| Message ID | 1411400116-33778-1-git-send-email-mike.holmes@linaro.org |
|---|---|
| State | Rejected |
| Headers | show |
Hi Mike, I would not prefer LOG functions to call abort() on behalf of the application, this design prevents the applications from calling its specific clean-up code. coz applications can be written in a way in which clean-up can be done in a single place at the end of the main function combining all the lock-release and clean-up for different calls. Is there any specific requirement for this design? Regards, Bala
On Mon, Sep 22, 2014 at 08:54:30AM -0700, Bala Manoharan wrote: >Hi Mike, > >I would not prefer LOG functions to call abort() on behalf of the application, >this design prevents the applications from calling its specific clean-up code. >coz applications can be written in a way in which clean-up can be done in a >single place at the end of the main function combining all the lock-release and >clean-up for different calls. I'd agree with that myself, yes, but Mike's just cleaning up existing code which calls LOG_ERR() then exit() in a number of cases as far as I can see. I'm a very stong believer in libraries *not* causing a process to exit. Cheers,
On 22 September 2014 11:54, Bala Manoharan <bala.manoharan@linaro.org> wrote: > Hi Mike, > > I would not prefer LOG functions to call abort() on behalf of the > application, > Np, If we want a new name that is fine, I wondered how many ODP_ERRs would be seriously considered ERR or a FATAL The purpose is to ensure there is no need of calls to exit in the implementations, so any ideas on a good name? From the above possibly ODP_FATAL? We also then need to update http://docs.opendataplane.org/arch/html/exception_handling.html to make it clearer. this design prevents the applications from calling its specific clean-up > code. coz applications can be written in a way in which clean-up can be > done in a single place at the end of the main function combining all the > lock-release and clean-up for different calls. > > Is there any specific requirement for this design? > > Regards, > Bala >
Yes. I would prefer ODP_ABORT("message") this call can be used by the
application so that there is no need to provide two separate calls for
ODP_ERR and abort().
Regards,
Bala
On 22 September 2014 09:14, Mike Holmes <mike.holmes@linaro.org> wrote:
>
>
> On 22 September 2014 11:54, Bala Manoharan <bala.manoharan@linaro.org>
> wrote:
>
>> Hi Mike,
>>
>> I would not prefer LOG functions to call abort() on behalf of the
>> application,
>>
> Np, If we want a new name that is fine, I wondered how many ODP_ERRs would
> be seriously considered ERR or a FATAL
> The purpose is to ensure there is no need of calls to exit in the
> implementations, so any ideas on a good name? From the above possibly
> ODP_FATAL?
> We also then need to update
> http://docs.opendataplane.org/arch/html/exception_handling.html to make
> it clearer.
>
> this design prevents the applications from calling its specific clean-up
>> code. coz applications can be written in a way in which clean-up can be
>> done in a single place at the end of the main function combining all the
>> lock-release and clean-up for different calls.
>>
>> Is there any specific requirement for this design?
>>
>> Regards,
>> Bala
>>
>
>
>
> --
> *Mike Holmes*
> Linaro Technical Manager / Lead
> LNG - ODP
>
diff --git a/platform/linux-generic/include/api/odp_debug.h b/platform/linux-generic/include/api/odp_debug.h index e8f6003..692a6b2 100644 --- a/platform/linux-generic/include/api/odp_debug.h +++ b/platform/linux-generic/include/api/odp_debug.h @@ -13,6 +13,7 @@ #define ODP_DEBUG_H_ #include <stdio.h> +#include <stdlib.h> #ifdef __cplusplus extern "C" { @@ -76,8 +77,10 @@ extern "C" { * Print output to stderr (file, line and function). */ #define ODP_ERR(fmt, ...) \ - fprintf(stderr, "%s:%d:%s(): " fmt, __FILE__, \ - __LINE__, __func__, ##__VA_ARGS__) +do { fprintf(stderr, "%s:%d:%s(): " fmt, __FILE__, \ + __LINE__, __func__, ##__VA_ARGS__); \ + abort(); \ +} while (0) #ifdef __cplusplus } diff --git a/platform/linux-generic/odp_buffer_pool.c b/platform/linux-generic/odp_buffer_pool.c index f54a0c4..b88cf88 100644 --- a/platform/linux-generic/odp_buffer_pool.c +++ b/platform/linux-generic/odp_buffer_pool.c @@ -100,7 +100,6 @@ static inline void set_handle(odp_buffer_hdr_t *hdr, if (pool_id >= ODP_CONFIG_BUFFER_POOLS) { ODP_ERR("set_handle: Bad pool handle %u\n", pool_hdl); - exit(0); } if (index > ODP_BUFFER_MAX_INDEX) @@ -220,13 +219,11 @@ static void check_align(pool_entry_t *pool, odp_buffer_hdr_t *hdr) if (!ODP_ALIGNED_CHECK_POWER_2(hdr->addr, pool->s.user_align)) { ODP_ERR("check_align: user data align error %p, align %zu\n", hdr->addr, pool->s.user_align); - exit(0); } if (!ODP_ALIGNED_CHECK_POWER_2(hdr, ODP_CACHE_LINE_SIZE)) { ODP_ERR("check_align: hdr align error %p, align %i\n", hdr, ODP_CACHE_LINE_SIZE); - exit(0); } } @@ -265,7 +262,6 @@ static void fill_hdr(void *ptr, pool_entry_t *pool, uint32_t index, break; default: ODP_ERR("Bad buffer type\n"); - exit(0); } memset(hdr, 0, size); @@ -313,7 +309,6 @@ static void link_bufs(pool_entry_t *pool) hdr_size = sizeof(odp_any_buffer_hdr_t); } else { ODP_ERR("odp_buffer_pool_create: Bad type %i\n", buf_type); - exit(0); } /* Chunk must fit into buffer data area.*/ diff --git a/platform/linux-generic/odp_time.c b/platform/linux-generic/odp_time.c index 181294a..b4b8650 100644 --- a/platform/linux-generic/odp_time.c +++ b/platform/linux-generic/odp_time.c @@ -60,7 +60,6 @@ uint64_t odp_time_get_cycles(void) if (ret != 0) { ODP_ERR("clock_gettime failed\n"); - exit(EXIT_FAILURE); } hz = odp_sys_cpu_hz();
Signed-off-by: Mike Holmes <mike.holmes@linaro.org> --- platform/linux-generic/include/api/odp_debug.h | 7 +++++-- platform/linux-generic/odp_buffer_pool.c | 5 ----- platform/linux-generic/odp_time.c | 1 - 3 files changed, 5 insertions(+), 8 deletions(-)