Skip to contentSkip to conversation

Fix-up series for x86/master

PATCH New Simon Glass 2 participants 16 messages Apr 27, 2020 – Apr 29, 2020

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

cover Fix-up series for x86/master

Note that this includes two patches which are in mainline but not in x86/master

Simon Glass (3):
test: Add the beginnings of some string tests
lib: Add a function to convert a string to upper case
acpi: Fix-up patch to correct sandbox test errors

 arch/x86/lib/acpi_table.c |   5 --
 include/test/suites.h     |   1 +
 include/vsprintf.h        |  12 ++++
 lib/acpi/acpi_table.c     |   7 ++-
 lib/strto.c               |   8 +++
 test/Makefile             |   1 +
 test/cmd_ut.c             |   5 ++
 test/str_ut.c             | 115 ++++++++++++++++++++++++++++++++++++++
 8 files changed, 148 insertions(+), 6 deletions(-)
 create mode 100644 test/str_ut.c
Signature · 1 line
2.26.2.303.gf8c07b1a785-goog
#1 lore raw

PATCH 1/3 test: Add the beginnings of some string tests

There are quite a few string functions in U-Boot with no tests. Make a start by adding a test for strtoul().

---

 include/test/suites.h |  1 +
 test/Makefile         |  1 +
 test/cmd_ut.c         |  5 ++++
 test/str_ut.c         | 67 +++++++++++++++++++++++++++++++++++++++++++
 4 files changed, 74 insertions(+)
 create mode 100644 test/str_ut.c
include/test/suites.h +1 -0
index 39ad81a90f..213e3cee77 100644
@@ -33,6 +33,7 @@ int do_ut_lib(cmd_tbl_t *cmdtp, int flag, int argc, char * const argv[]);
 int do_ut_log(cmd_tbl_t *cmdtp, int flag, int argc, char * const argv[]);
 int do_ut_optee(cmd_tbl_t *cmdtp, int flag, int argc, char * const argv[]);
 int do_ut_overlay(cmd_tbl_t *cmdtp, int flag, int argc, char * const argv[]);
+int do_ut_str(cmd_tbl_t *cmdtp, int flag, int argc, char *const argv[]);
 int do_ut_time(cmd_tbl_t *cmdtp, int flag, int argc, char * const argv[]);
 int do_ut_unicode(cmd_tbl_t *cmdtp, int flag, int argc, char * const argv[]);
 
test/Makefile +1 -0
index 2971d0d87f..bab8f1a5c2 100644
@@ -8,6 +8,7 @@ obj-$(CONFIG_UNIT_TEST) += ut.o
 obj-$(CONFIG_SANDBOX) += command_ut.o
 obj-$(CONFIG_SANDBOX) += compression.o
 obj-$(CONFIG_SANDBOX) += print_ut.o
+obj-$(CONFIG_SANDBOX) += str_ut.o
 obj-$(CONFIG_UT_TIME) += time_ut.o
 obj-$(CONFIG_UT_UNICODE) += unicode_ut.o
 obj-y += log/
test/cmd_ut.c +5 -0
index 7fdcdbb1a6..bd20a69c55 100644
@@ -74,6 +74,8 @@ static cmd_tbl_t cmd_ut_sub[] = {
 			 "", ""),
 	U_BOOT_CMD_MKENT(bloblist, CONFIG_SYS_MAXARGS, 1, do_ut_bloblist,
 			 "", ""),
+	U_BOOT_CMD_MKENT(str, CONFIG_SYS_MAXARGS, 1, do_ut_str,
+			 "", ""),
 #endif
 };
 
@@ -137,6 +139,9 @@ static char ut_help_text[] =
 #ifdef CONFIG_UT_OVERLAY
 	"ut overlay [test-name]\n"
 #endif
+#ifdef CONFIG_SANDBOX
+	"ut str - Basic test of string functions\n"
+#endif
 #ifdef CONFIG_UT_TIME
 	"ut time - Very basic test of time functions\n"
 #endif
test/str_ut.c new file +67 -0

67 lines not on this page · raw message

Signature · 1 line
2.26.2.303.gf8c07b1a785-goog
#2 lore raw

PATCH 2/3 lib: Add a function to convert a string to upper case

In reply to Simon Glass

Add a helper function for this operation. Update the strtoul() tests to check upper case as well.

---

 include/vsprintf.h | 12 +++++++
 lib/strto.c        |  8 +++++
 test/str_ut.c      | 78 +++++++++++++++++++++++++++++++++++++---------
 3 files changed, 83 insertions(+), 15 deletions(-)
include/vsprintf.h +12 -0
index 56844dd2de..d9fb68add0 100644
@@ -222,4 +222,16 @@ bool str2long(const char *p, ulong *num);
  * @hz: Value to convert
  */
 char *strmhz(char *buf, unsigned long hz);
+
+/**
+ * str_to_upper() - Convert a string to upper case
+ *
+ * This simply uses toupper() on each character of the string.
+ *
+ * @in: String to convert (must be large enough to hold the output string)
+ * @out: Buffer to put converted string
+ * @len: Number of bytes available in @out (SIZE_MAX for all)
+ */
+void str_to_upper(const char *in, char *out, size_t len);
+
 #endif
lib/strto.c +8 -0
index 1ac2b09c72..37e1fbe63f 100644
@@ -176,3 +176,11 @@ long trailing_strtol(const char *str)
 {
 	return trailing_strtoln(str, NULL);
 }
+
+void str_to_upper(const char *in, char *out, size_t len)
+{
+	for (; len > 0 && *in; len--)
+		*out++ = toupper(*in++);
+	if (len)
+		*out = '\0';
+}
test/str_ut.c +63 -15

99 lines not on this page · raw message

Signature · 1 line
2.26.2.303.gf8c07b1a785-goog
#3 lore raw

PATCH 3/3 acpi: Fix-up patch to correct sandbox test errors

In reply to Simon Glass

Move the alignment code into acpi_setup_base_tables() so that test and production code are in alignment.

This brings x86/master into line with patch series v8.

---

 arch/x86/lib/acpi_table.c | 5 -----
 lib/acpi/acpi_table.c     | 7 ++++++-
 2 files changed, 6 insertions(+), 6 deletions(-)
arch/x86/lib/acpi_table.c +0 -5
index 600bde2f5f..13f1409de8 100644
@@ -375,11 +375,6 @@ ulong write_acpi_tables(ulong start_addr)
 	debug("ACPI: Writing ACPI tables at %lx\n", start_addr);
 
 	acpi_setup_base_tables(ctx, start);
-	/*
-	 * Per ACPI spec, the FACS table address must be aligned to a 64 byte
-	 * boundary (Windows checks this, but Linux does not).
-	 */
-	acpi_align64(ctx);
 
 	debug("ACPI:    * FACS\n");
 	facs = ctx->current;
lib/acpi/acpi_table.c +6 -1
index 5abf1cad50..1c253af3bf 100644
@@ -145,7 +145,7 @@ int acpi_add_table(struct acpi_ctx *ctx, void *table)
 	}
 
 	if (i >= entries_num) {
-		debug("ACPI: Error: too many tables\n");
+		log_err("ACPI: Error: too many tables\n");
 		return -E2BIG;
 	}
 
@@ -256,4 +256,9 @@ void acpi_setup_base_tables(struct acpi_ctx *ctx, void *start)
 	acpi_write_rsdp(ctx->rsdp, ctx->rsdt, ctx->xsdt);
 	acpi_write_rsdt(ctx->rsdt);
 	acpi_write_xsdt(ctx->xsdt);
+	/*
+	 * Per ACPI spec, the FACS table address must be aligned to a 64 byte
+	 * boundary (Windows checks this, but Linux does not).
+	 */
+	acpi_align64(ctx);
 }
Signature · 1 line
2.26.2.303.gf8c07b1a785-goog
#4 lore raw

PATCH 3/3 acpi: Fix-up patch to correct sandbox test errors

Hi Simon,

Quoted 43 lines · Simon Glass

Move the alignment code into acpi_setup_base_tables() so that test and production code are in alignment.

This brings x86/master into line with patch series v8.

---

 arch/x86/lib/acpi_table.c | 5 -----
 lib/acpi/acpi_table.c     | 7 ++++++-
 2 files changed, 6 insertions(+), 6 deletions(-)
arch/x86/lib/acpi_table.c +0 -5
index 600bde2f5f..13f1409de8 100644
@@ -375,11 +375,6 @@ ulong write_acpi_tables(ulong start_addr)
        debug("ACPI: Writing ACPI tables at %lx\n", start_addr);
 
        acpi_setup_base_tables(ctx, start);
-       /*
-        * Per ACPI spec, the FACS table address must be aligned to a 64 byte
-        * boundary (Windows checks this, but Linux does not).
-        */
-       acpi_align64(ctx);
 
        debug("ACPI:    * FACS\n");
        facs = ctx->current;
lib/acpi/acpi_table.c +1 -1
index 5abf1cad50..1c253af3bf 100644
@@ -145,7 +145,7 @@ int acpi_add_table(struct acpi_ctx *ctx, void *table)
        }
 
        if (i >= entries_num) {
-               debug("ACPI: Error: too many tables\n");
+               log_err("ACPI: Error: too many tables\n");
                return -E2BIG;
        }
 

lib/acpi/acpi_table.c void acpi_setup_base_tables(struct acpi_ctx *ctx, void *start)

Quoted 10 lines
lib/acpi/acpi_table.c +5 -1
@@ -256,4 +256,9 @@ void acpi_setup_base_tables(struct acpi_ctx *ctx, void *start)
        acpi_write_rsdp(ctx->rsdp, ctx->rsdt, ctx->xsdt);
        acpi_write_rsdt(ctx->rsdt);
        acpi_write_xsdt(ctx->xsdt);
+       /*
+        * Per ACPI spec, the FACS table address must be aligned to a 64 byte
+        * boundary (Windows checks this, but Linux does not).
+        */
+       acpi_align64(ctx);
 }
--

Could you please point out which commit in u-boot-x86/master should squash in this patch to fix the build error on sandbox?

Regards,
Bin

#5 lore raw

PATCH 3/3 acpi: Fix-up patch to correct sandbox test errors

Hi Bin,

Quoted 46 lines · Bin Meng

Hi Simon,

Quoted 42 lines · Simon Glass

Move the alignment code into acpi_setup_base_tables() so that test and production code are in alignment.

This brings x86/master into line with patch series v8.

---

 arch/x86/lib/acpi_table.c | 5 -----
 lib/acpi/acpi_table.c     | 7 ++++++-
 2 files changed, 6 insertions(+), 6 deletions(-)
arch/x86/lib/acpi_table.c +0 -5
index 600bde2f5f..13f1409de8 100644
@@ -375,11 +375,6 @@ ulong write_acpi_tables(ulong start_addr)
        debug("ACPI: Writing ACPI tables at %lx\n", start_addr);
 
        acpi_setup_base_tables(ctx, start);
-       /*
-        * Per ACPI spec, the FACS table address must be aligned to a 64 byte
-        * boundary (Windows checks this, but Linux does not).
-        */
-       acpi_align64(ctx);
 
        debug("ACPI:    * FACS\n");
        facs = ctx->current;
lib/acpi/acpi_table.c +1 -1
index 5abf1cad50..1c253af3bf 100644
@@ -145,7 +145,7 @@ int acpi_add_table(struct acpi_ctx *ctx, void *table)
        }
 
        if (i >= entries_num) {
-               debug("ACPI: Error: too many tables\n");
+               log_err("ACPI: Error: too many tables\n");
                return -E2BIG;
        }
 
Quoted 14 lines

lib/acpi/acpi_table.c void acpi_setup_base_tables(struct acpi_ctx *ctx, void *start)

Quoted 11 lines
lib/acpi/acpi_table.c +5 -1
@@ -256,4 +256,9 @@ void acpi_setup_base_tables(struct acpi_ctx *ctx, void *start)
        acpi_write_rsdp(ctx->rsdp, ctx->rsdt, ctx->xsdt);
        acpi_write_rsdt(ctx->rsdt);
        acpi_write_xsdt(ctx->xsdt);
+       /*
+        * Per ACPI spec, the FACS table address must be aligned to a 64 byte
+        * boundary (Windows checks this, but Linux does not).
+        */
+       acpi_align64(ctx);
 }
--

Could you please point out which commit in u-boot-x86/master should squash in this patch to fix the build error on sandbox?

It might be easier to pick up v8 in that case. I think there are three patches that need to change because of conflicts caused by the first one.

So can you pick up the v8 patches? Also you do need to rebase on master because of the str_to_upper patches.

Regards,
Simon

#6 lore raw

PATCH 3/3 acpi: Fix-up patch to correct sandbox test errors

Hi Simon,

Quoted 65 lines · Simon Glass

Hi Bin,

Quoted 46 lines · Bin Meng

Hi Simon,

Quoted 42 lines · Simon Glass

Move the alignment code into acpi_setup_base_tables() so that test and production code are in alignment.

This brings x86/master into line with patch series v8.

---

 arch/x86/lib/acpi_table.c | 5 -----
 lib/acpi/acpi_table.c     | 7 ++++++-
 2 files changed, 6 insertions(+), 6 deletions(-)
arch/x86/lib/acpi_table.c +0 -5
index 600bde2f5f..13f1409de8 100644
@@ -375,11 +375,6 @@ ulong write_acpi_tables(ulong start_addr)
        debug("ACPI: Writing ACPI tables at %lx\n", start_addr);
 
        acpi_setup_base_tables(ctx, start);
-       /*
-        * Per ACPI spec, the FACS table address must be aligned to a 64 byte
-        * boundary (Windows checks this, but Linux does not).
-        */
-       acpi_align64(ctx);
 
        debug("ACPI:    * FACS\n");
        facs = ctx->current;
lib/acpi/acpi_table.c +1 -1
index 5abf1cad50..1c253af3bf 100644
@@ -145,7 +145,7 @@ int acpi_add_table(struct acpi_ctx *ctx, void *table)
        }
 
        if (i >= entries_num) {
-               debug("ACPI: Error: too many tables\n");
+               log_err("ACPI: Error: too many tables\n");
                return -E2BIG;
        }
 
Quoted 14 lines

lib/acpi/acpi_table.c void acpi_setup_base_tables(struct acpi_ctx *ctx, void *start)

Quoted 11 lines
lib/acpi/acpi_table.c +5 -1
@@ -256,4 +256,9 @@ void acpi_setup_base_tables(struct acpi_ctx *ctx, void *start)
        acpi_write_rsdp(ctx->rsdp, ctx->rsdt, ctx->xsdt);
        acpi_write_rsdt(ctx->rsdt);
        acpi_write_xsdt(ctx->xsdt);
+       /*
+        * Per ACPI spec, the FACS table address must be aligned to a 64 byte
+        * boundary (Windows checks this, but Linux does not).
+        */
+       acpi_align64(ctx);
 }
--

Could you please point out which commit in u-boot-x86/master should squash in this patch to fix the build error on sandbox?

Quoted 6 lines

It might be easier to pick up v8 in that case. I think there are three patches that need to change because of conflicts caused by the first one.

So can you pick up the v8 patches? Also you do need to rebase on master because of the str_to_upper patches.

But v8 is a complete new series for part B? I think we'd better re-tag v8 as v1. It's quite confusing.

Regards,
Bin

#7 lore raw

PATCH 3/3 acpi: Fix-up patch to correct sandbox test errors

Hi Bin,

Quoted 73 lines · Bin Meng

Hi Simon,

Quoted 69 lines · Simon Glass

Hi Bin,

Quoted 46 lines · Bin Meng

Hi Simon,

Quoted 42 lines · Simon Glass

Move the alignment code into acpi_setup_base_tables() so that test and production code are in alignment.

This brings x86/master into line with patch series v8.

---

 arch/x86/lib/acpi_table.c | 5 -----
 lib/acpi/acpi_table.c     | 7 ++++++-
 2 files changed, 6 insertions(+), 6 deletions(-)
arch/x86/lib/acpi_table.c +0 -5
index 600bde2f5f..13f1409de8 100644
@@ -375,11 +375,6 @@ ulong write_acpi_tables(ulong start_addr)
        debug("ACPI: Writing ACPI tables at %lx\n", start_addr);
 
        acpi_setup_base_tables(ctx, start);
-       /*
-        * Per ACPI spec, the FACS table address must be aligned to a 64 byte
-        * boundary (Windows checks this, but Linux does not).
-        */
-       acpi_align64(ctx);
 
        debug("ACPI:    * FACS\n");
        facs = ctx->current;
lib/acpi/acpi_table.c +1 -1
index 5abf1cad50..1c253af3bf 100644
@@ -145,7 +145,7 @@ int acpi_add_table(struct acpi_ctx *ctx, void *table)
        }
 
        if (i >= entries_num) {
-               debug("ACPI: Error: too many tables\n");
+               log_err("ACPI: Error: too many tables\n");
                return -E2BIG;
        }
 
Quoted 14 lines

lib/acpi/acpi_table.c void acpi_setup_base_tables(struct acpi_ctx *ctx, void *start)

Quoted 11 lines
lib/acpi/acpi_table.c +5 -1
@@ -256,4 +256,9 @@ void acpi_setup_base_tables(struct acpi_ctx *ctx, void *start)
        acpi_write_rsdp(ctx->rsdp, ctx->rsdt, ctx->xsdt);
        acpi_write_rsdt(ctx->rsdt);
        acpi_write_xsdt(ctx->xsdt);
+       /*
+        * Per ACPI spec, the FACS table address must be aligned to a 64 byte
+        * boundary (Windows checks this, but Linux does not).
+        */
+       acpi_align64(ctx);
 }
--

Could you please point out which commit in u-boot-x86/master should squash in this patch to fix the build error on sandbox?

It might be easier to pick up v8 in that case. I think there are three patches that need to change because of conflicts caused by the first one.

Quoted 5 lines
Quoted 2 lines

So can you pick up the v8 patches? Also you do need to rebase on master because of the str_to_upper patches.

But v8 is a complete new series for part B? I think we'd better re-tag v8 as v1. It's quite confusing.

No I mean the part A series. Let's get that applied and then we will be in a brave new world.

I could resent part B as v1 if you like, but note that it has some reviewed-by tags and some v3, etc. comments. The patches languished for quite a while which is why I decided to try to split them up.

Regards,
Simon

#8 lore raw

PATCH 3/3 acpi: Fix-up patch to correct sandbox test errors

Hi Simon,

Quoted 77 lines · Simon Glass

Hi Bin,

Quoted 73 lines · Bin Meng

Hi Simon,

Quoted 69 lines · Simon Glass

Hi Bin,

Quoted 46 lines · Bin Meng

Hi Simon,

Quoted 42 lines · Simon Glass

Move the alignment code into acpi_setup_base_tables() so that test and production code are in alignment.

This brings x86/master into line with patch series v8.

---

 arch/x86/lib/acpi_table.c | 5 -----
 lib/acpi/acpi_table.c     | 7 ++++++-
 2 files changed, 6 insertions(+), 6 deletions(-)
arch/x86/lib/acpi_table.c +0 -5
index 600bde2f5f..13f1409de8 100644
@@ -375,11 +375,6 @@ ulong write_acpi_tables(ulong start_addr)
        debug("ACPI: Writing ACPI tables at %lx\n", start_addr);
 
        acpi_setup_base_tables(ctx, start);
-       /*
-        * Per ACPI spec, the FACS table address must be aligned to a 64 byte
-        * boundary (Windows checks this, but Linux does not).
-        */
-       acpi_align64(ctx);
 
        debug("ACPI:    * FACS\n");
        facs = ctx->current;
lib/acpi/acpi_table.c +1 -1
index 5abf1cad50..1c253af3bf 100644
@@ -145,7 +145,7 @@ int acpi_add_table(struct acpi_ctx *ctx, void *table)
        }
 
        if (i >= entries_num) {
-               debug("ACPI: Error: too many tables\n");
+               log_err("ACPI: Error: too many tables\n");
                return -E2BIG;
        }
 
Quoted 14 lines

lib/acpi/acpi_table.c void acpi_setup_base_tables(struct acpi_ctx *ctx, void *start)

Quoted 11 lines
lib/acpi/acpi_table.c +5 -1
@@ -256,4 +256,9 @@ void acpi_setup_base_tables(struct acpi_ctx *ctx, void *start)
        acpi_write_rsdp(ctx->rsdp, ctx->rsdt, ctx->xsdt);
        acpi_write_rsdt(ctx->rsdt);
        acpi_write_xsdt(ctx->xsdt);
+       /*
+        * Per ACPI spec, the FACS table address must be aligned to a 64 byte
+        * boundary (Windows checks this, but Linux does not).
+        */
+       acpi_align64(ctx);
 }
--

Could you please point out which commit in u-boot-x86/master should squash in this patch to fix the build error on sandbox?

It might be easier to pick up v8 in that case. I think there are three patches that need to change because of conflicts caused by the first one.

Quoted 8 lines
Quoted 5 lines
Quoted 2 lines

So can you pick up the v8 patches? Also you do need to rebase on master because of the str_to_upper patches.

But v8 is a complete new series for part B? I think we'd better re-tag v8 as v1. It's quite confusing.

No I mean the part A series. Let's get that applied and then we will be in a brave new world.

Is the v8 this one?
http://patchwork.ozlabs.org/user/todo/uboot/?series=172777

The part A series are already applied, but it has a build error as I mentioned.

Quoted 4 lines

I could resent part B as v1 if you like, but note that it has some reviewed-by tags and some v3, etc. comments. The patches languished for quite a while which is why I decided to try to split them up.

Regards,
Bin

#9 lore raw

PATCH 3/3 acpi: Fix-up patch to correct sandbox test errors

Hi Bin,

Quoted 84 lines · Bin Meng

Hi Simon,

Quoted 80 lines · Simon Glass

Hi Bin,

Quoted 76 lines · Bin Meng

Hi Simon,

Quoted 71 lines · Simon Glass

Hi Bin,

Quoted 46 lines · Bin Meng

Hi Simon,

Quoted 42 lines · Simon Glass

Move the alignment code into acpi_setup_base_tables() so that test and production code are in alignment.

This brings x86/master into line with patch series v8.

---

 arch/x86/lib/acpi_table.c | 5 -----
 lib/acpi/acpi_table.c     | 7 ++++++-
 2 files changed, 6 insertions(+), 6 deletions(-)
arch/x86/lib/acpi_table.c +0 -5
index 600bde2f5f..13f1409de8 100644
@@ -375,11 +375,6 @@ ulong write_acpi_tables(ulong start_addr)
        debug("ACPI: Writing ACPI tables at %lx\n", start_addr);
 
        acpi_setup_base_tables(ctx, start);
-       /*
-        * Per ACPI spec, the FACS table address must be aligned to a 64 byte
-        * boundary (Windows checks this, but Linux does not).
-        */
-       acpi_align64(ctx);
 
        debug("ACPI:    * FACS\n");
        facs = ctx->current;
lib/acpi/acpi_table.c +1 -1
index 5abf1cad50..1c253af3bf 100644
@@ -145,7 +145,7 @@ int acpi_add_table(struct acpi_ctx *ctx, void *table)
        }
 
        if (i >= entries_num) {
-               debug("ACPI: Error: too many tables\n");
+               log_err("ACPI: Error: too many tables\n");
                return -E2BIG;
        }
 
Quoted 14 lines

lib/acpi/acpi_table.c void acpi_setup_base_tables(struct acpi_ctx *ctx, void *start)

Quoted 11 lines
lib/acpi/acpi_table.c +5 -1
@@ -256,4 +256,9 @@ void acpi_setup_base_tables(struct acpi_ctx *ctx, void *start)
        acpi_write_rsdp(ctx->rsdp, ctx->rsdt, ctx->xsdt);
        acpi_write_rsdt(ctx->rsdt);
        acpi_write_xsdt(ctx->xsdt);
+       /*
+        * Per ACPI spec, the FACS table address must be aligned to a 64 byte
+        * boundary (Windows checks this, but Linux does not).
+        */
+       acpi_align64(ctx);
 }
--

Could you please point out which commit in u-boot-x86/master should squash in this patch to fix the build error on sandbox?

It might be easier to pick up v8 in that case. I think there are three patches that need to change because of conflicts caused by the first one.

So can you pick up the v8 patches? Also you do need to rebase on master because of the str_to_upper patches.

Quoted 8 lines
Quoted 5 lines
Quoted 2 lines

But v8 is a complete new series for part B? I think we'd better re-tag v8 as v1. It's quite confusing.

No I mean the part A series. Let's get that applied and then we will be in a brave new world.

Is the v8 this one?
http://patchwork.ozlabs.org/user/todo/uboot/?series=172777

No it is this one:

http://patchwork.ozlabs.org/project/uboot/list/?series=172776

Quoted 2 lines

The part A series are already applied, but it has a build error as I mentioned.

Right, but the options are to do a fixup patch or to pick up the v8 series instead. Unfortunately there are merge conflicts in patches so if you just pick up one patch it won't apply cleanly.

Quoted 6 lines
Quoted 5 lines

I could resent part B as v1 if you like, but note that it has some reviewed-by tags and some v3, etc. comments. The patches languished for quite a while which is why I decided to try to split them up.

Regards,
Simon

#10 lore raw

PATCH 3/3 acpi: Fix-up patch to correct sandbox test errors

Hi Simon,

Quoted 99 lines · Simon Glass

Hi Bin,

Quoted 84 lines · Bin Meng

Hi Simon,

Quoted 80 lines · Simon Glass

Hi Bin,

Quoted 76 lines · Bin Meng

Hi Simon,

Quoted 71 lines · Simon Glass

Hi Bin,

Quoted 46 lines · Bin Meng

Hi Simon,

Quoted 42 lines · Simon Glass

Move the alignment code into acpi_setup_base_tables() so that test and production code are in alignment.

This brings x86/master into line with patch series v8.

---

 arch/x86/lib/acpi_table.c | 5 -----
 lib/acpi/acpi_table.c     | 7 ++++++-
 2 files changed, 6 insertions(+), 6 deletions(-)
arch/x86/lib/acpi_table.c +0 -5
index 600bde2f5f..13f1409de8 100644
@@ -375,11 +375,6 @@ ulong write_acpi_tables(ulong start_addr)
        debug("ACPI: Writing ACPI tables at %lx\n", start_addr);
 
        acpi_setup_base_tables(ctx, start);
-       /*
-        * Per ACPI spec, the FACS table address must be aligned to a 64 byte
-        * boundary (Windows checks this, but Linux does not).
-        */
-       acpi_align64(ctx);
 
        debug("ACPI:    * FACS\n");
        facs = ctx->current;
lib/acpi/acpi_table.c +1 -1
index 5abf1cad50..1c253af3bf 100644
@@ -145,7 +145,7 @@ int acpi_add_table(struct acpi_ctx *ctx, void *table)
        }
 
        if (i >= entries_num) {
-               debug("ACPI: Error: too many tables\n");
+               log_err("ACPI: Error: too many tables\n");
                return -E2BIG;
        }
 
Quoted 14 lines

lib/acpi/acpi_table.c void acpi_setup_base_tables(struct acpi_ctx *ctx, void *start)

Quoted 11 lines
lib/acpi/acpi_table.c +5 -1
@@ -256,4 +256,9 @@ void acpi_setup_base_tables(struct acpi_ctx *ctx, void *start)
        acpi_write_rsdp(ctx->rsdp, ctx->rsdt, ctx->xsdt);
        acpi_write_rsdt(ctx->rsdt);
        acpi_write_xsdt(ctx->xsdt);
+       /*
+        * Per ACPI spec, the FACS table address must be aligned to a 64 byte
+        * boundary (Windows checks this, but Linux does not).
+        */
+       acpi_align64(ctx);
 }
--

Could you please point out which commit in u-boot-x86/master should squash in this patch to fix the build error on sandbox?

It might be easier to pick up v8 in that case. I think there are three patches that need to change because of conflicts caused by the first one.

So can you pick up the v8 patches? Also you do need to rebase on master because of the str_to_upper patches.

Quoted 8 lines
Quoted 5 lines
Quoted 2 lines

But v8 is a complete new series for part B? I think we'd better re-tag v8 as v1. It's quite confusing.

No I mean the part A series. Let's get that applied and then we will be in a brave new world.

Is the v8 this one?
http://patchwork.ozlabs.org/user/todo/uboot/?series=172777

No it is this one:

Quoted 9 lines

http://patchwork.ozlabs.org/project/uboot/list/?series=172776

Quoted 2 lines

The part A series are already applied, but it has a build error as I mentioned.

Right, but the options are to do a fixup patch or to pick up the v8 series instead. Unfortunately there are merge conflicts in patches so if you just pick up one patch it won't apply cleanly.

But we still need maintain bisectablity right? To do that we need a fix patch that can be squashed into the u-boot-x86/master. Picking up the v8 series does not fix the bisectiablity I think.

Quoted 6 lines
Quoted 6 lines
Quoted 5 lines

I could resent part B as v1 if you like, but note that it has some reviewed-by tags and some v3, etc. comments. The patches languished for quite a while which is why I decided to try to split them up.

Regards,
Bin

#11 lore raw

PATCH 3/3 acpi: Fix-up patch to correct sandbox test errors

Hi Bin,

Quoted 108 lines · Bin Meng

Hi Simon,

Quoted 104 lines · Simon Glass

Hi Bin,

Quoted 84 lines · Bin Meng

Hi Simon,

Quoted 80 lines · Simon Glass

Hi Bin,

Quoted 76 lines · Bin Meng

Hi Simon,

Quoted 71 lines · Simon Glass

Hi Bin,

Quoted 46 lines · Bin Meng

Hi Simon,

Quoted 42 lines · Simon Glass

Move the alignment code into acpi_setup_base_tables() so that test and production code are in alignment.

This brings x86/master into line with patch series v8.

---

 arch/x86/lib/acpi_table.c | 5 -----
 lib/acpi/acpi_table.c     | 7 ++++++-
 2 files changed, 6 insertions(+), 6 deletions(-)
arch/x86/lib/acpi_table.c +0 -5
index 600bde2f5f..13f1409de8 100644
@@ -375,11 +375,6 @@ ulong write_acpi_tables(ulong start_addr)
        debug("ACPI: Writing ACPI tables at %lx\n", start_addr);
 
        acpi_setup_base_tables(ctx, start);
-       /*
-        * Per ACPI spec, the FACS table address must be aligned to a 64 byte
-        * boundary (Windows checks this, but Linux does not).
-        */
-       acpi_align64(ctx);
 
        debug("ACPI:    * FACS\n");
        facs = ctx->current;
lib/acpi/acpi_table.c +1 -1
index 5abf1cad50..1c253af3bf 100644
@@ -145,7 +145,7 @@ int acpi_add_table(struct acpi_ctx *ctx, void *table)
        }
 
        if (i >= entries_num) {
-               debug("ACPI: Error: too many tables\n");
+               log_err("ACPI: Error: too many tables\n");
                return -E2BIG;
        }
 
Quoted 14 lines

lib/acpi/acpi_table.c void acpi_setup_base_tables(struct acpi_ctx *ctx, void *start)

Quoted 11 lines
lib/acpi/acpi_table.c +5 -1
@@ -256,4 +256,9 @@ void acpi_setup_base_tables(struct acpi_ctx *ctx, void *start)
        acpi_write_rsdp(ctx->rsdp, ctx->rsdt, ctx->xsdt);
        acpi_write_rsdt(ctx->rsdt);
        acpi_write_xsdt(ctx->xsdt);
+       /*
+        * Per ACPI spec, the FACS table address must be aligned to a 64 byte
+        * boundary (Windows checks this, but Linux does not).
+        */
+       acpi_align64(ctx);
 }
--

Could you please point out which commit in u-boot-x86/master should squash in this patch to fix the build error on sandbox?

It might be easier to pick up v8 in that case. I think there are three patches that need to change because of conflicts caused by the first one.

So can you pick up the v8 patches? Also you do need to rebase on master because of the str_to_upper patches.

Quoted 8 lines
Quoted 5 lines
Quoted 2 lines

But v8 is a complete new series for part B? I think we'd better re-tag v8 as v1. It's quite confusing.

No I mean the part A series. Let's get that applied and then we will be in a brave new world.

Is the v8 this one?
http://patchwork.ozlabs.org/user/todo/uboot/?series=172777

No it is this one:

http://patchwork.ozlabs.org/project/uboot/list/?series=172776

Quoted 2 lines

The part A series are already applied, but it has a build error as I mentioned.

Quoted 8 lines
Quoted 4 lines

Right, but the options are to do a fixup patch or to pick up the v8 series instead. Unfortunately there are merge conflicts in patches so if you just pick up one patch it won't apply cleanly.

But we still need maintain bisectablity right? To do that we need a fix patch that can be squashed into the u-boot-x86/master. Picking up the v8 series does not fix the bisectiablity I think.

But my suggestion is to pick up the v8 patch *and drop what you currently have in x86/master*. Otherwise, as you say, bisect might not work.

I should have objected at the time to Andy refactoring patch, but I wasn't sure when my stuff would land so it did not seem reasonable to do so.

Regards,
SImon

#12 lore raw

PATCH 3/3 acpi: Fix-up patch to correct sandbox test errors

Hi Simon,

Quoted 117 lines · Simon Glass

Hi Bin,

Quoted 113 lines · Bin Meng

Hi Simon,

Quoted 108 lines · Simon Glass

Hi Bin,

Quoted 84 lines · Bin Meng

Hi Simon,

Quoted 80 lines · Simon Glass

Hi Bin,

Quoted 76 lines · Bin Meng

Hi Simon,

Quoted 71 lines · Simon Glass

Hi Bin,

Quoted 46 lines · Bin Meng

Hi Simon,

Quoted 42 lines · Simon Glass

Move the alignment code into acpi_setup_base_tables() so that test and production code are in alignment.

This brings x86/master into line with patch series v8.

---

 arch/x86/lib/acpi_table.c | 5 -----
 lib/acpi/acpi_table.c     | 7 ++++++-
 2 files changed, 6 insertions(+), 6 deletions(-)
arch/x86/lib/acpi_table.c +0 -5
index 600bde2f5f..13f1409de8 100644
@@ -375,11 +375,6 @@ ulong write_acpi_tables(ulong start_addr)
        debug("ACPI: Writing ACPI tables at %lx\n", start_addr);
 
        acpi_setup_base_tables(ctx, start);
-       /*
-        * Per ACPI spec, the FACS table address must be aligned to a 64 byte
-        * boundary (Windows checks this, but Linux does not).
-        */
-       acpi_align64(ctx);
 
        debug("ACPI:    * FACS\n");
        facs = ctx->current;
lib/acpi/acpi_table.c +1 -1
index 5abf1cad50..1c253af3bf 100644
@@ -145,7 +145,7 @@ int acpi_add_table(struct acpi_ctx *ctx, void *table)
        }
 
        if (i >= entries_num) {
-               debug("ACPI: Error: too many tables\n");
+               log_err("ACPI: Error: too many tables\n");
                return -E2BIG;
        }
 
Quoted 14 lines

lib/acpi/acpi_table.c void acpi_setup_base_tables(struct acpi_ctx *ctx, void *start)

Quoted 11 lines
lib/acpi/acpi_table.c +5 -1
@@ -256,4 +256,9 @@ void acpi_setup_base_tables(struct acpi_ctx *ctx, void *start)
        acpi_write_rsdp(ctx->rsdp, ctx->rsdt, ctx->xsdt);
        acpi_write_rsdt(ctx->rsdt);
        acpi_write_xsdt(ctx->xsdt);
+       /*
+        * Per ACPI spec, the FACS table address must be aligned to a 64 byte
+        * boundary (Windows checks this, but Linux does not).
+        */
+       acpi_align64(ctx);
 }
--

Could you please point out which commit in u-boot-x86/master should squash in this patch to fix the build error on sandbox?

It might be easier to pick up v8 in that case. I think there are three patches that need to change because of conflicts caused by the first one.

So can you pick up the v8 patches? Also you do need to rebase on master because of the str_to_upper patches.

Quoted 8 lines
Quoted 5 lines
Quoted 2 lines

But v8 is a complete new series for part B? I think we'd better re-tag v8 as v1. It's quite confusing.

No I mean the part A series. Let's get that applied and then we will be in a brave new world.

Is the v8 this one?
http://patchwork.ozlabs.org/user/todo/uboot/?series=172777

No it is this one:

http://patchwork.ozlabs.org/project/uboot/list/?series=172776

Quoted 2 lines

The part A series are already applied, but it has a build error as I mentioned.

Right, but the options are to do a fixup patch or to pick up the v8 series instead. Unfortunately there are merge conflicts in patches so if you just pick up one patch it won't apply cleanly.

Quoted 7 lines
Quoted 3 lines

But we still need maintain bisectablity right? To do that we need a fix patch that can be squashed into the u-boot-x86/master. Picking up the v8 series does not fix the bisectiablity I think.

But my suggestion is to pick up the v8 patch *and drop what you currently have in x86/master*. Otherwise, as you say, bisect might not work.

Oops, I missed the v8 patch in patchwork (not assigned to me yet)

Quoted 4 lines

I should have objected at the time to Andy refactoring patch, but I wasn't sure when my stuff would land so it did not seem reasonable to do so.

I will drop what is applied in x86/master, and apply the v8 part A series. But it's better to re-tag v8 part B as v1 part B. I got quite confused until now :)

Regards,
Bin

#13 lore raw

PATCH 3/3 acpi: Fix-up patch to correct sandbox test errors

Hi Bin,

Quoted 132 lines · Bin Meng

Hi Simon,

Quoted 117 lines · Simon Glass

Hi Bin,

Quoted 113 lines · Bin Meng

Hi Simon,

Quoted 108 lines · Simon Glass

Hi Bin,

Quoted 84 lines · Bin Meng

Hi Simon,

Quoted 80 lines · Simon Glass

Hi Bin,

Quoted 76 lines · Bin Meng

Hi Simon,

Quoted 71 lines · Simon Glass

Hi Bin,

Quoted 46 lines · Bin Meng

Hi Simon,

Quoted 42 lines · Simon Glass

Move the alignment code into acpi_setup_base_tables() so that test and production code are in alignment.

This brings x86/master into line with patch series v8.

---

 arch/x86/lib/acpi_table.c | 5 -----
 lib/acpi/acpi_table.c     | 7 ++++++-
 2 files changed, 6 insertions(+), 6 deletions(-)
arch/x86/lib/acpi_table.c +0 -5
index 600bde2f5f..13f1409de8 100644
@@ -375,11 +375,6 @@ ulong write_acpi_tables(ulong start_addr)
        debug("ACPI: Writing ACPI tables at %lx\n", start_addr);
 
        acpi_setup_base_tables(ctx, start);
-       /*
-        * Per ACPI spec, the FACS table address must be aligned to a 64 byte
-        * boundary (Windows checks this, but Linux does not).
-        */
-       acpi_align64(ctx);
 
        debug("ACPI:    * FACS\n");
        facs = ctx->current;
lib/acpi/acpi_table.c +1 -1
index 5abf1cad50..1c253af3bf 100644
@@ -145,7 +145,7 @@ int acpi_add_table(struct acpi_ctx *ctx, void *table)
        }
 
        if (i >= entries_num) {
-               debug("ACPI: Error: too many tables\n");
+               log_err("ACPI: Error: too many tables\n");
                return -E2BIG;
        }
 
Quoted 14 lines

lib/acpi/acpi_table.c void acpi_setup_base_tables(struct acpi_ctx *ctx, void *start)

Quoted 11 lines
lib/acpi/acpi_table.c +5 -1
@@ -256,4 +256,9 @@ void acpi_setup_base_tables(struct acpi_ctx *ctx, void *start)
        acpi_write_rsdp(ctx->rsdp, ctx->rsdt, ctx->xsdt);
        acpi_write_rsdt(ctx->rsdt);
        acpi_write_xsdt(ctx->xsdt);
+       /*
+        * Per ACPI spec, the FACS table address must be aligned to a 64 byte
+        * boundary (Windows checks this, but Linux does not).
+        */
+       acpi_align64(ctx);
 }
--

Could you please point out which commit in u-boot-x86/master should squash in this patch to fix the build error on sandbox?

It might be easier to pick up v8 in that case. I think there are three patches that need to change because of conflicts caused by the first one.

So can you pick up the v8 patches? Also you do need to rebase on master because of the str_to_upper patches.

Quoted 8 lines
Quoted 5 lines
Quoted 2 lines

But v8 is a complete new series for part B? I think we'd better re-tag v8 as v1. It's quite confusing.

No I mean the part A series. Let's get that applied and then we will be in a brave new world.

Is the v8 this one?
http://patchwork.ozlabs.org/user/todo/uboot/?series=172777

No it is this one:

http://patchwork.ozlabs.org/project/uboot/list/?series=172776

Quoted 2 lines

The part A series are already applied, but it has a build error as I mentioned.

Right, but the options are to do a fixup patch or to pick up the v8 series instead. Unfortunately there are merge conflicts in patches so if you just pick up one patch it won't apply cleanly.

Quoted 7 lines
Quoted 3 lines

But we still need maintain bisectablity right? To do that we need a fix patch that can be squashed into the u-boot-x86/master. Picking up the v8 series does not fix the bisectiablity I think.

But my suggestion is to pick up the v8 patch *and drop what you currently have in x86/master*. Otherwise, as you say, bisect might not work.

Oops, I missed the v8 patch in patchwork (not assigned to me yet)

Quoted 1 line
Quoted 7 lines
Quoted 3 lines

I should have objected at the time to Andy refactoring patch, but I wasn't sure when my stuff would land so it did not seem reasonable to do so.

I will drop what is applied in x86/master, and apply the v8 part A series. But it's better to re-tag v8 part B as v1 part B. I got quite confused until now :)

OK I see. Shall I send out part B as v1 now? Or wait until there are comments and then send v2?

Regards,
Simon

#14 lore raw

PATCH 3/3 acpi: Fix-up patch to correct sandbox test errors

Hi Simon,

Quoted 140 lines · Simon Glass

Hi Bin,

Quoted 136 lines · Bin Meng

Hi Simon,

Quoted 117 lines · Simon Glass

Hi Bin,

Quoted 113 lines · Bin Meng

Hi Simon,

Quoted 108 lines · Simon Glass

Hi Bin,

Quoted 84 lines · Bin Meng

Hi Simon,

Quoted 80 lines · Simon Glass

Hi Bin,

Quoted 76 lines · Bin Meng

Hi Simon,

Quoted 71 lines · Simon Glass

Hi Bin,

Quoted 46 lines · Bin Meng

Hi Simon,

Quoted 42 lines · Simon Glass

Move the alignment code into acpi_setup_base_tables() so that test and production code are in alignment.

This brings x86/master into line with patch series v8.

---

 arch/x86/lib/acpi_table.c | 5 -----
 lib/acpi/acpi_table.c     | 7 ++++++-
 2 files changed, 6 insertions(+), 6 deletions(-)
arch/x86/lib/acpi_table.c +0 -5
index 600bde2f5f..13f1409de8 100644
@@ -375,11 +375,6 @@ ulong write_acpi_tables(ulong start_addr)
        debug("ACPI: Writing ACPI tables at %lx\n", start_addr);
 
        acpi_setup_base_tables(ctx, start);
-       /*
-        * Per ACPI spec, the FACS table address must be aligned to a 64 byte
-        * boundary (Windows checks this, but Linux does not).
-        */
-       acpi_align64(ctx);
 
        debug("ACPI:    * FACS\n");
        facs = ctx->current;
lib/acpi/acpi_table.c +1 -1
index 5abf1cad50..1c253af3bf 100644
@@ -145,7 +145,7 @@ int acpi_add_table(struct acpi_ctx *ctx, void *table)
        }
 
        if (i >= entries_num) {
-               debug("ACPI: Error: too many tables\n");
+               log_err("ACPI: Error: too many tables\n");
                return -E2BIG;
        }
 
Quoted 14 lines

lib/acpi/acpi_table.c void acpi_setup_base_tables(struct acpi_ctx *ctx, void *start)

Quoted 11 lines
lib/acpi/acpi_table.c +5 -1
@@ -256,4 +256,9 @@ void acpi_setup_base_tables(struct acpi_ctx *ctx, void *start)
        acpi_write_rsdp(ctx->rsdp, ctx->rsdt, ctx->xsdt);
        acpi_write_rsdt(ctx->rsdt);
        acpi_write_xsdt(ctx->xsdt);
+       /*
+        * Per ACPI spec, the FACS table address must be aligned to a 64 byte
+        * boundary (Windows checks this, but Linux does not).
+        */
+       acpi_align64(ctx);
 }
--

Could you please point out which commit in u-boot-x86/master should squash in this patch to fix the build error on sandbox?

It might be easier to pick up v8 in that case. I think there are three patches that need to change because of conflicts caused by the first one.

So can you pick up the v8 patches? Also you do need to rebase on master because of the str_to_upper patches.

Quoted 8 lines
Quoted 5 lines
Quoted 2 lines

But v8 is a complete new series for part B? I think we'd better re-tag v8 as v1. It's quite confusing.

No I mean the part A series. Let's get that applied and then we will be in a brave new world.

Is the v8 this one?
http://patchwork.ozlabs.org/user/todo/uboot/?series=172777

No it is this one:

http://patchwork.ozlabs.org/project/uboot/list/?series=172776

Quoted 2 lines

The part A series are already applied, but it has a build error as I mentioned.

Right, but the options are to do a fixup patch or to pick up the v8 series instead. Unfortunately there are merge conflicts in patches so if you just pick up one patch it won't apply cleanly.

Quoted 7 lines
Quoted 3 lines

But we still need maintain bisectablity right? To do that we need a fix patch that can be squashed into the u-boot-x86/master. Picking up the v8 series does not fix the bisectiablity I think.

But my suggestion is to pick up the v8 patch *and drop what you currently have in x86/master*. Otherwise, as you say, bisect might not work.

Oops, I missed the v8 patch in patchwork (not assigned to me yet)

Quoted 4 lines

I should have objected at the time to Andy refactoring patch, but I wasn't sure when my stuff would land so it did not seem reasonable to do so.

Quoted 6 lines
Quoted 3 lines

I will drop what is applied in x86/master, and apply the v8 part A series. But it's better to re-tag v8 part B as v1 part B. I got quite confused until now :)

OK I see. Shall I send out part B as v1 now? Or wait until there are comments and then send v2?

Please resend it as v1. I suspect people might get confused like me.

Regards,
Bin

#15 lore raw

PATCH 3/3 acpi: Fix-up patch to correct sandbox test errors

Hi Bin,

Quoted 144 lines · Bin Meng

Hi Simon,

Quoted 140 lines · Simon Glass

Hi Bin,

Quoted 136 lines · Bin Meng

Hi Simon,

Quoted 117 lines · Simon Glass

Hi Bin,

Quoted 113 lines · Bin Meng

Hi Simon,

Quoted 108 lines · Simon Glass

Hi Bin,

Quoted 84 lines · Bin Meng

Hi Simon,

Quoted 80 lines · Simon Glass

Hi Bin,

Quoted 76 lines · Bin Meng

Hi Simon,

Quoted 71 lines · Simon Glass

Hi Bin,

Quoted 46 lines · Bin Meng

Hi Simon,

Quoted 42 lines · Simon Glass

Move the alignment code into acpi_setup_base_tables() so that test and production code are in alignment.

This brings x86/master into line with patch series v8.

---

 arch/x86/lib/acpi_table.c | 5 -----
 lib/acpi/acpi_table.c     | 7 ++++++-
 2 files changed, 6 insertions(+), 6 deletions(-)
arch/x86/lib/acpi_table.c +0 -5
index 600bde2f5f..13f1409de8 100644
@@ -375,11 +375,6 @@ ulong write_acpi_tables(ulong start_addr)
        debug("ACPI: Writing ACPI tables at %lx\n", start_addr);
 
        acpi_setup_base_tables(ctx, start);
-       /*
-        * Per ACPI spec, the FACS table address must be aligned to a 64 byte
-        * boundary (Windows checks this, but Linux does not).
-        */
-       acpi_align64(ctx);
 
        debug("ACPI:    * FACS\n");
        facs = ctx->current;
lib/acpi/acpi_table.c +1 -1
index 5abf1cad50..1c253af3bf 100644
@@ -145,7 +145,7 @@ int acpi_add_table(struct acpi_ctx *ctx, void *table)
        }
 
        if (i >= entries_num) {
-               debug("ACPI: Error: too many tables\n");
+               log_err("ACPI: Error: too many tables\n");
                return -E2BIG;
        }
 
Quoted 14 lines

lib/acpi/acpi_table.c void acpi_setup_base_tables(struct acpi_ctx *ctx, void *start)

Quoted 11 lines
lib/acpi/acpi_table.c +5 -1
@@ -256,4 +256,9 @@ void acpi_setup_base_tables(struct acpi_ctx *ctx, void *start)
        acpi_write_rsdp(ctx->rsdp, ctx->rsdt, ctx->xsdt);
        acpi_write_rsdt(ctx->rsdt);
        acpi_write_xsdt(ctx->xsdt);
+       /*
+        * Per ACPI spec, the FACS table address must be aligned to a 64 byte
+        * boundary (Windows checks this, but Linux does not).
+        */
+       acpi_align64(ctx);
 }
--

Could you please point out which commit in u-boot-x86/master should squash in this patch to fix the build error on sandbox?

It might be easier to pick up v8 in that case. I think there are three patches that need to change because of conflicts caused by the first one.

So can you pick up the v8 patches? Also you do need to rebase on master because of the str_to_upper patches.

Quoted 8 lines
Quoted 5 lines
Quoted 2 lines

But v8 is a complete new series for part B? I think we'd better re-tag v8 as v1. It's quite confusing.

No I mean the part A series. Let's get that applied and then we will be in a brave new world.

Is the v8 this one?
http://patchwork.ozlabs.org/user/todo/uboot/?series=172777

No it is this one:

http://patchwork.ozlabs.org/project/uboot/list/?series=172776

Quoted 2 lines

The part A series are already applied, but it has a build error as I mentioned.

Right, but the options are to do a fixup patch or to pick up the v8 series instead. Unfortunately there are merge conflicts in patches so if you just pick up one patch it won't apply cleanly.

Quoted 7 lines
Quoted 3 lines

But we still need maintain bisectablity right? To do that we need a fix patch that can be squashed into the u-boot-x86/master. Picking up the v8 series does not fix the bisectiablity I think.

But my suggestion is to pick up the v8 patch *and drop what you currently have in x86/master*. Otherwise, as you say, bisect might not work.

Oops, I missed the v8 patch in patchwork (not assigned to me yet)

Quoted 4 lines

I should have objected at the time to Andy refactoring patch, but I wasn't sure when my stuff would land so it did not seem reasonable to do so.

Quoted 8 lines
Quoted 6 lines
Quoted 3 lines

I will drop what is applied in x86/master, and apply the v8 part A series. But it's better to re-tag v8 part B as v1 part B. I got quite confused until now :)

OK I see. Shall I send out part B as v1 now? Or wait until there are comments and then send v2?

Please resend it as v1. I suspect people might get confused like me.

OK this is done. Hoping people can find time to review soon! I first posted all this in January.

Regards,
Simon