| Message ID | 20190502151548.11143-1-natechancellor@gmail.com (mailing list archive) |
|---|---|
| State | Changes Requested |
| Delegated to: | Kalle Valo |
| Headers | show |
| Series | rsi: Properly initialize data in rsi_sdio_ta_reset | expand |
On Thu, May 2, 2019 at 8:16 AM Nathan Chancellor <natechancellor@gmail.com> wrote: > > When building with -Wuninitialized, Clang warns: > > drivers/net/wireless/rsi/rsi_91x_sdio.c:940:43: warning: variable 'data' > is uninitialized when used here [-Wuninitialized] > put_unaligned_le32(TA_HOLD_THREAD_VALUE, data); > ^~~~ > drivers/net/wireless/rsi/rsi_91x_sdio.c:930:10: note: initialize the > variable 'data' to silence this warning > u8 *data; > ^ > = NULL > 1 warning generated. > > Using Clang's suggestion of initializing data to NULL wouldn't work out > because data will be dereferenced by put_unaligned_le32. Use kzalloc to > properly initialize data, which matches a couple of other places in this > driver. > > Fixes: e5a1ecc97e5f ("rsi: add firmware loading for 9116 device") > Link: https://github.com/ClangBuiltLinux/linux/issues/464 > Signed-off-by: Nathan Chancellor <natechancellor@gmail.com> > --- > drivers/net/wireless/rsi/rsi_91x_sdio.c | 21 ++++++++++++++------- > 1 file changed, 14 insertions(+), 7 deletions(-) > > diff --git a/drivers/net/wireless/rsi/rsi_91x_sdio.c b/drivers/net/wireless/rsi/rsi_91x_sdio.c > index f9c67ed473d1..b35728564c7b 100644 > --- a/drivers/net/wireless/rsi/rsi_91x_sdio.c > +++ b/drivers/net/wireless/rsi/rsi_91x_sdio.c > @@ -929,11 +929,15 @@ static int rsi_sdio_ta_reset(struct rsi_hw *adapter) > u32 addr; > u8 *data; > > + data = kzalloc(sizeof(u32), GFP_KERNEL); Something fishy is going on here. We allocate 4 B but declare data as a u8* (pointer to individual bytes)? In general, dynamically allocating that few bytes is a code smell; either you meant to just use the stack, or this memory's lifetime extends past the lifetime of this stackframe, at which point you probably just meant to stack allocate space in a higher parent frame and pass this preallocated memory down to the child frame to get filled in. Reading through this code, I don't think that the memory is meant to outlive the stack frame. Is there a reason why we can't just declare data as: u8 data [4]; then use ARRAY_SIZE(data) or RSI_9116_REG_SIZE in rsi_reset_chip(), getting rid of the kzalloc/kfree? (Sorry, I hate when a simple fixup becomes a "hey let's rewrite all this code" thus becoming "that guy.")
On Thu, May 02, 2019 at 11:18:01AM -0700, Nick Desaulniers wrote: > On Thu, May 2, 2019 at 8:16 AM Nathan Chancellor > <natechancellor@gmail.com> wrote: > > > > When building with -Wuninitialized, Clang warns: > > > > drivers/net/wireless/rsi/rsi_91x_sdio.c:940:43: warning: variable 'data' > > is uninitialized when used here [-Wuninitialized] > > put_unaligned_le32(TA_HOLD_THREAD_VALUE, data); > > ^~~~ > > drivers/net/wireless/rsi/rsi_91x_sdio.c:930:10: note: initialize the > > variable 'data' to silence this warning > > u8 *data; > > ^ > > = NULL > > 1 warning generated. > > > > Using Clang's suggestion of initializing data to NULL wouldn't work out > > because data will be dereferenced by put_unaligned_le32. Use kzalloc to > > properly initialize data, which matches a couple of other places in this > > driver. > > > > Fixes: e5a1ecc97e5f ("rsi: add firmware loading for 9116 device") > > Link: https://github.com/ClangBuiltLinux/linux/issues/464 > > Signed-off-by: Nathan Chancellor <natechancellor@gmail.com> > > --- > > drivers/net/wireless/rsi/rsi_91x_sdio.c | 21 ++++++++++++++------- > > 1 file changed, 14 insertions(+), 7 deletions(-) > > > > diff --git a/drivers/net/wireless/rsi/rsi_91x_sdio.c b/drivers/net/wireless/rsi/rsi_91x_sdio.c > > index f9c67ed473d1..b35728564c7b 100644 > > --- a/drivers/net/wireless/rsi/rsi_91x_sdio.c > > +++ b/drivers/net/wireless/rsi/rsi_91x_sdio.c > > @@ -929,11 +929,15 @@ static int rsi_sdio_ta_reset(struct rsi_hw *adapter) > > u32 addr; > > u8 *data; > > > > + data = kzalloc(sizeof(u32), GFP_KERNEL); > > Something fishy is going on here. We allocate 4 B but declare data as > a u8* (pointer to individual bytes)? In general, dynamically > allocating that few bytes is a code smell; either you meant to just > use the stack, or this memory's lifetime extends past the lifetime of > this stackframe, at which point you probably just meant to stack > allocate space in a higher parent frame and pass this preallocated > memory down to the child frame to get filled in. > > Reading through this code, I don't think that the memory is meant to > outlive the stack frame. Is there a reason why we can't just declare > data as: > > u8 data [4]; data was __le32 in rsi_reset_chip() before commit f700546682a6 ("rsi: fix nommu_map_sg overflow kernel panic"). I wonder if this would be okay for this function: ------------------------------------------------- diff --git a/drivers/net/wireless/rsi/rsi_91x_sdio.c b/drivers/net/wireless/rsi/rsi_91x_sdio.c index f9c67ed473d1..0330c50ab99c 100644 --- a/drivers/net/wireless/rsi/rsi_91x_sdio.c +++ b/drivers/net/wireless/rsi/rsi_91x_sdio.c @@ -927,7 +927,7 @@ static int rsi_sdio_ta_reset(struct rsi_hw *adapter) { int status; u32 addr; - u8 *data; + u8 data; status = rsi_sdio_master_access_msword(adapter, TA_BASE_ADDR); if (status < 0) { @@ -937,7 +937,7 @@ static int rsi_sdio_ta_reset(struct rsi_hw *adapter) } rsi_dbg(INIT_ZONE, "%s: Bring TA out of reset\n", __func__); - put_unaligned_le32(TA_HOLD_THREAD_VALUE, data); + put_unaligned_le32(TA_HOLD_THREAD_VALUE, &data); addr = TA_HOLD_THREAD_REG | RSI_SD_REQUEST_MASTER; status = rsi_sdio_write_register_multiple(adapter, addr, (u8 *)&data, @@ -947,7 +947,7 @@ static int rsi_sdio_ta_reset(struct rsi_hw *adapter) return status; } - put_unaligned_le32(TA_SOFT_RST_CLR, data); + put_unaligned_le32(TA_SOFT_RST_CLR, &data); addr = TA_SOFT_RESET_REG | RSI_SD_REQUEST_MASTER; status = rsi_sdio_write_register_multiple(adapter, addr, (u8 *)&data, @@ -957,7 +957,7 @@ static int rsi_sdio_ta_reset(struct rsi_hw *adapter) return status; } - put_unaligned_le32(TA_PC_ZERO, data); + put_unaligned_le32(TA_PC_ZERO, &data); addr = TA_TH0_PC_REG | RSI_SD_REQUEST_MASTER; status = rsi_sdio_write_register_multiple(adapter, addr, (u8 *)&data, @@ -967,7 +967,7 @@ static int rsi_sdio_ta_reset(struct rsi_hw *adapter) return -EINVAL; } - put_unaligned_le32(TA_RELEASE_THREAD_VALUE, data); + put_unaligned_le32(TA_RELEASE_THREAD_VALUE, &data); addr = TA_RELEASE_THREAD_REG | RSI_SD_REQUEST_MASTER; status = rsi_sdio_write_register_multiple(adapter, addr, (u8 *)&data, > > then use ARRAY_SIZE(data) or RSI_9116_REG_SIZE in rsi_reset_chip(), > getting rid of the kzalloc/kfree? > > (Sorry, I hate when a simple fixup becomes a "hey let's rewrite all > this code" thus becoming "that guy.") If we aren't actually improving the code, then why bother? :) Thank you for the review! Nathan > -- > Thanks, > ~Nick Desaulniers
Nick Desaulniers <ndesaulniers@google.com> writes: > On Thu, May 2, 2019 at 8:16 AM Nathan Chancellor > <natechancellor@gmail.com> wrote: >> >> When building with -Wuninitialized, Clang warns: >> >> drivers/net/wireless/rsi/rsi_91x_sdio.c:940:43: warning: variable 'data' >> is uninitialized when used here [-Wuninitialized] >> put_unaligned_le32(TA_HOLD_THREAD_VALUE, data); >> ^~~~ >> drivers/net/wireless/rsi/rsi_91x_sdio.c:930:10: note: initialize the >> variable 'data' to silence this warning >> u8 *data; >> ^ >> = NULL >> 1 warning generated. >> >> Using Clang's suggestion of initializing data to NULL wouldn't work out >> because data will be dereferenced by put_unaligned_le32. Use kzalloc to >> properly initialize data, which matches a couple of other places in this >> driver. >> >> Fixes: e5a1ecc97e5f ("rsi: add firmware loading for 9116 device") >> Link: https://github.com/ClangBuiltLinux/linux/issues/464 >> Signed-off-by: Nathan Chancellor <natechancellor@gmail.com> >> --- >> drivers/net/wireless/rsi/rsi_91x_sdio.c | 21 ++++++++++++++------- >> 1 file changed, 14 insertions(+), 7 deletions(-) >> >> diff --git a/drivers/net/wireless/rsi/rsi_91x_sdio.c b/drivers/net/wireless/rsi/rsi_91x_sdio.c >> index f9c67ed473d1..b35728564c7b 100644 >> --- a/drivers/net/wireless/rsi/rsi_91x_sdio.c >> +++ b/drivers/net/wireless/rsi/rsi_91x_sdio.c >> @@ -929,11 +929,15 @@ static int rsi_sdio_ta_reset(struct rsi_hw *adapter) >> u32 addr; >> u8 *data; >> >> + data = kzalloc(sizeof(u32), GFP_KERNEL); > > Something fishy is going on here. We allocate 4 B but declare data as > a u8* (pointer to individual bytes)? In general, dynamically > allocating that few bytes is a code smell; either you meant to just > use the stack, or this memory's lifetime extends past the lifetime of > this stackframe, at which point you probably just meant to stack > allocate space in a higher parent frame and pass this preallocated > memory down to the child frame to get filled in. > > Reading through this code, I don't think that the memory is meant to > outlive the stack frame. Is there a reason why we can't just declare > data as: > > u8 data [4]; > > then use ARRAY_SIZE(data) or RSI_9116_REG_SIZE in rsi_reset_chip(), > getting rid of the kzalloc/kfree? I haven't checked the details but AFAIK stack variables are not supposed to be used with DMA. So in that case I think it's ok alloc four bytes, unless the DMA rules have changed of course. But I didn't check if rsi is using DMA here, just a general comment.
On Fri, May 03, 2019 at 07:38:52AM +0300, Kalle Valo wrote: > Nick Desaulniers <ndesaulniers@google.com> writes: > > > On Thu, May 2, 2019 at 8:16 AM Nathan Chancellor > > <natechancellor@gmail.com> wrote: > >> > >> When building with -Wuninitialized, Clang warns: > >> > >> drivers/net/wireless/rsi/rsi_91x_sdio.c:940:43: warning: variable 'data' > >> is uninitialized when used here [-Wuninitialized] > >> put_unaligned_le32(TA_HOLD_THREAD_VALUE, data); > >> ^~~~ > >> drivers/net/wireless/rsi/rsi_91x_sdio.c:930:10: note: initialize the > >> variable 'data' to silence this warning > >> u8 *data; > >> ^ > >> = NULL > >> 1 warning generated. > >> > >> Using Clang's suggestion of initializing data to NULL wouldn't work out > >> because data will be dereferenced by put_unaligned_le32. Use kzalloc to > >> properly initialize data, which matches a couple of other places in this > >> driver. > >> > >> Fixes: e5a1ecc97e5f ("rsi: add firmware loading for 9116 device") > >> Link: https://github.com/ClangBuiltLinux/linux/issues/464 > >> Signed-off-by: Nathan Chancellor <natechancellor@gmail.com> > >> --- > >> drivers/net/wireless/rsi/rsi_91x_sdio.c | 21 ++++++++++++++------- > >> 1 file changed, 14 insertions(+), 7 deletions(-) > >> > >> diff --git a/drivers/net/wireless/rsi/rsi_91x_sdio.c b/drivers/net/wireless/rsi/rsi_91x_sdio.c > >> index f9c67ed473d1..b35728564c7b 100644 > >> --- a/drivers/net/wireless/rsi/rsi_91x_sdio.c > >> +++ b/drivers/net/wireless/rsi/rsi_91x_sdio.c > >> @@ -929,11 +929,15 @@ static int rsi_sdio_ta_reset(struct rsi_hw *adapter) > >> u32 addr; > >> u8 *data; > >> > >> + data = kzalloc(sizeof(u32), GFP_KERNEL); > > > > Something fishy is going on here. We allocate 4 B but declare data as > > a u8* (pointer to individual bytes)? In general, dynamically > > allocating that few bytes is a code smell; either you meant to just > > use the stack, or this memory's lifetime extends past the lifetime of > > this stackframe, at which point you probably just meant to stack > > allocate space in a higher parent frame and pass this preallocated > > memory down to the child frame to get filled in. > > > > Reading through this code, I don't think that the memory is meant to > > outlive the stack frame. Is there a reason why we can't just declare > > data as: > > > > u8 data [4]; > > > > then use ARRAY_SIZE(data) or RSI_9116_REG_SIZE in rsi_reset_chip(), > > getting rid of the kzalloc/kfree? > > I haven't checked the details but AFAIK stack variables are not supposed > to be used with DMA. So in that case I think it's ok alloc four bytes, > unless the DMA rules have changed of course. But I didn't check if rsi > is using DMA here, just a general comment. > > -- > Kalle Valo I don't think it is using the DMA API but it might be the same thing for SDIO. If passing that around on the stack is okay, great but we don't want what commit f700546682a6 ("rsi: fix nommu_map_sg overflow kernel panic") fixes to happen here. I can't answer that for sure though since I am not at all familiar with this driver or the SDIO APIs. Cheers, Nathan
On Thu, May 02, 2019 at 08:17:18PM -0700, Nathan Chancellor wrote: > On Thu, May 02, 2019 at 11:18:01AM -0700, Nick Desaulniers wrote: > > On Thu, May 2, 2019 at 8:16 AM Nathan Chancellor > > <natechancellor@gmail.com> wrote: > > > > > > When building with -Wuninitialized, Clang warns: > > > > > > drivers/net/wireless/rsi/rsi_91x_sdio.c:940:43: warning: variable 'data' > > > is uninitialized when used here [-Wuninitialized] > > > put_unaligned_le32(TA_HOLD_THREAD_VALUE, data); > > > ^~~~ > > > drivers/net/wireless/rsi/rsi_91x_sdio.c:930:10: note: initialize the > > > variable 'data' to silence this warning > > > u8 *data; > > > ^ > > > = NULL > > > 1 warning generated. > > > > > > Using Clang's suggestion of initializing data to NULL wouldn't work out > > > because data will be dereferenced by put_unaligned_le32. Use kzalloc to > > > properly initialize data, which matches a couple of other places in this > > > driver. > > > > > > Fixes: e5a1ecc97e5f ("rsi: add firmware loading for 9116 device") > > > Link: https://github.com/ClangBuiltLinux/linux/issues/464 > > > Signed-off-by: Nathan Chancellor <natechancellor@gmail.com> > > > --- > > > drivers/net/wireless/rsi/rsi_91x_sdio.c | 21 ++++++++++++++------- > > > 1 file changed, 14 insertions(+), 7 deletions(-) > > > > > > diff --git a/drivers/net/wireless/rsi/rsi_91x_sdio.c b/drivers/net/wireless/rsi/rsi_91x_sdio.c > > > index f9c67ed473d1..b35728564c7b 100644 > > > --- a/drivers/net/wireless/rsi/rsi_91x_sdio.c > > > +++ b/drivers/net/wireless/rsi/rsi_91x_sdio.c > > > @@ -929,11 +929,15 @@ static int rsi_sdio_ta_reset(struct rsi_hw *adapter) > > > u32 addr; > > > u8 *data; > > > > > > + data = kzalloc(sizeof(u32), GFP_KERNEL); > > > > Something fishy is going on here. We allocate 4 B but declare data as > > a u8* (pointer to individual bytes)? In general, dynamically > > allocating that few bytes is a code smell; either you meant to just > > use the stack, or this memory's lifetime extends past the lifetime of > > this stackframe, at which point you probably just meant to stack > > allocate space in a higher parent frame and pass this preallocated > > memory down to the child frame to get filled in. > > > > Reading through this code, I don't think that the memory is meant to > > outlive the stack frame. Is there a reason why we can't just declare > > data as: > > > > u8 data [4]; > > data was __le32 in rsi_reset_chip() before commit f700546682a6 ("rsi: > fix nommu_map_sg overflow kernel panic"). > > I wonder if this would be okay for this function: > > ------------------------------------------------- > > diff --git a/drivers/net/wireless/rsi/rsi_91x_sdio.c b/drivers/net/wireless/rsi/rsi_91x_sdio.c > index f9c67ed473d1..0330c50ab99c 100644 > --- a/drivers/net/wireless/rsi/rsi_91x_sdio.c > +++ b/drivers/net/wireless/rsi/rsi_91x_sdio.c > @@ -927,7 +927,7 @@ static int rsi_sdio_ta_reset(struct rsi_hw *adapter) > { > int status; > u32 addr; > - u8 *data; > + u8 data; > > status = rsi_sdio_master_access_msword(adapter, TA_BASE_ADDR); > if (status < 0) { > @@ -937,7 +937,7 @@ static int rsi_sdio_ta_reset(struct rsi_hw *adapter) > } > > rsi_dbg(INIT_ZONE, "%s: Bring TA out of reset\n", __func__); > - put_unaligned_le32(TA_HOLD_THREAD_VALUE, data); > + put_unaligned_le32(TA_HOLD_THREAD_VALUE, &data); > addr = TA_HOLD_THREAD_REG | RSI_SD_REQUEST_MASTER; > status = rsi_sdio_write_register_multiple(adapter, addr, > (u8 *)&data, > @@ -947,7 +947,7 @@ static int rsi_sdio_ta_reset(struct rsi_hw *adapter) > return status; > } > > - put_unaligned_le32(TA_SOFT_RST_CLR, data); > + put_unaligned_le32(TA_SOFT_RST_CLR, &data); > addr = TA_SOFT_RESET_REG | RSI_SD_REQUEST_MASTER; > status = rsi_sdio_write_register_multiple(adapter, addr, > (u8 *)&data, > @@ -957,7 +957,7 @@ static int rsi_sdio_ta_reset(struct rsi_hw *adapter) > return status; > } > > - put_unaligned_le32(TA_PC_ZERO, data); > + put_unaligned_le32(TA_PC_ZERO, &data); > addr = TA_TH0_PC_REG | RSI_SD_REQUEST_MASTER; > status = rsi_sdio_write_register_multiple(adapter, addr, > (u8 *)&data, > @@ -967,7 +967,7 @@ static int rsi_sdio_ta_reset(struct rsi_hw *adapter) > return -EINVAL; > } > > - put_unaligned_le32(TA_RELEASE_THREAD_VALUE, data); > + put_unaligned_le32(TA_RELEASE_THREAD_VALUE, &data); > addr = TA_RELEASE_THREAD_REG | RSI_SD_REQUEST_MASTER; > status = rsi_sdio_write_register_multiple(adapter, addr, > (u8 *)&data, > > > > > > then use ARRAY_SIZE(data) or RSI_9116_REG_SIZE in rsi_reset_chip(), > > getting rid of the kzalloc/kfree? > > > > (Sorry, I hate when a simple fixup becomes a "hey let's rewrite all > > this code" thus becoming "that guy.") > > If we aren't actually improving the code, then why bother? :) > > Thank you for the review! > Nathan > > > -- > > Thanks, > > ~Nick Desaulniers Hi all, Did any of the maintainers have any comments on what the correct solution is here to resolve this warning? It is one of the few left before we can turn on -Wuninitialized for the whole kernel. Thanks, Nathan
On Thu, May 23, 2019 at 3:54 AM Nathan Chancellor <natechancellor@gmail.com> wrote: > > On Thu, May 02, 2019 at 08:17:18PM -0700, Nathan Chancellor wrote: > > On Thu, May 02, 2019 at 11:18:01AM -0700, Nick Desaulniers wrote: > > > On Thu, May 2, 2019 at 8:16 AM Nathan Chancellor > > > u8 data [4]; > > > > data was __le32 in rsi_reset_chip() before commit f700546682a6 ("rsi: > > fix nommu_map_sg overflow kernel panic"). > > > > I wonder if this would be okay for this function: > > > > ------------------------------------------------- > > > > diff --git a/drivers/net/wireless/rsi/rsi_91x_sdio.c b/drivers/net/wireless/rsi/rsi_91x_sdio.c > > index f9c67ed473d1..0330c50ab99c 100644 > > --- a/drivers/net/wireless/rsi/rsi_91x_sdio.c > > +++ b/drivers/net/wireless/rsi/rsi_91x_sdio.c > > @@ -927,7 +927,7 @@ static int rsi_sdio_ta_reset(struct rsi_hw *adapter) > > { > > int status; > > u32 addr; > > - u8 *data; > > + u8 data; > > > > status = rsi_sdio_master_access_msword(adapter, TA_BASE_ADDR); > > if (status < 0) { > > @@ -937,7 +937,7 @@ static int rsi_sdio_ta_reset(struct rsi_hw *adapter) > > } > > > > rsi_dbg(INIT_ZONE, "%s: Bring TA out of reset\n", __func__); > > - put_unaligned_le32(TA_HOLD_THREAD_VALUE, data); > > + put_unaligned_le32(TA_HOLD_THREAD_VALUE, &data); > > addr = TA_HOLD_THREAD_REG | RSI_SD_REQUEST_MASTER; > > status = rsi_sdio_write_register_multiple(adapter, addr, > > (u8 *)&data, This is clearly not ok, as put_unaligned_le32() stores four bytes, and the local variable is only one byte! Also, sdio does use DMA for transfers, so the variable has to be dynamically allocated. I think your original patch was correct. The only change I'd possibly make would be to use RSI_9116_REG_SIZE instead of sizeof(u32). > Did any of the maintainers have any comments on what the correct > solution is here to resolve this warning? It is one of the few left > before we can turn on -Wuninitialized for the whole kernel. I would argue that this should not stop us from turning it on, as the warning is for a clear bug in the code that absolutely needs to be fixed, rather than a false-positive. Arnd
Nathan Chancellor <natechancellor@gmail.com> writes: > On Thu, May 02, 2019 at 08:17:18PM -0700, Nathan Chancellor wrote: >> On Thu, May 02, 2019 at 11:18:01AM -0700, Nick Desaulniers wrote: >> > On Thu, May 2, 2019 at 8:16 AM Nathan Chancellor >> > <natechancellor@gmail.com> wrote: >> > > >> > > When building with -Wuninitialized, Clang warns: >> > > >> > > drivers/net/wireless/rsi/rsi_91x_sdio.c:940:43: warning: variable 'data' >> > > is uninitialized when used here [-Wuninitialized] >> > > put_unaligned_le32(TA_HOLD_THREAD_VALUE, data); >> > > ^~~~ >> > > drivers/net/wireless/rsi/rsi_91x_sdio.c:930:10: note: initialize the >> > > variable 'data' to silence this warning >> > > u8 *data; >> > > ^ >> > > = NULL >> > > 1 warning generated. >> > > >> > > Using Clang's suggestion of initializing data to NULL wouldn't work out >> > > because data will be dereferenced by put_unaligned_le32. Use kzalloc to >> > > properly initialize data, which matches a couple of other places in this >> > > driver. >> > > >> > > Fixes: e5a1ecc97e5f ("rsi: add firmware loading for 9116 device") >> > > Link: https://github.com/ClangBuiltLinux/linux/issues/464 >> > > Signed-off-by: Nathan Chancellor <natechancellor@gmail.com> >> > > --- >> > > drivers/net/wireless/rsi/rsi_91x_sdio.c | 21 ++++++++++++++------- >> > > 1 file changed, 14 insertions(+), 7 deletions(-) >> > > >> > > diff --git a/drivers/net/wireless/rsi/rsi_91x_sdio.c b/drivers/net/wireless/rsi/rsi_91x_sdio.c >> > > index f9c67ed473d1..b35728564c7b 100644 >> > > --- a/drivers/net/wireless/rsi/rsi_91x_sdio.c >> > > +++ b/drivers/net/wireless/rsi/rsi_91x_sdio.c >> > > @@ -929,11 +929,15 @@ static int rsi_sdio_ta_reset(struct rsi_hw *adapter) >> > > u32 addr; >> > > u8 *data; >> > > >> > > + data = kzalloc(sizeof(u32), GFP_KERNEL); >> > >> > Something fishy is going on here. We allocate 4 B but declare data as >> > a u8* (pointer to individual bytes)? In general, dynamically >> > allocating that few bytes is a code smell; either you meant to just >> > use the stack, or this memory's lifetime extends past the lifetime of >> > this stackframe, at which point you probably just meant to stack >> > allocate space in a higher parent frame and pass this preallocated >> > memory down to the child frame to get filled in. >> > >> > Reading through this code, I don't think that the memory is meant to >> > outlive the stack frame. Is there a reason why we can't just declare >> > data as: >> > >> > u8 data [4]; >> >> data was __le32 in rsi_reset_chip() before commit f700546682a6 ("rsi: >> fix nommu_map_sg overflow kernel panic"). >> >> I wonder if this would be okay for this function: >> >> ------------------------------------------------- >> >> diff --git a/drivers/net/wireless/rsi/rsi_91x_sdio.c b/drivers/net/wireless/rsi/rsi_91x_sdio.c >> index f9c67ed473d1..0330c50ab99c 100644 >> --- a/drivers/net/wireless/rsi/rsi_91x_sdio.c >> +++ b/drivers/net/wireless/rsi/rsi_91x_sdio.c >> @@ -927,7 +927,7 @@ static int rsi_sdio_ta_reset(struct rsi_hw *adapter) >> { >> int status; >> u32 addr; >> - u8 *data; >> + u8 data; >> >> status = rsi_sdio_master_access_msword(adapter, TA_BASE_ADDR); >> if (status < 0) { >> @@ -937,7 +937,7 @@ static int rsi_sdio_ta_reset(struct rsi_hw *adapter) >> } >> >> rsi_dbg(INIT_ZONE, "%s: Bring TA out of reset\n", __func__); >> - put_unaligned_le32(TA_HOLD_THREAD_VALUE, data); >> + put_unaligned_le32(TA_HOLD_THREAD_VALUE, &data); >> addr = TA_HOLD_THREAD_REG | RSI_SD_REQUEST_MASTER; >> status = rsi_sdio_write_register_multiple(adapter, addr, >> (u8 *)&data, >> @@ -947,7 +947,7 @@ static int rsi_sdio_ta_reset(struct rsi_hw *adapter) >> return status; >> } >> >> - put_unaligned_le32(TA_SOFT_RST_CLR, data); >> + put_unaligned_le32(TA_SOFT_RST_CLR, &data); >> addr = TA_SOFT_RESET_REG | RSI_SD_REQUEST_MASTER; >> status = rsi_sdio_write_register_multiple(adapter, addr, >> (u8 *)&data, >> @@ -957,7 +957,7 @@ static int rsi_sdio_ta_reset(struct rsi_hw *adapter) >> return status; >> } >> >> - put_unaligned_le32(TA_PC_ZERO, data); >> + put_unaligned_le32(TA_PC_ZERO, &data); >> addr = TA_TH0_PC_REG | RSI_SD_REQUEST_MASTER; >> status = rsi_sdio_write_register_multiple(adapter, addr, >> (u8 *)&data, >> @@ -967,7 +967,7 @@ static int rsi_sdio_ta_reset(struct rsi_hw *adapter) >> return -EINVAL; >> } >> >> - put_unaligned_le32(TA_RELEASE_THREAD_VALUE, data); >> + put_unaligned_le32(TA_RELEASE_THREAD_VALUE, &data); >> addr = TA_RELEASE_THREAD_REG | RSI_SD_REQUEST_MASTER; >> status = rsi_sdio_write_register_multiple(adapter, addr, >> (u8 *)&data, >> >> >> > >> > then use ARRAY_SIZE(data) or RSI_9116_REG_SIZE in rsi_reset_chip(), >> > getting rid of the kzalloc/kfree? >> > >> > (Sorry, I hate when a simple fixup becomes a "hey let's rewrite all >> > this code" thus becoming "that guy.") >> >> If we aren't actually improving the code, then why bother? :) >> >> Thank you for the review! > > Did any of the maintainers have any comments on what the correct > solution is here to resolve this warning? It is one of the few left > before we can turn on -Wuninitialized for the whole kernel. I don't have any strong opinion, but as the commit log says that kzalloc() is also used in similar cases in the same driver I would happy to take this patch as is.
Arnd Bergmann <arnd@arndb.de> writes: >> > @@ -937,7 +937,7 @@ static int rsi_sdio_ta_reset(struct rsi_hw *adapter) >> > } >> > >> > rsi_dbg(INIT_ZONE, "%s: Bring TA out of reset\n", __func__); >> > - put_unaligned_le32(TA_HOLD_THREAD_VALUE, data); >> > + put_unaligned_le32(TA_HOLD_THREAD_VALUE, &data); >> > addr = TA_HOLD_THREAD_REG | RSI_SD_REQUEST_MASTER; >> > status = rsi_sdio_write_register_multiple(adapter, addr, >> > (u8 *)&data, > > This is clearly not ok, as put_unaligned_le32() stores four bytes, and > the local variable is only one byte! > > Also, sdio does use DMA for transfers, so the variable has to be > dynamically allocated. I think your original patch was correct. > The only change I'd possibly make would be to use > RSI_9116_REG_SIZE instead of sizeof(u32). Good point. Nathan please fix this and submit v2. >> Did any of the maintainers have any comments on what the correct >> solution is here to resolve this warning? It is one of the few left >> before we can turn on -Wuninitialized for the whole kernel. > > I would argue that this should not stop us from turning it on, as the > warning is for a clear bug in the code that absolutely needs to be > fixed, rather than a false-positive. I can queue v2 for v5.2, just remind me by adding "[PATCH v2 5.2]" to the subject.
diff --git a/drivers/net/wireless/rsi/rsi_91x_sdio.c b/drivers/net/wireless/rsi/rsi_91x_sdio.c index f9c67ed473d1..b35728564c7b 100644 --- a/drivers/net/wireless/rsi/rsi_91x_sdio.c +++ b/drivers/net/wireless/rsi/rsi_91x_sdio.c @@ -929,11 +929,15 @@ static int rsi_sdio_ta_reset(struct rsi_hw *adapter) u32 addr; u8 *data; + data = kzalloc(sizeof(u32), GFP_KERNEL); + if (!data) + return -ENOMEM; + status = rsi_sdio_master_access_msword(adapter, TA_BASE_ADDR); if (status < 0) { rsi_dbg(ERR_ZONE, "Unable to set ms word to common reg\n"); - return status; + goto err; } rsi_dbg(INIT_ZONE, "%s: Bring TA out of reset\n", __func__); @@ -944,7 +948,7 @@ static int rsi_sdio_ta_reset(struct rsi_hw *adapter) RSI_9116_REG_SIZE); if (status < 0) { rsi_dbg(ERR_ZONE, "Unable to hold TA threads\n"); - return status; + goto err; } put_unaligned_le32(TA_SOFT_RST_CLR, data); @@ -954,7 +958,7 @@ static int rsi_sdio_ta_reset(struct rsi_hw *adapter) RSI_9116_REG_SIZE); if (status < 0) { rsi_dbg(ERR_ZONE, "Unable to get TA out of reset\n"); - return status; + goto err; } put_unaligned_le32(TA_PC_ZERO, data); @@ -964,7 +968,8 @@ static int rsi_sdio_ta_reset(struct rsi_hw *adapter) RSI_9116_REG_SIZE); if (status < 0) { rsi_dbg(ERR_ZONE, "Unable to Reset TA PC value\n"); - return -EINVAL; + status = -EINVAL; + goto err; } put_unaligned_le32(TA_RELEASE_THREAD_VALUE, data); @@ -974,17 +979,19 @@ static int rsi_sdio_ta_reset(struct rsi_hw *adapter) RSI_9116_REG_SIZE); if (status < 0) { rsi_dbg(ERR_ZONE, "Unable to release TA threads\n"); - return status; + goto err; } status = rsi_sdio_master_access_msword(adapter, MISC_CFG_BASE_ADDR); if (status < 0) { rsi_dbg(ERR_ZONE, "Unable to set ms word to common reg\n"); - return status; + goto err; } rsi_dbg(INIT_ZONE, "***** TA Reset done *****\n"); - return 0; +err: + kfree(data); + return status; } static struct rsi_host_intf_ops sdio_host_intf_ops = {
When building with -Wuninitialized, Clang warns: drivers/net/wireless/rsi/rsi_91x_sdio.c:940:43: warning: variable 'data' is uninitialized when used here [-Wuninitialized] put_unaligned_le32(TA_HOLD_THREAD_VALUE, data); ^~~~ drivers/net/wireless/rsi/rsi_91x_sdio.c:930:10: note: initialize the variable 'data' to silence this warning u8 *data; ^ = NULL 1 warning generated. Using Clang's suggestion of initializing data to NULL wouldn't work out because data will be dereferenced by put_unaligned_le32. Use kzalloc to properly initialize data, which matches a couple of other places in this driver. Fixes: e5a1ecc97e5f ("rsi: add firmware loading for 9116 device") Link: https://github.com/ClangBuiltLinux/linux/issues/464 Signed-off-by: Nathan Chancellor <natechancellor@gmail.com> --- drivers/net/wireless/rsi/rsi_91x_sdio.c | 21 ++++++++++++++------- 1 file changed, 14 insertions(+), 7 deletions(-)