLEAR_ON_INIT
+	bool "Init with zeros the memory reserved for malloc (slow)"
+	default y
+	help
+	  This setting is enabled by default. The reserved malloc
+	  memory is initialized with zeros, so first malloc calls
+	  will return the pointer to the zeroed memory. But this
+	  slows the boot time.
+
+	  It is recommended to disable it, when CONFIG_SYS_MALLOC_LEN
+	  value, has more than few MiB, e.g. when uses bzip2 or bmp logo.
+	  Then the boot time can be significantly reduced.
+	  Warning:
+	  When disabling this, please check if malloc calls, maybe
+	  should be replaced by calloc - if expects zeroed memory.
+endif
 endmenu		# General setup
 
 menu "Boot images"
diff --git a/common/dlmalloc.c b/common/dlmalloc.c
index 6453ee9..b2ce063 100644
--- a/common/dlmalloc.c
+++ b/common/dlmalloc.c
@@ -1535,9 +1535,9 @@ void mem_malloc_init(ulong start, ulong size)
 
 	debug("using memory %#lx-%#lx for malloc()\n", mem_malloc_start,
 	      mem_malloc_end);
-
-	memset((void *)mem_malloc_start, 0, size);
-
+#ifdef CONFIG_SYS_MALLOC_CLEAR_ON_INIT
+	memset((void *)mem_malloc_start, 0x0, size);
+#endif
 	malloc_bin_reloc();
 }
 
@@ -2948,10 +2948,12 @@ Void_t* cALLOc(n, elem_size) size_t n; size_t elem_size;
 
 
   /* check if expand_top called, in which case don't need to clear */
+#ifdef CONFIG_SYS_MALLOC_CLEAR_ON_INIT
 #if MORECORE_CLEARS
   mchunkptr oldtop = top;
   INTERNAL_SIZE_T oldtopsize = chunksize(top);
 #endif
+#endif
   Void_t* mem = mALLOc (sz);
 
   if ((long)n < 0) return NULL;
@@ -2977,6 +2979,7 @@ Void_t* cALLOc(n, elem_size) size_t n; size_t elem_size;
 
     csz = chunksize(p);
 
+#ifdef CONFIG_SYS_MALLOC_CLEAR_ON_INIT
 #if MORECORE_CLEARS
     if (p == oldtop && csz > oldtopsize)
     {
@@ -2984,6 +2987,7 @@ Void_t* cALLOc(n, elem_size) size_t n; size_t elem_size;
       csz = oldtopsize;
     }
 #endif
+#endif
 
     MALLOC_ZERO(mem, csz - SIZE_SZ);
     return mem;
-- 
1.9.1

[U-Boot] [PATCH v6 05/10] trats2: defconfig: disable memset at malloc init

[Przemyslaw Marczak]

Reduce the boot time of Trats2 by disabling the memset
at malloc init.

This was tested on Trats2.
A quick test with trace. Boot time from start to main_loop() entry:
- ~464ms - before this change (arch memset enabled for .bss clear)
- ~341ms - after this change

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Reviewed-by: Simon Glass <sjg@chromium.org>

---
Changes V3
- none

Changes V4
- trats2_defconfig: remove CONFIG_EXPERT
- trats2_defconfig: disable CONFIG_SYS_MALLOC_INIT_DO_ZEROING

Changes v5:
- update disabled config name to CONFIG_SYS_MALLOC_CLEAR_ON_INIT

Changes v6:
- none
---
 configs/trats2_defconfig | 1 +
 1 file changed, 1 insertion(+)

diff --git a/configs/trats2_defconfig b/configs/trats2_defconfig
index 1b98b73..9359706 100644
--- a/configs/trats2_defconfig
+++ b/configs/trats2_defconfig
@@ -3,3 +3,4 @@ CONFIG_ARCH_EXYNOS=y
 CONFIG_TARGET_TRATS2=y
 CONFIG_OF_CONTROL=y
 CONFIG_DEFAULT_DEVICE_TREE="exynos4412-trats2"
+# CONFIG_SYS_MALLOC_CLEAR_ON_INIT is not set
-- 
1.9.1

[U-Boot] [PATCH v6 06/10] odroid: defconfig: disable memset at malloc init

[Przemyslaw Marczak]

Reduce the boot time of Odroid X2/U3 by disabling the memset
at malloc init.

This was tested on Odroid X2.
A quick test with checking gpio pin state using the oscilloscope.
Boot time from start to bootcmd (change gpio state by memory write command):
- ~228ms - before this change (arch memset enabled for .bss clear)
- ~100ms - after this change

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Reviewed-by: Simon Glass <sjg@chromium.org>

---
Changes V3
- update commit head

Changes V4
- odroid_defconfig: remove CONFIG_EXPERT
- odroid_defconfig: disable CONFIG_SYS_MALLOC_INIT_DO_ZEROING

Changes v5
- update disabled config name to CONFIG_SYS_MALLOC_CLEAR_ON_INIT

Changes v6:
- fix merge conflict with master
---
 configs/odroid_defconfig | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/configs/odroid_defconfig b/configs/odroid_defconfig
index 816a3fa..cfb29e0 100644
--- a/configs/odroid_defconfig
+++ b/configs/odroid_defconfig
@@ -2,6 +2,6 @@ CONFIG_ARM=y
 CONFIG_ARCH_EXYNOS=y
 CONFIG_TARGET_ODROID=y
 CONFIG_OF_CONTROL=y
-CONFIG_DEFAULT_DEVICE_TREE="exynos4412-odroid"
 CONFIG_DM_I2C=y
 CONFIG_DM_I2C_COMPAT=y
+# CONFIG_SYS_MALLOC_CLEAR_ON_INIT is not set
-- 
1.9.1

[U-Boot] [PATCH v6 07/10] odroid-xu3: defconfig: disable memset at malloc init

[Przemyslaw Marczak]

Reduce the boot time of Odroid XU3 by disabling the memset
at malloc init.

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
---
Changes v5
- new commit

Changes v6
- none

---
 configs/odroid-xu3_defconfig | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/configs/odroid-xu3_defconfig b/configs/odroid-xu3_defconfig
index 74aa0cf..0fb4623 100644
--- a/configs/odroid-xu3_defconfig
+++ b/configs/odroid-xu3_defconfig
@@ -2,3 +2,5 @@ CONFIG_ARM=y
 CONFIG_ARCH_EXYNOS=y
 CONFIG_TARGET_ODROID_XU3=y
 CONFIG_DEFAULT_DEVICE_TREE="exynos5422-odroidxu3"
+# CONFIG_SYS_MALLOC_CLEAR_ON_INIT is not set
+
-- 
1.9.1

[U-Boot] [PATCH v6 08/10] zynq-common: increase malloc pool len by dfu mmc file buffer size

[Przemyslaw Marczak]

The dfu mmc file buffer, which was static, now is allocated
by memalign(), so the malloc pool len should be also increased.

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Cc: Jagannadha Sutradharudu Teki <jaganna@xilinx.com>
Cc: Michal Simek <monstr@monstr.eu>
---
Changes v6:
- new commit
---
 include/configs/zynq-common.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/include/configs/zynq-common.h b/include/configs/zynq-common.h
index 864528a..485babd 100644
--- a/include/configs/zynq-common.h
+++ b/include/configs/zynq-common.h
@@ -255,7 +255,7 @@
 #define CONFIG_SYS_MEMTEST_START	CONFIG_SYS_SDRAM_BASE
 #define CONFIG_SYS_MEMTEST_END		(CONFIG_SYS_SDRAM_BASE + 0x1000)
 
-#define CONFIG_SYS_MALLOC_LEN		0xC00000
+#define CONFIG_SYS_MALLOC_LEN		0x1400000
 #define CONFIG_SYS_INIT_RAM_ADDR	CONFIG_SYS_SDRAM_BASE
 #define CONFIG_SYS_INIT_RAM_SIZE	CONFIG_SYS_MALLOC_LEN
 #define CONFIG_SYS_INIT_SP_ADDR		(CONFIG_SYS_INIT_RAM_ADDR + \
-- 
1.9.1

[U-Boot] [PATCH v6 09/10] ti-armv7-common: increase malloc pool len by dfu mmc file buffer size

[Przemyslaw Marczak]

The dfu mmc file buffer, which was static, now is allocated
by memalign(), so the malloc pool len should be also increased.

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Cc: Tom Rini <trini@konsulko.com>
---
Changes v6:
- new commit
---
 include/configs/ti_armv7_common.h | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/include/configs/ti_armv7_common.h b/include/configs/ti_armv7_common.h
index 2bd1164..f832ab3 100644
--- a/include/configs/ti_armv7_common.h
+++ b/include/configs/ti_armv7_common.h
@@ -127,7 +127,11 @@
  * we are on so we do not need to rely on the command prompt.  We set a
  * console baudrate of 115200 and use the default baud rate table.
  */
-#define CONFIG_SYS_MALLOC_LEN		(16 << 20)
+#ifdef CONFIG_DFU_MMC
+#define CONFIG_SYS_MALLOC_LEN	((16 << 20) + CONFIG_SYS_DFU_DATA_BUF_SIZE)
+#else
+#define CONFIG_SYS_MALLOC_LEN	(16 << 20)
+#endif
 #define CONFIG_SYS_HUSH_PARSER
 #define CONFIG_SYS_PROMPT		"U-Boot# "
 #define CONFIG_SYS_CONSOLE_INFO_QUIET
-- 
1.9.1

[U-Boot] [PATCH v6 10/10] tegra-common: increase malloc pool len by dfu mmc file buffer size

[Przemyslaw Marczak]

The dfu mmc file buffer, which was static, now is allocated
by memalign(), so the malloc pool len should be also increased.

Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Cc: Tom Warren <twarren.nvidia@gmail.com>

---
Changes v6:
- new commit
---
 include/configs/tegra-common.h | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/include/configs/tegra-common.h b/include/configs/tegra-common.h
index 005fc6a..3e86f6c 100644
--- a/include/configs/tegra-common.h
+++ b/include/configs/tegra-common.h
@@ -36,7 +36,12 @@
 /*
  * Size of malloc() pool
  */
+#ifdef CONFIG_DFU_MMC
+#define CONFIG_SYS_MALLOC_LEN		((4 << 20) + \
+					CONFIG_SYS_DFU_DATA_BUF_SIZE)
+#else
 #define CONFIG_SYS_MALLOC_LEN		(4 << 20)	/* 4MB  */
+#endif
 
 #define CONFIG_SYS_NONCACHED_MEMORY	(1 << 20)       /* 1 MiB */
 
-- 
1.9.1

[U-Boot] [PATCH v6 08/10] zynq-common: increase malloc pool len by dfu mmc file buffer size

[Michal Simek]

On 03/04/2015 02:01 PM, Przemyslaw Marczak wrote:
> The dfu mmc file buffer, which was static, now is allocated
> by memalign(), so the malloc pool len should be also increased.
> 
> Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
> Cc: Jagannadha Sutradharudu Teki <jaganna@xilinx.com>
> Cc: Michal Simek <monstr@monstr.eu>
> ---
> Changes v6:
> - new commit
> ---
>  include/configs/zynq-common.h | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/include/configs/zynq-common.h b/include/configs/zynq-common.h
> index 864528a..485babd 100644
> --- a/include/configs/zynq-common.h
> +++ b/include/configs/zynq-common.h
> @@ -255,7 +255,7 @@
>  #define CONFIG_SYS_MEMTEST_START	CONFIG_SYS_SDRAM_BASE
>  #define CONFIG_SYS_MEMTEST_END		(CONFIG_SYS_SDRAM_BASE + 0x1000)
>  
> -#define CONFIG_SYS_MALLOC_LEN		0xC00000
> +#define CONFIG_SYS_MALLOC_LEN		0x1400000
>  #define CONFIG_SYS_INIT_RAM_ADDR	CONFIG_SYS_SDRAM_BASE
>  #define CONFIG_SYS_INIT_RAM_SIZE	CONFIG_SYS_MALLOC_LEN
>  #define CONFIG_SYS_INIT_SP_ADDR		(CONFIG_SYS_INIT_RAM_ADDR + \
> 

Acked-by: Michal Simek <michal.simek@xilinx.com>

Thanks,
Michal

-- 
Michal Simek, Ing. (M.Eng), OpenPGP -> KeyID: FE3D1F91
w: www.monstr.eu p: +42-0-721842854
Maintainer of Linux kernel - Microblaze cpu - http://www.monstr.eu/fdt/
Maintainer of Linux kernel - Xilinx Zynq ARM architecture
Microblaze U-BOOT custodian and responsible for u-boot arm zynq platform


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: OpenPGP digital signature
URL: <http://lists.denx.de/pipermail/u-boot/attachments/20150304/80146e87/attachment.sig>

[U-Boot] [U-Boot, v6, 01/10] exynos: config: enable arch memcpy and arch memset

[Tom Rini]

On Wed, Mar 04, 2015 at 02:01:21PM +0100, Przemyslaw Marczak wrote:

> This commit enables the following configs:
> - CONFIG_USE_ARCH_MEMCPY
> - CONFIG_USE_ARCH_MEMSET
> This increases the performance of memcpy/memset
> and also reduces the boot time.
> 
> This was tested on Trats2.
> A quick test with trace. Boot time from start to main_loop() entry:
> - ~1527ms - before this change (arch memset enabled for .bss clear)
> - ~1384ms - after this change
> 
> Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
> Reviewed-by: Simon Glass <sjg@chromium.org>
> Cc: Minkyu Kang <mk7.kang@samsung.com>
> Cc: Akshay Saraswat <akshay.s@samsung.com>
> Cc: Simon Glass <sjg@chromium.org>
> Cc: Sjoerd Simons <sjoerd.simons@collabora.co.uk>

Applied to u-boot/master, thanks!

-- 
Tom
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.denx.de/pipermail/u-boot/attachments/20150309/f5c27ab9/attachment.sig>

[U-Boot] [U-Boot, v6, 02/10] arm: relocation: clear .bss section with arch memset if defined

[Tom Rini]

On Wed, Mar 04, 2015 at 02:01:22PM +0100, Przemyslaw Marczak wrote:

> For ARM architecture, enable the CONFIG_USE_ARCH_MEMSET/MEMCPY,
> will highly increase the memset/memcpy performance. This is able
> thanks to the ARM multiple register instructions.
> 
> Unfortunatelly the relocation is done without the cache enabled,
> so it takes some time, but zeroing the BSS memory takes much more
> longer, especially for the configs with big static buffers.
> 
> A quick test confirms, that the boot time improvement after using
> the arch memcpy for relocation has no significant meaning.
> The same test confirms that enable the memset for zeroing BSS,
> reduces the boot time.
> 
> So this patch enables the arch memset for zeroing the BSS after
> the relocation process. For ARM boards, this can be enabled
> in board configs by defining: 'CONFIG_USE_ARCH_MEMSET'.
> 
> This was tested on Trats2.
> A quick test with trace. Boot time from start to main_loop() entry:
> - ~1384ms - before this change
> -  ~888ms - after this change
> 
> Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
> Reviewed-by: Simon Glass <sjg@chromium.org>
> Cc: Albert Aribaud <albert.u.boot@aribaud.net>
> Cc: Tom Rini <trini@konsulko.com>

Applied to u-boot/master, thanks!

-- 
Tom
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.denx.de/pipermail/u-boot/attachments/20150309/87c16e05/attachment.sig>

[U-Boot] [U-Boot, v6, 03/10] dfu: mmc: file buffer: remove static allocation

[Tom Rini]

On Wed, Mar 04, 2015 at 02:01:23PM +0100, Przemyslaw Marczak wrote:

> For writing files, DFU implementation requires the file buffer
> with the len at least of file size. For big files it requires
> the same big buffer.
> 
> Previously the file buffer was allocated as a static variable,
> so it was a part of U-Boot .bss section. For 32MiB len of buffer
> we have 32MiB of additional space, required for this section.
> 
> The .bss needs to be cleared after the relocation.
> This introduces an additional boot delay at every start, but usually
> the dfu feature is not required at the standard boot, so the buffer
> should be allocated only if required.
> 
> This patch removes the static allocation of this buffer,
> and alloc it with memalign after first call of function:
> - dfu_fill_entity_mmc()
> and the buffer is freed on dfu_free_entity() call.
> 
> This was tested on Trats2.
> A quick test with trace. Boot time from start to main_loop() entry:
> - ~888ms - before this change (arch memset enabled for .bss clear)
> - ~464ms - after this change
> 
> Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
> Reviewed-by: Simon Glass <sjg@chromium.org>
> Cc: Lukasz Majewski <l.majewski@samsung.com>
> Cc: Stephen Warren <swarren@nvidia.com>
> Cc: Pantelis Antoniou <panto@antoniou-consulting.com>
> Cc: Tom Rini <trini@konsulko.com>
> Cc: Marek Vasut <marek.vasut@gmail.com>

Applied to u-boot/master, thanks!

-- 
Tom
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.denx.de/pipermail/u-boot/attachments/20150309/63e137c2/attachment.sig>

[U-Boot] [U-Boot, v6, 04/10] dlmalloc: do memset in malloc init as new default config

[Tom Rini]

On Wed, Mar 04, 2015 at 02:01:24PM +0100, Przemyslaw Marczak wrote:

> This commit introduces new config: CONFIG_SYS_MALLOC_CLEAR_ON_INIT.
> 
> This config is an expert option and is enabled by default.
> 
> The all amount of memory reserved for the malloc, is by default set
> to zero in mem_malloc_init(). When the malloc reserved memory exceeds
> few MiB, then the boot process can slow down.
> 
> So disabling this config, is an expert option to reduce the boot time,
> and can be disabled by Kconfig.
> 
> Note:
> After disable this option, only calloc() will return the pointer
> to the zeroed memory area. Previously, without this option,
> the memory pointed to untouched malloc memory region, was filled
> with zeros. So it means, that code with malloc() calls should
> be reexamined.
> 
> Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
> Reviewed-by: Simon Glass <sjg@chromium.org>

Applied to u-boot/master, thanks!

-- 
Tom
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.denx.de/pipermail/u-boot/attachments/20150309/027e8ad8/attachment.sig>

[U-Boot] [U-Boot, v6, 05/10] trats2: defconfig: disable memset at malloc init

[Tom Rini]

On Wed, Mar 04, 2015 at 02:01:25PM +0100, Przemyslaw Marczak wrote:

> Reduce the boot time of Trats2 by disabling the memset
> at malloc init.
> 
> This was tested on Trats2.
> A quick test with trace. Boot time from start to main_loop() entry:
> - ~464ms - before this change (arch memset enabled for .bss clear)
> - ~341ms - after this change
> 
> Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
> Reviewed-by: Simon Glass <sjg@chromium.org>

Applied to u-boot/master, thanks!

-- 
Tom
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.denx.de/pipermail/u-boot/attachments/20150309/c2cb7175/attachment.sig>

[U-Boot] [U-Boot, v6, 06/10] odroid: defconfig: disable memset at malloc init

[Tom Rini]

On Wed, Mar 04, 2015 at 02:01:26PM +0100, Przemyslaw Marczak wrote:

> Reduce the boot time of Odroid X2/U3 by disabling the memset
> at malloc init.
> 
> This was tested on Odroid X2.
> A quick test with checking gpio pin state using the oscilloscope.
> Boot time from start to bootcmd (change gpio state by memory write command):
> - ~228ms - before this change (arch memset enabled for .bss clear)
> - ~100ms - after this change
> 
> Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
> Reviewed-by: Simon Glass <sjg@chromium.org>

Applied to u-boot/master, thanks!

-- 
Tom
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.denx.de/pipermail/u-boot/attachments/20150309/cefd1b94/attachment.sig>

[U-Boot] [U-Boot, v6, 07/10] odroid-xu3: defconfig: disable memset at malloc init

[Tom Rini]

On Wed, Mar 04, 2015 at 02:01:27PM +0100, Przemyslaw Marczak wrote:

> Reduce the boot time of Odroid XU3 by disabling the memset
> at malloc init.
> 
> Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>

Applied to u-boot/master, thanks!

-- 
Tom
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.denx.de/pipermail/u-boot/attachments/20150309/03191836/attachment.sig>

[U-Boot] [U-Boot, v6, 08/10] zynq-common: increase malloc pool len by dfu mmc file buffer size

[Tom Rini]

On Wed, Mar 04, 2015 at 02:01:28PM +0100, Przemyslaw Marczak wrote:

> The dfu mmc file buffer, which was static, now is allocated
> by memalign(), so the malloc pool len should be also increased.
> 
> Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
> Cc: Jagannadha Sutradharudu Teki <jaganna@xilinx.com>
> Cc: Michal Simek <monstr@monstr.eu>
> Acked-by: Michal Simek <michal.simek@xilinx.com>

Applied to u-boot/master, thanks!

-- 
Tom
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.denx.de/pipermail/u-boot/attachments/20150309/2e563ec4/attachment.sig>

[U-Boot] [U-Boot, v6, 09/10] ti-armv7-common: increase malloc pool len by dfu mmc file buffer size

[Tom Rini]

On Wed, Mar 04, 2015 at 02:01:29PM +0100, Przemyslaw Marczak wrote:

> The dfu mmc file buffer, which was static, now is allocated
> by memalign(), so the malloc pool len should be also increased.
> 
> Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
> Cc: Tom Rini <trini@konsulko.com>

Applied to u-boot/master, thanks!

-- 
Tom
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.denx.de/pipermail/u-boot/attachments/20150309/834ed333/attachment.sig>

[U-Boot] [U-Boot, v6, 10/10] tegra-common: increase malloc pool len by dfu mmc file buffer size

[Tom Rini]

On Wed, Mar 04, 2015 at 02:01:30PM +0100, Przemyslaw Marczak wrote:

> The dfu mmc file buffer, which was static, now is allocated
> by memalign(), so the malloc pool len should be also increased.
> 
> Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
> Cc: Tom Warren <twarren.nvidia@gmail.com>

Applied to u-boot/master, thanks!

-- 
Tom
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.denx.de/pipermail/u-boot/attachments/20150309/11539ce9/attachment.sig>