Skip to contentSkip to conversation

btrfs addition to uboot

PATCH New Adnan Ali 4 participants 15 messages Feb 25, 2013 – Mar 1, 2013

Subsystems (3)
lore
Share & tools
lore

Adnan Ali

author
#0 lore raw

PATCH v3 1/1 btrfs addition to uboot

Adnan Ali (1):
Introduced btrfs file-system with btrload command

 Makefile                   |    1 +
 common/Makefile            |    1 +
 common/cmd_btr.c           |   53 +++
 fs/btrfs/Makefile          |   51 ++
 fs/btrfs/btrfs.c           | 1131 ++++++++++++++++++++++++++++++++++++++++++++
 fs/fs.c                    |   86 +++-
 include/btrfs.h            |  398 ++++++++++++++++
 include/config_fallbacks.h |    4 +
 include/crc32c.h           |   48 ++
 include/fs.h               |    1 +
 10 files changed, 1769 insertions(+), 5 deletions(-)
 create mode 100644 common/cmd_btr.c
 create mode 100644 fs/btrfs/Makefile
 create mode 100644 fs/btrfs/btrfs.c
 create mode 100644 include/btrfs.h
 create mode 100644 include/crc32c.h
Signature · 1 line
1.7.9.5
author
#1 lore raw

PATCH v3 1/1 Introduced btrfs file-system with btrload command

Introduces btrfs file-system to read file from volume/sub-volumes with btrload command. This implementation has read-only support.
This btrfs implementation is based on syslinux btrfs code, commit 269ebc845ebc8b46ef4b0be7fa0005c7fdb95b8d.

---
 Makefile                   |    1 +
 common/Makefile            |    1 +
 common/cmd_btr.c           |   53 +++
 fs/btrfs/Makefile          |   51 ++
 fs/btrfs/btrfs.c           | 1131 ++++++++++++++++++++++++++++++++++++++++++++
 fs/fs.c                    |   86 +++-
 include/btrfs.h            |  398 ++++++++++++++++
 include/config_fallbacks.h |    4 +
 include/crc32c.h           |   48 ++
 include/fs.h               |    1 +
 10 files changed, 1769 insertions(+), 5 deletions(-)
 create mode 100644 common/cmd_btr.c
 create mode 100644 fs/btrfs/Makefile
 create mode 100644 fs/btrfs/btrfs.c
 create mode 100644 include/btrfs.h
 create mode 100644 include/crc32c.h
Makefile +1 -0
index 3305e8c..cc35e7b 100644
@@ -261,6 +261,7 @@ endif
 LIBS-$(CONFIG_OF_EMBED) += dts/libdts.o
 LIBS-y += arch/$(ARCH)/lib/lib$(ARCH).o
 LIBS-y += fs/libfs.o \
+        fs/btrfs/libbtrfs.o \
 	fs/cbfs/libcbfs.o \
 	fs/cramfs/libcramfs.o \
 	fs/ext4/libext4fs.o \
common/Makefile +1 -0
index 54fcc81..093dd35 100644
@@ -73,6 +73,7 @@ COBJS-$(CONFIG_CMD_BEDBUG) += bedbug.o cmd_bedbug.o
 COBJS-$(CONFIG_CMD_BMP) += cmd_bmp.o
 COBJS-$(CONFIG_CMD_BOOTLDR) += cmd_bootldr.o
 COBJS-$(CONFIG_CMD_BOOTSTAGE) += cmd_bootstage.o
+COBJS-$(CONFIG_CMD_BTR) += cmd_btr.o
 COBJS-$(CONFIG_CMD_CACHE) += cmd_cache.o
 COBJS-$(CONFIG_CMD_CBFS) += cmd_cbfs.o
 COBJS-$(CONFIG_CMD_CONSOLE) += cmd_console.o
common/cmd_btr.c new file +53 -0

53 lines not on this page · raw message

fs/btrfs/Makefile new file +51 -0

51 lines not on this page · raw message

fs/btrfs/btrfs.c new file +1131 -0

1,131 lines not on this page · raw message

fs/fs.c +81 -5

139 lines not on this page · raw message

include/btrfs.h new file +398 -0

398 lines not on this page · raw message

include/config_fallbacks.h +4 -0
index bfb9680..28a0a66 100644
@@ -26,4 +26,8 @@
 #define CONFIG_EXT4_WRITE
 #endif
 
+#if defined(CONFIG_CMD_BTR) && !defined(CONFIG_FS_BTR)
+#define CONFIG_FS_BTR
+#endif
+
 #endif	/* __CONFIG_FALLBACKS_H */
include/crc32c.h new file +48 -0

48 lines not on this page · raw message

include/fs.h +1 -0
index 4f30a38..3a0e7d4 100644
@@ -21,6 +21,7 @@
 #define FS_TYPE_ANY	0
 #define FS_TYPE_FAT	1
 #define FS_TYPE_EXT	2
+#define FS_TYPE_BTR	3
 
 /*
  * Tell the fs layer which block device an partition to use for future
Signature · 1 line
1.7.9.5
#2 lore raw

PATCH v3 1/1 Introduced btrfs file-system with btrload command

Quoted 7 lines · Adnan Ali

Introduces btrfs file-system to read file from volume/sub-volumes with btrload command. This implementation has read-only support.
This btrfs implementation is based on syslinux btrfs code, commit 269ebc845ebc8b46ef4b0be7fa0005c7fdb95b8d.

A few things:
- In general in fs/btrfs/btrfs.c I see some coding style problems (lack of spacing, non-printf's longer than 80-wide). Do these come from syslinux and thus will make any re-syncs easier?
- It looks like you added support for CONFIG_CMD_FS_GENERIC, if so did you test that?
- Can you please enable this support code on at least one platform, preferably the one you've tested and developed with?

[snip]

Quoted 1 line
fs/fs.c +0 -0

[snip]

Quoted 1 line

+ //file handle is valid get the size of the file

/* Style comments only */

Quoted 1 line

+ len=filedata.size;

And spacing between variable and assignment.

Quoted 8 lines
fs/fs.c +0 -1
@@ -178,7 +248,6 @@ int fs_set_blk_dev(const char *ifname, const char *dev_part_str, int fstype)
 	for (i = 0; i < ARRAY_SIZE(fstypes); i++) {
 		if ((fstype != FS_TYPE_ANY) && (fstype != fstypes[i].fstype))
 			continue;
-
 		if (!fstypes[i].probe()) {
 			fs_type = fstypes[i].fstype;
 			return 0;

[snip]

fs/fs.c static void fs_close(void) · line 285

Quoted 8 lines
fs/fs.c +0 -1
@@ -208,7 +280,6 @@ static void fs_close(void)
 int fs_ls(const char *dirname)
 {
 	int ret;
-
 	switch (fs_type) {
 	case FS_TYPE_FAT:
 		ret = fs_ls_fat(dirname);

Unrelated, please drop.

fs/fs.c int fs_read(const char *filename, ulong addr, int offset, int len) · line 321

Quoted 13 lines
fs/fs.c +3 -1
@@ -237,11 +311,13 @@ int fs_read(const char *filename, ulong addr, int offset, int len)
 	case FS_TYPE_EXT:
 		ret = fs_read_ext(filename, addr, offset, len);
 		break;
+        case FS_TYPE_BTR:
+		ret = fs_read_btr(filename, addr, offset, len);
+		break; 
 	default:
 		ret = fs_read_unsupported(filename, addr, offset, len);
 		break;
 	}
-
 	fs_close();

And unrelated whitespace changes here as well.

Quoted 1 line
include/btrfs.h +0 -0

[snip]

include/btrfs.h line 46

Quoted 7 lines
include/btrfs.h +7 -0
+/*
+ * Extent structure: contains the mapping of some chunk of a file
+ * that is contiguous on disk.
+ */
+struct extent {
+    //sector_t    pstart;         /* Physical start sector */
+    __le64   pstart;

Fix please.

include/btrfs.h line 389

Quoted 7 lines
include/btrfs.h +7 -0
+/*
+ * Our definition of "not whitespace"
+ */
+static inline char not_whitespace(char c)
+{
+  return (unsigned char)c > ' ';
+}

Can't you just use isspace from <linux/ctypes.h> ?

Quoted 44 lines
include/crc32c.h new file +38 -0
new file mode 100644
index 0000000..d04916e
@@ -0,0 +1,48 @@
+/*
+ * Copied from Linux kernel crypto/crc32c.c
+ * Copyright (c) 2004 Cisco Systems, Inc.
+ * Copyright (c) 2008 Herbert Xu <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License as published by the Free
+ * Software Foundation; either version 2 of the License, or (at your option)
+ * any later version.
+ *
+ */
+
+/*
+ * This is the CRC-32C table
+ * Generated with:
+ * width = 32 bits
+ * poly = 0x1EDC6F41
+ * reflect input bytes = true
+ * reflect output bytes = true
+ */
+
+/*
+ * Steps through buffer one byte at at time, calculates reflected
+ * crc using table.
+ */
+
+static inline u32 crc32c_cal(u32 crc, const char *data, size_t length, u32 *crc32c_table)
+{
+	while (length--)
+		crc = crc32c_table[(u8)(crc ^ *data++)] ^ (crc >> 8);
+
+	return crc;
+}
+
+static inline void crc32c_init(u32 *crc32c_table, u32 pol)
+{
+	int i, j;
+	u32 v;

include/crc32c.h line 48

Quoted 10 lines
include/crc32c.h new file +10 -0
+	const u32 poly = pol; /* Bit-reflected CRC32C polynomial */
+
+	for (i = 0; i < 256; i++) {
+		v = i;
+		for (j = 0; j < 8; j++) {
+			v = (v >> 1) ^ ((v & 1) ? poly : 0);
+		}
+		crc32c_table[i] = v;
+	}
+}

Simon, since you've just been over all the crc32 recently, do we have these functions somewhere already that can be easily tapped in to? Thanks!

Signature · 8 lines
Tom
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.denx.de/pipermail/u-boot/attachments/20130225/e1521cf0/attachment.pgp>
author
#3 lore raw

PATCH v3 1/1 Introduced btrfs file-system with btrload command

Quoted 8 lines · Tom Rini
Quoted 6 lines · Adnan Ali

Introduces btrfs file-system to read file from volume/sub-volumes with btrload command. This implementation has read-only support.
This btrfs implementation is based on syslinux btrfs code, commit 269ebc845ebc8b46ef4b0be7fa0005c7fdb95b8d.

Quoted 5 lines
Quoted 1 line

A few things:
- In general in fs/btrfs/btrfs.c I see some coding style problems (lack of spacing, non-printf's longer than 80-wide). Do these come from syslinux and thus will make any re-syncs easier?

Most of the ported code is unmodified so its coding style should
be same as syslinux.

Quoted 2 lines

- It looks like you added support for CONFIG_CMD_FS_GENERIC, if so did you test that?

This command wasn't enabled in my configs. I haven't added
any command i.e btrls for this as this feature is not supported yet.

Quoted 2 lines

- Can you please enable this support code on at least one platform, preferably the one you've tested and developed with?

Even if do enable support for this, it will also debug 'Unsupported filesystem type.'

For the rest of changes you proposed i will change them and send as v4 patch.

Quoted 113 lines

[snip]

Quoted 1 line
fs/fs.c +0 -0

[snip]

Quoted 1 line

+ //file handle is valid get the size of the file

/* Style comments only */

Quoted 1 line

+ len=filedata.size;

And spacing between variable and assignment.

Quoted 8 lines
fs/fs.c +0 -1
@@ -178,7 +248,6 @@ int fs_set_blk_dev(const char *ifname, const char *dev_part_str, int fstype)
  	for (i = 0; i < ARRAY_SIZE(fstypes); i++) {
  		if ((fstype != FS_TYPE_ANY) && (fstype != fstypes[i].fstype))
  			continue;
-
  		if (!fstypes[i].probe()) {
  			fs_type = fstypes[i].fstype;
  			return 0;

[snip]

fs/fs.c static void fs_close(void) · line 285

Quoted 8 lines
fs/fs.c +0 -1
@@ -208,7 +280,6 @@ static void fs_close(void)
  int fs_ls(const char *dirname)
  {
  	int ret;
-
  	switch (fs_type) {
  	case FS_TYPE_FAT:
  		ret = fs_ls_fat(dirname);

Unrelated, please drop.

fs/fs.c int fs_read(const char *filename, ulong addr, int offset, int len) · line 321

Quoted 13 lines
fs/fs.c +3 -1
@@ -237,11 +311,13 @@ int fs_read(const char *filename, ulong addr, int offset, int len)
  	case FS_TYPE_EXT:
  		ret = fs_read_ext(filename, addr, offset, len);
  		break;
+        case FS_TYPE_BTR:
+		ret = fs_read_btr(filename, addr, offset, len);
+		break;
  	default:
  		ret = fs_read_unsupported(filename, addr, offset, len);
  		break;
  	}
-
  	fs_close();

And unrelated whitespace changes here as well.

Quoted 1 line
include/btrfs.h +0 -0

[snip]

include/btrfs.h line 46

Quoted 7 lines
include/btrfs.h +7 -0
+/*
+ * Extent structure: contains the mapping of some chunk of a file
+ * that is contiguous on disk.
+ */
+struct extent {
+    //sector_t    pstart;         /* Physical start sector */
+    __le64   pstart;

Fix please.

include/btrfs.h line 389

Quoted 7 lines
include/btrfs.h +7 -0
+/*
+ * Our definition of "not whitespace"
+ */
+static inline char not_whitespace(char c)
+{
+  return (unsigned char)c > ' ';
+}

Can't you just use isspace from <linux/ctypes.h> ?

Quoted 49 lines
include/crc32c.h new file +43 -0
new file mode 100644
index 0000000..d04916e
@@ -0,0 +1,48 @@
+/*
+ * Copied from Linux kernel crypto/crc32c.c
+ * Copyright (c) 2004 Cisco Systems, Inc.
+ * Copyright (c) 2008 Herbert Xu <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License as published by the Free
+ * Software Foundation; either version 2 of the License, or (at your option)
+ * any later version.
+ *
+ */
+
+/*
+ * This is the CRC-32C table
+ * Generated with:
+ * width = 32 bits
+ * poly = 0x1EDC6F41
+ * reflect input bytes = true
+ * reflect output bytes = true
+ */
+
+/*
+ * Steps through buffer one byte at at time, calculates reflected
+ * crc using table.
+ */
+
+static inline u32 crc32c_cal(u32 crc, const char *data, size_t length, u32 *crc32c_table)
+{
+	while (length--)
+		crc = crc32c_table[(u8)(crc ^ *data++)] ^ (crc >> 8);
+
+	return crc;
+}
+
+static inline void crc32c_init(u32 *crc32c_table, u32 pol)
+{
+	int i, j;
+	u32 v;
+	const u32 poly = pol; /* Bit-reflected CRC32C polynomial */
+
+	for (i = 0; i < 256; i++) {
+		v = i;
+		for (j = 0; j < 8; j++) {
Quoted 9 lines

include/crc32c.h line 48

Quoted 5 lines
include/crc32c.h new file +5 -0
+			v = (v >> 1) ^ ((v & 1) ? poly : 0);
+		}
+		crc32c_table[i] = v;
+	}
+}

Simon, since you've just been over all the crc32 recently, do we have these functions somewhere already that can be easily tapped in to? Thanks!

Thanks

Adnan

#4 lore raw

PATCH v3 1/1 Introduced btrfs file-system with btrload command

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Quoted 9 lines · Adnan Ali
Quoted 8 lines · Tom Rini
Quoted 6 lines · Adnan Ali

Introduces btrfs file-system to read file from volume/sub-volumes with btrload command. This implementation has read-only support. This btrfs implementation is based on syslinux btrfs code, commit 269ebc845ebc8b46ef4b0be7fa0005c7fdb95b8d.

Quoted 7 lines
Quoted 5 lines
Quoted 1 line

A few things: - In general in fs/btrfs/btrfs.c I see some coding style problems (lack of spacing, non-printf's longer than 80-wide). Do these come from syslinux and thus will make any re-syncs easier?

Most of the ported code is unmodified so its coding style should be same as syslinux.

That's what I thought.

Quoted 4 lines
Quoted 2 lines

- It looks like you added support for CONFIG_CMD_FS_GENERIC, if so did you test that?

This command wasn't enabled in my configs. I haven't added any command i.e btrls for this as this feature is not supported yet.

Can you then please? Thanks.

Quoted 4 lines
Quoted 2 lines

- Can you please enable this support code on at least one platform, preferably the one you've tested and developed with?

Even if do enable support for this, it will also debug 'Unsupported filesystem type.'

I don't understand.

Quoted 2 lines

For the rest of changes you proposed i will change them and send as v4 patch.

Thanks!

- --
Tom
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/

iQIcBAEBAgAGBQJRK5FtAAoJENk4IS6UOR1WLbYP/i/c8s10TIoya1UnLiCX87tt u/DRhip4/HBhyCtZmxCejLM6VrgJ8GBACq7hP6eXY3BE5zRB11rKnaypOdc1EA6h WG87W+6LDOQEvFPzuyXRpXIGvUwq5gHsa95rM0pq/k5WnUJwl4eRAyuhU0JFilQ8 p83GUbYH3Ss1ZX6sr2O+LNkGrIW/PNw1TVioJYyfUUQvcqVYw8CIBLSusxCkY445 ps2QD3Np4XvYXkGFoHUkb0t+Lp0m87PaHBAuABr+TqUEne+QT3GU1E6xN4jcoasX 2Bd0LQFnIhoU6hiH/EuAsQs8tt48ByEliug0Yvr7iK4Cb5sQen21LjXUagarvArF RKkg/SThdqwYStfkX5/+z73m2F3lcGGF9IobVuz9HWi2VvXEYM500+pGYcOzjP+3 V+deSp0PReW4lbFRnFgxcArETIumd4+deCQzX9XHUC20iC/Xf2N+s6L5VTGqQf0g 33+QDPaUo1SdCZmKiCpVuBa0hzZsQPRhFOHRUpGbEZRbf45iHl5Ewhm88K6dTiz4 vsr7YJDs4rJcph0q6tKSLVVGioPpa3EqpWZrXGB4JTwc6GHbluAvL7GXrHWxtbtX AZOxNEegmljq+l+6g+n2sZ+RDlyynU/x9NvKVEM8ucqmMGrcrTKG5fUc536U2sWL 8xEfGHfYMZ9AxbF0DEXY
=FPEv
-----END PGP SIGNATURE-----

author
#5 lore raw

PATCH v3 1/1 Introduced btrfs file-system with btrload command

Quoted 20 lines · Tom Rini

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Quoted 9 lines · Adnan Ali
Quoted 8 lines · Tom Rini
Quoted 6 lines · Adnan Ali

Introduces btrfs file-system to read file from volume/sub-volumes with btrload command. This implementation has read-only support. This btrfs implementation is based on syslinux btrfs code, commit 269ebc845ebc8b46ef4b0be7fa0005c7fdb95b8d.

Quoted 7 lines
Quoted 5 lines
Quoted 1 line

A few things: - In general in fs/btrfs/btrfs.c I see some coding style problems (lack of spacing, non-printf's longer than 80-wide). Do these come from syslinux and thus will make any re-syncs easier?

Most of the ported code is unmodified so its coding style should be same as syslinux.

Quoted 7 lines

That's what I thought.

Quoted 4 lines
Quoted 2 lines

- It looks like you added support for CONFIG_CMD_FS_GENERIC, if so did you test that?

This command wasn't enabled in my configs. I haven't added any command i.e btrls for this as this feature is not supported yet.

Can you then please? Thanks.

I have added support for this in my configs, build and flashed
on my mx53loco hardware it booted fine. When i type LS gives me ** Unrecognized filesystem type **. which is what i have done it in code. Any ls should print above message when its using btrfs.

Quoted 6 lines
Quoted 5 lines
Quoted 3 lines

- Can you please enable this support code on at least one platform, preferably the one you've tested and developed with?

Even if do enable support for this, it will also debug 'Unsupported filesystem type.'

I don't understand.

I'm assuming that you are asking me to test it on atleast one
hardware platform with CONFIG_CMD_FS_GENERIC enabled. Which i done it on mx53loco. If my assumption is wrong can you please rephrase your question.

Quoted 11 lines
Quoted 3 lines

For the rest of changes you proposed i will change them and send as v4 patch.

Thanks!

- --
Tom
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/

Quoted 14 lines

iQIcBAEBAgAGBQJRK5FtAAoJENk4IS6UOR1WLbYP/i/c8s10TIoya1UnLiCX87tt u/DRhip4/HBhyCtZmxCejLM6VrgJ8GBACq7hP6eXY3BE5zRB11rKnaypOdc1EA6h WG87W+6LDOQEvFPzuyXRpXIGvUwq5gHsa95rM0pq/k5WnUJwl4eRAyuhU0JFilQ8 p83GUbYH3Ss1ZX6sr2O+LNkGrIW/PNw1TVioJYyfUUQvcqVYw8CIBLSusxCkY445 ps2QD3Np4XvYXkGFoHUkb0t+Lp0m87PaHBAuABr+TqUEne+QT3GU1E6xN4jcoasX 2Bd0LQFnIhoU6hiH/EuAsQs8tt48ByEliug0Yvr7iK4Cb5sQen21LjXUagarvArF RKkg/SThdqwYStfkX5/+z73m2F3lcGGF9IobVuz9HWi2VvXEYM500+pGYcOzjP+3 V+deSp0PReW4lbFRnFgxcArETIumd4+deCQzX9XHUC20iC/Xf2N+s6L5VTGqQf0g 33+QDPaUo1SdCZmKiCpVuBa0hzZsQPRhFOHRUpGbEZRbf45iHl5Ewhm88K6dTiz4 vsr7YJDs4rJcph0q6tKSLVVGioPpa3EqpWZrXGB4JTwc6GHbluAvL7GXrHWxtbtX AZOxNEegmljq+l+6g+n2sZ+RDlyynU/x9NvKVEM8ucqmMGrcrTKG5fUc536U2sWL 8xEfGHfYMZ9AxbF0DEXY
=FPEv
-----END PGP SIGNATURE-----

Thanks

Adnan

#6 lore raw

PATCH v3 1/1 Introduced btrfs file-system with btrload command

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Quoted 1 line · Adnan Ali

On 25/02/13 16:29, Tom Rini wrote:

[snip]

Quoted 11 lines
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines
Quoted 2 lines

- It looks like you added support for
CONFIG_CMD_FS_GENERIC, if so did you test that?

This command wasn't enabled in my configs. I haven't added any command i.e btrls for this as this feature is not supported yet.

Can you then please? Thanks.

Quoted 4 lines

I have added support for this in my configs, build and flashed on my mx53loco hardware it booted fine. When i type LS gives me ** Unrecognized filesystem type **. which is what i have done it in code. Any ls should print above message when its using btrfs.

Quoted 10 lines
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines
Quoted 3 lines

- Can you please enable this support code on at least one platform, preferably the one you've tested and developed with?

Even if do enable support for this, it will also debug 'Unsupported filesystem type.'

I don't understand.

Quoted 4 lines

I'm assuming that you are asking me to test it on atleast one hardware platform with CONFIG_CMD_FS_GENERIC enabled. Which i done it on mx53loco. If my assumption is wrong can you please rephrase your question.

OK. Can you please add btrfs support to CONFIG_CMD_FS_GENERIC? Or try and point out that it's quite difficult / annoying / we need to fix our interface? You'd be the first person other than the author of the original changes to do so, so your feedback here would be quite useful. Thanks!

- --
Tom
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/

iQIcBAEBAgAGBQJRK5X2AAoJENk4IS6UOR1WFqAP/0DStonfADUwZTzyxwT22spZ xGrJGUCVBBpiZ+xoO3ZRuKYG/DAkD3kYATV2DqgM/kqVuusHyiPeh4z1qcO78tWk Quwce9sIU5uhMkQey0sCbIX5YHgxY0DbrcxpNLedQLKLo1NOZY3/uJUDRNt8uIeP B3PH5uNXR8hVwwC0ML8dC8ZK7/XbPBZvdr0jjhYR8RzfHVxbpFKNL1vjxsN9TZiS ZAp+DyIP389FPRQd2zt4bhzNHGuJ2TkwewEXi00MCO8vKkRKvuPZbV7hc0w9t+d7 m1VbAzLL/FZ9ld8r3CvowgBg91vP/y8Vjr5bnX9CaZzaf96vPTVldx/UmZ9xtWtb TzPxPaYkHXEmtdGfxfYI/v1CHu9rkjiDOrZMPbXNx2XzjYdwp5kk8Hlfqi+F8d+m wFt+q2kHZdAEPDoIcvsi4buAbiqyehH03K0QpaH1UaEH01lVvw8S9U1p9/iLcSrV c1kN4GMySKdPTIoTv97S4SpHUcijI1aMOaqc462M4YKroRg54iMMHCFkQBy0HF4U n6B+yZv1Yy5IWf4Z5b143V8FN8yOSvfzG+gEMICc/NZU09nT6fwfPRfMWj2lIPj8 wsDYpdL7amDAl2TUCLEB8vbue5ELjB2Bu2S25wKrsKrN9mPvh1LvFyHhO4HJ49Ly +1SAb04CSBOhgXOzO+SZ
=pYZr
-----END PGP SIGNATURE-----

#7 lore raw

PATCH v3 1/1 Introduced btrfs file-system with btrload command

Hi,

Quoted 28 lines · Tom Rini

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Quoted 1 line · Adnan Ali

On 25/02/13 16:29, Tom Rini wrote:

[snip]

Quoted 11 lines
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines
Quoted 2 lines

- It looks like you added support for
CONFIG_CMD_FS_GENERIC, if so did you test that?

This command wasn't enabled in my configs. I haven't added any command i.e btrls for this as this feature is not supported yet.

Can you then please? Thanks.

Quoted 4 lines

I have added support for this in my configs, build and flashed on my mx53loco hardware it booted fine. When i type LS gives me ** Unrecognized filesystem type **. which is what i have done it in code. Any ls should print above message when its using btrfs.

Quoted 10 lines
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines
Quoted 3 lines

- Can you please enable this support code on at least one platform, preferably the one you've tested and developed with?

Even if do enable support for this, it will also debug 'Unsupported filesystem type.'

I don't understand.

Quoted 4 lines

I'm assuming that you are asking me to test it on atleast one hardware platform with CONFIG_CMD_FS_GENERIC enabled. Which i done it on mx53loco. If my assumption is wrong can you please rephrase your question.

Quoted 5 lines

OK. Can you please add btrfs support to CONFIG_CMD_FS_GENERIC? Or try and point out that it's quite difficult / annoying / we need to fix our interface? You'd be the first person other than the author of the original changes to do so, so your feedback here would be quite useful. Thanks!

Actually please see my series starting here, which adds more support to that layer:

http://patchwork.ozlabs.org/patch/208215/

Perhaps btrfs could fit in there too?

Regards,
Simon

Quoted 21 lines

- --
Tom
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/

iQIcBAEBAgAGBQJRK5X2AAoJENk4IS6UOR1WFqAP/0DStonfADUwZTzyxwT22spZ xGrJGUCVBBpiZ+xoO3ZRuKYG/DAkD3kYATV2DqgM/kqVuusHyiPeh4z1qcO78tWk Quwce9sIU5uhMkQey0sCbIX5YHgxY0DbrcxpNLedQLKLo1NOZY3/uJUDRNt8uIeP B3PH5uNXR8hVwwC0ML8dC8ZK7/XbPBZvdr0jjhYR8RzfHVxbpFKNL1vjxsN9TZiS ZAp+DyIP389FPRQd2zt4bhzNHGuJ2TkwewEXi00MCO8vKkRKvuPZbV7hc0w9t+d7 m1VbAzLL/FZ9ld8r3CvowgBg91vP/y8Vjr5bnX9CaZzaf96vPTVldx/UmZ9xtWtb TzPxPaYkHXEmtdGfxfYI/v1CHu9rkjiDOrZMPbXNx2XzjYdwp5kk8Hlfqi+F8d+m wFt+q2kHZdAEPDoIcvsi4buAbiqyehH03K0QpaH1UaEH01lVvw8S9U1p9/iLcSrV c1kN4GMySKdPTIoTv97S4SpHUcijI1aMOaqc462M4YKroRg54iMMHCFkQBy0HF4U n6B+yZv1Yy5IWf4Z5b143V8FN8yOSvfzG+gEMICc/NZU09nT6fwfPRfMWj2lIPj8 wsDYpdL7amDAl2TUCLEB8vbue5ELjB2Bu2S25wKrsKrN9mPvh1LvFyHhO4HJ49Ly +1SAb04CSBOhgXOzO+SZ
=pYZr
-----END PGP SIGNATURE-----

#8 lore raw

PATCH v3 1/1 Introduced btrfs file-system with btrload command

In reply to Tom Rini

Hi,

Quoted 138 lines · Tom Rini
Quoted 7 lines · Adnan Ali

Introduces btrfs file-system to read file from volume/sub-volumes with btrload command. This implementation has read-only support.
This btrfs implementation is based on syslinux btrfs code, commit 269ebc845ebc8b46ef4b0be7fa0005c7fdb95b8d.

A few things:
- In general in fs/btrfs/btrfs.c I see some coding style problems (lack of spacing, non-printf's longer than 80-wide). Do these come from syslinux and thus will make any re-syncs easier?
- It looks like you added support for CONFIG_CMD_FS_GENERIC, if so did you test that?
- Can you please enable this support code on at least one platform, preferably the one you've tested and developed with?

[snip]

Quoted 1 line
fs/fs.c +0 -0

[snip]

Quoted 1 line

+ //file handle is valid get the size of the file

/* Style comments only */

Quoted 1 line

+ len=filedata.size;

And spacing between variable and assignment.

Quoted 8 lines
fs/fs.c +0 -1
@@ -178,7 +248,6 @@ int fs_set_blk_dev(const char *ifname, const char *dev_part_str, int fstype)
      for (i = 0; i < ARRAY_SIZE(fstypes); i++) {
              if ((fstype != FS_TYPE_ANY) && (fstype != fstypes[i].fstype))
                      continue;
-
              if (!fstypes[i].probe()) {
                      fs_type = fstypes[i].fstype;
                      return 0;

[snip]

fs/fs.c static void fs_close(void) · line 285

Quoted 8 lines
fs/fs.c +0 -1
@@ -208,7 +280,6 @@ static void fs_close(void)
 int fs_ls(const char *dirname)
 {
      int ret;
-
      switch (fs_type) {
      case FS_TYPE_FAT:
              ret = fs_ls_fat(dirname);

Unrelated, please drop.

fs/fs.c int fs_read(const char *filename, ulong addr, int offset, int len) · line 321

Quoted 13 lines
fs/fs.c +3 -1
@@ -237,11 +311,13 @@ int fs_read(const char *filename, ulong addr, int offset, int len)
      case FS_TYPE_EXT:
              ret = fs_read_ext(filename, addr, offset, len);
              break;
+        case FS_TYPE_BTR:
+             ret = fs_read_btr(filename, addr, offset, len);
+             break;
      default:
              ret = fs_read_unsupported(filename, addr, offset, len);
              break;
      }
-
      fs_close();

And unrelated whitespace changes here as well.

Quoted 1 line
include/btrfs.h +0 -0

[snip]

include/btrfs.h line 46

Quoted 7 lines
include/btrfs.h +7 -0
+/*
+ * Extent structure: contains the mapping of some chunk of a file
+ * that is contiguous on disk.
+ */
+struct extent {
+    //sector_t    pstart;         /* Physical start sector */
+    __le64   pstart;

Fix please.

include/btrfs.h line 389

Quoted 7 lines
include/btrfs.h +7 -0
+/*
+ * Our definition of "not whitespace"
+ */
+static inline char not_whitespace(char c)
+{
+  return (unsigned char)c > ' ';
+}

Can't you just use isspace from <linux/ctypes.h> ?

Quoted 50 lines
include/crc32c.h new file +44 -0
new file mode 100644
index 0000000..d04916e
@@ -0,0 +1,48 @@
+/*
+ * Copied from Linux kernel crypto/crc32c.c
+ * Copyright (c) 2004 Cisco Systems, Inc.
+ * Copyright (c) 2008 Herbert Xu <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License as published by the Free
+ * Software Foundation; either version 2 of the License, or (at your option)
+ * any later version.
+ *
+ */
+
+/*
+ * This is the CRC-32C table
+ * Generated with:
+ * width = 32 bits
+ * poly = 0x1EDC6F41
+ * reflect input bytes = true
+ * reflect output bytes = true
+ */
+
+/*
+ * Steps through buffer one byte at at time, calculates reflected
+ * crc using table.
+ */
+
+static inline u32 crc32c_cal(u32 crc, const char *data, size_t length, u32 *crc32c_table)
+{
+     while (length--)
+             crc = crc32c_table[(u8)(crc ^ *data++)] ^ (crc >> 8);
+
+     return crc;
+}
+
+static inline void crc32c_init(u32 *crc32c_table, u32 pol)
+{
+     int i, j;
+     u32 v;
+     const u32 poly = pol; /* Bit-reflected CRC32C polynomial */
+
+     for (i = 0; i < 256; i++) {
+             v = i;
+             for (j = 0; j < 8; j++) {
+                     v = (v >> 1) ^ ((v & 1) ? poly : 0);
Quoted 8 lines

include/crc32c.h line 48

Quoted 4 lines
include/crc32c.h new file +4 -0
+             }
+             crc32c_table[i] = v;
+     }
+}

Simon, since you've just been over all the crc32 recently, do we have these functions somewhere already that can be easily tapped in to? Thanks!

Should be - see lib/crc32.c. There is already at least one other copy (ubifs I think) so we should try to avoid adding more.

Maybe the polynomial is different here? But even so it should go with the existing code I think.

Regards,
Simon

Quoted 3 lines

--
Tom

author
#9 lore raw

PATCH v3 1/1 Introduced btrfs file-system with btrload command

Quoted 141 lines · Simon Glass

Hi,

Quoted 130 lines · Tom Rini
Quoted 7 lines · Adnan Ali

Introduces btrfs file-system to read file from volume/sub-volumes with btrload command. This implementation has read-only support.
This btrfs implementation is based on syslinux btrfs code, commit 269ebc845ebc8b46ef4b0be7fa0005c7fdb95b8d.

A few things:
- In general in fs/btrfs/btrfs.c I see some coding style problems (lack of spacing, non-printf's longer than 80-wide). Do these come from syslinux and thus will make any re-syncs easier?
- It looks like you added support for CONFIG_CMD_FS_GENERIC, if so did you test that?
- Can you please enable this support code on at least one platform, preferably the one you've tested and developed with?

[snip]

Quoted 1 line
fs/fs.c +0 -0

[snip]

Quoted 1 line

+ //file handle is valid get the size of the file

/* Style comments only */

Quoted 1 line

+ len=filedata.size;

And spacing between variable and assignment.

Quoted 8 lines
fs/fs.c +0 -1
@@ -178,7 +248,6 @@ int fs_set_blk_dev(const char *ifname, const char *dev_part_str, int fstype)
       for (i = 0; i < ARRAY_SIZE(fstypes); i++) {
               if ((fstype != FS_TYPE_ANY) && (fstype != fstypes[i].fstype))
                       continue;
-
               if (!fstypes[i].probe()) {
                       fs_type = fstypes[i].fstype;
                       return 0;

[snip]

fs/fs.c static void fs_close(void) · line 285

Quoted 8 lines
fs/fs.c +0 -1
@@ -208,7 +280,6 @@ static void fs_close(void)
  int fs_ls(const char *dirname)
  {
       int ret;
-
       switch (fs_type) {
       case FS_TYPE_FAT:
               ret = fs_ls_fat(dirname);

Unrelated, please drop.

fs/fs.c int fs_read(const char *filename, ulong addr, int offset, int len) · line 321

Quoted 13 lines
fs/fs.c +3 -1
@@ -237,11 +311,13 @@ int fs_read(const char *filename, ulong addr, int offset, int len)
       case FS_TYPE_EXT:
               ret = fs_read_ext(filename, addr, offset, len);
               break;
+        case FS_TYPE_BTR:
+             ret = fs_read_btr(filename, addr, offset, len);
+             break;
       default:
               ret = fs_read_unsupported(filename, addr, offset, len);
               break;
       }
-
       fs_close();

And unrelated whitespace changes here as well.

Quoted 1 line
include/btrfs.h +0 -0

[snip]

include/btrfs.h line 46

Quoted 7 lines
include/btrfs.h +7 -0
+/*
+ * Extent structure: contains the mapping of some chunk of a file
+ * that is contiguous on disk.
+ */
+struct extent {
+    //sector_t    pstart;         /* Physical start sector */
+    __le64   pstart;

Fix please.

include/btrfs.h line 389

Quoted 7 lines
include/btrfs.h +7 -0
+/*
+ * Our definition of "not whitespace"
+ */
+static inline char not_whitespace(char c)
+{
+  return (unsigned char)c > ' ';
+}

Can't you just use isspace from <linux/ctypes.h> ?

Quoted 49 lines
include/crc32c.h new file +43 -0
new file mode 100644
index 0000000..d04916e
@@ -0,0 +1,48 @@
+/*
+ * Copied from Linux kernel crypto/crc32c.c
+ * Copyright (c) 2004 Cisco Systems, Inc.
+ * Copyright (c) 2008 Herbert Xu <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License as published by the Free
+ * Software Foundation; either version 2 of the License, or (at your option)
+ * any later version.
+ *
+ */
+
+/*
+ * This is the CRC-32C table
+ * Generated with:
+ * width = 32 bits
+ * poly = 0x1EDC6F41
+ * reflect input bytes = true
+ * reflect output bytes = true
+ */
+
+/*
+ * Steps through buffer one byte at at time, calculates reflected
+ * crc using table.
+ */
+
+static inline u32 crc32c_cal(u32 crc, const char *data, size_t length, u32 *crc32c_table)
+{
+     while (length--)
+             crc = crc32c_table[(u8)(crc ^ *data++)] ^ (crc >> 8);
+
+     return crc;
+}
+
+static inline void crc32c_init(u32 *crc32c_table, u32 pol)
+{
+     int i, j;
+     u32 v;
+     const u32 poly = pol; /* Bit-reflected CRC32C polynomial */
+
+     for (i = 0; i < 256; i++) {
+             v = i;
+             for (j = 0; j < 8; j++) {
Quoted 8 lines

include/crc32c.h line 48

Quoted 5 lines
include/crc32c.h new file +5 -0
+                     v = (v >> 1) ^ ((v & 1) ? poly : 0);
+             }
+             crc32c_table[i] = v;
+     }
+}

Simon, since you've just been over all the crc32 recently, do we have these functions somewhere already that can be easily tapped in to? Thanks!

Quoted 5 lines

Should be - see lib/crc32.c. There is already at least one other copy (ubifs I think) so we should try to avoid adding more.

Maybe the polynomial is different here? But even so it should go with the existing code I think.

I have tried using crc code lib/crc32.c but it always failed even
though i did change the polynomial but still result is search for file on btrfs fails due to bad crc calculation. I have also enable dynamic table creation but still result is same. so should add my my crc code under ifdef in lib/crc32.c what do you suggest?

Quoted 6 lines

Regards,
Simon

Quoted 2 lines

--
Tom

#10 lore raw

PATCH v3 1/1 Introduced btrfs file-system with btrload command

Hi Adnan,

Quoted 170 lines · Adnan Ali
Quoted 169 lines · Simon Glass

Hi,

Quoted 153 lines · Tom Rini
Quoted 7 lines · Adnan Ali

Introduces btrfs file-system to read file from volume/sub-volumes with btrload command. This implementation has read-only support.
This btrfs implementation is based on syslinux btrfs code, commit 269ebc845ebc8b46ef4b0be7fa0005c7fdb95b8d.

A few things:
- In general in fs/btrfs/btrfs.c I see some coding style problems (lack of spacing, non-printf's longer than 80-wide). Do these come from syslinux and thus will make any re-syncs easier?
- It looks like you added support for CONFIG_CMD_FS_GENERIC, if so did you test that?
- Can you please enable this support code on at least one platform, preferably the one you've tested and developed with?

[snip]

Quoted 2 lines
fs/fs.c +0 -0

[snip]

Quoted 2 lines

+ //file handle is valid get the size of the file

/* Style comments only */

Quoted 1 line

+ len=filedata.size;

And spacing between variable and assignment.

Quoted 10 lines
fs/fs.c +0 -1
@@ -178,7 +248,6 @@ int fs_set_blk_dev(const char *ifname, const char
 *dev_part_str, int fstype)
       for (i = 0; i < ARRAY_SIZE(fstypes); i++) {
               if ((fstype != FS_TYPE_ANY) && (fstype !=
 fstypes[i].fstype))
                       continue;
-
               if (!fstypes[i].probe()) {
                       fs_type = fstypes[i].fstype;
                       return 0;

[snip]

fs/fs.c static void fs_close(void) · line 285

Quoted 9 lines
fs/fs.c +0 -1
@@ -208,7 +280,6 @@ static void fs_close(void)
  int fs_ls(const char *dirname)
  {
       int ret;
-
       switch (fs_type) {
       case FS_TYPE_FAT:
               ret = fs_ls_fat(dirname);

Unrelated, please drop.

Quoted 4 lines
fs/fs.c +0 -0
@@ -237,11 +311,13 @@ int fs_read(const char *filename, ulong addr, int
 offset, int len)
       case FS_TYPE_EXT:
               ret = fs_read_ext(filename, addr, offset, len);

fs/fs.c int fs_read(const char *filename, ulong addr, int · line 322

Quoted 10 lines
fs/fs.c +3 -1
               break;
+        case FS_TYPE_BTR:
+             ret = fs_read_btr(filename, addr, offset, len);
+             break;
       default:
               ret = fs_read_unsupported(filename, addr, offset, len);
               break;
       }
-
       fs_close();

And unrelated whitespace changes here as well.

Quoted 1 line
include/btrfs.h +0 -0

[snip]

include/btrfs.h line 46

Quoted 8 lines
include/btrfs.h +7 -0
+/*
+ * Extent structure: contains the mapping of some chunk of a file
+ * that is contiguous on disk.
+ */
+struct extent {
+    //sector_t    pstart;         /* Physical start sector */
+    __le64   pstart;

Fix please.

include/btrfs.h line 389

Quoted 7 lines
include/btrfs.h +7 -0
+/*
+ * Our definition of "not whitespace"
+ */
+static inline char not_whitespace(char c)
+{
+  return (unsigned char)c > ' ';
+}

Can't you just use isspace from <linux/ctypes.h> ?

Quoted 54 lines
include/crc32c.h new file +6 -0
new file mode 100644
index 0000000..d04916e
@@ -0,0 +1,48 @@
+/*
+ * Copied from Linux kernel crypto/crc32c.c
+ * Copyright (c) 2004 Cisco Systems, Inc.
+ * Copyright (c) 2008 Herbert Xu <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or modify

it
+ * under the terms of the GNU General Public License as published by the Free
+ * Software Foundation; either version 2 of the License, or (at your option)
+ * any later version.
+ *
+ */

include/crc32c.h new file +16 -0
+
+/*
+ * This is the CRC-32C table
+ * Generated with:
+ * width = 32 bits
+ * poly = 0x1EDC6F41
+ * reflect input bytes = true
+ * reflect output bytes = true
+ */
+
+/*
+ * Steps through buffer one byte at at time, calculates reflected
+ * crc using table.
+ */
+
+static inline u32 crc32c_cal(u32 crc, const char *data, size_t length,

u32 *crc32c_table)

include/crc32c.h new file +17 -0
+{
+     while (length--)
+             crc = crc32c_table[(u8)(crc ^ *data++)] ^ (crc >> 8);
+
+     return crc;
+}
+
+static inline void crc32c_init(u32 *crc32c_table, u32 pol)
+{
+     int i, j;
+     u32 v;
+     const u32 poly = pol; /* Bit-reflected CRC32C polynomial */
+
+     for (i = 0; i < 256; i++) {
+             v = i;
+             for (j = 0; j < 8; j++) {
+                     v = (v >> 1) ^ ((v & 1) ? poly : 0);
Quoted 8 lines

include/crc32c.h line 48

Quoted 4 lines
include/crc32c.h new file +4 -0
+             }
+             crc32c_table[i] = v;
+     }
+}

Simon, since you've just been over all the crc32 recently, do we have these functions somewhere already that can be easily tapped in to? Thanks!

Should be - see lib/crc32.c. There is already at least one other copy (ubifs I think) so we should try to avoid adding more.

Quoted 8 lines
Quoted 2 lines

Maybe the polynomial is different here? But even so it should go with the existing code I think.

I have tried using crc code lib/crc32.c but it always failed even
though i did change the polynomial but still result is search for file on btrfs fails due to bad crc calculation. I have also enable dynamic table creation but still result is same. so should add my my crc code under ifdef in lib/crc32.c what do you suggest?

I suggest you pull the two functions out and try to see what is different. You might need to define DYNAMIC_CRC_TABLE if the polynomial has changed. 'sandbox' might help here (make sandbox_config) since you can build and run on your linux box.

If you have changed the polynomial and it still doesn't work then something is wrong...

I think adding the code to lib/crc32 with a new CONFIG is OK - but probably better if you put it in a new file like crc32_c.c or similar. Hopefully there is some common element to the algorithms?

Regards,
Simon

Quoted 9 lines
Quoted 7 lines

Regards,
Simon

Quoted 2 lines

--
Tom

#11 lore raw

PATCH v3 1/1 Introduced btrfs file-system with btrload command

In reply to Adnan Ali

Dear Adnan Ali,

Quoted 7 lines · In message <[email protected]> you…

Introduces btrfs file-system to read file from volume/sub-volumes with btrload command. This implementation has read-only support.
This btrfs implementation is based on syslinux btrfs code, commit 269ebc845ebc8b46ef4b0be7fa0005c7fdb95b8d.

Checkpatch says:

total: 410 errors, 527 warnings, 1849 lines checked

Can you fix these, please?

Also, you say this is patch version 3, but I see no history of changes anywhere. Please read
http://www.denx.de/wiki/view/U-Boot/Patches#Sending_updated_patch_versions

Best regards,

Wolfgang Denk

Signature · 5 lines
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
Let's say the docs present a simplified view of reality...    :-)
                      - Larry Wall in  <[email protected]>
#12 lore raw

PATCH v3 1/1 Introduced btrfs file-system with btrload command

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Quoted 10 lines · Wolfgang Denk

Dear Adnan Ali,

In message <[email protected]> you

Quoted 5 lines · wrote:

Introduces btrfs file-system to read file from volume/sub-volumes with btrload command. This implementation has read-only support. This btrfs implementation is based on syslinux btrfs code, commit 269ebc845ebc8b46ef4b0be7fa0005c7fdb95b8d.

Quoted 8 lines
Quoted 1 line

Checkpatch says:

total: 410 errors, 527 warnings, 1849 lines checked

Can you fix these, please?

Some number of these are following upstream errors which we should take so future resyncs aren't as painful.

- --
Tom
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/

iQIcBAEBAgAGBQJRMTI1AAoJENk4IS6UOR1WKtAQAJoMEuQM3HL3amkzbZfWA8TD wRr0bWMdWPUgzp+DqVggubYdKe7Yj3hrjlw+aXrjmFJmJNuO+vub+jAhMBOasMOG SU0A5rxjIebRwcGzwZcdCHZ0wc5960uJs/ZJa13D8TPUUlG30lL8dUVVpkrfiLnX frrGhraaL2xBaf3PXiZaFrQ1bicDrMIwwPdGadUl3h5/xmukFvmuteQZgMk4c/Sv 8R4r7G1zD+wmbIOY49GUCuetfncy4mQhtaUIXbicMUs7+RLEyA9PieknxxHbKx70 fm9utmQyzXyU/lE/cpxOpz112TzuQPFn+qanr6PzTBYsd46YqaJpw7CVO6u1xv31 KzBslg8/SSZmGEbGrkgrXPO0Ey/T3Gw9tfOaWM/2SPEumBsS1KDQWuo2yPH1LiZs O/b3VbjBiUTWvc/Djlov/9ebVzi1kTbqd8I0HSKC8qF2FwqmKU4O9qxHC2Gr9ioX 0E6rOA4w7OXN0wsfn3X7woq+l05dtT9ZOlFyNEm9BVhmeytge3RSWrvMkj6Ll+pv WF4hS6GW/xwY1Wttg2RHoNAGpkKa8pnk9BC0g3XTYyXdP7Qgb66wjoTfIK5TuXOB pKYMwiPb3b2XG3qY+zraXpS+dXed3/BD9rAbpov4YHrPDEnOAYIa+sQVgA6UZsod mJkTQrcYujpB5RFfHgh2
=RQIq
-----END PGP SIGNATURE-----

#13 lore raw

PATCH v3 1/1 Introduced btrfs file-system with btrload command

Dear Tom,

Quoted 5 lines · In message <[email protected]> you wrote:
Quoted 2 lines

Can you fix these, please?

Some number of these are following upstream errors which we should take so future resyncs aren't as painful.

Maybe - but then this should be explained in the patch, at least.

Best regards,

Wolfgang Denk

Signature · 5 lines
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
Doubt is a pain too lonely to know that faith is his twin brother.
- Kahlil Gibran
#14 lore raw

PATCH v3 1/1 Introduced btrfs file-system with btrload command

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Quoted 10 lines · Wolfgang Denk

Dear Tom,

Quoted 5 lines · In message <[email protected]> you wrote:
Quoted 2 lines

Can you fix these, please?

Some number of these are following upstream errors which we should take so future resyncs aren't as painful.

Maybe - but then this should be explained in the patch, at least.

Agreed.

- --
Tom
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/

iQIcBAEBAgAGBQJRMTWXAAoJENk4IS6UOR1WJ7EP/A8WCfiwQ6jpyvv39rJ8xp6h grJ7TNsCMApIMD5geor3ttifOw6omo2uOEPslvwqsBGkSrwMpeDTmpe75pWvb/AS iTStKIGiPeWJh09V5yyaT98WY8p3QCMrKjGPACOQX9LK/LgYuvtE2K5L+s6CZKXE S0WsHwslGlqjVia6Z0KCInLIiKmNLgC70Zyd9ASinGLWedhsSm6UE3l2rNxZKvXU H9D/PyDewlkLerFA6mcFAA5H+YQMBOHIjVpf3x7k/QYUHfQwU4LOls3Wn9Mn21sO ksOXnOCjYXv8jEhyZQSuW4CerBBwl210WeclDpMRLRaCzp3ayCyoR9shxqYoOOR+ f40mfrfkgmd9GYunorSC1KZ6TYlaRm3k//okk0ELRt0HjlcX3pvrRBJosmJW0Rcv pu3f268wc9lusTnJTcQ+XSLraxb21SzjHCeAoJd0EUwVirnrC6i0o7Na10VVFE56 h631W5kiyNYMevsCvZ282Eb+EFB8GAMkoroI8A2JHVjvh79c53D5WPRPQ1Sg8Fjs mG58gA17TGQgNFWM438vuppRKHqhumMQtjuzY8pAR9yweR3/KKjFD+/EUeVNd0qS NG9QW4WzaNkrMAuUIYFuSQy3Qth/nW0nk8IQplHCousNSRpWFMOwEU8tr0miQIsU EOL0UYD1QTdl4E6OLHOC
=JR43
-----END PGP SIGNATURE-----