Re: [PATCH] drm/v3d: remove duplicated kfree in v3d_submit_cl_ioctl

[yukuai (C)]

On 2019/12/29 4:45, Markus Elfring wrote:
> > v3d_submit_cl_ioctl call kfree() with variable 'bin' twice.
>
> I would prefer a wording like âkfree() was called for the same variable twice
> within an if branch.â.
>
>
> > Fix it by removing the latter one.
>
> I find the wording âDelete a duplicate function call.â more appropriate.
Thank you for your advise, I'll make changes in V2 patch.

> Please add the tag âFixesâ to your change description.

I got the results from "git blame":
git blame -L 570,575 drivers/gpu/drm/v3d/v3d_gem.c
a783a09ee76d6 (Eric Anholt        2019-04-16 15:58:53 -0700 570) 
       if (ret) {
0d352a3a8a1f2 (Iago Toral Quiroga 2019-09-16 09:11:25 +0200 571) 
               kfree(bin);
a783a09ee76d6 (Eric Anholt        2019-04-16 15:58:53 -0700 572) 
               v3d_job_put(&render->base);
29cd13cfd7624 (Navid Emamdoost    2019-10-21 13:52:49 -0500 573) 
               kfree(bin);
a783a09ee76d6 (Eric Anholt        2019-04-16 15:58:53 -0700 574) 
               return ret;
a783a09ee76d6 (Eric Anholt        2019-04-16 15:58:53 -0700 575) 
       }

The first kfree belong to the patch 0d352a3a8a1f2 :
commit 0d352a3a8a1f26168d09f7073e61bb4b328e3bb9
Author: Iago Toral Quiroga <itoral@xxxxxxxxxx>
Date:   Mon Sep 16 09:11:25 2019 +0200

    drm/v3d: don't leak bin job if v3d_job_init fails.

    If the initialization of the job fails we need to kfree() it
    before returning.

    Signed-off-by: Iago Toral Quiroga <itoral@xxxxxxxxxx>
    Signed-off-by: Eric Anholt <eric@xxxxxxxxxx>
    Link: 
https://patchwork.freedesktop.org/patch/msgid/20190916071125.5255-1-itoral@xxxxxxxxxx
    Fixes: a783a09ee76d ("drm/v3d: Refactor job management.")
    Reviewed-by: Eric Anholt <eric@xxxxxxxxxx>

diff --git a/drivers/gpu/drm/v3d/v3d_gem.c b/drivers/gpu/drm/v3d/v3d_gem.c
index 5d80507b539b..fb32cda18ffe 100644
--- a/drivers/gpu/drm/v3d/v3d_gem.c
+++ b/drivers/gpu/drm/v3d/v3d_gem.c
@@ -563,6 +563,7 @@ v3d_submit_cl_ioctl(struct drm_device *dev, void *data,
                ret = v3d_job_init(v3d, file_priv, &bin->base,
                                   v3d_job_free, args->in_sync_bcl);
                if (ret) {
+                       kfree(bin);
                        v3d_job_put(&render->base);
                        return ret;
                }

And the second belong to 29cd13cfd7624:
commit 29cd13cfd7624726d9e6becbae9aa419ef35af7f
Author: Navid Emamdoost <navid.emamdoost@xxxxxxxxx>
Date:   Mon Oct 21 13:52:49 2019 -0500

    drm/v3d: Fix memory leak in v3d_submit_cl_ioctl

    In the impelementation of v3d_submit_cl_ioctl() there are two memory
    leaks. One is when allocation for bin fails, and the other is when bin
    initialization fails. If kcalloc fails to allocate memory for bin then
    render->base should be put. Also, if v3d_job_init() fails to initialize
    bin->base then allocated memory for bin should be released.

    Fixes: a783a09ee76d ("drm/v3d: Refactor job management.")
    Signed-off-by: Navid Emamdoost <navid.emamdoost@xxxxxxxxx>
    Reviewed-by: Eric Anholt <eric@xxxxxxxxxx>
    Signed-off-by: Daniel Vetter <daniel.vetter@xxxxxxxx>
    Link: 
https://patchwork.freedesktop.org/patch/msgid/20191021185250.26130-1-navid.emamdoost@xxxxxxxxx

diff --git a/drivers/gpu/drm/v3d/v3d_gem.c b/drivers/gpu/drm/v3d/v3d_gem.c
index 5d80507b539b..19c092d75266 100644
--- a/drivers/gpu/drm/v3d/v3d_gem.c
+++ b/drivers/gpu/drm/v3d/v3d_gem.c
@@ -557,13 +557,16 @@ v3d_submit_cl_ioctl(struct drm_device *dev, void 
*data,

        if (args->bcl_start != args->bcl_end) {
                bin = kcalloc(1, sizeof(*bin), GFP_KERNEL);
-               if (!bin)
+               if (!bin) {
+                       v3d_job_put(&render->base);
                        return -ENOMEM;
+               }

                ret = v3d_job_init(v3d, file_priv, &bin->base,
                                   v3d_job_free, args->in_sync_bcl);
                if (ret) {
                        v3d_job_put(&render->base);
+                       kfree(bin);
                        return ret;
                }

It seems the two patches fix the same memory leak, but I have no idea 
how thet get together without conflict.

Thanks
Yu Kuai