[U-Boot] [PATCH v3 4/5] efi_loader: EFI_UNICODE_COLLATION_PROTOCOL

Alexander Graf agraf at suse.de
Mon Sep 3 21:09:07 UTC 2018



On 03.09.18 18:35, Heinrich Schuchardt wrote:
> The patch implements the EFI_UNICODE_COLLATION_PROTOCOL.
> 
> Signed-off-by: Heinrich Schuchardt <xypron.glpk at gmx.de>
> ---
> v3:
> 	no change
> v2:
> 	add more comments
> ---
>  MAINTAINERS                            |   2 +
>  include/cp1250.h                       |  40 +++
>  include/cp437.h                        |  40 +++
>  include/efi_api.h                      |  21 ++
>  include/efi_loader.h                   |   5 +
>  lib/efi_loader/Makefile                |  18 +-
>  lib/efi_loader/efi_boottime.c          |   6 +
>  lib/efi_loader/efi_unicode_collation.c | 329 +++++++++++++++++++++++++
>  8 files changed, 457 insertions(+), 4 deletions(-)
>  create mode 100644 include/cp1250.h
>  create mode 100644 include/cp437.h
>  create mode 100644 lib/efi_loader/efi_unicode_collation.c
> 
> diff --git a/MAINTAINERS b/MAINTAINERS
> index 8c9cd83347..67b015205c 100644
> --- a/MAINTAINERS
> +++ b/MAINTAINERS
> @@ -382,6 +382,8 @@ F:	doc/README.uefi
>  F:	doc/README.iscsi
>  F:	Documentation/efi.rst
>  F:	include/capitalization.h
> +F:	include/cp1250.h
> +F:	include/cp437.h
>  F:	include/efi*
>  F:	include/pe.h
>  F:	include/asm-generic/pe.h
> diff --git a/include/cp1250.h b/include/cp1250.h
> new file mode 100644
> index 0000000000..adacf8a958
> --- /dev/null
> +++ b/include/cp1250.h
> @@ -0,0 +1,40 @@
> +/* SPDX-License-Identifier: GPL-2.0+ */
> +
> +/*
> + * Constant CP1250 contains the Unicode code points for characters 0x80 - 0xff
> + * of the code page 1250.
> + */
> +#define CP1250 { \
> +	0x20ac, 0x0000, 0x201a, 0x0000, \
> +	0x201e, 0x2026, 0x2020, 0x2021, \
> +	0x0000, 0x2030, 0x0160, 0x2039, \
> +	0x015a, 0x0164, 0x017d, 0x0179, \
> +	0x0000, 0x2018, 0x2019, 0x201c, \
> +	0x201d, 0x2022, 0x2013, 0x2014, \
> +	0x0000, 0x2122, 0x0161, 0x203a, \
> +	0x015b, 0x0165, 0x017e, 0x017a, \
> +	0x00a0, 0x02c7, 0x02d8, 0x0141, \
> +	0x00a4, 0x0104, 0x00a6, 0x00a7, \
> +	0x00a8, 0x00a9, 0x015e, 0x00ab, \
> +	0x00ac, 0x00ad, 0x00ae, 0x017b, \
> +	0x00b0, 0x00b1, 0x02db, 0x0142, \
> +	0x00b4, 0x00b5, 0x00b6, 0x00b7, \
> +	0x00b8, 0x0105, 0x015f, 0x00bb, \
> +	0x013d, 0x02dd, 0x013e, 0x017c, \
> +	0x0154, 0x00c1, 0x00c2, 0x0102, \
> +	0x00c4, 0x0139, 0x0106, 0x00c7, \
> +	0x010c, 0x00c9, 0x0118, 0x00cb, \
> +	0x011a, 0x00cd, 0x00ce, 0x010e, \
> +	0x0110, 0x0143, 0x0147, 0x00d3, \
> +	0x00d4, 0x0150, 0x00d6, 0x00d7, \
> +	0x0158, 0x016e, 0x00da, 0x0170, \
> +	0x00dc, 0x00dd, 0x0162, 0x00df, \
> +	0x0155, 0x00e1, 0x00e2, 0x0103, \
> +	0x00e4, 0x013a, 0x0107, 0x00e7, \
> +	0x010d, 0x00e9, 0x0119, 0x00eb, \
> +	0x011b, 0x00ed, 0x00ee, 0x010f, \
> +	0x0111, 0x0144, 0x0148, 0x00f3, \
> +	0x00f4, 0x0151, 0x00f6, 0x00f7, \
> +	0x0159, 0x016f, 0x00fa, 0x0171, \
> +	0x00fc, 0x00fd, 0x0163, 0x02d9, \
> +}
> diff --git a/include/cp437.h b/include/cp437.h
> new file mode 100644
> index 0000000000..0b2b97132e
> --- /dev/null
> +++ b/include/cp437.h
> @@ -0,0 +1,40 @@
> +/* SPDX-License-Identifier: GPL-2.0+ */
> +
> +/*
> + * Constant CP437 contains the Unicode code points for characters 0x80 - 0xff
> + * of the code page 437.
> + */
> +#define CP437 { \
> +	0x00c7, 0x00fc, 0x00e9, 0x00e2, \
> +	0x00e4, 0x00e0, 0x00e5, 0x00e7, \
> +	0x00ea, 0x00eb, 0x00e8, 0x00ef, \
> +	0x00ee, 0x00ec, 0x00c4, 0x00c5, \
> +	0x00c9, 0x00e6, 0x00c6, 0x00f4, \
> +	0x00f6, 0x00f2, 0x00fb, 0x00f9, \
> +	0x00ff, 0x00d6, 0x00dc, 0x00a2, \
> +	0x00a3, 0x00a5, 0x20a7, 0x0192, \
> +	0x00e1, 0x00ed, 0x00f3, 0x00fa, \
> +	0x00f1, 0x00d1, 0x00aa, 0x00ba, \
> +	0x00bf, 0x2310, 0x00ac, 0x00bd, \
> +	0x00bc, 0x00a1, 0x00ab, 0x00bb, \
> +	0x2591, 0x2592, 0x2593, 0x2502, \
> +	0x2524, 0x2561, 0x2562, 0x2556, \
> +	0x2555, 0x2563, 0x2551, 0x2557, \
> +	0x255d, 0x255c, 0x255b, 0x2510, \
> +	0x2514, 0x2534, 0x252c, 0x251c, \
> +	0x2500, 0x253c, 0x255e, 0x255f, \
> +	0x255a, 0x2554, 0x2569, 0x2566, \
> +	0x2560, 0x2550, 0x256c, 0x2567, \
> +	0x2568, 0x2564, 0x2565, 0x2559, \
> +	0x2558, 0x2552, 0x2553, 0x256b, \
> +	0x256a, 0x2518, 0x250c, 0x2588, \
> +	0x2584, 0x258c, 0x2590, 0x2580, \
> +	0x03b1, 0x00df, 0x0393, 0x03c0, \
> +	0x03a3, 0x03c3, 0x00b5, 0x03c4, \
> +	0x03a6, 0x0398, 0x03a9, 0x03b4, \
> +	0x221e, 0x03c6, 0x03b5, 0x2229, \
> +	0x2261, 0x00b1, 0x2265, 0x2264, \
> +	0x2320, 0x2321, 0x00f7, 0x2248, \
> +	0x00b0, 0x2219, 0x00b7, 0x221a, \
> +	0x207f, 0x00b2, 0x25a0, 0x00a0, \
> +}
> diff --git a/include/efi_api.h b/include/efi_api.h
> index ebf2a3bc18..1efc448184 100644
> --- a/include/efi_api.h
> +++ b/include/efi_api.h
> @@ -31,6 +31,7 @@ enum efi_timer_delay {
>  	EFI_TIMER_RELATIVE = 2
>  };
>  
> +#define efi_intn_t ssize_t
>  #define efi_uintn_t size_t
>  typedef uint16_t *efi_string_t;
>  
> @@ -958,4 +959,24 @@ struct efi_driver_binding_protocol {
>  	efi_handle_t driver_binding_handle;
>  };
>  
> +#define EFI_UNICODE_COLLATION_PROTOCOL2_GUID \
> +	EFI_GUID(0xa4c751fc, 0x23ae, 0x4c3e, \
> +		 0x92, 0xe9, 0x49, 0x64, 0xcf, 0x63, 0xf3, 0x49)
> +struct efi_unicode_collation_protocol {
> +	efi_intn_t (EFIAPI *stri_coll)(
> +		struct efi_unicode_collation_protocol *this, u16 *s1, u16 *s2);
> +	bool (EFIAPI *metai_match)(struct efi_unicode_collation_protocol *this,
> +				   const u16 *string, const u16 *patter);
> +	void (EFIAPI *str_lwr)(struct efi_unicode_collation_protocol
> +			       *this, u16 *string);
> +	void (EFIAPI *str_upr)(struct efi_unicode_collation_protocol *this,
> +			       u16 *string);
> +	void (EFIAPI *fat_to_str)(struct efi_unicode_collation_protocol *this,
> +				  efi_uintn_t fat_size, char *fat, u16 *string);
> +	bool (EFIAPI *str_to_fat)(struct efi_unicode_collation_protocol *this,
> +				  const u16 *string, efi_uintn_t fat_size,
> +				  char *fat);
> +	char *supported_languages;
> +};
> +
>  #endif
> diff --git a/include/efi_loader.h b/include/efi_loader.h
> index b46babf931..736e82c506 100644
> --- a/include/efi_loader.h
> +++ b/include/efi_loader.h
> @@ -98,6 +98,9 @@ extern const struct efi_device_path_to_text_protocol efi_device_path_to_text;
>  /* implementation of the EFI_DEVICE_PATH_UTILITIES_PROTOCOL */
>  extern const struct efi_device_path_utilities_protocol
>  					efi_device_path_utilities;
> +/* Implementation of the EFI_UNICODE_COLLATION_PROTOCOL */
> +extern const struct efi_unicode_collation_protocol
> +					efi_unicode_collation_protocol;
>  
>  uint16_t *efi_dp_str(struct efi_device_path *dp);
>  
> @@ -127,6 +130,8 @@ extern const efi_guid_t efi_file_info_guid;
>  /* GUID for file system information */
>  extern const efi_guid_t efi_file_system_info_guid;
>  extern const efi_guid_t efi_guid_device_path_utilities_protocol;
> +/* GUID of the Unicode collation protocol */
> +extern const efi_guid_t efi_guid_unicode_collation_protocol;
>  
>  extern unsigned int __efi_runtime_start, __efi_runtime_stop;
>  extern unsigned int __efi_runtime_rel_start, __efi_runtime_rel_stop;
> diff --git a/lib/efi_loader/Makefile b/lib/efi_loader/Makefile
> index 1ffbf52a89..0769df20f7 100644
> --- a/lib/efi_loader/Makefile
> +++ b/lib/efi_loader/Makefile
> @@ -17,10 +17,20 @@ always += helloworld.efi
>  endif
>  
>  obj-$(CONFIG_CMD_BOOTEFI_HELLO) += helloworld_efi.o
> -obj-y += efi_image_loader.o efi_boottime.o efi_runtime.o efi_console.o
> -obj-y += efi_memory.o efi_device_path_to_text.o efi_device_path.o
> -obj-y += efi_device_path_utilities.o efi_file.o efi_variable.o efi_bootmgr.o
> -obj-y += efi_watchdog.o
> +obj-y += \
> +efi_bootmgr.o \
> +efi_boottime.o \
> +efi_console.o \
> +efi_device_path.o \
> +efi_device_path_to_text.o \
> +efi_device_path_utilities.o \
> +efi_file.o \
> +efi_image_loader.o \
> +efi_memory.o \
> +efi_runtime.o \
> +efi_unicode_collation.o \
> +efi_variable.o \
> +efi_watchdog.o

Instead of line breaking them, just make each line an obj-y += line.

>  obj-$(CONFIG_LCD) += efi_gop.o
>  obj-$(CONFIG_DM_VIDEO) += efi_gop.o
>  obj-$(CONFIG_PARTITIONS) += efi_disk.o
> diff --git a/lib/efi_loader/efi_boottime.c b/lib/efi_loader/efi_boottime.c
> index cd5fdb7aa8..f9f59b93f9 100644
> --- a/lib/efi_loader/efi_boottime.c
> +++ b/lib/efi_loader/efi_boottime.c
> @@ -1527,6 +1527,12 @@ efi_status_t efi_setup_loaded_image(
>  	if (ret != EFI_SUCCESS)
>  		goto failure;
>  
> +	ret = efi_add_protocol(obj->handle,
> +			       &efi_guid_unicode_collation_protocol,
> +			       (void *)&efi_unicode_collation_protocol);
> +	if (ret != EFI_SUCCESS)
> +		goto failure;
> +
>  	return ret;
>  failure:
>  	printf("ERROR: Failure to install protocols for loaded image\n");
> diff --git a/lib/efi_loader/efi_unicode_collation.c b/lib/efi_loader/efi_unicode_collation.c
> new file mode 100644
> index 0000000000..2c1e819174
> --- /dev/null
> +++ b/lib/efi_loader/efi_unicode_collation.c
> @@ -0,0 +1,329 @@
> +// SPDX-License-Identifier: GPL-2.0+
> +/*
> + * EFI Unicode collation protocol
> + *
> + * Copyright (c) 2018 Heinrich Schuchardt <xypron.glpk at gmx.de>
> + */
> +
> +#include <common.h>
> +#include <charset.h>
> +#include <cp1250.h>
> +#include <cp437.h>
> +#include <efi_loader.h>
> +
> +/* Characters that may not be used in file names */
> +static const char illegal[] = "<>:\"/\\|?*";
> +
> +/*
> + * EDK2 assumes codepage 1250 when creating FAT 8.3 file names.
> + * Linux defaults to codepage 437 for FAT 8.3 file names.
> + */
> +#if CONFIG_FAT_DEFAULT_CODEPAGE == 1250
> +/* Unicode code points for code page 1250 characters 0x80 - 0xff */
> +static const u16 codepage[] = CP1250;
> +#else
> +/* Unicode code points for code page 437 characters 0x80 - 0xff */
> +static const u16 codepage[] = CP437;
> +#endif
> +
> +/* GUID of the EFI_UNICODE_COLLATION_PROTOCOL */
> +const efi_guid_t efi_guid_unicode_collation_protocol =
> +	EFI_UNICODE_COLLATION_PROTOCOL2_GUID;
> +
> +/**
> + * efi_stri_coll() - compare utf-16 strings case-insenitively
> + *
> + * @this:	unicode collation protocol instance
> + * @s1:		first string
> + * @s2:		second string
> + *
> + * This function implements the StriColl() service of the
> + * EFI_UNICODE_COLLATION_PROTOCOL.
> + *
> + * See the Unified Extensible Firmware Interface (UEFI) specification for
> + * details.
> + *
> + * TODO:
> + * The implementation does not follow the Unicode collation algorithm.
> + * For ASCII characters it results in the same sort order as EDK2.
> + * We could use table UNICODE_CAPITALIZATION_TABLE for better results.
> + *
> + * Return:	0: s1 == s2, > 0: s1 > s2, < 0: s1 < s2
> + */
> +static efi_intn_t EFIAPI efi_stri_coll(
> +		struct efi_unicode_collation_protocol *this, u16 *s1, u16 *s2)
> +{
> +	s32 c1, c2;
> +	efi_intn_t ret = 0;
> +
> +	EFI_ENTRY("%p, %ls, %ls", this, s1, s2);
> +	for (; *s1 | *s2; ++s1, ++s2) {
> +		c1 = utf_to_upper(*s1);
> +		c2 = utf_to_upper(*s2);
> +		if (c1 < c2) {
> +			ret = -1;
> +			goto out;
> +		} else if (c1 > c2) {
> +			ret = 1;
> +			goto out;
> +		}
> +	}
> +out:
> +	EFI_EXIT(EFI_SUCCESS);
> +	return ret;
> +}
> +
> +/**
> + * metai_match() - compare utf-16 string with a pattern string case-insenitively
> + *
> + * @s:		string to compare
> + * @p:		pattern string
> + *
> + * The pattern string may use these:
> + *	- * matches >= 0 characters
> + *	- ? matches 1 character
> + *	- [<char1><char2>...<charN>] match any character in the set
> + *	- [<char1>-<char2>] matches any character in the range
> + *
> + * This function is called my efi_metai_match().
> + *
> + * For '*' pattern searches this function calls itself recursively.
> + * Performance-wise this is suboptimal, especially for multiple '*' wildcards.
> + * But it results in simple code.
> + *
> + * Return:	true if the string is matched.
> + */
> +static bool metai_match(const u16 *s, const u16 *p)
> +{
> +	u16 first;
> +
> +	for (; *s && *p; ++s, ++p) {
> +		switch (*p) {
> +		case '*':
> +			/* Match 0 or more characters */
> +			++p;
> +			for (;; ++s) {
> +				if (metai_match(s, p))
> +					return true;
> +				if (!*s)
> +					return false;
> +			}
> +		case '?':
> +			/* Match any one character */
> +			break;
> +		case '[':
> +			/* Match any character in the set */
> +			++p;
> +			first = *p;
> +			if (first == ']')
> +				/* Empty set */
> +				return false;
> +			++p;
> +			if (*p == '-') {
> +				/* Range */
> +				++p;
> +				if (*s < first || *s > *p)
> +					return false;
> +				++p;
> +				if (*p != ']')
> +					return false;
> +			} else {
> +				/* Set */
> +				bool hit = false;
> +
> +				if (*s == first)
> +					hit = true;
> +				for (; *p && *p != ']'; ++p) {
> +					if (*p == *s)
> +						hit = true;
> +				}
> +				if (!hit || *p != ']')
> +					return false;
> +			}
> +			break;
> +		default:
> +			/* Match one character */
> +			if (*p != *s)
> +				return false;
> +		}
> +	}
> +	if (!*p && !*s)
> +		return true;
> +	return false;
> +}
> +
> +/**
> + * efi_metai_match() - compare utf-16 string with a pattern string
> + *		       case-insenitively
> + *
> + * @this:	unicode collation protocol instance
> + * @s:		string to compare
> + * @p:		pattern string
> + *
> + * The pattern string may use these:
> + *	- * matches >= 0 characters
> + *	- ? matches 1 character
> + *	- [<char1><char2>...<charN>] match any character in the set
> + *	- [<char1>-<char2>] matches any character in the range
> + *
> + * This function implements the MetaMatch() service of the
> + * EFI_UNICODE_COLLATION_PROTOCOL.
> + *
> + * Return:	true if the string is matched.
> + */
> +static bool EFIAPI efi_metai_match(struct efi_unicode_collation_protocol *this,
> +				   const u16 *string, const u16 *pattern)
> +{
> +	bool ret;
> +
> +	EFI_ENTRY("%p, %ls, %ls", this, string, pattern);
> +	ret =  metai_match(string, pattern);
> +	EFI_EXIT(EFI_SUCCESS);
> +	return ret;
> +}
> +
> +/**
> + * efi_str_lwr() - convert to lower case
> + *
> + * @this:	unicode collation protocol instance
> + * @string:	string to convert
> + * @p:		pattern string
> + *
> + * The conversion is done in place. As long as upper and lower letters use the
> + * same number of words this does not pose a problem.
> + *
> + * This function implements the StrLwr() service of the
> + * EFI_UNICODE_COLLATION_PROTOCOL.
> + */
> +static void EFIAPI efi_str_lwr(struct efi_unicode_collation_protocol *this,
> +			       u16 *string)
> +{
> +	EFI_ENTRY("%p, %ls", this, string);
> +	for (; *string; ++string)
> +		*string = utf_to_lower(*string);
> +	EFI_EXIT(EFI_SUCCESS);
> +}
> +
> +/**
> + * efi_str_upr() - convert to upper case
> + *
> + * @this:	unicode collation protocol instance
> + * @string:	string to convert
> + * @p:		pattern string
> + *
> + * The conversion is done in place. As long as upper and lower letters use the
> + * same number of words this does not pose a problem.
> + *
> + * This function implements the StrUpr() service of the
> + * EFI_UNICODE_COLLATION_PROTOCOL.
> + */
> +static void EFIAPI efi_str_upr(struct efi_unicode_collation_protocol *this,
> +			       u16 *string)
> +{
> +	EFI_ENTRY("%p, %ls", this, string);
> +	for (; *string; ++string)
> +		*string = utf_to_upper(*string);
> +	EFI_EXIT(EFI_SUCCESS);
> +}
> +
> +/**
> + * efi_fat_to_str() - convert an 8.3 file name from an OEM codepage to Unicode
> + *
> + * @this:	unicode collation protocol instance
> + * @fat_size:	size of the string to convert
> + * @fat:	string to convert
> + * @string:	converted string
> + *
> + * This function implements the FatToStr() service of the
> + * EFI_UNICODE_COLLATION_PROTOCOL.
> + */
> +static void EFIAPI efi_fat_to_str(struct efi_unicode_collation_protocol *this,
> +				  efi_uintn_t fat_size, char *fat, u16 *string)
> +{
> +	efi_uintn_t i;
> +	u16 c;
> +
> +	EFI_ENTRY("%p, %zu, %s, %p", this, fat_size, fat, string);
> +	for (i = 0; i < fat_size; ++i) {
> +		c = (unsigned char)fat[i];
> +		if (c > 0x80)
> +			c = codepage[i - 0x80];
> +		string[i] = c;
> +		if (!c)
> +			break;
> +	}
> +	string[i] = 0;
> +	EFI_EXIT(EFI_SUCCESS);
> +}
> +
> +/**
> + * efi_fat_to_str() - convert a utf-16 string to legal characters for a FAT
> + *                    file name in an OEM code page
> + *
> + * @this:	unicode collation protocol instance
> + * @string:	Unicode string to convert
> + * @fat_size:	size of the target buffer
> + * @fat:	converted string
> + *
> + * This function implements the StrToFat() service of the
> + * EFI_UNICODE_COLLATION_PROTOCOL.
> + *
> + * Return:	true if an illegal character was substituted by '_'.
> + */
> +static bool EFIAPI efi_str_to_fat(struct efi_unicode_collation_protocol *this,
> +				  const u16 *string, efi_uintn_t fat_size,
> +				  char *fat)
> +{
> +	efi_uintn_t i;
> +	s32 c;
> +	bool ret = false;
> +
> +	EFI_ENTRY("%p, %ls, %zu, %p", this, string, fat_size, fat);
> +	for (i = 0; i < fat_size;) {
> +		c = utf16_get(&string);
> +		switch (c) {
> +		/* Ignore period and space */
> +		case '.':
> +		case ' ':
> +			continue;
> +		case 0:
> +			break;
> +		}
> +		c = utf_to_upper(c);
> +		if (c >= 0x80) {
> +			int j;
> +
> +			/* Look for codepage translation */
> +			for (j = 0; j < 0x80; ++j) {
> +				if (c == codepage[j]) {
> +					c = j + 0x80;
> +					break;
> +				}
> +			}
> +			if (j >= 0x80) {
> +				c = '_';
> +				ret = true;
> +			}
> +		} else if (c && (c < 0x20 || strchr(illegal, c))) {
> +			c = '_';
> +			ret = true;
> +		}
> +
> +		fat[i] = c;
> +		if (!c)
> +			break;
> +		++i;
> +	}
> +	EFI_EXIT(EFI_SUCCESS);
> +	return ret;
> +}
> +
> +const struct efi_unicode_collation_protocol efi_unicode_collation_protocol = {
> +	.stri_coll = efi_stri_coll,
> +	.metai_match = efi_metai_match,
> +	.str_lwr = efi_str_lwr,
> +	.str_upr = efi_str_upr,
> +	.fat_to_str = efi_fat_to_str,
> +	.str_to_fat = efi_str_to_fat,
> +	.supported_languages = "en-US",

I have to admit I'm slightly puzzled on the supported language field.
The spec says quite clearly in appendix M what to do, but everyone seems
to interpret the specs differently:

  edk2 provides "en"
  appendix M example is "en-us"
  your patch provides "en-US"

I'm not sure which one of them is correct :).


Alex


More information about the U-Boot mailing list