x86/gfx: early boot improvements
v3 cover x86/gfx: early boot improvements
Hello,
The following series contains some fixes and improvements related to graphics usage when booting Xen.
Proposed patches fix some shortcomings when using multiboot2, like the ignoring of the vga= parameter and forcefully switching the console to the maximum supported resolution.
Thanks, Roger.
Roger Pau Monne (3):
multiboot2: parse vga= option when setting GOP mode
multiboot2: do not set StdOut mode unconditionally
cmdline: parse multiple instances of the vga option
docs/misc/xen-command-line.pandoc | 3 ++ xen/arch/x86/boot/cmdline.c | 85 +++++++++++++++---------------- xen/arch/x86/boot/head.S | 13 ++++- xen/arch/x86/efi/efi-boot.h | 69 +++++++++++++++++++++++-- xen/arch/x86/x86_64/asm-offsets.c | 1 + 5 files changed, 123 insertions(+), 48 deletions(-)
Signature · 1 line
2.40.0
PATCH v3 1/3 multiboot2: parse vga= option when setting GOP mode
Introduce support for passing the command line to the efi_multiboot2() helper, and parse the vga= option if present.
Add support for the 'gfx' and 'current' vga options, ignore the 'keep' option, and print a warning message about other options not being currently implemented.
---
Changes since v2:
- Do not parse console=.
- Allow width or height to be 0 as long as the gfx- option is well formed.
Changes since v1:
- Do not return the last occurrence of a command line.
- Rearrange the code for assembly processing of the cmdline and use lea.
- Merge patches handling console= and vga= together.
--- xen/arch/x86/boot/head.S | 13 ++++++- xen/arch/x86/efi/efi-boot.h | 61 ++++++++++++++++++++++++++++++- xen/arch/x86/x86_64/asm-offsets.c | 1 + 3 files changed, 71 insertions(+), 4 deletions(-)
xen/arch/x86/boot/head.S +11 -2
@@ -226,9 +226,10 @@ __efi64_mb2_start: jmp x86_32_switch .Lefi_multiboot2_proto: - /* Zero EFI SystemTable and EFI ImageHandle addresses. */ + /* Zero EFI SystemTable, EFI ImageHandle addresses and cmdline. */ xor %esi,%esi xor %edi,%edi + xor %edx,%edx /* Skip Multiboot2 information fixed part. */ lea (MB2_fixed_sizeof+MULTIBOOT2_TAG_ALIGN-1)(%rbx),%ecx
@@ -266,6 +267,13 @@ __efi64_mb2_start: cmove MB2_efi64_ih(%rcx),%rdi je .Lefi_mb2_next_tag + /* Get command line from Multiboot2 information. */ + cmpl $MULTIBOOT2_TAG_TYPE_CMDLINE,MB2_tag_type(%rcx) + jne .Lno_cmdline + lea MB2_tag_string(%rcx),%rdx + jmp .Lefi_mb2_next_tag +.Lno_cmdline: + /* Is it the end of Multiboot2 information? */ cmpl $MULTIBOOT2_TAG_TYPE_END,MB2_tag_type(%rcx) je .Lrun_bs
@@ -329,7 +337,8 @@ __efi64_mb2_start: /* * efi_multiboot2() is called according to System V AMD64 ABI: - * - IN: %rdi - EFI ImageHandle, %rsi - EFI SystemTable. + * - IN: %rdi - EFI ImageHandle, %rsi - EFI SystemTable, + * %rdx - MB2 cmdline */ call efi_multiboot2
xen/arch/x86/efi/efi-boot.h +59 -2
73 lines not on this page · raw message
xen/arch/x86/x86_64/asm-offsets.c +1 -0
@@ -175,6 +175,7 @@ void __dummy__(void)OFFSET (MB2_mem_lower, multiboot2_tag_basic_meminfo_t, mem_lower);OFFSET (MB2_efi64_st, multiboot2_tag_efi64_t, pointer);OFFSET (MB2_efi64_ih, multiboot2_tag_efi64_ih_t, pointer); +OFFSET (MB2_tag_string, multiboot2_tag_string_t, string);BLANK (); OFFSET(DOMAIN_vm_assist, struct domain, vm_assist);
Signature · 1 line
2.40.0
Re: [PATCH v3 1/3] multiboot2: parse vga= option when setting GOP mode
xen/arch/x86/boot/head.S __efi64_mb2_start: · line 232
Quoted 11 lines · Roger Pau Monne
xen/arch/x86/boot/head.S +2
-1@@ -226,9 +226,10 @@ __efi64_mb2_start: jmp x86_32_switch .Lefi_multiboot2_proto: - /* Zero EFI SystemTable and EFI ImageHandle addresses. */ + /* Zero EFI SystemTable, EFI ImageHandle addresses and cmdline. */ xor %esi,%esi xor %edi,%edi + xor %edx,%edx
While perhaps better to leave this as you have it, ...
xen/arch/x86/boot/head.S __efi64_mb2_start: · line 275
Quoted 10 lines
xen/arch/x86/boot/head.S +6
-0@@ -266,6 +267,13 @@ __efi64_mb2_start: cmove MB2_efi64_ih(%rcx),%rdi je .Lefi_mb2_next_tag + /* Get command line from Multiboot2 information. */ + cmpl $MULTIBOOT2_TAG_TYPE_CMDLINE,MB2_tag_type(%rcx) + jne .Lno_cmdline + lea MB2_tag_string(%rcx),%rdx + jmp .Lefi_mb2_next_tag +.Lno_cmdline:
... in new blocks of code I think it would be nice if commas were followed by visually separating blanks.
xen/arch/x86/boot/head.S __efi64_mb2_start: · line 343
Quoted 9 lines
xen/arch/x86/boot/head.S +2
-1@@ -329,7 +337,8 @@ __efi64_mb2_start: /* * efi_multiboot2() is called according to System V AMD64 ABI: - * - IN: %rdi - EFI ImageHandle, %rsi - EFI SystemTable. + * - IN: %rdi - EFI ImageHandle, %rsi - EFI SystemTable, + * %rdx - MB2 cmdline */ call efi_multiboot2
All you obtain is a pointer to the string, which is fine from what I was able to establish, but that's not entirely obvious: The MBI structure used has a size field, so it could have been that the string isn't nul-terminated, and hence the size would also need passing. May I ask that this be mentioned at least in the description?
Quoted 8 lines
xen/arch/x86/efi/efi-boot.h +1
-1@@ -786,7 +786,30 @@ static bool __init efi_arch_use_config_file(EFI_SYSTEM_TABLE *SystemTable)static void __init efi_arch_flush_dcache_area (const void *vaddr,UINTN size) { } -void __init efi_multiboot2 (EFI_HANDLE ImageHandle,EFI_SYSTEM_TABLE *SystemTable) +/* Return the first occurrence of opt in cmd. */
xen/arch/x86/efi/efi-boot.h static bool __init efi_arch_use_config_file(EFI_SYSTEM_TABLE *SystemTable) · line 799
Quoted 10 lines
xen/arch/x86/efi/efi-boot.h +10
-0+static const char __init *get_option (const char *cmd,const char *opt) +{ +const char *s = cmd, *o =NULL ; + +if ( !cmd || !opt ) +return NULL ; + +while ( (s =strstr (s, opt)) !=NULL ) + { +if ( s == cmd || *(s -1 ) ==' ' )
Iirc I had asked before: Not allowing for at least tab? (See cmdline.c:delim_chars_comma[] for what the non-EFI parsing permits, which in turn might be going a little too far especially with permitting comma as well.)
xen/arch/x86/efi/efi-boot.h static bool __init efi_arch_use_config_file(EFI_SYSTEM_TABLE *SystemTable) · line 801
Quoted 2 lines
xen/arch/x86/efi/efi-boot.h +2
-0+ { + o = s +strlen (opt);
I don't think the comment ahead of the function describes this behavior, i.e. in particular the adding of the length of the option.
xen/arch/x86/efi/efi-boot.h void __init efi_multiboot2(EFI_HANDLE ImageHandle, EFI_SYSTEM_TABLE *SystemTable · line 833
Quoted 6 lines
xen/arch/x86/efi/efi-boot.h +1
-1@@ -807,7 +830,41 @@ void __init efi_multiboot2(EFI_HANDLE ImageHandle, EFI_SYSTEM_TABLE *SystemTableif ( gop ) { - gop_mode =efi_find_gop_mode (gop,0 ,0 ,0 ); +const char *last = cmdline;
Nit: Maybe better "cur" than "last"?
Quoted 12 lines
xen/arch/x86/efi/efi-boot.h +12
-0+unsigned int width =0 , height =0 , depth =0 ; +bool keep_current = false; + +while ( (last =get_option (last,"vga=" )) !=NULL ) + { +if ( !strncmp (last,"gfx-" ,4 ) ) + { + width =simple_strtoul (last +4 , &last,10 ); +if ( *last =='x' ) + height =simple_strtoul (last +1 , &last,10 ); +if ( *last =='x' ) + depth =simple_strtoul (last +1 , &last,10 );
xen/arch/x86/efi/efi-boot.h void __init efi_multiboot2(EFI_HANDLE ImageHandle, EFI_SYSTEM_TABLE *SystemTable · line 855
Quoted 10 lines
xen/arch/x86/efi/efi-boot.h +10
-0+if ( *last !=' ' && *last !='\t' && *last !='\0' && + *last !=',' ) + width = height = depth =0 ; + keep_current = false; + } +else if ( !strncmp (last,"current" ,7 ) ) + keep_current = true; +else if ( !strncmp (last,"keep" ,4 ) ) + { +/* Ignore. */
Maybe "Ignore here, handled in ..."?
Jan
Re: [PATCH v3 1/3] multiboot2: parse vga= option when setting GOP mode
Quoted 5 lines · Jan Beulich
Quoted 4 lines · Roger Pau Monne
xen/arch/x86/boot/head.S +0
-0@@ -329,7 +337,8 @@ __efi64_mb2_start:
Quoted 13 lines
xen/arch/x86/boot/head.S __efi64_mb2_start: · line 343
Quoted 7 lines
xen/arch/x86/boot/head.S +2
-1/* * efi_multiboot2() is called according to System V AMD64 ABI: - * - IN: %rdi - EFI ImageHandle, %rsi - EFI SystemTable. + * - IN: %rdi - EFI ImageHandle, %rsi - EFI SystemTable, + * %rdx - MB2 cmdline */ call efi_multiboot2All you obtain is a pointer to the string, which is fine from what I was able to establish, but that's not entirely obvious: The MBI structure used has a size field, so it could have been that the string isn't nul-terminated, and hence the size would also need passing. May I ask that this be mentioned at least in the description?
Sure. The multiboot2 specification already states that the string must be zero terminated. It's my understanding the size field is part of all the tags, so that a parser can find the next tag even if it doesn't know how to parse the current one.
Quoted 15 lines
Quoted 15 lines
xen/arch/x86/efi/efi-boot.h +8
-1@@ -786,7 +786,30 @@ static bool __init efi_arch_use_config_file(EFI_SYSTEM_TABLE *SystemTable)static void __init efi_arch_flush_dcache_area (const void *vaddr,UINTN size) { } -void __init efi_multiboot2 (EFI_HANDLE ImageHandle,EFI_SYSTEM_TABLE *SystemTable) +/* Return the first occurrence of opt in cmd. */ +static const char __init *get_option (const char *cmd,const char *opt) +{ +const char *s = cmd, *o =NULL ; + +if ( !cmd || !opt ) +return NULL ; +
Quoted 8 lines
xen/arch/x86/efi/efi-boot.h static bool __init efi_arch_use_config_file(EFI_SYSTEM_TABLE *SystemTable) · line 799
Quoted 3 lines
xen/arch/x86/efi/efi-boot.h +3
-0+while ( (s =strstr (s, opt)) !=NULL ) + { +if ( s == cmd || *(s -1 ) ==' ' )Iirc I had asked before: Not allowing for at least tab? (See cmdline.c:delim_chars_comma[] for what the non-EFI parsing permits, which in turn might be going a little too far especially with permitting comma as well.)
I've fixed this when parsing the gfx- option but not here, sorry.
Quoted 7 lines
xen/arch/x86/efi/efi-boot.h static bool __init efi_arch_use_config_file(EFI_SYSTEM_TABLE *SystemTable) · line 801
Quoted 3 lines
xen/arch/x86/efi/efi-boot.h +2
-0+ { + o = s +strlen (opt);I don't think the comment ahead of the function describes this behavior, i.e. in particular the adding of the length of the option.
I've changed the comment to:
Return a pointer to the character after the first occurrence of opt in cmd.
Thanks, Roger.
Re: [PATCH v3 1/3] multiboot2: parse vga= option when setting GOP mode
In reply to Roger Pau MonneQuoted 10 lines · Roger Pau Monne
xen/arch/x86/efi/efi-boot.h +5
-1@@ -807,7 +830,41 @@ void __init efi_multiboot2(EFI_HANDLE ImageHandle, EFI_SYSTEM_TABLE *SystemTableif ( gop ) { - gop_mode =efi_find_gop_mode (gop,0 ,0 ,0 ); +const char *last = cmdline; +unsigned int width =0 , height =0 , depth =0 ; +bool keep_current = false; + +while ( (last =get_option (last,"vga=" )) !=NULL )
xen/arch/x86/efi/efi-boot.h void __init efi_multiboot2(EFI_HANDLE ImageHandle, EFI_SYSTEM_TABLE *SystemTable · line 847
Quoted 10 lines
xen/arch/x86/efi/efi-boot.h +10
-0+ { +if ( !strncmp (last,"gfx-" ,4 ) ) + { + width =simple_strtoul (last +4 , &last,10 ); +if ( *last =='x' ) + height =simple_strtoul (last +1 , &last,10 ); +if ( *last =='x' ) + depth =simple_strtoul (last +1 , &last,10 ); +if ( *last !=' ' && *last !='\t' && *last !='\0' && + *last !=',' )
You check for an appropriate terminator here.
xen/arch/x86/efi/efi-boot.h void __init efi_multiboot2(EFI_HANDLE ImageHandle, EFI_SYSTEM_TABLE *SystemTable · line 851
Quoted 4 lines
xen/arch/x86/efi/efi-boot.h +4
-0+ width = height = depth =0 ; + keep_current = false; + } +else if ( !strncmp (last,"current" ,7 ) )
Don't you also need to do so here, and maybe even ...
xen/arch/x86/efi/efi-boot.h void __init efi_multiboot2(EFI_HANDLE ImageHandle, EFI_SYSTEM_TABLE *SystemTable · line 855
Quoted 4 lines
xen/arch/x86/efi/efi-boot.h +4
-0+ keep_current = true; + else if ( !strncmp(last, "keep", 4) ) + { +/* Ignore. */
... here?
Jan
Re: [PATCH v3 1/3] multiboot2: parse vga= option when setting GOP mode
Quoted 29 lines · Jan Beulich
Quoted 10 lines · Roger Pau Monne
xen/arch/x86/efi/efi-boot.h +5
-1@@ -807,7 +830,41 @@ void __init efi_multiboot2(EFI_HANDLE ImageHandle, EFI_SYSTEM_TABLE *SystemTableif ( gop ) { - gop_mode =efi_find_gop_mode (gop,0 ,0 ,0 ); +const char *last = cmdline; +unsigned int width =0 , height =0 , depth =0 ; +bool keep_current = false; + +while ( (last =get_option (last,"vga=" )) !=NULL )xen/arch/x86/efi/efi-boot.h void __init efi_multiboot2(EFI_HANDLE ImageHandle, EFI_SYSTEM_TABLE *SystemTable · line 847
Quoted 10 lines
xen/arch/x86/efi/efi-boot.h +10
-0+ { +if ( !strncmp (last,"gfx-" ,4 ) ) + { + width =simple_strtoul (last +4 , &last,10 ); +if ( *last =='x' ) + height =simple_strtoul (last +1 , &last,10 ); +if ( *last =='x' ) + depth =simple_strtoul (last +1 , &last,10 ); +if ( *last !=' ' && *last !='\t' && *last !='\0' && + *last !=',' )You check for an appropriate terminator here.
xen/arch/x86/efi/efi-boot.h void __init efi_multiboot2(EFI_HANDLE ImageHandle, EFI_SYSTEM_TABLE *SystemTable · line 851
Quoted 4 lines
xen/arch/x86/efi/efi-boot.h +4
-0+ width = height = depth =0 ; + keep_current = false; + } +else if ( !strncmp (last,"current" ,7 ) )
Quoted 8 lines
Don't you also need to do so here, and maybe even ...
xen/arch/x86/efi/efi-boot.h void __init efi_multiboot2(EFI_HANDLE ImageHandle, EFI_SYSTEM_TABLE *SystemTable · line 855
Quoted 4 lines
xen/arch/x86/efi/efi-boot.h +4
-0+ keep_current = true; + else if ( !strncmp(last, "keep", 4) ) + { +/* Ignore. */ ... here?
Hm, quite possibly for correctness. I felt it was relevant in gfx- as to avoid things like: gfx-1024x786x32x64 being handled, but the same could apply to passing on option like current-bar.
Will try to generalize the terminator parsing so it applies to all options.
Thanks, Roger.
PATCH v3 2/3 multiboot2: do not set StdOut mode unconditionally
In reply to Roger Pau MonneOnly initialize StdOut if the current StdOut mode is unusable. This avoids forcefully switching StdOut to the maximum supported resolution, and thus very likely changing the GOP mode without having first parsed the command line options.
---
The code is very similar to the approach suggested by Jan, please let me know if you would be OK with your suggested-by tag added.
---
Changes since v2:
- Use approach suggested by Jan.
Changes since v1:
- New in this version.
--- xen/arch/x86/efi/efi-boot.h | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-)
xen/arch/x86/efi/efi-boot.h +7 -1
@@ -820,7 +820,13 @@ void __init efi_multiboot2(EFI_HANDLE ImageHandle, EFI_SYSTEM_TABLE *SystemTableefi_init (ImageHandle, SystemTable); -efi_console_set_mode (); +if ( StdOut->QueryMode (StdOut, StdOut->Mode->Mode, + &cols, &rows) !=EFI_SUCCESS ) +/* +* If active StdOut mode is invalid init ConOut (StdOut) to the max +* supported size. +*/ +efi_console_set_mode ();if ( StdOut->QueryMode (StdOut, StdOut->Mode->Mode, &cols, &rows) ==EFI_SUCCESS )
Signature · 1 line
2.40.0
Re: [PATCH v3 2/3] multiboot2: do not set StdOut mode unconditionally
Quoted 9 lines · Roger Pau Monne
Only initialize StdOut if the current StdOut mode is unusable. This avoids forcefully switching StdOut to the maximum supported resolution, and thus very likely changing the GOP mode without having first parsed the command line options.
Signed-off-by Roger Pau Monné <[email protected]>---
The code is very similar to the approach suggested by Jan, please let me know if you would be OK with your suggested-by tag added.
I'm okay either way; I only suggested the "how", not the "that" after all. In any event
Jan
PATCH v3 3/3 cmdline: parse multiple instances of the vga option
In reply to Roger Pau MonneParse all instances of the vga= option on the command line, in order to always enforce the last selection on the command line.
Note that it's not safe to parse just the last occurrence of the vga= option, as then a command line with vga=current vga=keep would result in current being ignored.
Adjust the command line documentation to describe the new behavior.
---
Changes since v2:
- New in this version.
---
Build tested only, as I don't have a system that does legacy boot and has VGA output I can check.
It's mostly encapsulating the current code inside of a while loop and adding an extra else if for the "ask" option, there's a lot of indentation changes.
--- docs/misc/xen-command-line.pandoc | 3 ++ xen/arch/x86/boot/cmdline.c | 85 +++++++++++++++---------------- 2 files changed, 45 insertions(+), 43 deletions(-)
docs/misc/xen-command-line.pandoc +3 -0
@@ -2628,6 +2628,9 @@ with the specified width, height and depth. `ask` option. (N.B menu modes are displayed in hex, so `<mode>` should be a hexadecimal number) +Note that all the occurrences of the vga option in the command line are parsed, +and hence later occurrences can overwrite selections done by prior ones. + The optional `keep` parameter causes Xen to continue using the vga console even after dom0 has been started. The default behaviour is to relinquish control to dom0.
xen/arch/x86/boot/cmdline.c +42 -43
101 lines not on this page · raw message
Signature · 1 line
2.40.0
Re: [PATCH v3 3/3] cmdline: parse multiple instances of the vga option
docs/misc/xen-command-line.pandoc with the specified width, height and depth. · line 2632
Quoted 8 lines · Roger Pau Monne
docs/misc/xen-command-line.pandoc +2
-0@@ -2628,6 +2628,9 @@ with the specified width, height and depth. `ask` option. (N.B menu modes are displayed in hex, so `<mode>` should be a hexadecimal number) +Note that all the occurrences of the vga option in the command line are parsed, +and hence later occurrences can overwrite selections done by prior ones.
I'm not a native speaker, but is it perhaps more "override" that you mean?
Quoted 6 lines
xen/arch/x86/boot/cmdline.c +0
-0@@ -277,59 +277,58 @@ static u16 rows2vmode(unsigned int rows)static void vga_parse (const char *cmdline,early_boot_opts_t *ebo) {
xen/arch/x86/boot/cmdline.c static u16 rows2vmode(unsigned int rows) · line 288
Quoted 10 lines
xen/arch/x86/boot/cmdline.c +1
-8-const char *c; -unsigned int tmp, vesa_depth, vesa_height, vesa_width; - - c =find_opt (cmdline,"vga=" , true); - -if ( !c ) -return ; +const char *c = cmdline; - ebo->boot_vid_mode =ASK_VGA ;
I think this needs to stay here along with the addition of the related "else if" below. Otherwise I expect behavior for e.g. a sole "vga=keep" on the command line would change (in no longer prompting for the mode to use).
Jan
Re: [PATCH v3 3/3] cmdline: parse multiple instances of the vga option
Quoted 11 lines · Jan Beulich
docs/misc/xen-command-line.pandoc with the specified width, height and depth. · line 2632
Quoted 8 lines · Roger Pau Monne
docs/misc/xen-command-line.pandoc +2
-0@@ -2628,6 +2628,9 @@ with the specified width, height and depth. `ask` option. (N.B menu modes are displayed in hex, so `<mode>` should be a hexadecimal number) +Note that all the occurrences of the vga option in the command line are parsed, +and hence later occurrences can overwrite selections done by prior ones.I'm not a native speaker, but is it perhaps more "override" that you mean?
I don't think this is appropriate to say in the cmdline docs. Absent bugs, all command line options get parsed left->right with latest-takes-precedence.
It is a bug that vga= didn't behave like this.
It might be ok to say "Note, prior to Xen 4.18, only the first instance of vga= got parsed, with subsequent instances getting ignored".
Explicitly calling out the default behaviour in one option does nothing but confuse things.
~Andrew
Re: [PATCH v3 3/3] cmdline: parse multiple instances of the vga option
In reply to Jan BeulichQuoted 7 lines · Jan Beulich
Quoted 6 lines · Roger Pau Monne
docs/misc/xen-command-line.pandoc +0
-0@@ -2628,6 +2628,9 @@ with the specified width, height and depth. `ask` option. (N.B menu modes are displayed in hex, so `<mode>` should be a hexadecimal number)
Quoted 5 lines
docs/misc/xen-command-line.pandoc with the specified width, height and depth. · line 2632
Quoted 2 lines
docs/misc/xen-command-line.pandoc +2
-0+Note that all the occurrences of the vga option in the command line are parsed, +and hence later occurrences can overwrite selections done by prior ones.I'm not a native speaker, but is it perhaps more "override" that you mean?
TBH I always get confused with overwrite vs override. I will remove the line as requested by Andrew.
Quoted 15 lines
Quoted 15 lines
xen/arch/x86/boot/cmdline.c +1
-7@@ -277,59 +277,58 @@ static u16 rows2vmode(unsigned int rows)static void vga_parse (const char *cmdline,early_boot_opts_t *ebo) { -const char *c; -unsigned int tmp, vesa_depth, vesa_height, vesa_width; - - c =find_opt (cmdline,"vga=" , true); - -if ( !c ) -return ; +const char *c = cmdline;
Quoted 6 lines
xen/arch/x86/boot/cmdline.c static u16 rows2vmode(unsigned int rows) · line 288
Quoted 1 line
xen/arch/x86/boot/cmdline.c +0
-1- ebo->boot_vid_mode =ASK_VGA ;I think this needs to stay here along with the addition of the related "else if" below. Otherwise I expect behavior for e.g. a sole "vga=keep" on the command line would change (in no longer prompting for the mode to use).
Done.
Thanks, Roger.