firmware: give a protection when map page failed
authorzhang jun <jun.zhang@intel.com>
Thu, 13 Feb 2014 07:18:47 +0000 (15:18 +0800)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Sat, 15 Feb 2014 19:31:55 +0000 (11:31 -0800)
so, we need give a protection and return a error value.
[ 7341.474236] [drm:do_intel_finish_page_flip] *ERROR* invalid or inactive unpin_work!
[ 7341.494464] atomisp-css2400b0_v21 0000:00:03.0: unhandled css stored event: 0x20
[ 7341.503627] vmap allocation for size 208896 failed: use vmalloc=<size> to increase size.<=================== map failed
[ 7341.507135] [drm:do_intel_finish_page_flip] *ERROR* invalid or inactive unpin_work!
[ 7341.503848] BUG: unable to handle kernel NULL pointer dereference at   (null)
[ 7341.520394] IP: [<c18f5c1b>] sst_load_all_modules_elf+0x1bb/0x850
[ 7341.527216] *pdpt = 0000000030dfe001 *pde = 0000000000000000
[ 7341.533640] Oops: 0000 [#1] PREEMPT SMP
[ 7341.540360] [drm:do_intel_finish_page_flip] *ERROR* invalid or inactive unpin_work!
[ 7341.538037] Modules linked in: atomisp_css2400b0_v21 lm3554 ov2722 imx1x5 atmel_mxt_ts vxd392 videobuf_vmalloc videobuf_core lm_dump(O) bcm_bt_lpm hdmi_audio bcm4334x(O)
[ 7341.563531] CPU: 1 PID: 525 Comm: mediaserver Tainted: G        W  O 3.10.20-262518-ga83c053 #1
[ 7341.573253] task: f0994ec0 ti: f09f0000 task.ti: f09f0000
[ 7341.579284] EIP: 0060:[<c18f5c1b>] EFLAGS: 00010246 CPU: 1
[ 7341.585415] EIP is at sst_load_all_modules_elf+0x1bb/0x850
[ 7341.591541] EAX: 00000000 EBX: e3595ba0 ECX: 00000000 EDX: 00031c1c
[ 7341.598541] ESI: e04a0000 EDI: 00000000 EBP: f09f1d80 ESP: f09f1cf4
[ 7341.605542]  DS: 007b ES: 007b FS: 00d8 GS: 003b SS: 0068
[ 7341.611573] CR0: 80050033 CR2: 00000000 CR3: 30db4000 CR4: 001007f0
[ 7341.618573] DR0: 00000000 DR1: 00000000 DR2: 00000000 DR3: 00000000
[ 7341.625575] DR6: ffff0ff0 DR7: 00000400
[ 7341.629856] Stack:
[ 7341.632097]  f09f1d57 00000019 c1d656d7 c1d658d3 c1d56409 00000f28 c1d64af9 18000103
[ 7341.640766]  01000001 00080000 c1f910a0 f326f4c8 00000034 f326f520 00000002 e04a02bc
[ 7341.649465]  00000001 f326e014 c1f910b0 e04a0000 c0080100 00031c1c e3595ba0 c0080100
[ 7341.658149] Call Trace:
[ 7341.660888]  [<c18f6308>] sst_post_download_byt+0x58/0xb0
[ 7341.666925]  [<c18f4fbc>] sst_load_fw+0xdc/0x510
[ 7341.672086]  [<c1a7b2c0>] ? __mutex_lock_slowpath+0x250/0x370
[ 7341.678507]  [<c1a80e05>] ? sub_preempt_count+0x55/0xe0
[ 7341.684346]  [<c18f1294>] sst_download_fw+0x14/0x60
[ 7341.689796]  [<c1a7b403>] ? mutex_lock+0x23/0x30
[ 7341.694954]  [<c18f191c>] intel_sst_check_device+0x6c/0x120
[ 7341.701181]  [<c18f1d08>] sst_set_generic_params+0x1b8/0x4a0
[ 7341.707504]  [<c1a80e05>] ? sub_preempt_count+0x55/0xe0
[ 7341.713341]  [<c1a80e05>] ? sub_preempt_count+0x55/0xe0
[ 7341.719178]  [<c1a7b2c0>] ? __mutex_lock_slowpath+0x250/0x370
[ 7341.725600]  [<c1320d84>] ? __kmalloc_track_caller+0xe4/0x1d0
[ 7341.732022]  [<c18e35f5>] sst_set_mixer_param+0x25/0x40
[ 7341.737859]  [<c18e3853>] lpe_mixer_ihf_set+0xb3/0x160
[ 7341.743602]  [<c1855b99>] snd_ctl_ioctl+0xa89/0xb40
[ 7341.749052]  [<c1331e65>] ? path_openat+0xa5/0x3d0
[ 7341.754409]  [<c1447857>] ? avc_has_perm_flags+0xc7/0x170
[ 7341.760441]  [<c1855110>] ? snd_ctl_elem_add_user+0x540/0x540
[ 7341.766862]  [<c1334047>] do_vfs_ioctl+0x77/0x5e0
[ 7341.772117]  [<c144842a>] ? inode_has_perm.isra.42.constprop.79+0x3a/0x50
[ 7341.779705]  [<c14490a0>] ? file_has_perm+0xa0/0xb0
[ 7341.785155]  [<c14493b8>] ? selinux_file_ioctl+0x48/0xe0
[ 7341.791090]  [<c1334628>] SyS_ioctl+0x78/0x90
[ 7341.795958]  [<c1a7dde8>] syscall_call+0x7/0xb
[ 7341.800925]  [<c1a70000>] ? mm_fault_error+0x13c/0x198

Signed-off-by: zhang jun <jun.zhang@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drivers/base/firmware_class.c

index 8a97ddfa61222db79506dcecebc619ab5a333865..e5808eb63b67321de4b16e4a2e1b643e0e4f0d4b 100644 (file)
@@ -649,7 +649,9 @@ static ssize_t firmware_loading_store(struct device *dev,
                         * see the mapped 'buf->data' once the loading
                         * is completed.
                         * */
-                       fw_map_pages_buf(fw_buf);
+                       if (fw_map_pages_buf(fw_buf))
+                               dev_err(dev, "%s: map pages failed\n",
+                                       __func__);
                        list_del_init(&fw_buf->pending_list);
                        complete_all(&fw_buf->completion);
                        break;
@@ -908,6 +910,8 @@ static int _request_firmware_load(struct firmware_priv *fw_priv,
        wait_for_completion(&buf->completion);
 
        cancel_delayed_work_sync(&fw_priv->timeout_work);
+       if (!buf->data)
+               retval = -ENOMEM;
 
        device_remove_file(f_dev, &dev_attr_loading);
 err_del_bin_attr: