Date: Mon, 26 Jan 2009 14:04:27 +0100 From: Lukasz Jaroszewski <lvj@nietykalni.org> To: Christoph Mallon <christoph.mallon@gmx.de> Cc: freebsd-hackers@freebsd.org, Hans Petter Selasky <hselasky@c2i.net> Subject: Re: write(2) to /dev/ad4 = EINVAL Message-ID: <3930e0780901260504m44f43958m82f2ba8902558906@mail.gmail.com> In-Reply-To: <497D9116.1040408@gmx.de> References: <3930e0780901260135j1428e77fmfe7392770fcd31fc@mail.gmail.com> <200901261109.43376.hselasky@c2i.net> <3930e0780901260224ye4acbabx7a629ec8da195138@mail.gmail.com> <497D9116.1040408@gmx.de>
next in thread | previous in thread | raw e-mail | index | archive | help
2009/1/26 Christoph Mallon <christoph.mallon@gmx.de>:
> Lukasz Jaroszewski schrieb:
>>
>> 2009/1/26 Hans Petter Selasky <hselasky@c2i.net>:
>>>
>>> On Monday 26 January 2009, Lukasz Jaroszewski wrote:
>>>>
>>>> Hi,
>>>> after opening /dev/ad4 with success for O_RDWR, I am getting [EINVAL]
>>>> from write(2), which according to man 2 write, means
>>>> `` [EINVAL] The pointer associated with d was
>>>> negative.'', as you can see below it is not true, I have tried
>>>> different block sizes, with same result.
>>>>
>>>> How can I write one byte to /dev/adN ?
>>>
>>> Hi,
>>>
>>> You cannot write one byte to /dev/adN, I think. Harddisks are block
>>> based.
>>> Please see:
>>>
>>> diskinfo -v ad4
>>>
>>> And especially the "sector size". When you seek and transfer data the
>>> offset
>>> and length must be a factor of the "sector size" or "block size". See
>>> also
>>> LBA, logical block address.
>>
>>
>> Yes, I have tried with bs=512 and multiplies, no luck.
>>
>> root@~(0) diskinfo -v ad4
>> ad4
>> 512 # sectorsize
>> 2029805568 # mediasize in bytes (1.9G)
>> 3964464 # mediasize in sectors
>> 3933 # Cylinders according to firmware.
>> 16 # Heads according to firmware.
>> 63 # Sectors according to firmware.
>> ad:CFx20CARDx200000190C # Disk ident.
>
> It would be helpful, if you showed the actual code, instead of letting us
> guess what you are doing.
Ok, I solved that, I were trying to dd with bs correctly set, tho
if=./s wasn't big enough plus I had code like below, where I didn't
check write for errors, just called perror, which gave me wrong errno.
So to change one byte, I need to read it, change it, and write it in
512 chunks? or play with kernel and add smth like ``write_byte'' to
ata?
I am sorry for bothering with my silly errors.
<includes>
#define S 0x200
int
main(int argc, char **argv){
int fd,i=0;
char *s;
s = malloc(S);
for(i=0;i<S;i++)
s[i] = 'B';
if((fd = open("/dev/ad4",
O_RDWR|O_CREAT|O_TRUNC,S_IRUSR|S_IWUSR|S_IRGRP|S_IWGRP|S_IROTH|S_IWOTH))==-1)
perror("open(2)");
lseek(fd, 0, SEEK_SET);
write(fd, s,S);
perror("write");
close(fd);
return 0;
}
Best Regards
Lukasz
>
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3930e0780901260504m44f43958m82f2ba8902558906>