search for: __export

...msg_vga(void) { NextCharJump = msg_filename; - VGAFilePtr = (uint16_t *)VGAFileBuf; + VGAFilePtr = VGAFileBuf; } static void msg_normal(uint8_t data) diff --git a/core/graphics.c b/core/graphics.c index 834372f..1604ab4 100644 --- a/core/graphics.c +++ b/core/graphics.c @@ -27,9 +27,9 @@ __export uint8_t UsingVGA = 0; uint16_t VGAPos; /* Pointer into VGA memory */ -__export uint16_t *VGAFilePtr; /* Pointer into VGAFileBuf */ __export uint16_t VGAFontSize = 16; /* Defaults to 16 byte font */ +__export char *VGAFilePtr; /* Pointer into VGAFileBuf */ __export char VGAFile...

...00644 core/multifs.c diff --git a/core/fs/cache.c b/core/fs/cache.c index 3b21fc2..b8ed8c2 100644 --- a/core/fs/cache.c +++ b/core/fs/cache.c @@ -16,7 +16,7 @@ * implementation since the block(cluster) size in FAT is a bit big. * */ -void cache_init(struct device *dev, int block_size_shift) +__export void cache_init(struct device *dev, int block_size_shift) { struct cache *prev, *cur; char *data = dev->cache_data; diff --git a/core/fs/diskio.c b/core/fs/diskio.c index 7d95d67..466b6a8 100644 --- a/core/fs/diskio.c +++ b/core/fs/diskio.c @@ -21,13 +21,27 @@ void getoneblk(struct di...

Hmm... well adding the __export stopped the "undefined symbol" error from showing up when I add "SENDCOOKIES" in the config, however I'm still not seeing any cookies being sent to my webserver. Any ideas on how to further debug? (if you want to send me the debug compile options privately that's okay t...

On Fri, 2015-08-21 at 12:33 -0700, Patrick Masotta wrote: > I think probably this patch is buggy: > > You have redefined VGAFilePtr > > -__export uint16_t *VGAFilePtr; /* Pointer into VGAFileBuf */ > +__export char *VGAFilePtr; /* Pointer into VGAFileBuf */ > > but you did not redefine i.e. > > *VGAFilePtr++ = data; > > Incrementing a pointer to "uint16_t" implies moving the pointer forward...

We recently have upgraded syslinux and have noticed that the boot logo code via boot.txt no longer works. I have a couple patches that get it close to working but I appear to be having trouble with color maps. Is there someone would want to help me finish figuring this out? Any interest in the patches I have to fix the rle decoder?

> Hi, > > the display of LSS16 files from a DISPLAY file (as documented in > http://www.syslinux.org/wiki/index.php/SYSLINUX#Display_graphic_from_filename:) > seems heavily broken since syslinux has been converted from assembler to > C. I already discovered one bug in core/include/graphics.h (and > core/graphics.c): the pointer VGAFilePtr is of type uint16_t*, but > should

Hi, the display of LSS16 files from a DISPLAY file (as documented in http://www.syslinux.org/wiki/index.php/SYSLINUX#Display_graphic_from_filename:) seems heavily broken since syslinux has been converted from assembler to C. I already discovered one bug in core/include/graphics.h (and core/graphics.c): the pointer VGAFilePtr is of type uint16_t*, but should be plain char*. This bug causes the

I think my oven is broken, but maybe I have the recipe wrong. I've been trying to utilize the lwIP stack and dmi information to avoid iPXE use. I can do a SYSAPPEND 0x00080 and it shows up in /proc/cmdline after booting the kernel, but tcpdump is not revealing the cookies as described during HTTP calls in the documentation. Trying the magic "SENDCOOKIES" parameter outside of an

...LINUX problems, but it should at least restore the behaviour from before -pre14. --- >From 129c3c31f1a7c2d58d72493d8aed5e6d30d79eb4 Mon Sep 17 00:00:00 2001 From: Matt Fleming <matt.fleming at intel.com> Date: Sat, 21 Jun 2014 19:30:53 +0100 Subject: [PATCH] efi: Tag __syslinux_adv* with __export commit 415d571 ("adv: Remove double defintion") introduced a regression for the EFI boot loader. The commit is correct, but should have tagged the __syslinux_adv_ptr and __syslinux_adv_size data objects in the EFI core as __export. This change is required because symbols in the EFI code...

...inux/execute.c index 653c880..3955571 100644 --- a/com32/elflink/ldlinux/execute.c +++ b/com32/elflink/ldlinux/execute.c @@ -44,6 +44,7 @@ const struct image_types image_boot_types[] = { { NULL, 0 }, }; +extern jmp_buf __return_to_command_prompt; extern int create_args_and_load(char *); __export void execute(const char *cmdline, uint32_t type, bool sysappend) @@ -136,7 +137,8 @@ __export void execute(const char *cmdline, uint32_t type, bool sysappend) /* Restore the console */ ldlinux_console_init(); - ldlinux_enter_command(); + /* Jump back to the main to call ldlinux_enter_comm...

...d __weak http_bake_cookies(void); #endif /* _SYSLINUX_PXE_API_H */ --- syslinux-6.02/core/fs/pxe/http.c 2013-10-13 13:59:03.000000000 -0400 +++ syslinux-6.02-changed/core/fs/pxe/http.c 2013-11-21 17:18:40.026705000 -0500 @@ -121,7 +121,7 @@ return n; } -void http_bake_cookies(void) +__export void http_bake_cookies(void) { if (cookie_buf) free(cookie_buf); -----Original Message----- From: Syslinux [mailto:syslinux-bounces at zytor.com] On Behalf Of Santillanes, Russel [Tech] Sent: Thursday, November 21, 2013 12:42 AM To: 'For discussion of Syslinux and tftp-hpa'...

From: Sylvain Gault <sylvain.gault at gmail.com> Hello there, I propose 2 patches that fix two possible stack overflows either when running a COM32 module or when loading a new config file. I didn't find a better way to do this than to use the infamous setjmp/longjmp functions to restore the stack to a previous state. This makes the logic a bit more complex, but the behavior is not

diff --git a/core/localboot.c b/core/localboot.c index 0b8769e4..30bfb272 100644 --- a/core/localboot.c +++ b/core/localboot.c @@ -63,7 +63,6 @@ __export void local_boot(int16_t ax) ireg.eax.w[0] = 0; /* Reset drive */ __intcall(0x13, &ireg, NULL); - memset(&ireg, 0, sizeof(ireg)); ireg.eax.w[0] = 0x0201; /* Read one sector */ ireg.ecx.w[0] = 0x0001; /* C/H/S = 0/0/1 (first sector) */ ireg.ebx.w[0] = OFFS(trackbuf); 21.01.20...

On Wed, 12 Jun, at 11:17:44AM, Gerardo Exequiel Pozzi wrote: > Cool thanks!. Now looks better, but still not work. > > For some reason, "ldlinux.c32" is apparently sent but "Failed to load" > by PXELINUX and few seconds later, dnsmasq shows an error message > "failed sending": Argh! The patch was broken. I missed the new core/path.c file. My bad.

So last week I was wondering if XFS was still working -- even with its last on-disk structure changes -- and it _suprisingly_ worked as expected. Right, now I can finally get rid of GRUB and use Syslinux to boot my Linux on EFI from a rootfs with xfs. Shit, I have two partitions (the first one being the required ESP) so there is no way to access the other partitions since because Syslinux does not

...start_ldlinux(1, argv); } out: - free(PATH); writestr("\nFailed to load ldlinux.c32"); } diff --git a/core/fs/fs.c b/core/fs/fs.c index 1cb4b00..b6ee19c 100644 --- a/core/fs/fs.c +++ b/core/fs/fs.c @@ -10,8 +10,6 @@ #include "fs.h" #include "cache.h" -__export char *PATH; - /* The currently mounted filesystem */ __export struct fs_info *this_fs = NULL; /* Root filesystem */ diff --git a/core/include/fs.h b/core/include/fs.h index c7d0fd7..b5c7f0d 100644 --- a/core/include/fs.h +++ b/core/include/fs.h @@ -1,6 +1,7 @@ #ifndef FS_H #define FS_H +#i...

>>>> VGAFilePtr is used as a pointer into VGAFileBuf, so it should be the same type. Signed-off-by: Chas Williams <ciwillia at brocade.com> --- .... <<<< I think probably this patch is buggy: You have redefined VGAFilePtr -__export uint16_t *VGAFilePtr; /* Pointer into VGAFileBuf */ +__export char *VGAFilePtr; /* Pointer into VGAFileBuf */ but you did not redefine i.e. *VGAFilePtr++ = data; Incrementing a pointer to "uint16_t" implies moving the pointer forward 2 bytes while incrementing a point...

...(-) diff --git a/core/elflink/load_env32.c b/core/elflink/load_env32.c index 492cc09..db19c7a 100644 --- a/core/elflink/load_env32.c +++ b/core/elflink/load_env32.c @@ -55,7 +55,7 @@ void init_module_subsystem(struct elf_module *module) list_add(&module->list, &modules_head); } -__export int start_ldlinux(int argc, char **argv) +static int _start_ldlinux(int argc, char **argv) { int rv; @@ -96,6 +96,62 @@ again: return rv; } +__export int start_ldlinux(int argc, char **argv) +{ + /* These variables are static to survive the longjmp. */ + static int has_jmpbuf = 0; + stati...

...ip = (ip << 8) | part; + part = 0; + p++; + } + p--; + + *res = htonl(ip); + return *p == '\0'; +} + + +/* + * Similar implementation as core/fs/pxe/dnsresolv.c + * + * Allow ipv4 host names to be used. Do not support + * DNS names or ipv6. + * + */ + __export uint32_t pxe_dns(const char *name) { + + uint32_t ip; + /* * Return failure on an empty input... this can happen during * some types of URL parsing, and this is the easiest place to @@ -45,6 +89,11 @@ __export uint32_t pxe_dns(const char *name) if (!name || !*name)...

..."SENDCOOKIES" parameter outside of an inline SYSAPPEND call just gives me an "Error: An undefined symbol was referenced" > > I'm using 6.02 and lpxelinux.0; Loading lpxelinux.0 over tftp, kernel/initrd over http (see below) > Hmm... I wonder if we are missing an __export in http.c: __export void http_bake_cookies(void) If you have the ability to rebuild Syslinux, could you try adding the above and see if it helps? -hpa