search for: fs_ops

Check if fs_ops->readdir is available before calling it. At least PXELINUX doesn't implement it. Signed-off-by: Sebastian Herbszt <herbszt at gmx.de> diff --git a/core/fs/readdir.c b/core/fs/readdir.c index d2b112b..d20fc33 100644 --- a/core/fs/readdir.c +++ b/core/fs/readdir.c @@ -28,8 +28,11 @@ st...

...+++ b/efi/main.c > @@ -1177,13 +1177,10 @@ static void efi_setcwd(CHAR16 *dp) > > EFI_STATUS efi_main(EFI_HANDLE image, EFI_SYSTEM_TABLE *table) > { > + EFI_PXE_BASE_CODE *pxe; > EFI_LOADED_IMAGE *info; > EFI_STATUS status = EFI_SUCCESS; > -#if 0 > - const struct fs_ops *ops[] = { &vfat_fs_ops, NULL }; > -#else > - const struct fs_ops *ops[] = { &pxe_fs_ops, NULL }; > -#endif > + const struct fs_ops *ops[] = { NULL, NULL }; Probably need to drop "const" here, no? -hpa

...form factor */ + SYSAPPEND_FSUUID, /* Boot filesystem UUID */ SYSAPPEND_MAX /* Total number of strings */ }; diff --git a/core/fs/btrfs/btrfs.c b/core/fs/btrfs/btrfs.c index 16386cc..dfa9ad7 100644 --- a/core/fs/btrfs/btrfs.c +++ b/core/fs/btrfs/btrfs.c @@ -674,5 +674,6 @@ const struct fs_ops btrfs_fs_ops = { .next_extent = btrfs_next_extent, .readdir = btrfs_readdir, .chdir_start = generic_chdir_start, - .open_config = generic_open_config + .open_config = generic_open_config, + .fs_uuid = NULL, }; diff --git a/core/fs/ext2/ext2.c b/core/fs/e...

...* Root filesystem */ +/* root_fs means the file system where ldlinux.sys lives in. */ +__export struct fs_info *root_fs = NULL; +/* this_fs means the file system being currently used. */ +__export struct fs_info *this_fs = NULL; +/* export p_ops to be used outside the core */ +__export const struct fs_ops **p_ops = NULL; /* Actual file structures (we don't have malloc yet...) */ __export struct file files[MAX_OPEN]; @@ -344,6 +349,9 @@ __export int open_file(const char *name, int flags, struct com32_filedata *filed dprintf("open_file %s\n", name); + if (switch_fs(&name...

Hi hpa, I just added the generic path_lookup function for EXTLINUX. hpa, It changed a lot, so I want hear something from you. First , here is the main skeleton of the generic path_lookup function: > if (*name == '/') { > inode = this_fs->fs_ops->iget_root(); // The way to get the root > inode is different from different fs > while(*name == '/') > name++; > } else { > inode = this_inode; // pwd > } > parent = inode; > > while (*name) { &g...

In 'core/fs/fs.c', 'fs.root' is set, but the 'iget_root' function pointer call might return a null pointer and we don't check for it. A symptom was that QEmu crashed as EIP landed outside of memory. - Shao /* start out in the root directory */ if (fs.fs_ops->iget_root) { fs.root = fs.fs_ops->iget_root(&fs); /* Maybe we should check 'fs.root' here */ fs.cwd = get_inode(fs.root); }

...MULTIDISK_UTILS_H + +#include <syslinux/partiter.h> +#include "fs.h" + +struct part_node { + int partition; + struct fs_info *fs; + struct part_node *next; +}; + +struct queue_head { + struct part_node *first; + struct part_node *last; +}; + +/* + * Needs to keep ROOT_FS_OPS after fs_init() + * to be used by multidisk + */ +extern const struct fs_ops **p_ops; + +/* + * Used to initialize MultiFS support + */ +extern void enable_multifs(void *); +extern void init_multifs(void); + +#endif /* MULTIDISK_UTILS_H */ \ No newline at end of file diff --git a/com32/lib/syslinux...

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

Hello, the following patches should get multidisk access working. The syntax accepted is the following: (hdx,y)/path/to/file where x is the disk number and start at 0 and the y is the partition number starting at 1. So (hd0,1) is the first partition of the first disk. the other accepted syntax is using MBR's 32 bits disk signature so for example: (mbr:0x12345678,2)/foo/bar would address

...>>> +{ >>> + ufs_debug("ufs_readlink\n"); >>> + return inode->size; >>> +} >> >> Something missing here? > Yes, implementation. It's just a placeholder until I implement it. You > can see that I didn't even set it into fs_ops. At that time, I think > Matt removed it himself as compiler complained about it being unused. >> Yes, it did. However, it would be better to implement it. As far as I know, UFS symlinks are very similar to ext* symlinks (not surprising since ext2 was heavily inspired by UFS): symlinks...

..., + EXT2_SB(fs)->s_uuid[11], + EXT2_SB(fs)->s_uuid[12], + EXT2_SB(fs)->s_uuid[13], + EXT2_SB(fs)->s_uuid[14], + EXT2_SB(fs)->s_uuid[15] + ) < 0) { + free(uuid); + return NULL; + } + + return uuid; +} + const struct fs_ops ext2_fs_ops = { .fs_name = "ext2", .fs_flags = FS_THISIND | FS_USEMEM, @@ -336,5 +374,5 @@ const struct fs_ops ext2_fs_ops = { .readlink = ext2_readlink, .readdir = ext2_readdir, .next_extent = ext2_next_extent, - .fs_uuid = NULL,...

On 05/29/2014 07:20 AM, Raphael S.Carvalho wrote: > +static int ufs_readlink(struct inode *inode, char *buf) > +{ > + ufs_debug("ufs_readlink\n"); > + return inode->size; > +} Something missing here? -hpa

...t;ufs_readlink\n"); >>>>> + return inode->size; >>>>> +} >>>> >>>> Something missing here? >>> Yes, implementation. It's just a placeholder until I implement it. You >>> can see that I didn't even set it into fs_ops. At that time, I think >>> Matt removed it himself as compiler complained about it being unused. >>>> >> >> Yes, it did. However, it would be better to implement it. As far as I >> know, UFS symlinks are very similar to ext* symlinks (not surprising >>...

...ns pm_fs_init() which I believe is the >> first protected-mode code. >> >> void pm_fs_init(com32sys_t *regs) >> { >> static struct bios_disk_private priv; >> >> writechr('^'); >> priv.regs = regs; >> fs_init((const struct fs_ops **)regs->eax.l, (void *)&priv); >> } >> >> >> With the above, I see the caret. >> >> The call to this function in ISOLINUX is in core/isolinux.asm:1163 >> >> .common: >> mov ecx,[Hidden] >> mov ebx,[Hidden+4]...

On Fri, Jul 3, 2015 at 12:50 AM, poma via Syslinux <syslinux at zytor.com> wrote: > - "unsigned char c;" does not solve the problem > > - "c >= 0 && c <= ' '" solves the problem for the current git Could you try the following patch? Feel free to only apply the change to readconfig.c if you want. -- -Gene diff --git

...gt; core/fs/diskio_bios.c:395 contains pm_fs_init() which I believe is the > first protected-mode code. > > void pm_fs_init(com32sys_t *regs) > { > static struct bios_disk_private priv; > > writechr('^'); > priv.regs = regs; > fs_init((const struct fs_ops **)regs->eax.l, (void *)&priv); > } > > > With the above, I see the caret. > > The call to this function in ISOLINUX is in core/isolinux.asm:1163 > > .common: > mov ecx,[Hidden] > mov ebx,[Hidden+4] > mov si,[bsHeads] &...

...4; dev.cache_data = malloc(dev.cache_size); + dev.cache_init = 0; /* Explicitly set cache as uninitialized */ return &dev; } diff --git a/core/fs/fs.c b/core/fs/fs.c index 8c1feea..d6da8a5 100644 --- a/core/fs/fs.c +++ b/core/fs/fs.c @@ -422,8 +422,9 @@ void fs_init(const struct fs_ops **ops, void *priv) } this_fs = &fs; - /* initialize the cache */ - if (fs.fs_dev && fs.fs_dev->cache_data) + /* initialize the cache only if it wasn't already initialized + * by the fs driver */ + if (fs.fs_dev && fs.fs_dev->cache_data &...

...k(struct inode *inode, char *buf) >> +{ >> + ufs_debug("ufs_readlink\n"); >> + return inode->size; >> +} > > Something missing here? Yes, implementation. It's just a placeholder until I implement it. You can see that I didn't even set it into fs_ops. At that time, I think Matt removed it himself as compiler complained about it being unused. > > -hpa > > -- Raphael S. Carvalho

...> + ufs_debug("ufs_readlink\n"); >>>> + return inode->size; >>>> +} >>> >>> Something missing here? >> Yes, implementation. It's just a placeholder until I implement it. You >> can see that I didn't even set it into fs_ops. At that time, I think >> Matt removed it himself as compiler complained about it being unused. >>> > > Yes, it did. However, it would be better to implement it. As far as I > know, UFS symlinks are very similar to ext* symlinks (not surprising > since ext2 was heavily...

...> + ufs_debug("ufs_readlink\n"); >>>> + return inode->size; >>>> +} >>> >>> Something missing here? >> Yes, implementation. It's just a placeholder until I implement it. You >> can see that I didn't even set it into fs_ops. At that time, I think >> Matt removed it himself as compiler complained about it being unused. >>> > > Yes, it did. However, it would be better to implement it. As far as I > know, UFS symlinks are very similar to ext* symlinks (not surprising > since ext2 was heavily...