| Message ID | 20170201124800.13865-9-linus.walleij@linaro.org |
|---|---|
| State | New |
| Headers | show |
| Series | Further MMC core/block cleanups | expand |
[...] > > static void mmc_blk_rw_cmd_abort(struct mmc_card *card, struct request *req) > { > - int ret = 1; > - > if (mmc_card_removed(card)) > req->rq_flags |= RQF_QUIET; > - while (ret) > - ret = blk_end_request(req, -EIO, > - blk_rq_cur_bytes(req)); > + while (blk_end_request(req, -EIO, blk_rq_cur_bytes(req))) > + { > + } These brackets isn't needed, so am going to change this before applying. while (blk_end_request(req, -EIO, blk_rq_cur_bytes(req))); [...] Kind regards Uffe -- To unsubscribe from this list: send the line "unsubscribe linux-mmc" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Fri, Feb 3, 2017 at 10:52 AM, Ulf Hansson <ulf.hansson@linaro.org> wrote: > [...] >> static void mmc_blk_rw_cmd_abort(struct mmc_card *card, struct request *req) >> { >> - int ret = 1; >> - >> if (mmc_card_removed(card)) >> req->rq_flags |= RQF_QUIET; >> - while (ret) >> - ret = blk_end_request(req, -EIO, >> - blk_rq_cur_bytes(req)); >> + while (blk_end_request(req, -EIO, blk_rq_cur_bytes(req))) >> + { >> + } > > These brackets isn't needed, so am going to change this before applying. > > while (blk_end_request(req, -EIO, blk_rq_cur_bytes(req))); Aha you know your C syntax better than me. Looks like something that checkpatch should catch actually, but I ran this through checkpatch and it didn't complain. Yours, Linus Walleij -- To unsubscribe from this list: send the line "unsubscribe linux-mmc" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Fri, Feb 3, 2017 at 1:57 PM, Linus Walleij <linus.walleij@linaro.org> wrote: >>> if (mmc_card_removed(card)) >>> req->rq_flags |= RQF_QUIET; >>> - while (ret) >>> - ret = blk_end_request(req, -EIO, >>> - blk_rq_cur_bytes(req)); >>> + while (blk_end_request(req, -EIO, blk_rq_cur_bytes(req))) >>> + { >>> + } >> >> These brackets isn't needed, so am going to change this before applying. >> >> while (blk_end_request(req, -EIO, blk_rq_cur_bytes(req))); > > Aha you know your C syntax better than me. Looks like something that > checkpatch should catch actually, but I ran this through checkpatch and > it didn't complain. IIRC gcc-7 will warn about Ulf's version, as it is a bit misleading and can lead to subtle bugs like int ret = 0; while (!ret); ret = do_something(); which would pass a casual review but is actually an endless loop. Arnd -- To unsubscribe from this list: send the line "unsubscribe linux-mmc" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On 6 February 2017 at 10:07, Arnd Bergmann <arnd@arndb.de> wrote: > On Fri, Feb 3, 2017 at 1:57 PM, Linus Walleij <linus.walleij@linaro.org> wrote: >>>> if (mmc_card_removed(card)) >>>> req->rq_flags |= RQF_QUIET; >>>> - while (ret) >>>> - ret = blk_end_request(req, -EIO, >>>> - blk_rq_cur_bytes(req)); >>>> + while (blk_end_request(req, -EIO, blk_rq_cur_bytes(req))) >>>> + { >>>> + } >>> >>> These brackets isn't needed, so am going to change this before applying. >>> >>> while (blk_end_request(req, -EIO, blk_rq_cur_bytes(req))); >> >> Aha you know your C syntax better than me. Looks like something that >> checkpatch should catch actually, but I ran this through checkpatch and >> it didn't complain. > > IIRC gcc-7 will warn about Ulf's version, as it is a bit misleading and can > lead to subtle bugs like > > int ret = 0; > while (!ret); > ret = do_something(); > > which would pass a casual review but is actually an endless loop. > > Arnd Okay, seems reasonable. However mine version is already being used at some places in the kernel. Linus version, was actually complained by checkpatch. Perhaps me an Linus ran different versions of check patch. So, then which version is preferred here? Perhaps something like this is the best? while (1) { ret = do_something(); if (!ret) break; } Br Uffe -- To unsubscribe from this list: send the line "unsubscribe linux-mmc" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Mon, Feb 6, 2017 at 10:54 AM, Ulf Hansson <ulf.hansson@linaro.org> wrote: > On 6 February 2017 at 10:07, Arnd Bergmann <arnd@arndb.de> wrote: >> On Fri, Feb 3, 2017 at 1:57 PM, Linus Walleij <linus.walleij@linaro.org> wrote: >>>>> if (mmc_card_removed(card)) >>>>> req->rq_flags |= RQF_QUIET; >>>>> - while (ret) >>>>> - ret = blk_end_request(req, -EIO, >>>>> - blk_rq_cur_bytes(req)); >>>>> + while (blk_end_request(req, -EIO, blk_rq_cur_bytes(req))) >>>>> + { >>>>> + } >>>> >>>> These brackets isn't needed, so am going to change this before applying. >>>> >>>> while (blk_end_request(req, -EIO, blk_rq_cur_bytes(req))); >>> >>> Aha you know your C syntax better than me. Looks like something that >>> checkpatch should catch actually, but I ran this through checkpatch and >>> it didn't complain. >> >> IIRC gcc-7 will warn about Ulf's version, as it is a bit misleading and can >> lead to subtle bugs like >> >> int ret = 0; >> while (!ret); >> ret = do_something(); >> >> which would pass a casual review but is actually an endless loop. >> >> Arnd > > Okay, seems reasonable. However mine version is already being used at > some places in the kernel. > > Linus version, was actually complained by checkpatch. Perhaps me an > Linus ran different versions of check patch. I might have missed it. It complained when I had: while (cond) { }; about the semicolon in the end... maybe there was another complaint too. > So, then which version is preferred here? > > Perhaps something like this is the best? > > while (1) { > ret = do_something(); > if (!ret) > break; > } Actually in this case the whole while() thing looks a bit bogus: it is just hammering the block layer to end the request. I guess since there is no way to return any unfinished status upward. True, I preserved that strange semantic, but what should we really do? Timeout and complain in dmesg? It just looks dangerous. Anyways let's address any issues with follow-up patches... Yours, Linus Walleij -- To unsubscribe from this list: send the line "unsubscribe linux-mmc" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
>> Perhaps something like this is the best? >> >> while (1) { >> ret = do_something(); >> if (!ret) >> break; >> } > > Actually in this case the whole while() thing looks a bit bogus: > it is just hammering the block layer to end the request. > > I guess since there is no way to return any unfinished status > upward. > > True, I preserved that strange semantic, but what should we > really do? Timeout and complain in dmesg? > > It just looks dangerous. > > Anyways let's address any issues with follow-up patches... Agreed! Kind regards Uffe -- To unsubscribe from this list: send the line "unsubscribe linux-mmc" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/drivers/mmc/core/block.c b/drivers/mmc/core/block.c index 91d506b37024..92f7772ca56d 100644 --- a/drivers/mmc/core/block.c +++ b/drivers/mmc/core/block.c @@ -1566,11 +1566,13 @@ static void mmc_blk_rw_rq_prep(struct mmc_queue_req *mqrq, mmc_queue_bounce_pre(mqrq); } -static int mmc_blk_cmd_err(struct mmc_blk_data *md, struct mmc_card *card, - struct mmc_blk_request *brq, struct request *req, - int ret) +static bool mmc_blk_rw_cmd_err(struct mmc_blk_data *md, struct mmc_card *card, + struct mmc_blk_request *brq, struct request *req, + bool old_req_pending) { struct mmc_queue_req *mq_rq; + bool req_pending; + mq_rq = container_of(brq, struct mmc_queue_req, brq); /* @@ -1586,24 +1588,23 @@ static int mmc_blk_cmd_err(struct mmc_blk_data *md, struct mmc_card *card, int err; err = mmc_sd_num_wr_blocks(card, &blocks); - if (!err) { - ret = blk_end_request(req, 0, blocks << 9); - } + if (err) + req_pending = old_req_pending; + else + req_pending = blk_end_request(req, 0, blocks << 9); } else { - ret = blk_end_request(req, 0, brq->data.bytes_xfered); + req_pending = blk_end_request(req, 0, brq->data.bytes_xfered); } - return ret; + return req_pending; } static void mmc_blk_rw_cmd_abort(struct mmc_card *card, struct request *req) { - int ret = 1; - if (mmc_card_removed(card)) req->rq_flags |= RQF_QUIET; - while (ret) - ret = blk_end_request(req, -EIO, - blk_rq_cur_bytes(req)); + while (blk_end_request(req, -EIO, blk_rq_cur_bytes(req))) + { + } } /** @@ -1634,12 +1635,13 @@ static void mmc_blk_issue_rw_rq(struct mmc_queue *mq, struct request *new_req) struct mmc_blk_data *md = mq->blkdata; struct mmc_card *card = md->queue.card; struct mmc_blk_request *brq; - int ret = 1, disable_multi = 0, retry = 0, type, retune_retry_done = 0; + int disable_multi = 0, retry = 0, type, retune_retry_done = 0; enum mmc_blk_status status; struct mmc_queue_req *mq_rq; struct request *old_req; struct mmc_async_req *new_areq; struct mmc_async_req *old_areq; + bool req_pending = true; if (!new_req && !mq->mqrq_prev->req) return; @@ -1693,15 +1695,14 @@ static void mmc_blk_issue_rw_rq(struct mmc_queue *mq, struct request *new_req) */ mmc_blk_reset_success(md, type); - ret = blk_end_request(old_req, 0, - brq->data.bytes_xfered); - + req_pending = blk_end_request(old_req, 0, + brq->data.bytes_xfered); /* * If the blk_end_request function returns non-zero even * though all data has been transferred and no errors * were returned by the host controller, it's a bug. */ - if (status == MMC_BLK_SUCCESS && ret) { + if (status == MMC_BLK_SUCCESS && req_pending) { pr_err("%s BUG rq_tot %d d_xfer %d\n", __func__, blk_rq_bytes(old_req), brq->data.bytes_xfered); @@ -1710,13 +1711,13 @@ static void mmc_blk_issue_rw_rq(struct mmc_queue *mq, struct request *new_req) } break; case MMC_BLK_CMD_ERR: - ret = mmc_blk_cmd_err(md, card, brq, old_req, ret); + req_pending = mmc_blk_rw_cmd_err(md, card, brq, old_req, req_pending); if (mmc_blk_reset(md, card->host, type)) { mmc_blk_rw_cmd_abort(card, old_req); mmc_blk_rw_try_restart(mq, new_req); return; } - if (!ret) { + if (!req_pending) { mmc_blk_rw_try_restart(mq, new_req); return; } @@ -1758,9 +1759,9 @@ static void mmc_blk_issue_rw_rq(struct mmc_queue *mq, struct request *new_req) * time, so we only reach here after trying to * read a single sector. */ - ret = blk_end_request(old_req, -EIO, - brq->data.blksz); - if (!ret) { + req_pending = blk_end_request(old_req, -EIO, + brq->data.blksz); + if (!req_pending) { mmc_blk_rw_try_restart(mq, new_req); return; } @@ -1777,7 +1778,7 @@ static void mmc_blk_issue_rw_rq(struct mmc_queue *mq, struct request *new_req) return; } - if (ret) { + if (req_pending) { /* * In case of a incomplete request * prepare it again and resend. @@ -1788,7 +1789,7 @@ static void mmc_blk_issue_rw_rq(struct mmc_queue *mq, struct request *new_req) &mq_rq->areq, NULL); mq_rq->brq.retune_retry_done = retune_retry_done; } - } while (ret); + } while (req_pending); } void mmc_blk_issue_rq(struct mmc_queue *mq, struct request *req)
The return value from blk_end_request() is a bool but is treated like an int. This is generally safe, but the variable also has the opaque name "ret" and gets returned from the helper function mmc_blk_cmd_err(). - Switch the variable to a bool, applies everywhere. - Return a bool from mmc_blk_cmd_err() and rename the function mmc_blk_rw_cmd_err() to indicate through the namespace that this is a helper for mmc_blk_issue_rw_rq(). - Rename the variable from "ret" to "req_pending" inside the while() loop inside mmc_blk_issue_rq_rq(), which finally makes it very clear what this while loop is waiting for. - Augment the argument "ret" to mmc_blk_rq_cmd_err() to old_req_pending so it becomes evident that this is an older state, and it is returned only if we fail to get the number of written blocks from an SD card in the function mmc_sd_num_wr_blocks(). - Augment the while() loop in mmc_blk_rq_cmd_abort(): it is evident now that we know this is a bool variable, that the function is just spinning waiting for blk_end_request() to return false. Signed-off-by: Linus Walleij <linus.walleij@linaro.org> --- drivers/mmc/core/block.c | 51 ++++++++++++++++++++++++------------------------ 1 file changed, 26 insertions(+), 25 deletions(-) -- 2.9.3 -- To unsubscribe from this list: send the line "unsubscribe linux-mmc" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html