| Message ID | 1669470654-45828-1-git-send-email-wangyufen@huawei.com (mailing list archive) |
|---|---|
| State | Superseded |
| Headers | show |
| Series | [v3,1/2] RDMA/hfi1: Fix error return code in parse_platform_config() | expand |
On Sat, Nov 26, 2022 at 09:50:53PM +0800, Wang Yufen wrote: > In the previous while loop, "ret" may be assigned zero, so the error > return code may be incorrectly set to 0 instead of -EINVAL. > Add bail_with_einval goto label and covert all "goto bail;" to "goto > bail_with_einval:" where it's appropriate. Add dropping some duplicative > "ret = -EINVAL;" lines, as Andy suggessted. > > Fixes: 97167e813415 ("staging/rdma/hfi1: Tune for unknown channel if configuration file is absent") > Signed-off-by: Wang Yufen <wangyufen@huawei.com> > Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> > --- > drivers/infiniband/hw/hfi1/firmware.c | 20 +++++++++++--------- > 1 file changed, 11 insertions(+), 9 deletions(-) <...> > +bail_with_einval: > + ret = -EINVAL; Sorry for being late, but no. It can be seen as anti-pattern as it causes to the situations where unrelated code changes can potentially overwrite return value. Please set valid return code before calling to goto. Thanks > bail: > memset(pcfgcache, 0, sizeof(struct platform_config_cache)); > return ret; > -- > 1.8.3.1 >
在 2022/11/28 19:57, Leon Romanovsky 写道: > On Sat, Nov 26, 2022 at 09:50:53PM +0800, Wang Yufen wrote: >> In the previous while loop, "ret" may be assigned zero, so the error >> return code may be incorrectly set to 0 instead of -EINVAL. >> Add bail_with_einval goto label and covert all "goto bail;" to "goto >> bail_with_einval:" where it's appropriate. Add dropping some duplicative >> "ret = -EINVAL;" lines, as Andy suggessted. >> >> Fixes: 97167e813415 ("staging/rdma/hfi1: Tune for unknown channel if configuration file is absent") >> Signed-off-by: Wang Yufen <wangyufen@huawei.com> >> Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> >> --- >> drivers/infiniband/hw/hfi1/firmware.c | 20 +++++++++++--------- >> 1 file changed, 11 insertions(+), 9 deletions(-) > > <...> > > >> +bail_with_einval: >> + ret = -EINVAL; > > Sorry for being late, but no. It can be seen as anti-pattern as it > causes to the situations where unrelated code changes can potentially > overwrite return value. Please set valid return code before calling to > goto. OK, will change in v4. Thanks > > Thanks > >> bail: >> memset(pcfgcache, 0, sizeof(struct platform_config_cache)); >> return ret; >> -- >> 1.8.3.1 >>
diff --git a/drivers/infiniband/hw/hfi1/firmware.c b/drivers/infiniband/hw/hfi1/firmware.c index 1d77514..44d8628 100644 --- a/drivers/infiniband/hw/hfi1/firmware.c +++ b/drivers/infiniband/hw/hfi1/firmware.c @@ -1730,7 +1730,7 @@ int parse_platform_config(struct hfi1_devdata *dd) u32 *ptr = NULL; u32 header1 = 0, header2 = 0, magic_num = 0, crc = 0, file_length = 0; u32 record_idx = 0, table_type = 0, table_length_dwords = 0; - int ret = -EINVAL; /* assume failure */ + int ret; /* * For integrated devices that did not fall back to the default file, @@ -1743,7 +1743,7 @@ int parse_platform_config(struct hfi1_devdata *dd) if (!dd->platform_config.data) { dd_dev_err(dd, "%s: Missing config file\n", __func__); - goto bail; + goto bail_with_einval; } ptr = (u32 *)dd->platform_config.data; @@ -1751,7 +1751,7 @@ int parse_platform_config(struct hfi1_devdata *dd) ptr++; if (magic_num != PLATFORM_CONFIG_MAGIC_NUM) { dd_dev_err(dd, "%s: Bad config file\n", __func__); - goto bail; + goto bail_with_einval; } /* Field is file size in DWORDs */ @@ -1774,7 +1774,7 @@ int parse_platform_config(struct hfi1_devdata *dd) if (file_length > dd->platform_config.size) { dd_dev_info(dd, "%s:File claims to be larger than read size\n", __func__); - goto bail; + goto bail_with_einval; } else if (file_length < dd->platform_config.size) { dd_dev_info(dd, "%s:File claims to be smaller than read size, continuing\n", @@ -1794,7 +1794,7 @@ int parse_platform_config(struct hfi1_devdata *dd) dd_dev_err(dd, "%s: Failed validation at offset %ld\n", __func__, (ptr - (u32 *) dd->platform_config.data)); - goto bail; + goto bail_with_einval; } record_idx = *ptr & @@ -1837,7 +1837,7 @@ int parse_platform_config(struct hfi1_devdata *dd) __func__, table_type, (ptr - (u32 *) dd->platform_config.data)); - goto bail; /* We don't trust this file now */ + goto bail_with_einval; /* We don't trust this file now */ } pcfgcache->config_tables[table_type].table = ptr; } else { @@ -1856,7 +1856,7 @@ int parse_platform_config(struct hfi1_devdata *dd) __func__, table_type, (ptr - (u32 *)dd->platform_config.data)); - goto bail; /* We don't trust this file now */ + goto bail_with_einval; /* We don't trust this file now */ } pcfgcache->config_tables[table_type].table_metadata = ptr; @@ -1873,8 +1873,7 @@ int parse_platform_config(struct hfi1_devdata *dd) dd_dev_err(dd, "%s: Failed CRC check at offset %ld\n", __func__, (ptr - (u32 *)dd->platform_config.data)); - ret = -EINVAL; - goto bail; + goto bail_with_einval; } /* Jump the CRC DWORD */ ptr++; @@ -1882,6 +1881,9 @@ int parse_platform_config(struct hfi1_devdata *dd) pcfgcache->cache_valid = 1; return 0; + +bail_with_einval: + ret = -EINVAL; bail: memset(pcfgcache, 0, sizeof(struct platform_config_cache)); return ret;
In the previous while loop, "ret" may be assigned zero, so the error return code may be incorrectly set to 0 instead of -EINVAL. Add bail_with_einval goto label and covert all "goto bail;" to "goto bail_with_einval:" where it's appropriate. Add dropping some duplicative "ret = -EINVAL;" lines, as Andy suggessted. Fixes: 97167e813415 ("staging/rdma/hfi1: Tune for unknown channel if configuration file is absent") Signed-off-by: Wang Yufen <wangyufen@huawei.com> Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> --- drivers/infiniband/hw/hfi1/firmware.c | 20 +++++++++++--------- 1 file changed, 11 insertions(+), 9 deletions(-)