Brought back the code removed in that commit so that we can bring back
Cygwin support.
Small changes are done in the process of resolving conflicts against the
current head.
Some checks have not been reverted, because they were not necessary
anymore.
Signed-off-by: b-aaz <b-aazbsd@proton.me>
It is patch 1/2 of a series that provides a convenient way to specify
module search paths on a per-driver basis.
This patch allows to set ModulePath for certain modules, in particular,
GPU drivers.
Signed-off-by: Oleh Nykyforchyn <oleh.nyk@gmail.com>
See: https://github.com/gentoo/gentoo/blob/master/x11-base/xorg-server/files/xorg-server-1.12-unloadsubmodule.patch
See: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=686152#14
Verbatim copy of https://github.com/X11Libre/xserver/issues/319#issuecomment-3033729517 ,which gives more context for this patch:
I took a closer look at that patch.
It is logically equivalent to:
```
diff --git a/hw/xfree86/loader/loadmod.c b/hw/xfree86/loader/loadmod.c
index 2cdf91fd2..49785fdc8 100644
--- a/hw/xfree86/loader/loadmod.c
+++ b/hw/xfree86/loader/loadmod.c
@@ -885,6 +885,7 @@ RemoveChild(ModuleDescPtr child)
parent = child->parent;
if (parent->child == child) {
parent->child = child->sib;
+ child->sib = NULL;
return;
}
```
RemoveChild is a static function that is only called in UnloadSubModule:
```
void
UnloadSubModule(ModuleDescPtr mod)
{
/* Some drivers are calling us on built-in submodules, ignore them */
if (mod == (ModuleDescPtr) 1)
return;
RemoveChild(mod);
UnloadModule(mod);
}
```
Whether or not child->sib is NULL tells UnloadModule if it should recursively unload child->sib or not:
```
if (mod->child)
UnloadModule(mod->child);
if (mod->sib)
UnloadModule(mod->sib);
free(mod);
```
Looking at the source, the module loader uses some weird kind of tree-like structure,
where every node has at most one child and one sibling (but then, if foo has child bar, and bar has sibling baz, shouldn't baz also be foo's child?).
```
typedef struct module_desc {
struct module_desc *child;
struct module_desc *sib;
struct module_desc *parent;
void *handle;
ModuleSetupProc SetupProc;
ModuleTearDownProc TearDownProc;
void *TearDownData; /* returned from SetupProc */
const XF86ModuleVersionInfo *VersionInfo;
} ModuleDesc, *ModuleDescPtr;
```
All in all, this patch makes UnloadSubModule to never unload the sibling of the unloaded module, whereas
as it is now, UnloadSubModule would also unload the module's sibling if `child->parent == child->parent->child`
(master child?).
I don't see how this patch changed the behavior on ia64, or any other arch.
@metux Could you tell me what kind of data structure this is, and whether or not this patch is right?
Fixes: https://github.com/X11Libre/xserver/issues/319
Signed-off-by: stefan11111 <stefan11111@shitposting.expert>
Only specify the top module directories since the loader is recursing
into the subdirectories either way.
Fixes: #991
Signed-off-by: callmetango <callmetango@users.noreply.github.com>
It is patch 1/7 of a series that provides a convenient way to specify
IgnoreABI and module search paths on a per-driver basis.
It adds #ifdef CONFIG_LEGACY_NVIDIA_PADDING wherever it should be but
is missed.
Signed-off-by: Oleh Nykyforchyn <oleh.nyk@gmail.com>
and new versions of the proprietary nvidia DDX driver
The 470 driver expects older abi, while the 570 driver
uses different code at runtime depending of the abi version.
This commit tells the new nvdia driver to use the older
abi that the 470 driver expects.
Signed-off-by: stefan11111 <stefan11111@shitposting.expert>
* switch major version from 25.0 to 25
* make loader respect legacy 25.0 module dirs (to be removed in version
26)
Fixes: #646
Signed-off-by: callmetango <callmetango@users.noreply.github.com>
IgnoreABI option is kept in a single bit of an unsigned long variable
LoaderOptions that has no other use.
This patch replaces it with a variable named LoaderIgnoreAbi
and a proc for setting it.
Inspired by b61b35a0b3
Signed-off-by: stefan11111 <stefan11111@shitposting.expert>
In CheckVersion() the errtype variable is used in two separate scopes,
but not globally, so it's cleaner to have it only in the scopes that
are actually using it.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Modules are now placed into a sub-directory by major Xserver release,
so we have less hassle with trying to load drivers w/ incompatible ABI.
The legacy directories are still searched (after the versioned ones)
for backwards compat with badly maintained proprietary drivers (Nvidia).
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Only key difference that calloc(), in contrast to rellocarray(),
is zero-initializing. The overhead is hard to measure on today's
machines, and it's safer programming practise to always allocate
zero-initialized, so one can't forget to do it explicitly.
Cocci rule:
@@
expression COUNT;
expression LEN;
@@
- xallocarray(COUNT,LEN)
+ calloc(COUNT,LEN)
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Using calloc() instead of malloc() as preventive measure, so there
never can be any hidden bugs or leaks due uninitialized memory.
The extra cost of using this compiler intrinsic should be practically
impossible to measure - in many cases a good compiler can even deduce
if certain areas really don't need to be zero'd (because they're written
to right after allocation) and create more efficient machine code.
The code pathes in question are pretty cold anyways, so it's probably
not worth even thinking about potential extra runtime costs.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
There's (remote) chance that the (internal) module name could become
NULL (eg. allocation failure). Even though chances to hit it are very
low, it's still better to have a check here (that doesn't cost us much),
just in case. Assert fail is still better than segfault, since we're
at least getting some hint what might have happened.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
This has been nothing but an alias for two decades now (somewhere in R6.6),
so there doesn't seem to be any practical need for this indirection.
The macro still needs to remain, as long as (external) drivers still using it.
Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Part-of: <https://gitlab.freedesktop.org/xorg/xserver/-/merge_requests/1529>
There's a free(name) at the end of the function.
GCC warned about this:
../hw/xfree86/loader/loadmod.c: In function ‘LoadModule’:
../hw/xfree86/loader/loadmod.c:702:18: warning: assignment discards ‘const’ qualifier from pointer target type [-Wdiscarded-qualifiers]
702 | m = name = "int10";
| ^
The thing you get back from xf86LoadSubModule is a ModuleDescPtr, not a
dlsym handle. We don't expose ModuleDescPtr to the drivers, so change
LoaderSymbolFromModule to cast its void * argument to a ModuleDescPtr.
Reviewed-by: Michel Dänzer <mdaenzer@redhat.com>
Roundhouse kick replacing the various (sizeof(foo)/sizeof(foo[0])) with
the ARRAY_SIZE macro from dix.h when possible. A semantic patch for
coccinelle has been used first. Additionally, a few macros have been
inlined as they had only one or two users.
Signed-off-by: Daniel Martin <consume.noise@gmail.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Because we can use a shared lib in dlsym() for the symbol loader test,
but we can't use the executable.
v2: Drop a stray ' ' and rebase on 820a4cbe9f, by anholt
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Eric Anholt <eric@anholt.net>
This is a work in progress that builds Xvfb, Xephyr, Xwayland, Xnest,
and Xdmx so far. The outline of Xquartz/Xwin support is in tree, but
hasn't been built yet. The unit tests are also not done.
The intent is to build this as a complete replacement for the
autotools system, then eventually replace autotools. meson is faster
to generate the build, faster to run the bulid, shorter to write the
build files in, and less error-prone than autotools.
v2: Fix indentation nits, move version declaration to project(), use
existing meson_options for version-config.h's vendor name/web.
Signed-off-by: Eric Anholt <eric@anholt.net>
Acked-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
We mostly use #ifdef throughout the tree, and this lets the generated
config.h files just be #define TOKEN instead of #define TOKEN 1.
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Keith Packard <keithp@keithp.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Eric Anholt <eric@anholt.net>
This can happen when a module fails to load:
Program received signal SIGSEGV, Segmentation fault.
UnloadModule (_mod=0x5555559d9280) at ../../../../hw/xfree86/loader/loadmod.c:848
848 name = mod->VersionInfo->modname;
(gdb) bt
#0 UnloadModule (_mod=0x5555559d9280) at ../../../../hw/xfree86/loader/loadmod.c:848
#1 0x00005555555ddd1b in LoadModule (module=module@entry=0x5555559c7ce0 "fbdev", options=0x0, modreq=modreq@entry=0x0, errmaj=errmaj@entry=0x7fffffffe8ec) at ../../../../hw/xfree86/loader/loadmod.c:824
#2 0x00005555555edfe9 in xf86LoadModules (list=list@entry=0x5555559dcf50, optlist=optlist@entry=0x0) at ../../../../hw/xfree86/common/xf86Init.c:1506
#3 0x00005555555ee7bc in InitOutput (pScreenInfo=pScreenInfo@entry=0x5555559abf80 <screenInfo>, argc=argc@entry=4, argv=argv@entry=0x7fffffffeb18) at ../../../../hw/xfree86/common/xf86Init.c:484
#4 0x00005555555a885c in dix_main (argc=4, argv=0x7fffffffeb18, envp=<optimized out>) at ../../dix/main.c:197
#5 0x00007ffff5d582b1 in __libc_start_main (main=0x555555593130 <main>, argc=4, argv=0x7fffffffeb18, init=<optimized out>, fini=<optimized out>, rtld_fini=<optimized out>, stack_end=0x7fffffffeb08) at ../csu/libc-start.c:291
#6 0x000055555559316a in _start ()
Fixes: 8e83eacb9e ("loader: Remove unused path and name from ModuleDescPtr")
Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Just a waste of memory. Path was never referenced at all, and name was
only used when unloading the module; we can just as well get the
module's internal idea of its name from VersionInfo.
Reviewed-by: Julien Cristau <jcristau@debian.org>
Signed-off-by: Adam Jackson <ajax@redhat.com>
Nobody was ever calling this with a non-null argument for subdir list or
pattern list. Having done this, InitSubdirs is only ever called with a
NULL argument, so it's really just a complicated way of duplicating the
default list; we can remove that and just walk the list directly.
The minor error code was only ever used to distinguish among two cases
of LDR_BADUSAGE. Whatever.
Reviewed-by: Julien Cristau <jcristau@debian.org>
Signed-off-by: Adam Jackson <ajax@redhat.com>
Callers only ever use this for a single directory anyway.
While we're at it, also move xf86DriverListFromCompile near its only
user in the X -configure code (and inline it out of existence), and
remove LoaderFreeDirList as it's unused (since X -configure is just
going to exit anyway, none of that code cares about cleanup).
Reviewed-by: Julien Cristau <jcristau@debian.org>
Signed-off-by: Adam Jackson <ajax@redhat.com>
Everybody using this functionality specifies a major version, which
makes sense. If you don't care about a minor version, that's equivalent
to saying you require minor >= 0, so just say so; likewise patch level.
Likewise ABI class is always specified.
Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Julien Cristau <jcristau@debian.org>
Signed-off-by: Adam Jackson <ajax@redhat.com>
The enum has been unused since at least the removal of elfloader.
Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Julien Cristau <jcristau@debian.org>
Signed-off-by: Adam Jackson <ajax@redhat.com>
