| Message ID | 1252587402-7382-2-git-send-email-albert_herranz@yahoo.es (mailing list archive) |
|---|---|
| State | Not Applicable, archived |
| Headers | show |
On Thu, 10 Sep 2009 14:56:42 +0200 Albert Herranz <albert_herranz@yahoo.es> wrote: > Some manufacturers provide vendor information in non-vendor specific CIS > tuples. For example, Broadcom uses an Extended Function tuple to provide > the MAC address on some of their network cards, as in the case of the > Nintendo Wii WLAN daughter card. > > This patch allows passing correct tuples unknown to the SDIO core to > a matching SDIO driver instead of rejecting them and failing. > This looks leaky to me. : if (i < ARRAY_SIZE(cis_tpl_list)) { : const struct cis_tpl *tpl = cis_tpl_list + i; : if (tpl_link < tpl->min_size) { : printk(KERN_ERR : "%s: bad CIS tuple 0x%02x" : " (length = %u, expected >= %u)\n", : mmc_hostname(card->host), : tpl_code, tpl_link, tpl->min_size); : ret = -EINVAL; ret == -EINVAL : } else if (tpl->parse) { : ret = tpl->parse(card, func, : this->data, tpl_link); : } : /* already successfully parsed, not needed anymore */ : if (!ret) : kfree(this); `this' doesn't get freed : } else { : /* unknown tuple */ : ret = -EILSEQ; : } : : if (ret == -EILSEQ) { `this' doesn't get remembered. : /* this tuple is unknown to the core */ : this->next = NULL; : this->code = tpl_code; : this->size = tpl_link; : *prev = this; : prev = &this->next; : pr_debug("%s: queuing CIS tuple 0x%02x length %u\n", : mmc_hostname(card->host), tpl_code, tpl_link); : /* keep on analyzing tuples */ : ret = 0; : } : : ptr += tpl_link; `this' leaks. : } while (!ret); Please check? -- To unsubscribe from this list: send the line "unsubscribe linux-wireless" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Thu, 10 Sep 2009 14:56:42 +0200 Albert Herranz <albert_herranz@yahoo.es> wrote: > Some manufacturers provide vendor information in non-vendor specific CIS > tuples. For example, Broadcom uses an Extended Function tuple to provide > the MAC address on some of their network cards, as in the case of the > Nintendo Wii WLAN daughter card. > > This patch allows passing correct tuples unknown to the SDIO core to > a matching SDIO driver instead of rejecting them and failing. > > Signed-off-by: Albert Herranz <albert_herranz@yahoo.es> > --- The description for this patch should be made clearer. The title suggests it adds functionality that's already in place. It should be something along the lines of "Also pass malformed tuples to card drivers". In the sake of sanity, you might want to add this behaviour to all parsers, not just the FUNCE one. I'm also unclear on how this is supposed to work. What does the broadcom tuple look like? This patch looks like it will silence a lot of legitimate warnings, and possibly pollute the card structures with bogus data. > diff --git a/drivers/mmc/core/sdio_cis.c b/drivers/mmc/core/sdio_cis.c > index 963f293..87934ac 100644 > --- a/drivers/mmc/core/sdio_cis.c > +++ b/drivers/mmc/core/sdio_cis.c > @@ -123,8 +123,9 @@ static int cistpl_funce_func(struct sdio_func *func, > vsn = func->card->cccr.sdio_vsn; > min_size = (vsn == SDIO_SDIO_REV_1_00) ? 28 : 42; > > + /* let the SDIO driver take care of unknown tuples */ > if (size < min_size || buf[0] != 1) Misleading comment, the tuple is not unknown. > - return -EINVAL; > + return -EILSEQ; > What does this change improve? > /* TPLFE_MAX_BLK_SIZE */ > func->max_blksize = buf[12] | (buf[13] << 8); > @@ -154,13 +155,7 @@ static int cistpl_funce(struct mmc_card *card, struct sdio_func *func, > else > ret = cistpl_funce_common(card, buf, size); > > - if (ret) { > - printk(KERN_ERR "%s: bad CISTPL_FUNCE size %u " > - "type %u\n", mmc_hostname(card->host), size, buf[0]); > - return ret; > - } > - > - return 0; > + return ret; > } > > typedef int (tpl_parse_t)(struct mmc_card *, struct sdio_func *, Silencing a legitimate error. > + if (ret == -EILSEQ) { > + /* this tuple is unknown to the core */ Misleading comment, the tuple might be known but malformed. Rgds
--- El vie, 11/9/09, Andrew Morton <akpm@linux-foundation.org> escribió: > > Some manufacturers provide vendor information in > non-vendor specific CIS > > tuples. For example, Broadcom uses an Extended > Function tuple to provide > > the MAC address on some of their network cards, as in > the case of the > > Nintendo Wii WLAN daughter card. > > > > This patch allows passing correct tuples unknown to > the SDIO core to > > a matching SDIO driver instead of rejecting them and > failing. > > > > This looks leaky to me. > Hi Andrew, thanks for the review. I hope I can clarify a bit what the patch does in the next comments. > > :       if (i < ARRAY_SIZE(cis_tpl_list)) { > :          const struct cis_tpl *tpl = cis_tpl_list + i; > :          if (tpl_link < tpl->min_size) { > :             printk(KERN_ERR > :                  "%s: bad CIS tuple 0x%02x" > :                  " (length = %u, expected >= %u)\n", > :             mmc_hostname(card->host), > :                  tpl_code, tpl_link, tpl->min_size); > :             ret = -EINVAL; > > ret == -EINVAL > At this point ret is not -EINVAL. If it was -EINVAL the code would have had exit at this snipped before: if (ret) { kfree(this); break; } > :          } else if (tpl->parse) { > :             ret = tpl->parse(card, func, > :               this->data, tpl_link); > :          } > :          /* already successfully parsed, not needed anymore */ > :          if (!ret) > :             kfree(this); > > `this' doesn't get freed > Yes, that's the whole point of the patch. It must be freed _only_ if the SDIO core parsed it. If the SDIO core cannot parse it then it gets passed to the SDIO driver via the SDIO func struct tuple list (later). > :       } else { > :          /* unknown tuple */ > :          ret = -EILSEQ; > :       } > : > :       if (ret == -EILSEQ) { > > `this' doesn't get remembered. > When ret is -EILSEQ `this' is linked to the SDIO func tuple list (later). > :          /* this tuple is unknown to the core */ > :          this->next = NULL; > :          this->code = tpl_code; > :          this->size = tpl_link; > :          *prev = this; > :          prev = &this->next; > :          pr_debug("%s: queuing CIS tuple 0x%02x length %u\n", > :               mmc_hostname(card->host), tpl_code, tpl_link); > :          /* keep on analyzing tuples */ > :          ret = 0; > :       } > : > :       ptr += tpl_link; > > `this' leaks. > `this' doesn't leak. `this' has been linked to the SDIO func tuple list in: *prev = this; > :    } while (!ret); > > Please check? > Thanks a lot for you comments, Albert -- To unsubscribe from this list: send the line "unsubscribe linux-wireless" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Fri, 11 Sep 2009 07:52:11 +0000 (GMT) Albert Herranz <albert_herranz@yahoo.es> wrote: > --- El vie, 11/9/09, Andrew Morton <akpm@linux-foundation.org> escribi__: > > > Some manufacturers provide vendor information in > > non-vendor specific CIS > > > tuples. For example, Broadcom uses an Extended > > Function tuple to provide > > > the MAC address on some of their network cards, as in > > the case of the > > > Nintendo Wii WLAN daughter card. > > > > > > This patch allows passing correct tuples unknown to > > the SDIO core to > > > a matching SDIO driver instead of rejecting them and > > failing. > > > > > > > This looks leaky to me. > > > > Hi Andrew, thanks for the review. > I hope I can clarify a bit what the patch does in the next comments. > > > > > : ______ ______ if (i < ARRAY_SIZE(cis_tpl_list)) { argh, now yahoo mail is converting tabs to 0xa0's as well. Despair. > > : ______ ______ ______ const struct cis_tpl *tpl = cis_tpl_list + i; > > : ______ ______ ______ if (tpl_link < tpl->min_size) { > > : ______ ______ ______ ______ printk(KERN_ERR > > : ______ ______ ______ ______ __ __ ______"%s: bad CIS tuple 0x%02x" > > : ______ ______ ______ ______ __ __ ______" (length = %u, expected >= %u)\n", > > : ______ ______ ______ ______mmc_hostname(card->host), > > : ______ ______ ______ ______ __ __ ______tpl_code, tpl_link, tpl->min_size); > > : ______ ______ ______ ______ ret = -EINVAL; > > > > ret == -EINVAL > > > > At this point ret is not -EINVAL. Yes it is. We just did ret = -EINVAL; If that assignment happens, we leak `this'. -- To unsubscribe from this list: send the line "unsubscribe linux-wireless" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
--- El vie, 11/9/09, Pierre Ossman <pierre@ossman.eu> escribió: > The description for this patch should be made clearer. The title > suggests it adds functionality that's already in place. It should be > something along the lines of "Also pass malformed tuples to > card drivers". Hi Pierre, Thanks for your patch review. I didn't want to use "malformed" in the first place. I used "unknown" as "unknown to the SDIO core". The SDIO core in Linux only knows about FUNCE tuples of type 1 (with a sane length) as described in the SDIO Simplified Spec V2.00. I think we just have a language issue here, but if you prefer the "malformed" wording I'm ok with that. > In the sake of sanity, you might want to add this behaviour to all > parsers, not just the FUNCE one. I didn't find an application for the other parsers yet, so I tried to stick to the strictly necessary and just did the FUNCE one which has a direct application for Broadcom cards. > > I'm also unclear on how this is supposed to work. What does the > broadcom tuple look like? This patch looks like it will silence a lot > of legitimate warnings, and possibly pollute the card structures with > bogus data. The contents of the Broadcom FUNCE (type 0x22) tuple that contains the MAC address of a card looks like the following (tuple size 8): 04 06 00 1d bc 62 79 fd ^^ ^^ ^^^^^^^^^^^^^^^^^ | | | | | +--- MAC address | +--------------------- length (6 bytes for a ethernet MAC address) +------------------------ type 4 (CISTPL_FUNCE_LAN_NODE_ID) If you prefer it, instead of passing al "unknown" (or "malformed") FUNCE tuples to SDIO drivers, I can let through only a subset of whitelisted FUNCE types (starting with type 4). > > > diff --git a/drivers/mmc/core/sdio_cis.c b/drivers/mmc/core/sdio_cis.c > > index 963f293..87934ac 100644 > > --- a/drivers/mmc/core/sdio_cis.c > > +++ b/drivers/mmc/core/sdio_cis.c > > @@ -123,8 +123,9 @@ static int cistpl_funce_func(struct sdio_func *func, > >    vsn = func->card->cccr.sdio_vsn; > >    min_size = (vsn == SDIO_SDIO_REV_1_00) ? 28 : 42; > > > > +   /* let the SDIO driver take care of unknown tuples */ > >    if (size < min_size || buf[0] != 1) > > Misleading comment, the tuple is not unknown. > Same language issue described before. > > -      return -EINVAL; > > +      return -EILSEQ; > > > > What does this change improve? -EILSEQ is used to indicate that the tuple was not parsed by the SDIO core and should be passed to the SDIO driver via the SDIO func tuple list. > > >    /* TPLFE_MAX_BLK_SIZE */ > >    func->max_blksize = > buf[12] | (buf[13] << 8); > > @@ -154,13 +155,7 @@ static int cistpl_funce(struct mmc_card *card, struct sdio_func *func, > >    else > >       ret = cistpl_funce_common(card, buf, size); > > > > -   if (ret) { > > -      printk(KERN_ERR "%s: bad CISTPL_FUNCE size %u " > > -           "type %u\n", mmc_hostname(card->host), size, buf[0]); > > -      return ret; > > -   } > > - > > -   return 0; > > +   return ret; > > } > > > > typedef int (tpl_parse_t)(struct mmc_card *, struct sdio_func *, > > Silencing a legitimate error. > Yes, I see your point. I think we can keep this code but prevent displaying the error if ret == -EILSEQ (i.e. the tuple is "unknown"/"malformed" BUT should be passed to the SDIO driver for parsing). > > +      if (ret == -EILSEQ) { > > +      >    /* this tuple is unknown to the core */ > > Misleading comment, the tuple might be known but malformed. Same languange issue again. > > Rgds > -- >     -- Pierre Ossman Thanks a lot for your comments, Albert -- To unsubscribe from this list: send the line "unsubscribe linux-wireless" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
--- El vie, 11/9/09, Andrew Morton <akpm@linux-foundation.org> escribió: > > > > > > ret == -EINVAL > > > > > > > At this point ret is not -EINVAL. > > Yes it is. We just did > >    ret = -EINVAL; > > > If that assignment happens, we leak `this'. > Hi Andrew, I misunderstood you. I thought that you were trying to imply on your original comment that retval was _already_ -EINVAL at that point. Now I see the issue. `this' should be freed if successfully parsed (!ret) or if invalid and not going to be passed to a SDIO driver (ret == -EINVAL). Thanks for catching that. I'll send an updated patch. Cheers, Albert -- To unsubscribe from this list: send the line "unsubscribe linux-wireless" 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/sdio_cis.c b/drivers/mmc/core/sdio_cis.c index 963f293..87934ac 100644 --- a/drivers/mmc/core/sdio_cis.c +++ b/drivers/mmc/core/sdio_cis.c @@ -123,8 +123,9 @@ static int cistpl_funce_func(struct sdio_func *func, vsn = func->card->cccr.sdio_vsn; min_size = (vsn == SDIO_SDIO_REV_1_00) ? 28 : 42; + /* let the SDIO driver take care of unknown tuples */ if (size < min_size || buf[0] != 1) - return -EINVAL; + return -EILSEQ; /* TPLFE_MAX_BLK_SIZE */ func->max_blksize = buf[12] | (buf[13] << 8); @@ -154,13 +155,7 @@ static int cistpl_funce(struct mmc_card *card, struct sdio_func *func, else ret = cistpl_funce_common(card, buf, size); - if (ret) { - printk(KERN_ERR "%s: bad CISTPL_FUNCE size %u " - "type %u\n", mmc_hostname(card->host), size, buf[0]); - return ret; - } - - return 0; + return ret; } typedef int (tpl_parse_t)(struct mmc_card *, struct sdio_func *, @@ -253,21 +248,12 @@ static int sdio_read_cis(struct mmc_card *card, struct sdio_func *func) for (i = 0; i < ARRAY_SIZE(cis_tpl_list); i++) if (cis_tpl_list[i].code == tpl_code) break; - if (i >= ARRAY_SIZE(cis_tpl_list)) { - /* this tuple is unknown to the core */ - this->next = NULL; - this->code = tpl_code; - this->size = tpl_link; - *prev = this; - prev = &this->next; - printk(KERN_DEBUG - "%s: queuing CIS tuple 0x%02x length %u\n", - mmc_hostname(card->host), tpl_code, tpl_link); - } else { + if (i < ARRAY_SIZE(cis_tpl_list)) { const struct cis_tpl *tpl = cis_tpl_list + i; if (tpl_link < tpl->min_size) { printk(KERN_ERR - "%s: bad CIS tuple 0x%02x (length = %u, expected >= %u)\n", + "%s: bad CIS tuple 0x%02x" + " (length = %u, expected >= %u)\n", mmc_hostname(card->host), tpl_code, tpl_link, tpl->min_size); ret = -EINVAL; @@ -275,7 +261,25 @@ static int sdio_read_cis(struct mmc_card *card, struct sdio_func *func) ret = tpl->parse(card, func, this->data, tpl_link); } - kfree(this); + /* already successfully parsed, not needed anymore */ + if (!ret) + kfree(this); + } else { + /* unknown tuple */ + ret = -EILSEQ; + } + + if (ret == -EILSEQ) { + /* this tuple is unknown to the core */ + this->next = NULL; + this->code = tpl_code; + this->size = tpl_link; + *prev = this; + prev = &this->next; + pr_debug("%s: queuing CIS tuple 0x%02x length %u\n", + mmc_hostname(card->host), tpl_code, tpl_link); + /* keep on analyzing tuples */ + ret = 0; } ptr += tpl_link;
Some manufacturers provide vendor information in non-vendor specific CIS tuples. For example, Broadcom uses an Extended Function tuple to provide the MAC address on some of their network cards, as in the case of the Nintendo Wii WLAN daughter card. This patch allows passing correct tuples unknown to the SDIO core to a matching SDIO driver instead of rejecting them and failing. Signed-off-by: Albert Herranz <albert_herranz@yahoo.es> --- v1 - fixed typo in commit message - CC'd akpm as suggested by mb - required by commit 4ea602e183ca20a7577ebe253323d0e5d0f9847 in net-next-2.6 drivers/mmc/core/sdio_cis.c | 46 +++++++++++++++++++++++------------------- 1 files changed, 25 insertions(+), 21 deletions(-)