drm_legacy_ctxbitmap_next() returns idr_alloc() which can return -ENOMEM, -EINVAL or -ENOSPC none of which are -1 . but the call sites of drm_legacy_ctxbitmap_next() seem to be assuming that the error case would be -1 (original return of drm_ctxbitmap_next() prior to 2.6.23 was actually -1). Thus reenable error handling by checking for < 0.
Signed-off-by: Nicholas Mc Guire hofrat@osadl.org Fixes: 62968144e673 ("drm: convert drm context code to use Linux idr") ---
Problem located with experimental coccinelle script
The noted Fixes tag is one of the commits that removed the -1 return value in 2.6.23 in drm_ctxbitmap_next() but thats not the only change that must have taken place to invalidate the error check.
Patch was compile tested with: x86_64_defconfig
Patch is against 4.18-rc4 (localversion-next is next-20180713)
drivers/gpu/drm/drm_context.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/drm_context.c b/drivers/gpu/drm/drm_context.c index 3c4000f..f973d28 100644 --- a/drivers/gpu/drm/drm_context.c +++ b/drivers/gpu/drm/drm_context.c @@ -372,7 +372,7 @@ int drm_legacy_addctx(struct drm_device *dev, void *data, ctx->handle = drm_legacy_ctxbitmap_next(dev); } DRM_DEBUG("%d\n", ctx->handle); - if (ctx->handle == -1) { + if (ctx->handle < 0) { DRM_DEBUG("Not enough free contexts.\n"); /* Should this return -EBUSY instead? */ return -ENOMEM;